Joint Industry Plan; Notice of Filing of the National Market System Plan Governing the Consolidated Audit Trail

Download PDF
Federal RegisterMay 17, 2016
81 Fed. Reg. 30613 (May. 17, 2016)
April 27, 2016.

Table of Contents

I. Introduction

II. Background

III. Description of the Plan

A. Statement of Purpose and Request for Comment

1. Background

2. Request for Exemption from Certain Requirements under Rule 613

3. Requirements Pursuant to Rule 608(a)

B. Summary of Additional CAT NMS Plan Provisions and Request for Comment

1. Reporting Procedures

2. Timeliness of Data Reporting

3. Uniform Format

4. Clock Synchronization

5. Time Stamp Granularity

6. CAT-Reporter-ID

7. Customer-ID

8. Order Allocation Information

9. Options Market Maker Quotes

10. Error Rates

11. Regulatory Access

12. Security, Confidentiality, and Use of Data

IV. Economic Analysis

A. Introduction

B. Summary of Expected Economic Effects

C. Framework for Economic Analysis

1. Economic Framework

2. Existing Uncertainties

3. Request for Comment on the Framework

D. Baseline

1. Current State of Regulatory Activities

2. Current State of Trade and Order Data

3. Request for Comment on the Baseline

E. Benefits

1. Improvements in Data Qualities

2. Improvements to Regulatory Activities

3. Other Provisions of the CAT NMS Plan

4. Request for Comment on the Benefits

F. Costs

1. Analysis of Expected Costs

2. Aggregate Costs to Industry

3. Further Analysis of Costs

4. Second-Order Effects and Other Security-related Costs

5. Request for Comment on the Costs

G. Efficiency, Competition, and Capital Formation

1. Competition

2. Efficiency

3. Capital Formation

4. Related Considerations Affecting Competition, Efficiency and Capital Formation

5. Request for Comment on Efficiency, Competition, and Capital Formation

H. Alternatives

1. Alternatives to the Approaches the Exemption Order Permitted to be Included in the Plan

2. Alternatives to Certain Specific Approaches in the CAT NMS Plan

3. Alternatives to the Scope of Certain Specific Elements in the CAT NMS Plan

4. Alternatives to the CAT NMS Plan

5. Request for Comment on the Alternatives

I. Request for Comment on the Economic Analysis

V. Paperwork Reduction Act

A. Summary of Collection of Information under Rule 613

1. Central Repository

2. Data Collection and Reporting

3. Collection and Retention of NBBO, Last Sale Data and Transaction Reports

4. Surveillance

5. Participant Rule Filings

6. Written Assessment of Operation of the Consolidated Audit Trail

7. Document on Expansion to Other Securities

B. Proposed Use of Information

1. Central Repository

2. Data Collection and Reporting

3. Collection and Retention of NBBO, Last Sale Data and Transaction Reports

4. Surveillance

5. Written Assessment of Operation of the Consolidated Audit Trail

6. Document on Expansion to Other Securities

C. Respondents

1. National Securities Exchanges and National Securities Associations

2. Members of National Securities Exchanges and National Securities Association

D. Total Initial and Annual Reporting and Recordkeeping Burden

1. Burden on National Securities Exchanges and National Securities Associations

2. Burden on Members of National Securities Exchanges and National Securities Associations

E. Collection of Information is Mandatory

F. Confidentiality

G. Recordkeeping Requirements

H. Request for Comments

VI. Solicitation of Comments

I. Introduction

Pursuant to Section 11A of the Securities Exchange Act of 1934 (the “Act”) and Rule 608 thereunder, notice is hereby given that on February 27, 2015, BATS Exchange, Inc., BATS-Y Exchange, Inc., BOX Options Exchange LLC, C2 Options Exchange, Incorporated, Chicago Board Options Exchange, Incorporated, Chicago Stock Exchange, Inc., EDGA Exchange, Inc., EDGX Exchange, Inc., Financial Industry Regulatory Authority, Inc., International Securities Exchange, LLC, ISE Gemini, LLC, Miami International Securities Exchange LLC, NASDAQ OMX BX, Inc., NASDAQ OMX PHLX LLC, The NASDAQ Stock Market LLC, National Stock Exchange, Inc., New York Stock Exchange LLC, NYSE MKT LLC, and NYSE Arca, Inc. (collectively, “SROs” or “Participants”), filed with the Securities and Exchange Commission (the “Commission” or “SEC”) a National Market System Plan Governing the Consolidated Audit Trail (the “CAT NMS Plan” or “Plan”). On December 24, 2015, the SROs submitted an Amendment to the CAT NMS Plan. A copy of the CAT NMS Plan, as modified by the Amendment, is attached as Exhibit A hereto. The Commission is publishing this Notice to solicit comments on the CAT NMS Plan. The Commission also is publishing notice of, and soliciting comment on, an analysis of the potential economic effects of implementing the CAT NMS Plan, as set forth in Section IV of this Notice, and the collection of information requirements in the CAT NMS Plan as required by the Paperwork Reduction Act, as set forth in Section V of this Notice.

15 U.S.C. 78k-1.

17 CFR 242.608.

See Letter from Participants to Brent J. Fields, Secretary, Commission, dated February 27, 2015. Pursuant to Rule 613, the SROs were required to file the CAT NMS Plan on or before April 28, 2013. At the SROs' request, the Commission granted exemptions to extend the deadline for filing the CAT NMS Plan to December 6, 2013, and then to September 30, 2014. See Securities Exchange Act Release Nos. 69060 (March 7, 2013), 78 FR 15771 (March 12, 2013); 71018 (December 6, 2013), 78 FR 75669 (December 12, 2013). The SROs filed the CAT NMS Plan on September 30, 2014 (the “Initial CAT NMS Plan”). See Letter from the SROs, to Brent J. Fields, Secretary, Commission, dated September 30, 2014. The CAT NMS Plan filed on February 27, 2015, was an amendment to and replacement of the Initial CAT NMS Plan (the “Amended and Restated CAT NMS Plan”). On December 24, 2015, the SROs submitted an Amendment to the Amended and Restated CAT NMS Plan. See Letter from Participants to Brent J. Fields, Secretary, Commission, dated December 23, 2015 (the “Amendment”). On February 9, 2016, the Participants filed with the Commission an identical, but unmarked, version of the Amended and Restated CAT NMS Plan, dated February 27, 2015, as modified by the Amendment, as well as a copy of the request for proposal issued by the Participants to solicit Bids from parties interested in serving as the Plan Processor for the consolidated audit trail. See Exhibit A and infra note 29. Unless the context otherwise requires, the “CAT NMS Plan” shall refer to the Amended and Restated CAT NMS Plan, as modified by the Amendment. The Commission notes that the application of ISE Mercury, LLC for registration as a national securities exchange was granted on January 29, 2016. See Securities Exchange Act Release No. 76998 (January 29, 2016), 81 FR 6066 (February 4, 2016). The Commission understands that ISE Mercury, LLC will become a Participant in the CAT NMS Plan and thus is accounted for as a Participant for purposes of this Notice.

See Amendment, supra note 3.

II. Background

The Commission believes that the regulatory data infrastructure on which the SROs and the Commission currently must rely generally is outdated and inadequate to effectively oversee a complex, dispersed, and highly automated national market system. In performing their oversight responsibilities, regulators today must attempt to cobble together disparate data from a variety of existing information systems lacking in completeness, accuracy, accessibility, and/or timeliness—a model that neither supports the efficient aggregation of data from multiple trading venues nor yields the type of complete and accurate market activity data needed for robust market oversight.

Currently, FINRA and some of the exchanges maintain their own separate audit trail systems for certain segments of this trading activity, which vary in scope, required data elements and format. In performing their market oversight responsibilities, SRO and Commission Staffs today must rely heavily on data from these various SRO audit trails. However, as noted in Section IV.D below, there are shortcomings in the completeness, accuracy, accessibility, and timeliness of these existing audit trail systems. Some of these shortcomings are a result of the disparate nature of the systems, which make it impractical, for example, to follow orders through their entire lifecycle as they may be routed, aggregated, re-routed, and disaggregated across multiple markets. The lack of key information in the audit trails that would be useful for regulatory oversight, such as the identity of the customers who originate orders, or even the fact that two sets of orders may have been originated by the same customer, is another shortcoming.

The Commission notes that the SROs have taken steps in recent years to update their audit trail requirements. For example, NYSE, NYSE Amex LLC (n/k/a “NYSE MKT LLC”) (“NYSE Amex”), and NYSE ARCA, Inc. (“NYSE Arca”) have adopted audit trail rules that coordinate with FINRA's OATS requirements. See Securities Exchange Act Release No. 65523 (October 7, 2011), 76 FR 64154 (October 17, 2011) (concerning NYSE); Securities Exchange Act Release No. 65524 (October 7, 2011), 76 FR 64151 (October 17, 2011) (concerning NYSE Amex); Securities Exchange Act Release No. 65544 (October 12, 2011), 76 FR 64406 (October 18, 2011) (concerning NYSE Arca). This allows the SROs to submit their data to FINRA pursuant to a Regulatory Service Agreement (“RSA”), which FINRA can then reformat and combine with OATS data. Despite these efforts, however, significant deficiencies remain. See Section IV.D.2, infra.

Though SRO and Commission Staff also have access to sources of market activity data other than SRO audit trails, these systems each suffer their own drawbacks. For example, data obtained from the electronic blue sheet (“EBS”) system and equity cleared reports comprise only trade executions, and not orders or quotes. In addition, like data from existing audit trails, data from these sources lacks key elements important to regulators, such as the identity of the customer in the case of equity cleared reports. Furthermore, recent experience with implementing incremental improvements to the EBS system has illustrated some of the overall limitations of the current technologies and mechanisms used by the industry to collect, record, and make available market activity data for regulatory purposes.

EBSs are trading records requested by the Commission and SROs from broker-dealers that are used in regulatory investigations to identify buyers and sellers of specific securities.

The Commission uses the National Securities Clearing Corporation's (“NSCC”) equity cleared report for initial regulatory inquiries. This report is generated on a daily basis by the SROs and is provided to the NSCC in a database accessible by the Commission, and shows the number of trades and daily volume of all equity securities in which transactions took place, sorted by clearing member. The information provided is end-of-day data and is searchable by security name and CUSIP number.

See Securities Exchange Act Release No. 64976 (July 27, 2011), 76 FR 46960 (August 3, 2011) (“Large Trader Release”).

Recognizing these shortcomings, on July 11, 2012, the Commission adopted Rule 613 of Regulation NMS under the Act. Rule 613 required the SROs to submit a national market system (“NMS”) plan to create, implement, and maintain a consolidated audit trail (“CAT”) that would capture customer and order event information for orders in NMS securities, across all markets, from the time of order inception through routing, cancellation, modification, or execution in a single, consolidated data source. On February 27, 2015, the SROs submitted the CAT NMS Plan.

See Securities Exchange Act Release No. 67457 (July 18, 2012), 77 FR 45722 (August 1, 2012) (“Adopting Release”); see also Securities Exchange Act Release No. 62174 (May 26, 2010), 75 FR 32556 (June 8, 2010) (“Proposing Release”).

See 17 CFR 242.613(a)(1), (c)(1), (c)(7).

See supra note 3.

The SROs also submitted a separate NMS plan and an exemptive request letter related to the CAT NMS Plan. Specifically, on September 3, 2013, the SROs filed an NMS Plan pursuant to Rule 608 governing the SROs' review, evaluation, and ultimate selection of the Plan Processor for the consolidated audit trail (the “Selection Plan”). The Selection Plan was published for comment in the Federal Register on November 21, 2013 and approved by the Commission on February 21, 2014. Subsequently, the SROs filed three amendments to the Selection Plan, two of which were approved by the Commission on June 17, 2015 and September 24, 2015 The CAT NMS Plan reflects the process approved by the Commission for reviewing, evaluating and ultimately selecting the Plan Processor, as set forth in the Selection Plan, as amended. Second, on January 30, 2015, the SROs filed an application, pursuant to Rule 0-12 under the Act, requesting that the Commission grant exemptions from certain requirements of Rule 613. The Commission granted the exemptions on March 1, 2016. The CAT NMS Plan published for comment in this Notice reflects the exemptive relief granted by the Commission.

As set forth in Section 1.1 of the CAT NMS Plan, the Plan Processor “means the Initial Plan Processor or any other Person selected by the Operating Committee pursuant to SEC Rule 613 and Sections 4.3(b)(i) and 6.1, and with regard to the Initial Plan Processor, the Selection Plan, to perform the CAT processing functions required by SEC Rule 613 and set forth in [the CAT NMS Plan].”

See Securities Exchange Act Release No. 70892 (November 15, 2013), 78 FR 69910 (November 21, 2013) (“Selection Plan Notice”).

See id.; see also Securities Exchange Act Release No. 71596, 79 FR 11152 (February 27, 2014) (“Selection Plan Approval Order”).

See Securities Exchange Act Release Nos. 75192 (June 17, 2015), 80 FR 36028 (June 23, 2015) (Order Approving Amendment No. 1 to the Selection Plan); 75980 (September 24, 2015), 80 FR 58796 (September 30, 2015) (Order Approving Amendment No. 2 to the Selection Plan); Letter from SROs to Brent J. Fields, Secretary, Commission, dated March 29, 2016; see also Securities Exchange Act Release Nos. 74223 (February 6, 2015), 80 FR 7654 (February 11, 2015) (Notice of Amendment No. 1 to the Selection Plan); 75193 (June 17, 2015), 80 FR 36006 (June 23, 2015) (Notice of Amendment No. 2 to the Selection Plan).

See Letter from Participants to Brent J. Fields, Secretary, Commission, dated January 30, 2015 (“Exemptive Request Letter”). Specifically, the SROs request exemptive relief from the Rule's requirements related to: (1) The reporting of Options Market Maker quotations, as required under Rule 613(c)(7)(ii) and (iv); (2) the reporting and use of the Customer-ID under Rule 613(c)(7)(i)(A), (iv)(F), (viii)(B) and 613(c)(8); (3) the reporting of the CAT-Reporter-ID, as required under Rule 613(c)(7)(i)(C), (ii)(D), (ii)(E), (iii)(D), (iii)(E), (iv)(F), (v)(F), (vi)(B), and (c)(8); (4) the linking of executions to specific subaccount allocations, as required under Rule 613(c)(7)(vi)(A); and (5) the time stamp granularity requirement of Rule 613(d)(3) for certain manual order events subject to reporting under Rule 613(c)(7)(i)(E), (ii)(C), (iii)(C) and (iv)(C). On April 3, 2015, the SROs filed a supplement related to the requested exemption for Rule 613(c)(7)(vi)(A). See Letter from Robert Colby, FINRA, on behalf of the SROs, to Brent J. Fields, Secretary, Commission, dated April 3, 2015 (“April 2015 Supplement”). This supplement provided examples of how the proposed relief related to allocations would operate. On September 2, 2015, the SROs filed a second supplement to the Exemptive Request Letter. See Letter from the SROs to Brent J. Fields, Secretary, Commission, dated September 2, 2015 (“September 2015 Supplement”). This supplement to the Exemptive Request Letter further addressed the use of an “effective date” in lieu of a “date account opened.” Unless the context otherwise requires, the “Exemption Request” shall refer to the Exemptive Request Letter, as supplemented by the April 2015 Supplement and the September 2015 Supplement.

17 CFR 240.0-12.

See Securities Exchange Act Release No. 77265 (March 1, 2016), 81 FR 11856 (March 7, 2016) (“Exemption Order”). The Commission requests comment specifically on the advantages and disadvantages of each aspect of the relief granted in the Exemption Order and whether the approaches permitted by the Exemption Order to be included in the CAT NMS Plan are preferable to those originally permitted by Rule 613. See Request for Comment Nos. 168-170 (Options Market Maker Quotes), 135-161 (Customer ID), 128-134 (CAT-Reporter-ID), 162-167 (Linking Order Executions to Allocations) and 114-127 (Time Stamp Granularity), infra.

III. Description of the Plan

As described further in this Section III of this Notice, the SROs propose to conduct the activities of the CAT through CAT NMS, LLC, a jointly owned limited liability company formed under Delaware state law; and to that end, the SROs submitted the CAT NMS, LLC's limited liability company agreement (the “LLC Agreement”), including exhibits and appendices attached thereto, to the Commission as the CAT NMS Plan. The SROs also submitted a cover letter that included a description of the CAT NMS Plan, along with the information required by Rule 608(a)(4) and (5) under the Act, which is set forth below in Section III.A of this Notice as substantially prepared and submitted by the SROs. Set forth in Section III.B is a summary of additional CAT NMS Plan provisions and requests for comment.

17 CFR 242.608(a)(4) and (a)(5).

All capitalized terms not otherwise defined herein shall have the meaning ascribed to them in Rule 613, the Adopting Release, or the CAT NMS Plan, as applicable.

The LLC Agreement, attached hereto as Exhibit A, sets forth a governing structure, whereby the Operating Committee will manage the CAT NMS, LLC, and each SRO will be a member of, and have one vote within, the Operating Committee. The LLC Agreement details the Operating Committee's procedures for selecting the Plan Processor, who will be contracted to build the CAT, as well as the functions and activities of the Plan Processor. The LLC Agreement also sets forth the responsibilities of the Central Repository which, under the oversight of the Plan Processor, will receive, consolidate and retain the CAT Data. The LLC Agreement also lists the requirements regarding the recording and reporting of CAT Data by the SROs as well as by broker-dealers, the security and confidentiality safeguards for CAT Data, surveillance requirements, fees and costs associated with operating the CAT, as well as other reporting and Technical Specifications and requirements.

See CAT NMS Plan, supra note 3, at Article IV.

See id. at Article V; see also Order Approving Amendment No. 1 to the Selection Plan and Order Approving Amendment No. 2 to the Selection Plan, supra note 15.

See CAT NMS Plan, supra note 3, at Article VI.

See id.

In Appendix C to the LLC Agreement, the SROs address the considerations listed in Rule 613(a)(1), providing information and analysis regarding the specific features, details, costs, and processes related to the CAT NMS Plan. Appendix D to the LLC Agreement provides an outline of the CAT's minimum functional and technical requirements for the Plan Processor.

A. Statement of Purpose and Request for Comment

The following statement of purpose provided herein is substantially as prepared and submitted by the SROs to the Commission. Throughout the statement of purpose, the Commission has inserted requests for comment. The portion of this Notice prepared by the Commission will re-commence in Section III.B.

See CAT NMS Plan, supra note 3.

1. Background

On July 11, 2012, the Commission adopted Rule 613 to require the national securities exchanges and national securities association to jointly submit a national market system plan to create, implement, and maintain a consolidated audit trail and central repository. Rule 613 outlines a broad framework for the creation, implementation, and maintenance of the consolidated audit trail, including the minimum elements the Commission believes are necessary for an effective consolidated audit trail.

17 CFR 242.613.

17 CFR 242.613(a)(1).

See Adopting Release, supra note 9, at 45743.

Since the adoption of Rule 613, the Participants have worked to formulate an effective Plan. To this end, the Participants have, among other things, developed a plan for selecting the Plan Processor, solicited and evaluated Bids, and engaged diverse industry participants in the development of the Plan. Throughout, the Participants have sought to implement a process that is fair, transparent, and consistent with the standards and considerations in Rule 613.

a. The Request for Proposal and Selection Plan

On February 26, 2013, the Participants published a request for proposal (“RFP”) soliciting Bids from parties interested in serving as the Plan Processor. The Participants concluded that publication of an RFP was necessary to ensure that potential alternative solutions to creating the Plan and the CAT could be presented and considered, and that a detailed and meaningful cost-benefit analysis could be performed. The Participants asked any potential bidders to notify the Participants of their intent to bid by March 5, 2013. Initially, 31 firms submitted intentions to bid, four of which were Participants or affiliates of Participants. In the following weeks and months, the Participants engaged with potential bidders with respect to, among other things, the selection process, selection criteria, and potential bidders' questions and concerns.

See Appendix A of the CAT NMS Plan for the Consolidated Audit Trail National Market System Plan Request for Proposal (issued February 26, 2013, version 3.0 updated March 4, 2014). Other materials related to the RFP are available at http://catnmsplan.com/process/ .

In an effort to ensure Bidders were aware of all information provided in response to Bidders' questions related to the RFP, the Participants published answers to questions received from Bidders available at http://catnmsplan.com/process/ .

On September 4, 2013, the Participants filed with the Commission a national market system plan to govern the process for Participant review of the Bids submitted in response to the RFP, the procedure for evaluating the Bids, and, ultimately, selection of the Plan Processor (the “Selection Plan”). The Commission approved the Selection Plan as filed on February 21, 2014. On March 21, 2014, the Participants received ten Bids in response to the RFP.

See Selection Plan Notice, supra note 13.

See Selection Plan Approval Order, supra note 14.

The Selection Plan divides the review and evaluation of Bids, and the selection of the Plan Processor, into various stages, certain of which have been completed to date. Specifically, pursuant to the Selection Plan, the Selection Committee reviewed all Bids and determined which Bids contained sufficient information to allow the Participants to meaningfully assess and evaluate the Bids. The ten submitted Bids were deemed “Qualified Bids,” and so passed to the next stage, in which each Bidder presented its Bids in person to the Participants on a confidential basis. On July 1, 2014, after conducting careful analysis and comparison of the Bids, the Selection Committee voted and selected six Shortlisted Bidders, thus eliminating four Bidders from continuing in the process. The Selection Committee, subject to applicable recusal provisions in the Selection Plan, will determine whether Shortlisted Bidders will be provided the opportunity to revise their Bids. After the Selection Committee further assesses and evaluates the Shortlisted Bids, including any permitted revisions to the Bids, the Selection Committee will select the Plan Processor via two rounds of voting by the Senior Voting Officers as specified in the Plan.

See, e.g., id. at 11154.

A list of Qualified Bidders is available at http://catnmsplan.com/web/groups/catnms/@catnms/documents/appsupportdocs/p493591.pdf . The Commission notes that this Web site address has been updated to http://www.catnmsplan.com/process/p493591.pdf .

The announcement and list of the Shortlisted Bidders is available at http://catnmsplan.com/web/groups/catnms/@catnms/documents/appsupportdocs/p542077.pdf . The Commission notes that this Web site address has been updated to http://www.catnmsplan.com/pastevents/p542077.pdf . Additionally, the Commission notes that the Selection Committee further narrowed the list of Shortlisted Bidders to three Shortlisted Bidders. See Participants, SROs Reduce Short List Bids from Six to Three for Consolidated Audit Trail (November 16, 2015), available at http://www.catnmsplan.com/pastevents/catnms_release_downselect_111615.pdf .

See Selection Plan Approval Order, supra note 14, at 11154. The SEC published a notice of an amendment to the Selection Plan, which proposed to amend the Selection Plan in two ways. First, the Participants proposed to provide opportunities to accept revised Bids prior to approval of the CAT NMS Plan, and second, to allow the list of Shortlisted Bids to be narrowed prior to Commission approval of the CAT NMS Plan. See Notice of Amendment No. 1 to the Selection Plan, supra note 15. In addition, the Participants filed a second amendment to the Selection Plan, which would require the recusal of a Bidding Participant in a vote in any round by the Selection Committee to select the Plan Processor from among the Shortlisted Bidders if such Bidding Participant's Bid, a Bid submitted by an Affiliate of such Bidding Participant, or a Bid including such Bidding Participant or its Affiliate is also considered in that round. See Notice of Amendment No. 2 to the Selection Plan, supra note 15. The prior Selection Plan required recusal of a Bidding Participant under such circumstances in the vote in only the second round by the Selection Committee to select the Plan Processor from among the Shortlisted Bidders. The Commission notes that Amendment Nos. 1 and 2 have been approved. See Order Approving Amendment No. 1 to the Selection Plan and Order Approving Amendment No. 2 to the Selection Plan, supra note 15.

b. Selection Plan Governance and Operations

The Selection Plan established an Operating Committee responsible for formulating, drafting, and filing with the Commission the Plan and for ensuring that the Participants' joint obligations under Rule 613 were met in a timely and efficient manner. Each Participant selected one individual and one substitute to serve on the Operating Committee, with other representatives of each Participant permitted to attend Operating Committee meetings. In formulating the Plan, the Participants also engaged multiple persons across a wide range of roles and expertise, engaged the consulting firm Deloitte & Touche LLP as a project manager, and engaged the law firm Wilmer Cutler Pickering Hale and Dorr LLP to serve as legal counsel in drafting the Plan. Within this structure, the Participants focused on, among other things, comparative analyses of the proposed technologies and operating models, development of funding models to support the building and operation of the CAT, and detailed review of governance considerations. Since July 2012, the Participants have held approximately 608 meetings related to the CAT. These governance and organizational structures will continue to be in effect until the Commission's final approval of the Plan.

Id.

Id.

Additional information regarding these meetings can be found at http://catnmsplan.com/ . The Commission notes that the number of meetings in the SROs' statement is as of February 27, 2015. See CAT NMS Plan, supra note 3.

See Selection Plan Approval Order, supra note 14, at 11155.

c. Engagement With Industry Participants

Throughout the process of developing the Plan, the Participants consistently have been engaged in meaningful dialogue with industry participants with respect to the development of the CAT. From the outset of this process, the Participants have recognized that industry input is a critical component in the creation of the Plan. To this end, the Participants created a Web site to update the public on the progress of the Plan, published requests for comment on multiple issues related to the Plan, held multiple public events to inform the industry of the progress of the CAT and to address inquiries, and formed, and later expanded, a Development Advisory Group (the “DAG”) to solicit more input from a representative industry group.

The Web site is available at http://catnmsplan.com/ .

The DAG conducted 43 meetings to discuss, among other things, technical and operational aspects the Participants were considering for the Plan. The Participants twice issued press releases soliciting participants for the DAG, and a wide spectrum of firms was deliberately chosen to provide insight from various industry segments affected by the CAT. The DAG currently consists of the Participants, and 27 diverse firms and organizations (including broker-dealers of varying sizes, the Options Clearing Corporation, a service bureau and three industry trade associations) with a variety of subject matter expertise. The DAG meetings have included discussions of topics such as Options Market Maker quote reporting, requirements for capturing Customer-IDs, time stamps and clock synchronization, reporting requirements for order handling scenarios, cost and funding, error handling and corrections, and potential elimination of Rules made redundant by the CAT.

In addition to these meetings, DAG subcommittee meetings also were held. The Commission notes that the number of meetings in the SROs' statement is as of February 27, 2015. See CAT NMS Plan, supra note 3.

For a list of DAG members, see Summary of the Consolidated Audit Trail Initiative at 13 (Jan. 2015), available at http://catnmsplan.com/web/groups/catnms/@catnms/documents/appsupportdocs/p571933.pdf . The Commission notes that the list of DAG members appears on page 6 of the linked document, which is dated May 2015.

The list of current DAG members is available at http://catnmsplan.com/PastEvents/ .

See, e.g., Summary of the Consolidated Audit Trail Initiative, supra note 43, at 14.

In addition, the CAT Web site includes a variety of resources for the public with respect to the development of the CAT. The site contains an overview of the process, an expression of the guiding principles behind the Plan development, links to relevant regulatory actions, gap analyses comparing the requirements of Rule 613 with current reporting systems, the CAT implementation timeline, a summary of the RFP process, a set of frequently-asked questions (updated on an ongoing basis), questions for comment from the industry, industry feedback on the development of the Plan, and announcements and notices of upcoming events. This Web site, along with the requests for comments and many public events (announced on the site), have been a venue for public communication with respect to the development of the Plan.

2. Request for Exemption From Certain Requirements Under Rule 613

Following multiple discussions between the Participants and both the DAG and the Bidders, as well as among the Participants themselves, the Participants recognized that some provisions of Rule 613 would not permit certain solutions to be included in the Plan that the Participants determined advisable to effectuate the most efficient and cost-effective CAT. Consequently, on January 30, 2015, the Participants submitted to the Commission a request for exemptive relief from certain provisions of Rule 613 regarding: (1) Options Market Maker quotes; (2) Customer-IDs; (3) CAT-Reporter-IDs; (4) linking of executions to specific subaccount allocations on Allocation Reports; and (5) time stamp granularity for manual order events. Specifically, the Participants requested that the Commission grant an exemption from:

See Exemptive Request Letter, supra note 16.

Rule 613(c)(7)(ii) and (iv) for Options Market Makers with regard to their options quotes;

Rule 613(c)(7)(i)(A), (c)(7)(iv)(F), (c)(7)(viii)(B) and (c)(8) which relate to the requirements for Customer-IDs;

Rule 613(c)(7)(i)(C), (c)(7)(ii)(D), (c)(7)(ii)(E), (c)(7)(iii)(D), (c)(7)(iii)(E), (c)(7)(iv)(F), (c)(7)(v)(F), (c)(7)(vi)(B) and (c)(8) which relate to the requirements for CAT-Reporter-IDs;

Rule 613(c)(7)(vi)(A), which requires CAT Reporters to record and report the account number of any subaccounts to which the execution is allocated; and

The millisecond time stamp granularity requirement in Rule 613(d)(3) for certain manual order events subject to time stamp reporting under Rules 613(c)(7)(i)(E), 613(c)(7)(ii)(C), 613(c)(7)(iii)(C), and 613(c)(7)(iv)(C).

The Participants believe that the requested relief is critical to the development of a cost-effective approach to the CAT.

The Commission notes the Participants' request for exemptive relief was granted on March 1, 2016. See Exemption Order, supra note 18.

3. Requirements Pursuant to Rule 608(a)

a. Description of Plan

Rule 613 requires the Participants to “jointly file . . . a national market system plan to govern the creation, implementation, and maintenance of a consolidated audit trail and Central Repository.” The purpose of the Plan, and the creation, implementation and maintenance of a comprehensive audit trail for the U.S. securities market described therein, is to “substantially enhance the ability of the SROs and the Commission to oversee today's securities markets and fulfill their responsibilities under the federal securities laws.” It “will allow for the prompt and accurate recording of material information about all orders in NMS securities, including the identity of customers, as these orders are generated and then routed throughout the U.S. markets until execution, cancellation, or modification. This information will be consolidated and made readily available to regulators in a uniform electronic format.” The SROs note that the following summarizes various provisions of the Plan, which is set forth in full as Exhibit A to this Notice.

See Adopting Release, supra note 9, at 45726.

Id. Note that the Plan also includes certain recording and reporting obligations for OTC Equity Securities.

(1) LLC Agreement

The Participants propose to conduct the activities related to the CAT in a Delaware limited liability company pursuant to a limited liability company agreement, entitled the Limited Liability Company Agreement of CAT NMS, LLC (“Company”). The Participants will jointly own on an equal basis the Company. The Company will create, implement and maintain the CAT. The limited liability company agreement (“LLC Agreement”) itself, including its appendices, is the proposed Plan, which would be a national market system plan as defined in Rule 600(b)(43) of NMS.

(2) Participants

Each national securities exchange and national securities association currently registered with the Commission would be a Participant in the Plan. The names and addresses of each Participant are set forth in Exhibit A to the Plan. Article III of the Plan provides that any entity approved by the Commission as a national securities exchange or national securities association under the Exchange Act after the Effective Date may become a Participant by submitting to the Company a completed application in the form provided by the Company and satisfying each of the following requirements: (1) Executing a counterpart of the LLC Agreement as then in effect; and (2) paying a fee to the Company in an amount determined by a Majority Vote of the Operating Committee as fairly and reasonably compensating the Company and the Participants for costs incurred in creating, implementing and maintaining the CAT (including such costs incurred in evaluating and selecting the Initial Plan Processor and any subsequent Plan Processor) and for costs the Company incurs in providing for the prospective Participant's participation in the Company, including after consideration of certain factors identified in Section 3.3(b) of the Agreement (“Participation Fee”). The amendment of the Plan reflecting the admission of a new Participant will be effective only when: (1) It is approved by the SEC in accordance with Rule 608 or otherwise becomes effective pursuant to Rule 608; and (2) the prospective Participant pays the Participation Fee.

A number of factors are relevant to the determination of a Participation Fee. Such factors include: (1) The portion of costs previously paid by the Company for the development, expansion and maintenance of the CAT which, under GAAP, would have been treated as capital expenditures and would have been amortized over the five years preceding the admission of the prospective Participant; (2) an assessment of costs incurred and to be incurred by the Company for modifying the CAT or any part thereof to accommodate the prospective Participant, which costs are not otherwise required to be paid or reimbursed by the prospective Participant; (3) Participation Fees paid by other Participants admitted as such after the Effective Date; (4) elapsed time from the Effective Date to the anticipated date of admittance of the prospective Participant; and (5) such other factors, if any, as may be determined to be appropriate by the Operating Committee and approved by the Commission. In the event that the Company and a prospective Participant do not agree on the amount of the Participation Fee, such amount will be subject to review by the SEC pursuant to Section 11A(b)(5) of the Exchange Act.

An applicant for participation in the Company may apply for limited access to the CAT System for planning and testing purposes pending its admission as a Participant by submitting to the Company a completed Application for Limited Access to the CAT System in a form provided by the Company, accompanied by payment of a deposit in the amount established by the Company, which will be applied or refunded as described in such application. To be eligible to apply for such limited access, the applicant must have been approved by the SEC as a national securities exchange or national securities association under the Exchange Act but the applicant has not yet become a Participant of the Plan, or the SEC must have published such applicant's Form 1 Application or From [sic] X-15AA-1 Application to become a national securities exchange or a national securities association, respectively.

All Company Interests will have the same rights, powers, preferences and privileges and be subject to the same restrictions, qualifications and limitations. Once admitted, each Participant will be entitled to one vote on any matter presented to Participants for their consideration and to participate equally in any distribution made by the Company (other than a distribution made pursuant to Section 10.2 of the Plan). Each Participant will have a Company Interest equal to that of each other Participant.

Article III also describes a Participant's ability to Transfer a Company Interest. A Participant may only Transfer any Company Interest to a national securities exchange or national securities association that succeeds to the business of such Participant as a result of a merger or consolidation with such Participant or the Transfer of all or substantially all of the assets or equity of such Participant (“Permitted Transferee”). A Participant may not Transfer any Company Interest to a Permitted Transferee unless: (1) Such Permitted Transferee executes a counterpart of the Plan; and (2) the amendment to the Plan reflecting the Transfer is approved by the SEC in accordance with Rule 608 or otherwise becomes effective pursuant to Rule 608.

In addition, Article III addresses the voluntary resignation and termination of participation in the Plan. Any Participant may voluntarily resign from the Company, and thereby withdraw from and terminate its right to any Company Interest, only if: (1) A Permitted Legal Basis for such action exists; and (2) such Participant provides to the Company and each other Participant no less than thirty days prior to the effective date of such action written notice specifying such Permitted Legal Basis, including appropriate documentation evidencing the existence of such Permitted Legal Basis, and, to the extent applicable, evidence reasonably satisfactory to the Company and other Participants that any orders or approvals required from the SEC in connection with such action have been obtained. A validly withdrawing Participant will have the rights and obligations discussed below with regard to termination of participation.

A Participant's participation in the Company, and its right to any Company Interest, will terminate as of the earliest of: (1) The effective date specified in a valid resignation notice; (2) such time as such Participant is no longer registered as a national securities exchange or national securities association; or (3) the date of termination for failure to pay fees. With regard to the payment of fees, each Participant is required to pay all fees or other amounts required to be paid under the Plan within thirty days after receipt of an invoice or other notice indicating payment is due (unless a longer payment period is otherwise indicated) (the “Payment Date”). If a Participant fails to make such a required payment by the Payment Date, any balance in the Participant's Capital Account will be applied to the outstanding balance. If a balance still remains with respect to any such required payment, the Participant will pay interest on the outstanding balance from the Payment Date until such fee or amount is paid at a per annum rate equal to the lesser of: (1) The Prime Rate plus 300 basis points; or (2) the maximum rate permitted by applicable law. If any such remaining outstanding balance is not paid within thirty days after the Payment Date, the Participants will file an amendment to the Plan requesting the termination of the participation in the Company of such Participant, and its right to any Company Interest, with the SEC. Such amendment will be effective only when it is approved by the SEC in accordance with Rule 608 or otherwise becomes effective pursuant to Rule 608.

From and after the effective date of termination of a Participant's participation in the Company, profits and losses of the Company will cease to be allocated to the Capital Account of the Participant. A terminated Participant will be entitled to receive the balance in its Capital Account as of the effective date of termination adjusted for profits and losses through that date, payable within ninety days of the effective date of termination, and will remain liable for its proportionate share of costs and expenses allocated to it for the period during which it was a Participant, for obligations under Section 3.8(c) regarding the return of amounts previously distributed (if required by a court of competent jurisdiction), for its indemnification obligations pursuant to Section 4.1, and for obligations under Section 9.6 regarding confidentiality, but it will have no other obligations under the Plan following the effective date of termination. The Plan will be amended to reflect any termination of participation in the Company of a Participant, provided that such amendment will be effective only when it is approved by the SEC in accordance with Rule 608 or otherwise becomes effective pursuant to Rule 608.

Request for Comment

1. Do Commenters believe that the process for a national securities exchange and national securities association to become a Participant pursuant to and under the CAT NMS Plan is clearly and adequately set forth in the CAT NMS Plan? Do Commenters believe that the process for, and the circumstances under which a Participant could voluntarily terminate its participation as a Participant to the CAT NMS Plan is clearly and adequately set forth in the CAT NMS Plan? If not, what additional details should be provided? Do Commenters believe that these two processes are appropriate and reasonable?

2. Do Commenters believe that the process and enumerated factors for determining the Participation Fee are clear and reasonable under the CAT NMS Plan? If not, what additional modifications, if any, should be considered in the Participation Fee determination process?

3. Are restrictions on the transfer of a Company Interest appropriate and reasonable? If not, why not? What additional limitations or factors, if any, should be imposed on such transfers? Please explain.

4. Do Commenters believe that permitting the termination of a Participant that continues to be a registered national securities exchange or national securities association from participation in the Company is an appropriate recourse for failure to pay Participant fees? If not, can Commenters recommend an alternative remedy? Please explain.

5. Are there other circumstances that should trigger termination of participation in the Company? If yes, what are they?

(3) Management

Article IV of the Plan establishes the overall governance structure for the management of the Company. Specifically, the Participants propose that the Company be managed by an Operating Committee.

The Operating Committee will manage the Company except for situations in which the approval of the Participants is required by the Plan or by non-waivable provisions of applicable law.

The Operating Committee will consist of one voting member representing each Participant and one alternate voting member representing each Participant who will have a right to vote only in the absence of the Participant's voting member of the Operating Committee. Each of the voting and alternate voting members of the Operating Committee will be appointed by the Participant that he or she represents, will serve at the will of the Participant appointing such member and will be subject to the confidentiality obligations of the Participant that he or she represents as set forth in Section 9.6. One individual may serve as the voting member of the Operating Committee for multiple Affiliated Participants, and such individual will have the right to vote on behalf of each such Affiliated Participant.

The Operating Committee will elect, by Majority Vote, one of its members to act as Chair for a term of two years. No Person may serve as Chair for more than two successive full terms, and no Person then appointed to the Operating Committee by a Participant that then serves, or whose Affiliate then serves, as the Plan Processor will be eligible to serve as the Chair. The Chair will preside at all meetings of the Operating Committee, designate a Person to act as Secretary, and perform such other duties and possess such other powers as the Operating Committee may from time to time prescribe. The Chair will not be entitled to a tie-breaking vote at any meeting of the Operating Committee.

Each of the members of the Operating Committee, including the Chair, will be authorized to cast one vote for each Participant that he or she represents on all matters voted upon by the Operating Committee. Action of the Operating Committee will be authorized by Majority Vote (except under certain designated circumstances), subject to the approval of the SEC whenever such approval is required under the Exchange Act and the rules thereunder. For example, the Plan specifically notes that a Majority Vote of the Operating Committee is required to: (1) Select the Chair; (2) select the members of the Advisory Committee (as described below); (3) interpret the Plan (unless otherwise noted therein); (4) approve any recommendation by the Chief Compliance Officer pursuant to Section 6.2(a)(v)(A); (5) determine to hold an Executive Session of the Operating Committee; (6) determine the appropriate funding-related policies, procedures and practices consistent with Article XI; and (7) any other matter specified elsewhere in the Plan (which includes the Appendices to the Plan) as requiring a vote, approval or other action of the Operating Committee (other than those matters expressly requiring a Supermajority Vote or a different vote of the Operating Committee).

Article IV requires a Supermajority Vote of the Operating Committee, subject to the approval of the SEC when required, for the following: (1) Selecting a Plan Processor, other than the Initial Plan Processor selected in accordance with Article V of the Plan; (2) terminating the Plan Processor without cause in accordance with Section 6.1(p); (3) approving the Plan Processor's appointment or removal of the Chief Information Security Officer, Chief Compliance Officer, or any Independent Auditor in accordance with Section 6.1(b); (4) entering into, modifying or terminating any Material Contract (if the Material Contract is with a Participant or an Affiliate of a Participant, such Participant and Affiliated Participant will be recused from any vote); (5) making any Material Systems Change; (6) approving the initial Technical Specifications or any Material Amendment to the Technical Specifications proposed by the Plan Processor; (7) amending the Technical Specifications on its own motion; and (8) any other matter specified elsewhere in the Plan (which includes the Appendices to the Plan) as requiring a vote, approval or other action of the Operating Committee by a Supermajority Vote.

A member of the Operating Committee or any Subcommittee thereof (as discussed below) shall recuse himself or herself from voting on any matter under consideration by the Operating Committee or such Subcommittee if such member determines that voting on such matter raises a Conflict of Interest. In addition, if the members of the Operating Committee or any Subcommittee (excluding the member thereof proposed to be recused) determine by Supermajority Vote that any member voting on a matter under consideration by the Operating Committee or such Subcommittee raises a Conflict of Interest, such member shall be recused from voting on such matter. No member of the Operating Committee or any Subcommittee will be automatically recused from voting on any matter except matters involving Material Contracts as discussed in the prior paragraph, as otherwise specified in the Plan, and as follows: (1) If a Participant is a Bidding Participant whose Bid remains under consideration, members appointed to the Operating Committee or any Subcommittee by such Participant or any of its Affiliated Participants will be recused from any vote concerning: (a) Whether another Bidder may revise its Bid; (b) the selection of a Bidder; or (c) any contract to which such Participant or any of its Affiliates would be a party in its capacity as Plan Processor; and (2) if a Participant is then serving as Plan Processor, is an Affiliate of the Person then serving as Plan Processor, or is an Affiliate of an entity that is a Material Subcontractor to the Plan Processor, then in each case members appointed to the Operating Committee or any Subcommittee by such Participant or any of its Affiliated Participants shall be recused from any vote concerning: (a) The proposed removal of such Plan Processor; or (b) any contract between the Company and such Plan Processor.

Article IV also addresses meetings of the Operating Committee. Meetings of the Operating Committee may be attended by each Participant's voting Representative and its alternate voting Representative and by a maximum of two nonvoting Representatives of each Participant, by members of the Advisory Committee, by the Chief Compliance Officer, by other Representatives of the Company and the Plan Processor, by Representatives of the SEC and by such other Persons that the Operating Committee may invite to attend. The Operating Committee, however, may, where appropriate, determine to meet in Executive Session during which only voting members of the Operating Committee will be present. The Operating Committee, however, may invite other Representatives of the Participants, of the Company, of the Plan Processor (including the Chief Compliance Officer and the Chief Information Security Officer) or the SEC, or such other Persons that the Operating Committee may invite to attend, to be present during an Executive Session. Any determination of the Operating Committee to meet in an Executive Session will be made upon a Majority Vote and will be reflected in the minutes of the meeting. In addition, any Person that is not a Participant but for which the SEC has published a Form 1 Application or Form X-15AA-1 to become a national securities exchange or national securities association, respectively, will be permitted to appoint one primary Representative and one alternate Representative to attend regularly scheduled Operating Committee meetings in the capacity of a non-voting observer, but will not be permitted to have any Representative attend a special meeting, emergency meeting or meeting held in Executive Session of the Operating Committee.

Article IV also addresses, among other things, different types of Operating Committee meetings (regular, special and emergency), frequency of such meetings, how to call such meetings, the location of the meetings, the role of the Chair, and notice regarding such meetings.

The Operating Committee may, by Majority Vote, designate by resolution one or more Subcommittees it deems necessary or desirable in furtherance of the management of the business and affairs of the Company. For any Subcommittee, any member of the Operating Committee who wants to serve thereon may so serve. If Affiliated Participants have collectively appointed one member to the Operating Committee to represent them, then such Affiliated Participants may have only that member serve on the Subcommittee or may decide not to have only that collectively appointed member serve on the Subcommittee. Such member may designate an individual other than himself or herself who is also an employee of the Participant or Affiliated Participants that appointed such member to serve on a Subcommittee in lieu of the particular member. Subject to the requirements of the Plan and non-waivable provisions of Delaware law, a Subcommittee may exercise all the powers and authority of the Operating Committee in the management of the business and affairs of the Company as so specified in the resolution of the Operating Committee designating such Subcommittee.

Article IV requires that the Operating Committee maintain a Compliance Subcommittee for the purpose of aiding the Chief Compliance Officer as necessary, including with respect to issues involving: (1) The maintenance of the confidentiality of information submitted to the Plan Processor or Central Repository pursuant to Rule 613, applicable law, or the Plan by Participants and Industry Members; (2) the timeliness, accuracy, and completeness of information submitted pursuant to Rule 613, applicable law or the Plan by Participants and Industry Members; and (3) the manner and extent to which each Participant is meeting its obligations under Rule 613, Section 3.11, and as set forth elsewhere in the Plan and ensuring the consistency of the Plan's enforcement as to all Participants.

Article IV also sets forth the requirements for the formation and functioning of an Advisory Committee, which will advise the Participants on the implementation, operation and administration of the Central Repository, including possible expansion of the Central Repository to other securities and other types of transactions.

Article IV describes the composition of the Advisory Committee. No member of the Advisory Committee may be employed by or affiliated with any Participant or any of its Affiliates or facilities. The Operating Committee will select one member from representatives of each of the following categories to serve on the Advisory Committee on behalf of himself or herself individually and not on behalf of the entity for which the individual is then currently employed: (1) A broker-dealer with no more than 150 Registered Persons; (2) a broker-dealer with at least 151 and no more than 499 Registered Persons; (3) a broker-dealer with 500 or more Registered Persons; (4) a broker-dealer with a substantial wholesale customer base; (5) a broker-dealer that is approved by a national securities exchange: (a) To effect transactions on an exchange as a specialist, market maker or floor broker; or (b) to act as an institutional broker on an exchange; (6) a proprietary-trading broker-dealer; (7) a clearing firm; (8) an individual who maintains a securities account with a registered broker or dealer but who otherwise has no material business relationship with a broker or dealer or with a Participant; (9) a member of academia with expertise in the securities industry or any other industry relevant to the operation of the CAT System; (10) an institutional investor trading on behalf of a public entity or entities; (11) an institutional investor trading on behalf of a private entity or entities; and (12) an individual with significant and reputable regulatory expertise. The members selected to represent categories (1) through (12) above must include, in the aggregate, representatives of no fewer than three broker-dealers that are active in the options business and representatives of no fewer than three broker-dealers that are active in the equities business. In addition, upon a change in employment of any such selected member, a Majority Vote of the Operating Committee will be required for such member to be eligible to continue to serve on the Advisory Committee. Furthermore, the SEC's Chief Technology Officer (or the individual then currently employed in a comparable position providing equivalent services) will serve as an observer of the Advisory Committee (but not be a member). The members of the Advisory Committee will have a term of three years.

Four of the initial twelve members of the Advisory Committee will have an initial term of one year, and another four of the initial twelve members of the Advisory Committee will have an initial term of two years.

Members of the Advisory Committee will have the right to attend meetings of the Operating Committee or any Subcommittee, to receive information concerning the operation of the Central Repository, and to submit their views to the Operating Committee or any Subcommittee on matters pursuant to the Plan prior to a decision by the Operating Committee on such matters. A member of the Advisory Committee will not have a right to vote on any matter considered by the Operating Committee or any Subcommittee. In addition, the Operating Committee or any Subcommittee may meet in Executive Session if the Operating Committee or Subcommittee determines by Majority Vote that such an Executive Session is advisable. Although members of the Advisory Committee will have the right to receive information concerning the operation of the Central Repository, the Operating Committee retains the authority to determine the scope and content of information supplied to the Advisory Committee, which will be limited to that information that is necessary and appropriate for the Advisory Committee to fulfill its functions. Any information received by members of the Advisory Committee will remain confidential unless otherwise specified by the Operating Committee.

The Operating Committee may solicit and consider views on the operation of the Central Repository in addition to those of the Advisory Committee.

Article IV also describes the appointment of Officers for the Company. Specifically, the Chief Compliance Officer and the Chief Information Security Officer, each of whom will be employed solely by the Plan Processor and neither of whom will be deemed or construed in any way to be an employee of the Company, will be Officers of the Company. Neither such Officer will receive or be entitled to any compensation from the Company or any Participant by virtue of his or her service in such capacity (other than if a Participant is then serving as the Plan Processor, compensation paid to such Officer as an employee of such Participant). Each such Officer will report directly to the Operating Committee. The Chief Compliance Officer will work on a regular and frequent basis with the Compliance Subcommittee and/or other Subcommittees as may be determined by the Operating Committee. Except to the extent otherwise provided in the Plan, including Section 6.2, each such Officer will have such fiduciary and other duties with regard to the Plan Processor as imposed by the Plan Processor on such individual by virtue of his or her employment by the Plan Processor.

In addition, the Plan Processor will inform the Operating Committee of the individual who has direct management responsibility for the Plan Processor's performance of its obligations with respect to the CAT. Subject to approval by the Operating Committee of such individual, the Operating Committee will appoint such individual as an Officer. In addition, the Operating Committee by Supermajority Vote may appoint other Officers as it shall from time to time deem necessary. Any Officer appointed pursuant to Section 4.6(b) will have only such duties and responsibilities as set forth in the Plan, or as the Operating Committee shall from time to time expressly determine. No such Officer shall have any authority to bind the Company (which authority is vested solely in the Operating Committee) or be an employee of the Company, unless in each case the Operating Committee, by Supermajority Vote, expressly determines otherwise. No person subject to a “statutory disqualification” (as defined in Section 3(a)(39) of the Exchange Act) may serve as an Officer. It is the intent of the Participants that the Company have no employees.

Request for Comment

6. Do Commenters believe that the organizational, governance and/or managerial structure of CAT NMS, LLC is in the public interest? Why or why not?

7. Do Commenters believe that the organizational, governance, and/or managerial structure set forth in the CAT NMS Plan, including the role of the Operating Committee, is appropriate and reasonable? If not, please explain.

8. The CAT NMS Plan specifies the corporate actions that require a Majority Vote and the corporate actions that require a Supermajority Vote. Do Commenters believe that such voting procedures are appropriate and reasonable? Should any corporate actions require a higher or lower voting threshold than specified in the Plan? Are there any corporate actions that should require a Supermajority Vote? Please explain.

9. Do Commenters believe that the CAT NMS Plan should explicitly or more clearly specify who should determine whether a systems change or amendment is “material”? If so, who? Please explain.

10. Do Commenters believe that two successive full terms is an appropriate and reasonable term limit for a Person to serve as chair of the Operating Committee? If not, please explain.

11. Section 1.1 defines Conflict of Interest to mean that the interest of a Participant (e.g., commercial, reputational, regulatory, or otherwise) in the matter that is subject to the vote; (a) interferes, or would be reasonably likely to interfere with that Participant's objective consideration of the matter; and (b) is, or is reasonably likely to be, inconsistent with the purpose and objectives of the Company, and the CAT, taking into account all relevant considerations, including whether a Participant that may otherwise have a conflict of interest has established appropriate safeguards to eliminate such conflicts of interest and taking into account the other guiding principles set forth in the LLC Agreement. Do Commenters believe this definition of “Conflict of Interest” is appropriate and reasonable? Please explain.

12. Do Commenters believe that the definition of Conflict of Interest of the CAT NMS Plan properly reflects the business interests of each Participant and the Operating Committee? If not, please explain. Do Commenters believe that the CAT NMS Plan governing procedures on Conflicts of Interest and recusals contained in Section 4.3(d) of the CAT NMS Plan, reasonably and adequately address Conflicts of Interest? If not, please explain. Are there other conflicts of interest that may arise for any Participant that are not addressed in the CAT NMS Plan definitions or governing procedures? If so, what?

13. Is the CAT NMS Plan clear and reasonable regarding whether it permits the Operating Committee to delegate the authority to vote on matters to a Subcommittee? If so, in what circumstances? Are there any circumstances in which a Subcommittee would or should be prohibited from voting in place of the Operating Committee? Please explain.

14. Do Commenters believe that the Advisory Committee structure and provisions set forth in the CAT NMS Plan are appropriate and reasonable? Is the size of the Advisory Committee as contemplated by the Plan appropriate and reasonable? Are the Advisory Committee member categories reasonable and adequately representative of entities impacted by the CAT NMS Plan? Would expanding membership on the Advisory Committee to any additional types of entities enhance the quality of the input it would provide to the Operating Committee? Please explain.

15. Is the mechanism for determining who serves on the Advisory Committee (i.e., selection by the Operating Committee) appropriate and reasonable? Should Participants be required to publicly solicit Advisory Committee membership interest? Should the Advisory Committee be able to self-nominate replacement candidates? Please explain.

16. Do Commenters believe that the CAT NMS Plan's requirement that Advisory Committee members serve on the Advisory Committee in their personal capacities, and that the Operating Committee members serve on the Operating Committee as representatives of their employers who are the Plan Participants create different incentives for members of the Advisory Committee and members of the Operating Committee? If so, in what ways? Do Commenters believe that these differing incentives would impact the regulatory objective of the CAT? If so, in what ways?

17. The CAT NMS Plan outlines the size, tenure and membership categories of the Advisory Committee members. Do Commenters believe there are any additional or alternative factors that should be taken into consideration in structuring the Advisory Committee that would benefit the operation of the CAT? If so, what are those additional or alternative factors? How would these factors benefit the operation of the CAT?

18. Are the roles and responsibilities of the Advisory Committee clearly and adequately set forth in the CAT NMS Plan? If not, why not? Should additional details on these roles and responsibilities be provided? If so, what additional details should be provided?

19. Are there any alternatives for involvement by the Advisory Committee that could increase the effectiveness of the Advisory Committee? For example, should the Advisory Committee be given a vote in connection with decisions regarding the CAT NMS Plan, equivalent to the vote each Participant has? If so, please specifically identify the alternatives for involvement and how those alternatives could increase the effectiveness of the CAT.

20. Do Commenters believe that the Advisory Committee is structured in a way that would allow industry to provide meaningful input on the implementation, operation, and administration of the CAT? If not, please explain and/or provide specific suggestions for improving the Advisory Committee structure. Should additional authority be given to the Advisory Committee, for example allowing it to initiate its own recommendations? Should additional mechanisms through which the industry or others could provide input be included in the CAT NMS Plan? Should the Operating Committee be required to respond to the Advisory Committee's views, formally or informally, in advance of or following a decision by the Operating Committee? Should the Operating Committee be required to include Advisory Committee views in filings with the Commission? Please explain.

See Section IV.E.4, infra, for additional requests for comment on the Advisory Committee.

21. Do Commenters believe that the Plan's provision that prohibits the Advisory Committee from attending any Executive Session of the Operating Committee is appropriate and reasonable?

22. Do Commenters believe that the CAT NMS Plan adequately sets forth provisions regarding the scope, authority, and duties of the Officers of the CAT, as well as the scope and authority of the Plan Processor generally? If not, what further provisions should the CAT NMS Plan set forth with respect to Officers and the Plan Processor and why?

23. Do Commenters believe that the Operating Committee and the proposed CAT NMS Plan governance structure would ensure effective corporate governance, process and action? Why or why not?

24. The CAT NMS Plan provides that emergency meetings of the Operating Committee may be called at the request of two or more Participants, and may be held as soon as practical after such a meeting is called. Do Commenters believe that there should be a different method for the Operating Committee to meet and take action in the event of an emergency? Should the CAT NMS Plan denote certain emergency situations in which the Operating Committee must be required to take action on an expedited basis? If so, what time period would be reasonable to require action by the Operating Committee and what mechanisms or processes should the Operating Committee be required to follow?

25. What, if any, impact on the Operating Committee's governance and voting do Affiliated Participant groups have? Do Commenters believe that the Operating Committee's governance and voting provisions set forth in the CAT NMS Plan, including the definitions of Supermajority Vote and Majority Vote, are appropriate and reasonable in light of these Affiliated Participant groups? What, if any, additional governance and voting provisions or protections should be included? Is there an alternative model for voting rights that would be more appropriate and reasonable, for example distributing votes using a measure other than exchange licenses?

26. Do Commenters believe the use of Executive Session is appropriate and reasonable? Is a Majority Vote the appropriate mechanism for the Operating Committee to go into Executive Session? Should the CAT NMS Plan specify particular scenarios for which an Executive Session is or is not appropriate?

27. Do Commenters believe that the provisions in the CAT NMS Plan regarding the mechanics of voting by the Operating Committee, the Selection Committee, or other entities are appropriate and reasonable? Does the CAT NMS Plan include sufficient detail on when voting should be carried out openly (e.g., in the presence of other attendees at a committee meeting) as opposed to when voting may be conducted by secret ballot or by some other confidential method? What are the advantages and disadvantages of different voting methodologies? Would particular actions or decisions regarding CAT be better suited to one voting methodology over others? Please explain.

28. Are there any other matters relating to the operation and administration of the Plan that should be included in the Plan for the Commission's consideration? If so, please identify such matters and explain why and how they should be addressed in the Plan.

(4) Initial Plan Processor Selection

Article V of the Plan sets forth the process for the Participants' evaluation of Bids and the selection process for narrowing down the Bids and choosing the Initial Plan Processor. The initial steps in the evaluation and selection process were and will be performed pursuant to the Selection Plan; the final two rounds of evaluation and voting, as well as the final selection of the Initial Plan Processor, will be performed pursuant to the Plan.

By its terms, the Selection Plan will terminate upon Commission approval of the Plan.

As discussed above, the Selection Committee has selected the Shortlisted Bids pursuant to the Selection Plan. After reviewing the Shortlisted Bids, the Participants have identified the optimal proposed solutions for the CAT and, to the extent possible, included such solutions in the Plan. The Selection Committee will determine, by majority vote, whether Shortlisted Bidders will have the opportunity to revise their Bids. To reduce potential conflicts of interest, no Bidding Participant may vote on whether a Shortlisted Bidder will be permitted to revise its Bid if a Bid submitted by or including the Participant or an Affiliate of the Participant is a Shortlisted Bid. The Selection Committee will review and evaluate all Shortlisted Bids, including any permitted revisions submitted by Shortlisted Bidders. In performing this review and evaluation, the Selection Committee may consult with the Advisory Committee and such other Persons as the Selection Committee deems appropriate, which may include the DAG until the Advisory Committee is formed.

As noted above, the Participants stated their belief that certain exemptive relief is necessary to include in the Plan all of the provisions the Participants believe are part of the optimal solution for the CAT. The Commission notes that the request for exemptive relief was granted on March 1, 2016. See Exemption Order, supra note 18.

After receipt of any permitted revisions, the Selection Committee will select the Initial Plan Processor from the Shortlisted Bids in two rounds of voting where each Participant has one vote via its Voting Senior Officer in each round. No Bidding Participant, however, will be entitled to vote in any round if the Participant's Bid, a Bid submitted by an Affiliate of the Participant, or a Bid including the Participant or an Affiliate of the Participant is considered in such round. In the first round, each Voting Senior Officer, subject to the recusal provision in Section 5.2(e)(ii), will select a first and second choice, with the first choice receiving two points and the second choice receiving one point. The two Shortlisted Bids receiving the highest cumulative scores in the first round will advance to the second round. In the event of a tie, the tie will be broken by assigning one point per vote to the tied Shortlisted Bids, and the Shortlisted Bid with the most votes will advance. If this procedure fails to break the tie, a revote will be taken on the tied Bids with each vote receiving one point. If the tie persists, the Participants will identify areas for discussion, and revotes will be taken until the tie is broken.

If the proposed amendment to the Selection Plan is approved, the Selection Committee may determine to narrow the number of Shortlisted Bids prior to the two rounds of voting.

This recusal provision is included in the Plan, as well as in an amendment to the Selection Plan. See Order Approving Amendment No. 2 to the Selection Plan, supra note 15.

Each round of voting throughout the Plan is independent of other rounds.

Once two Shortlisted Bids have been chosen, the Voting Senior Officers of the Participants (other than those subject to recusal) will vote for a single Shortlisted Bid from the final two to determine the Initial Plan Processor. If the tie persists, the Participants will identify areas for discussion and, following these discussions, revotes will be taken until the tie is broken. As set forth in Article VI of the Plan, following the selection of the Initial Plan Processor, the Participants will file with the Commission a statement identifying the Initial Plan Processor and including the information required by Rule 608.

(5) Functions and Activities of CAT System

A. Plan Processor

Article VI describes the responsibilities of the selected Plan Processor. The Company, under the direction of the Operating Committee, will enter into one or more agreements with the Plan Processor obligating the Plan Processor to perform the functions and duties contemplated by the Plan to be performed by the Plan Processor, as well as such other functions and duties the Operating Committee deems necessary or appropriate.

As set forth in the Plan, the Plan Processor is required to develop and, with the prior approval of the Operating Committee, implement policies, procedures, and control structures related to the CAT System that are consistent with Rule 613(e)(4), Appendix C and Appendix D. The Plan Processor will: (1) Comply with applicable provisions of 15 U.S. Code § 78u-6 (Securities Whistleblower Incentives and Protection) and the recordkeeping requirements of Rule 613(e)(8); (2) consistent with Appendix D, Central Repository Requirements, ensure the effective management and operation of the Central Repository; (3) consistent with Appendix D, Data Management, ensure the accuracy of the consolidation of the CAT Data reported to the Central Repository; and (4) consistent with Appendix D, Upgrade Process and Development of New Functionality, design and implement appropriate policies and procedures governing the determination to develop new functionality for the CAT including, among other requirements, a mechanism by which changes can be suggested by Advisory Committee members, Participants, or the SEC. Such policies and procedures also shall: (1) Provide for the escalation of reviews of proposed technological changes and upgrades to the Operating Committee; and (2) address the handling of surveillance, including coordinated, Rule 17d-2 under the Exchange Act or Regulatory Surveillance Agreement(s) (RSA) surveillance queries and requests for data. Any policy, procedure or standard (and any material modification or amendment thereto) applicable primarily to the performance of the Plan Processor's duties as the Plan Processor (excluding any policies, procedures or standards generally applicable to the Plan Processor's operations and employees) will become effective only upon approval by the Operating Committee. The Plan Processor also will, subject to the prior approval of the Operating Committee, establish appropriate procedures for escalation of matters to the Operating Committee. In addition to other policies, procedures and standards generally applicable to the Plan Processor's employees and contractors, the Plan Processor will have hiring standards and will conduct and enforce background checks (e.g., fingerprint-based) for all of its employees and contractors to ensure the protection, safeguarding and security of the facilities, systems, networks, equipment and data of the CAT System, and will have an insider and external threat policy to detect, monitor and remedy cyber and other threats.

The Plan Processor will enter into appropriate Service Level Agreements (“SLAs”) governing the performance of the Central Repository, as generally described in Appendix D, Functionality of the CAT System, with the prior approval of the Operating Committee. The Plan Processor in conjunction with the Operating Committee will regularly review and, as necessary, update the SLAs, in accordance with the terms of the SLAs. As further contemplated in Appendix C, System Service Level Agreements (SLAs), and in Appendix D, System SLAs, the Plan Processor may enter into appropriate service level agreements with third parties applicable to the Plan Processor's functions related to the CAT System (“Other SLAs”), with the prior approval of the Operating Committee. The Chief Compliance Officer and/or the Independent Auditor will, in conjunction with the Plan Processor and as necessary the Operating Committee, regularly review and, as necessary, update the Other SLAs, in accordance with the terms of the applicable Other SLA. In addition, the Plan Processor: (1) Will, on an ongoing basis and consistent with any applicable policies and procedures, evaluate and implement potential system changes and upgrades to maintain and improve the normal day-to-day operating function of the CAT System; (2) in consultation with the Operating Committee, will, on an as needed basis and consistent with any applicable operational and escalation policies and procedures, implement such material system changes and upgrades as may be required to ensure effective functioning of the CAT System; and (3) in consultation with the Operating Committee, will, on an as needed basis, implement system changes and upgrades to the CAT System to ensure compliance with applicable laws, regulations or rules (including those promulgated by the SEC or any Participant). Furthermore, the Plan Processor will develop and, with the prior approval of the Operating Committee, implement a securities trading policy, as well as necessary procedures, control structures and tools to enforce this policy.

In addition, the Plan Processor will provide the Operating Committee regular reports on the CAT System's operation and maintenance. Furthermore, upon request of the Operating Committee or any Subcommittee, the Plan Processor will attend any meetings of the Operating Committee or such Subcommittee.

The Plan Processor may appoint such officers of the Plan Processor as it deems necessary and appropriate to perform its functions under the Plan and Rule 613. The Plan Processor, however, will be required to appoint, at a minimum, the Chief Compliance Officer, the Chief Information Security Officer, and the Independent Auditor. The Operating Committee, by Supermajority Vote, will approve any appointment or removal of the Chief Compliance Officer, Chief Information Security Officer, or the Independent Auditor.

The Plan Processor will designate an employee of the Plan Processor to serve, subject to the approval of the Operating Committee by Supermajority Vote, as the Chief Compliance Officer. The Plan Processor will also designate at least one other employee (in addition to the person then serving as Chief Compliance Officer), which employee the Operating Committee has previously approved, to serve temporarily as the Chief Compliance Officer if the employee then serving as the Chief Compliance Officer becomes unavailable or unable to serve in such capacity (including by reason of injury or illness). Any person designated to serve as the Chief Compliance Officer (including to serve temporarily) will be appropriately qualified to serve in such capacity based on the duties and responsibilities assigned to the Chief Compliance Officer and will dedicate such person's entire working time to such service (or temporary service) (except for any time required to attend to any incidental administrative matters related to such person's employment with the Plan Processor that do not detract in any material respect from such person's service as the Chief Compliance Officer). Article VI sets forth various responsibilities of the Chief Compliance Officer. With respect to all of his or her duties and responsibilities in such capacity (including those as set forth in the Plan), the Chief Compliance Officer will be directly responsible and will directly report to the Operating Committee, notwithstanding that she or he is employed by the Plan Processor. The Plan Processor, subject to the oversight of the Operating Committee, will ensure that the Chief Compliance Officer has appropriate resources to fulfill his or her obligations under the Plan and Rule 613. The compensation (including base salary and bonus) of the Chief Compliance Officer will be payable by the Plan Processor, but be subject to review and approval by the Operating Committee. The Operating Committee will render the Chief Compliance Officer's annual performance review.

The Plan Processor also will designate an employee of the Plan Processor to serve, subject to the approval of the Operating Committee by Supermajority Vote, as the Chief Information Security Officer. The Plan Processor will also designate at least one other employee (in addition to the person then serving as Chief Information Security Officer), which employee the Operating Committee has previously approved, to serve temporarily as the Chief Information Security Officer if the employee then serving as the Chief Information Security Officer becomes unavailable or unable to serve in such capacity (including by reason of injury or illness). Any person designated to serve as the Chief Information Security Officer (including to serve temporarily) will be appropriately qualified to serve in such capacity based on the duties and responsibilities assigned to the Chief Information Security Officer under the Plan and will dedicate such person's entire working time to such service (or temporary service) (except for any time required to attend to any incidental administrative matters related to such person's employment with the Plan Processor that do not detract in any material respect from such person's service as the Chief Information Security Officer).

The Plan Processor, subject to the oversight of the Operating Committee, will ensure that the Chief Information Security Officer has appropriate resources to fulfill the obligations of the Chief Information Security Officer set forth in Rule 613 and in the Plan, including providing appropriate responses to questions posed by the Participants and the SEC. In performing such obligations, the Chief Information Security Officer will be directly responsible and directly report to the Operating Committee, notwithstanding that he or she is employed by the Plan Processor. The compensation (including base salary and bonus) of the Chief Information Security Officer will be payable by the Plan Processor, but be subject to review and approval by the Operating Committee, and the Operating Committee will render the Chief Information Security Officer's annual performance review. Consistent with Appendices C and D, the Chief Information Security Officer will be responsible for creating and enforcing appropriate policies, procedures, standards, control structures and real time tools to monitor and address data security issues for the Plan Processor and the Central Repository, as described in the Plan. At regular intervals, to the extent that such information is available to the Company, the Chief Information Security Officer will report to the Operating Committee the activities of the Financial Services Information Sharing and Analysis Center (“FS-ISAC”) or comparable bodies to the extent that the Company has joined FS-ISAC or other comparable body.

The Plan Processor will afford to Participants and the Commission such access to the Representatives of the Plan Processor as any Participant or the Commission may reasonably request solely for the purpose of performing such Person's regulatory and oversight responsibilities pursuant to the federal securities laws, rules, and regulations or any contractual obligations. The Plan Processor will direct such Representatives to reasonably cooperate with any inquiry, investigation, or proceeding conducted by or on behalf of any Participant or the Commission related to such purpose.

The Operating Committee will review the Plan Processor's performance under the Plan at least once each year, or more often than once each year upon the request of two Participants that are not Affiliated Participants. The Operating Committee will notify the SEC of any determination made by the Operating Committee concerning the continuing engagement of the Plan Processor as a result of the Operating Committee's review of the Plan Processor and will provide the SEC with a copy of any reports that may be prepared in connection therewith.

The Operating Committee, by Supermajority Vote, may remove the Plan Processor from such position at any time. However, the Operating Committee, by Majority Vote, may remove the Plan Processor from such position at any time if it determines that the Plan Processor has failed to perform its functions in a reasonably acceptable manner in accordance with the provisions of the Plan or that the Plan Processor's expenses have become excessive and are not justified. In making such a determination, the Operating Committee will consider, among other factors: (1) The reasonableness of the Plan Processor's response to requests from Participants or the Company for technological changes or enhancements; (2) results of any assessments performed pursuant to Section 6.6; (3) the timeliness of conducting preventative and corrective information technology system maintenance for reliable and secure operations; (4) compliance with requirements of Appendix D; and (5) such other factors related to experience, technological capability, quality and reliability of service, costs, back-up facilities, failure to meet service level agreement(s) and regulatory considerations as the Operating Committee may determine to be appropriate.

In addition, the Plan Processor may resign upon two year's (or such other shorter period as may be determined by the Operating Committee by Supermajority Vote) prior written notice. The Operating Committee will fill any vacancy in the Plan Processor position by Supermajority Vote, and will establish a Plan Processor Selection Subcommittee to evaluate and review Bids and make a recommendation to the Operating Committee with respect to the selection of the successor Plan Processor.

Request for Comment

29. The CAT NMS Plan, Section 6.1 (Plan Processor) sets forth details regarding the Plan Processor's responsibilities. Do Commenters believe that the enumerated responsibilities of the Plan Processor are appropriate and reasonable? Please explain.

30. Do Commenters believe that the CAT NMS Plan provides the Operating Committee with sufficient authority to maintain oversight of the Plan Processor? Is the Plan Processor given too much discretion? Too little? Please explain.

31. The CAT NMS Plan provides in Section 6.1(s) that a Plan Processor may resign upon giving two years notice of such resignation. Do Commenters believe that two years is a sufficient amount of notice to ensure a replacement Plan Processor could be selected? Is two years too long a period to require notice of resignation? Why or why not?

32. The CAT NMS Plan includes two provisions governing removal of the Plan Processor. Section 6.1(q) allows the Operating Committee to remove the Plan Processor at any time by a Supermajority Vote. Do Commenters believe it is appropriate for the Operating Committee to have authority to remove the Plan Processor without cause upon a Supermajority Vote? Why or why not?

33. Section 6.1(r) of the CAT NMS Plan allows the Operating Committee to remove the Plan Processor by a Majority Vote if it determines that the Plan Processor has failed to perform its functions in a reasonably acceptable manner in accordance with the provisions of the CAT LLC Agreement or that the Plan Processor's expenses have become excessive and are not justified. Do Commenters believe it is appropriate and reasonable for the Operating Committee to have the authority to remove the Plan Processor on these bases using a Majority Vote? Why or why not, and with respect to which of these bases? Do Commenters believe there are other grounds upon which the Operating Committee should have the ability to remove the Plan Processor upon a Majority Vote?

34. The CAT NMS Plan states that the Plan Processor must implement policies and procedures consistent with Rule 613(e)(4). Further, Rule 613(e)(4) requires that the CAT NMS Plan include policies and procedures to be used by the Plan Processor to ensure: (1) The security and confidentiality of all information reported to the Central Repository; (2) the timeliness, accuracy, integrity, and completeness of the data provided to the Central Repository; and (3) the accuracy of the consolidation by the Plan Processor of the data provided to the Central Repository. Do Commenters believe that such policies and procedures are adequately described in Appendix D of the CAT NMS Plan? Do Commenters believe such policies and procedures are appropriate and reasonable? Do Commenters believe that additions or deletions should be made to the policies and procedures? If so, please describe.

35. The CAT NMS Plan provides that the CCO and CISO, while Officers of CAT NMS, LLC, would be employees of the Plan Processor. Do Commenters believe that this arrangement creates any conflicts of interest that could undermine the ability of the CCO and CISO to effectively carry out their responsibilities under the CAT NMS Plan? Please describe any such conflicts of interest and explain how they could affect the performance of the CCO or CISO's CAT-related duties.

36. The CAT NMS Plan provides that the Operating Committee must approve the CCO and CISO selected by the Plan Processor by Supermajority Vote, that the CCO and CISO shall dedicate their entire working time to their service as CCO or CISO, that the Operating Committee shall have oversight over the Plan Processor's compensation of and provision of resources to the CCO and CISO, and that the CCO and CISO shall report directly to and receive annual performance reviews from the Operating Committee. Do Commenters believe that these provisions adequately address any conflicts of interest resulting from the CCO and CISO being employees of the Plan Processor? Are there additional steps that could be taken to insulate the CCO and CISO from being unduly influenced by the Plan Processor?

See CAT NMS Plan, supra note 3, at Sections 6.2(a)(i)-(iv), b(i)-(iv).

37. The CAT NMS Plan provides that the CCO and CISO would not, to the extent permitted under applicable law, have fiduciary or similar duties to CAT NMS, LLC, but that they may have fiduciary or similar duties to the Plan Processor to the extent that their employment with the Plan Processor entails such duties. Do Commenters believe that these provisions could affect the ability of the CCO and CISO to carry out their CAT-related duties? Would any alternative provisions be preferable? For example, should the Plan remain silent regarding the CCO and CISO's fiduciary or other duties to the Plan Processor and CAT NMS, LLC? Should the Plan require the CCO and CISO to affirmatively undertake fiduciary or similar duties to CAT NMS, LLC? Should the Plan Processor be required to select individuals who do not have fiduciary or similar duties to the Plan Processor to be the CCO or CISO? What are the advantages and disadvantages to each approach?

See id. at Section 4.6(a), 4.7(c).

38. Is the mechanism by which changes to CAT functionality can be suggested to the Plan Processor by the Advisory Committee members, Participants, or the SEC appropriate and reasonable? Why or why not?

39. Is the Operating Committee's role in the hiring of the CCO, CISO, and Independent Auditor appropriate and reasonable? Should the Advisory Committee be consulted on these decisions? Why or why not?

B. Central Repository

The Central Repository, under the oversight of the Plan Processor, and consistent with Appendix D, Central Repository Requirements, will receive, consolidate, and retain all CAT Data. The Central Repository will collect (from a SIP or pursuant to an NMS Plan) and retain on a current and continuing basis, in a format compatible with the Participant Data and Industry Member Data, all data, including the following: (1) Information, including the size and quote condition, on quotes, including the National Best Bid and National Best Offer for each NMS Security; (2) Last Sale Reports and transaction reports reported pursuant to an effective transaction reporting plan filed with the SEC pursuant to, and meeting the requirements of, Rules 601 and 608; (3) trading halts, LULD price bands and LULD indicators; and (4) summary data.

In the CAT NMS Plan as attached hereto as Exhibit A, Section 6.5(a)(ii)(D) was amended to clarify that “summary data” refers to “summary data or reports described in the specifications for each of the SIPs and disseminated by the respective SIP.”

Consistent with Appendix D, Data Retention Requirements, the Central Repository will retain the information collected pursuant to paragraphs (c)(7) and (e)(7) of Rule 613 in a convenient and usable standard electronic data format that is directly available and searchable electronically without any manual intervention by the Plan Processor for a period of not less than six years. Such data when available to the Participant regulatory Staff and the SEC will be linked. In addition, the Plan Processor will implement and comply with the records retention policy contemplated by Section 6.1(d)(i).

Consistent with Appendix D, Data Access, the Plan Processor will provide Participants and the SEC access to the Central Repository (including all systems operated by the Central Repository), and access to and use of the CAT Data stored in the Central Repository, solely for the purpose of performing their respective regulatory and oversight responsibilities pursuant to the federal securities laws, rules and regulations or any contractual obligations. The Plan Processor will create and maintain a method of access to the CAT Data stored in the Central Repository that includes the ability to run searches and generate reports. The method in which the CAT Data is stored in the Central Repository will allow the ability to return results of queries that are complex in nature including market reconstruction and the status of order books at varying time intervals. The Plan Processor will, at least annually and at such earlier time promptly following a request by the Operating Committee, certify to the Operating Committee that only the Participants and the SEC have access to the Central Repository (other than access provided to any Industry Member for the purpose of correcting CAT Data previously reported to the Central Repository by such Industry Member).

See CAT NMS Plan, supra note 3, at Appendix C, The Security and Confidentiality of Information Reported to the Central Repository, and Appendix D, Data Security, describe the security and confidentiality of the CAT Data, including how access to the Central Repository is controlled.

Request for Comment

40. Do Commenters believe that the requirements presented in Appendix D, Central Repository Requirements, are sufficiently detailed to guide the Plan Processor in how to build and operate the Central Repository with regard to receiving, consolidating, and retaining data? If not, what additional information should the requirements contain? Are there any requirements that should be eliminated? Will such provisions give the Plan Processor too much discretion or flexibility in how to build and operate the Central Repository with regard to receiving, consolidating, and retaining data? Please identify and explain why such requirements are not necessary or appropriate.

41. Do Commenters believe that the information provided in Appendix D, Data Access, is sufficiently detailed to inform the Plan Processor and regulators how access to data will be granted? Are the controls and security provisions related to regulatory access to data appropriate and reasonable? Should additional provisions be included? If so, please identify and explain why such provisions are necessary. Should any provisions be modified or eliminated? Will such provisions give the Plan Processor too much discretion or flexibility in how to build and operate the Central Repository with regard to regulator access to the data? If so, please identify and explain why such provisions should be modified or not included in the CAT NMS Plan.

42. The CAT NMS Plan does not mandate a specific method for primary data storage of CAT Data, but does require that the storage solution would meet the security, reliability, and accessibility requirements for the CAT, including storage of personally identifiable information (“PII”) data, separately. The CAT NMS Plan also indicates several considerations in the selection of a storage solution including maturity, cost, complexity, and reliability of the storage method. The Commission requests comment on whether the CAT NMS Plan should mandate a particular data storage method. Why or why not? What are the advantages and disadvantages for CAT of the various storage methods?

C. Data Recording and Reporting by Participants

The Plan also sets forth the requirements regarding the data recording and reporting by Participants. Each Participant will record and electronically report to the Central Repository the following details for each order and each Reportable Event, as applicable (“Participant Data”; also referred to as “Recorded Industry Member Data”, as discussed in the next Section):

Participants may, but are not required to, coordinate compliance with the recording and reporting efforts through the use of regulatory services agreements and/or agreements adopted pursuant to Rule 17d-2 under the Exchange Act.

The CAT NMS Plan defines “Reportable Event” as “includ[ing], but . . . not limited to, the original receipt or origination, modification, cancellation, routing, execution (in whole or in part) and allocation of an order, and receipt of a routed order.” See CAT NMS Plan, supra note 3, at Section 1.1.

for original receipt or origination of an order: (1) Firm Designated ID(s) (FDIs) for each customer; (2) CAT-Order-ID; (3) SRO-Assigned Market Participant Identifier of the Industry Member receiving or originating the order; (4) date of order receipt or origination; (5) time of order receipt or origination (using time stamps pursuant to Section 6.8); (6) the Material Terms of the Order; and (7) other information as may be determined by the Operating Committee.

For a discussion of the Material Terms of the Order required by Rule 613, see Adopting Release, supra note 9, at 45750-52. The Commission notes that the Participants include in the Plan a requirement for the reporting of the OTC equity security symbol as one of the “Material Terms of the Order.” See CAT NMS Plan, supra note 3, at Section 1.1.

In the CAT NMS Plan as attached hereto as Exhibit A, the provisions of Section 6.3 enabling the Operating Committee to require Participants to record and report “other information” were removed.

for the routing of an order: (1) CAT-Order-ID; (2) date on which the order is routed; (3) time at which the order is routed (using time stamps pursuant to Section 6.8); (4) SRO-Assigned Market Participant Identifier of the Industry Member or Participant routing the order; (5) SRO-Assigned Market Participant Identifier of the Industry Member or Participant to which the order is being routed; (6) if routed internally at the Industry Member, the identity and nature of the department or desk to which the order is routed; (7) the Material Terms of the Order; and (8) other information as may be determined by the Operating Committee.

Id.

for the receipt of an order that has been routed, the following information: (1) CAT-Order-ID; (2) date on which the order is received; (3) time at which the order is received (using time stamps pursuant to Section 6.8); (4) SRO-Assigned Market Participant Identifier of the Industry Member or Participant receiving the order; (5) SRO-Assigned Market Participant Identifier of the Industry Member or Participant routing the order; (6) the Material Terms of the Order; and (7) other information as may be determined by the Operating Committee.

Id.

if the order is modified or cancelled: (1) CAT-Order-ID; (2) date the modification or cancellation is received or originated; (3) time at which the modification or cancellation is received or originated (using time stamps pursuant to Section 6.8); (4) price and remaining size of the order, if modified; (5) other changes in Material Terms, if modified; (6) whether the modification or cancellation instruction was given by the Customer, or was initiated by the Industry Member or Participant; and (7) other information as may be determined by the Operating Committee.

Id.

if the order is executed, in whole or in part: (1) CAT-Order-ID; (2) date of execution; (3) time of execution (using time stamps pursuant to Section 6.8); (4) execution capacity (principal, agency or riskless principal); (5) execution price and size; (6) the SRO-Assigned Market Participant Identifier of the Participant or Industry Member executing the order; and (7) whether the execution was reported pursuant to an effective transaction reporting plan or the Plan for Reporting of Consolidated Options Last Sale Reports and Quotation Information; and

other information or additional events as may be determined by the Operating Committee or otherwise prescribed in Appendix D, Reporting and Linkage Requirements.

Id.

As contemplated in Appendix D, Data Types and Sources, each Participant will report Participant Data to the Central Repository for consolidation and storage in a format specified by the Plan Processor, approved by the Operating Committee and compliant with Rule 613. As further described in Appendix D, Reporting and Linkage Requirements, each Participant is required to record the Participant Data contemporaneously with the Reportable Event. In addition, each Participant must report the Participant Data to the Central Repository by 8:00 a.m. Eastern Time on the Trading Day following the day that the Participant recorded the Participant Data. Participants may voluntarily report the Participant Data prior to the 8:00 a.m. Eastern Time deadline.

Each Participant that is a national securities exchange is required to comply with the above recording and reporting requirements for each NMS Security registered or listed for trading on such exchange or admitted to unlisted trading privileges on such exchange. Each Participant that is a national securities association is required to comply with the above recording and reporting requirements for each Eligible Security for which transaction reports are required to be submitted to the association.

D. Data Reporting and Recording by Industry Members

The Plan also sets forth the data reporting and recording requirements for Industry Members. Specifically, subject to Section 6.4(c), and Section 6.4(d)(iii) with respect to Options Market Makers, and consistent with Appendix D, Reporting and Linkage Requirements, each Participant, through its Compliance Rule, will require its Industry Members to record and electronically report to the Central Repository for each order and each Reportable Event the information referred to in Section 6.3(d), as applicable (“Recorded Industry Member Data”)—that is, Participant Data discussed above. In addition, subject to Section 6.4(c), and Section 6.4(d)(iii) with respect to Options Market Makers, and consistent with Appendix D, Reporting and Linkage Requirements, each Participant, through its Compliance Rule, will require its Industry Members to record and report to the Central Repository the following (“Received Industry Member Data” and, collectively with the Recorded Industry Member Data, “Industry Member Data”): (1) If the order is executed, in whole or in part: (a) An Allocation Report that includes the Firm Designated ID when an execution is allocated (in whole or in part); (b) SRO-Assigned Market Participant Identifier of the clearing broker or prime broker, if applicable; and (c) CAT-Order-ID of any contra-side order(s); (2) if the trade is cancelled, a cancelled trade indicator; and (3) for original receipt or origination of an order, information of sufficient detail to identify the Customer.

In the Amendment to the CAT NMS Plan, language in Section 6.4(d) that read, “that includes the Firm Designated ID when an execution is allocated (in whole or in part)” was removed because the definition of “Allocation Report” includes this information.

With respect to the reporting obligations of an Options Market Maker with regard to its quotes in Listed Options, Reportable Events required pursuant to Section 6.3(d)(ii) and (iv) will be reported to the Central Repository by an Options Exchange in lieu of the reporting of such information by the Options Market Maker. Each Participant that is an Options Exchange will, through its Compliance Rule, require its Industry Members that are Options Market Makers to report to the Options Exchange the time at which a quote in a Listed Option is sent to the Options Exchange (and, if applicable, any subsequent quote modifications and/or cancellation time when such modification or cancellation is originated by the Options Market Maker). Such time information also will be reported to the Central Repository by the Options Exchange in lieu of reporting by the Options Market Maker.

See Section III.B.9, infra, and accompanying requests for comment.

Each Participant will, through its Compliance Rule, require its Industry Members to record and report to the Central Repository other information or additional events as prescribed in Appendix D, Reporting and Linkage Requirements.

As contemplated in Appendix D, Data Types and Sources, each Participant will require its Industry Members to report Industry Member Data to the Central Repository for consolidation and storage in a format(s) specified by the Plan Processor, approved by the Operating Committee and compliant with Rule 613. As further described in Appendix D, Reporting and Linkage Requirements, each Participant will require its Industry Members to record Recorded Industry Member Data contemporaneously with the applicable Reportable Event. In addition, consistent with Appendix D, Reporting and Linkage Requirements, each Participant will require its Industry Members to report: (1) Recorded Industry Member Data to the Central Repository by 8:00 a.m. Eastern Time on the Trading Day following the day the Industry Member records such Recorded Industry Member Data; and (2) Received Industry Member Data to the Central Repository by 8:00 a.m. Eastern Time on the Trading Day following the day the Industry Member receives such Received Industry Member Data. Each Participant will permit its Industry Members to voluntarily report Industry Member Data prior to the applicable 8:00 a.m. Eastern Time deadline.

See Section III.B.2, infra, and accompanying requests for comment.

Each Participant that is a national securities exchange must require its Industry Members to report Industry Member Data for each NMS Security registered or listed for trading on such exchange or admitted to unlisted trading privileges on such exchange. Each Participant that is a national securities association must require its Industry Members to report Industry Member Data for each Eligible Security for which transaction reports are required to be submitted to the association.

Request for Comment

43. Sections 6.3(d) and 6.4(d) of the CAT NMS Plan set forth the details that Participants and Industry Members must report to the Central Repository. Do Commenters believe that these details will be sufficient to allow the Central Repository to link information to accurately reflect the lifecycle of an order? If not, what additional information should be required to be reported for this purpose?

44. Sections 6.3 and 6.4 of the CAT NMS Plan require Participants and Industry Members to record and report to the Central Repository other information or additional events as may be prescribed in Appendix D, Reporting and Linkage Requirements. Do Commenters believe that the CAT NMS Plan is sufficiently clear regarding the “other information or additional events as may be prescribed in Appendix D” that may be required? Please explain. Are these “other information or additional events prescribed in Appendix D” appropriate and reasonable? Please explain.

45. The CAT NMS Plan does not specify the format in which CAT Reporters must submit data, and states the Plan Processor will specify the format. Do Commenters believe that the CAT NMS Plan should specify a particular format? If so, what format? Please explain.

E. Regular Written Assessment

As described in Article VI, the Participants are required to provide the Commission with a written assessment of the operation of the CAT that meets the requirements set forth in Rule 613, Appendix D, and the Plan at least every two years or more frequently in connection with any review of the Plan Processor's performance under the Plan pursuant to Section 6.1(m). The Chief Compliance Officer will oversee this assessment and will provide the Participants a reasonable time to review and comment upon the written assessment prior to its submission to the SEC. In no case will the written assessment be changed or amended in response to a comment from a Participant; rather any comment by a Participant will be provided to the SEC at the same time as the written assessment.

The Commission notes that the applicable provision in the Amendment is Section 6.1(n).

Request for Comment

46. Do Commenters believe that the details and requirements regarding the regular written assessment of the operation of the CAT provided in Section 6.6 of the CAT NMS Plan are appropriate and reasonable? Would additional details or requirements for this assessment be beneficial?

47. Do Commenters believe that the Chief Compliance Officer should oversee the regular written assessment, as is required by Section 6.6? If not, would another party be better suited to this role?

F. Time Stamps and Synchronization of Business Clocks

Section 6.8 of the Plan discusses time stamps and the synchronization of Business Clocks. Each Participant is required to synchronize its Business Clocks (other than such Business Clocks used solely for Manual Order Events) at a minimum to within 50 milliseconds of the time maintained by the National Institute of Standards and Technology, consistent with industry standards. In addition, each Participant must, through its Compliance Rule, require its Industry Members to: (1) Synchronize their respective Business Clocks (other than such Business Clocks used solely for Manual Order Events) at a minimum to within 50 milliseconds of the time maintained by the National Institute of Standards and Technology, and maintain such a synchronization; (2) certify periodically that their Business Clocks meet the requirements of the Compliance Rule; and (3) report to the Plan Processor and the Participant any violation of the Compliance Rule pursuant to the thresholds set by the Operating Committee. Furthermore, each Participant is required to synchronize its Business Clocks and, through its Compliance Rule, require its Industry Members to synchronize their Business Clocks used solely for Manual Order Events at a minimum to within one second of the time maintained by the National Institute of Standards and Technology, consistent with industry standards, and maintain such synchronization. Each Participant will require its Industry Members to certify periodically (according to a schedule defined by the Operating Committee) that their Business Clocks used solely for Manual Order Events meet the requirements of the Compliance Rule. The Compliance Rule of a Participant shall require its Industry Members using Business Clocks solely for Manual Order Events to report to the Plan Processor any violation of the Compliance Rule pursuant to the thresholds set by the Operating Committee. The Participants stated their belief that pursuant to Rule 613(d)(1) that these synchronization standards are consistent with current industry standards.

Each Participant shall, and through its Compliance Rule require its Industry Members to, report information required by Rule 613 and this Agreement to the Central Repository in milliseconds. To the extent that any Participant utilizes time stamps in increments finer than the minimum required by the Plan, the Participant is required to make reports to the Central Repository utilizing such finer increment when reporting CAT Data to the Central Repository so that all Reportable Events reported to the Central Repository could be adequately sequenced. Each Participant will, through its Compliance Rule: (1) Require that, to the extent that its Industry Members utilize time stamps in increments finer than the minimum required in the Plan, such Industry Members will utilize such finer increment when reporting CAT Data to the Central Repository; and (2) provide that a pattern or practice of reporting events outside of the required clock synchronization time period without reasonable justification or exceptional circumstances may be considered a violation of SEC Rule 613 and the Plan. Notwithstanding the preceding sentences, each Participant and Industry Member will be permitted to record and report Manual Order Events to the Central Repository in increments up to and including one second, provided that Participants and Industry Members will be required to record and report the time when a Manual Order Event has been captured electronically in an order handling and execution system of such Participant or Industry Member in milliseconds. In conjunction with Participants' and other appropriate Industry Member advisory groups, the Chief Compliance Officer will annually evaluate and make a recommendation to the Operating Committee as to whether industry standards have evolved such that the required synchronization should be shortened or the required time stamp should be in finer increments. The Operating Committee will make determinations regarding the need to revise the synchronization and time stamp requirements.

Request for Comment

See Sections III.B.4 and III.B.5, infra, for additional requests for comment on clock synchronization and time stamp granularity.

48. Do Commenters believe that the CAT NMS Plan's requirement that Participants and Industry Members synchronize their Business Clocks to within 50 milliseconds of the time maintained by the National Institute of Standards and Technology (“NIST”) is appropriate and reasonable? Do Commenters agree with the Participants that this clock offset tolerance represents current industry standards? Would a tighter clock offset tolerance be feasible?

49. Do Commenters believe that the CAT NMS Plan's requirement that Participants and Industry Members report information to the Central Repository in milliseconds is appropriate and reasonable? Would a more granular time stamp requirement be feasible? Do Commenters agree with the Participants that time stamp granularity to the millisecond represents current industry standards?

50. How should “industry standard,” for purposes of the CAT NMS Plan's clock synchronization and time stamping requirements, be determined? Do Commenters believe that “industry standard” should be based on current industry practice? If not, how should “industry standard” be defined? What other factors, if any, should be considered in defining such “industry standards”?

G. Technical Specifications

Section 6.9 of the Plan establishes the requirements involving the Plan Processor's Technical Specifications. The Plan Processor will publish Technical Specifications that are at a minimum consistent with Appendices C and D, and updates thereto as needed, providing detailed instructions regarding the submission of CAT Data by Participants and Industry Members to the Plan Processor for entry into the Central Repository. The Technical Specifications will be made available on a publicly available Web site to be developed and maintained by the Plan Processor. The initial Technical Specifications and any Material Amendments thereto will require the approval of the Operating Committee by Supermajority Vote.

The Technical Specifications will include a detailed description of the following: (1) The specifications for the layout of files and records submitted to the Central Repository; (2) the process for the release of new data format specification changes; (3) the process for industry testing for any changes to data format specifications; (4) the procedures for obtaining feedback about and submitting corrections to information submitted to the Central Repository; (5) each data element, including permitted values, in any type of report submitted to the Central Repository; (6) any error messages generated by the Plan Processor in the course of validating the data; (7) the process for file submissions (and re-submissions for corrected files); (8) the storage and access requirements for all files submitted; (9) metadata requirements for all files submitted to the CAT System; (10) any required secure network connectivity; (11) data security standards, which will, at a minimum: (a) Satisfy all applicable regulations regarding database security, including provisions of Regulation Systems Compliance and Integrity under the Exchange Act (“Reg SCI”); (b) to the extent not otherwise provided for under the Plan (including Appendix C thereto), set forth such provisions as may be necessary or appropriate to comply with Rule 613(e)(4); and (c) comply with industry best practices; and (12) any other items reasonably deemed appropriate by the Plan Processor and approved by the Operating Committee.

Amendments to the Technical Specifications may be made only in accordance with Section 6.9(c). The process for amending the Technical Specifications varies depending on whether the change is material. An amendment will be deemed “material” if it would require a Participant or an Industry Member to engage in significant changes to the coding necessary to submit information to the Central Repository pursuant to the Plan, or if it is required to safeguard the security or confidentiality of the CAT Data. Except for Material Amendments to the Technical Specifications, the Plan Processor will have the sole discretion to amend and publish interpretations regarding the Technical Specifications; however, all non-Material Amendments made to the Technical Specifications and all published interpretations will be provided to the Operating Committee in writing at least ten days before being published. Such non-Material Amendments and published interpretations will be deemed approved ten days following provision to the Operating Committee unless two unaffiliated Participants call for a vote to be taken on the proposed amendment or interpretation. If an amendment or interpretation is called for a vote by two or more unaffiliated Participants, the proposed amendment must be approved by Majority Vote of the Operating Committee. Once a non-Material Amendment has been approved or deemed approved by the Operating Committee, the Plan Processor will be responsible for determining the specific changes to the Central Repository and providing technical documentation of those changes, including an implementation timeline.

Material Amendments to the Technical Specifications require approval of the Operating Committee by Supermajority Vote. The Operating Committee, by Supermajority Vote, may amend the Technical Specifications on its own motion.

Request for Comment

51. Do Commenters believe that the list of items to be included in the Technical Specifications, as set forth in Section 6.9(b) of the CAT NMS Plan, is appropriate and reasonable? Do Commenters believe that detailed descriptions of any of the listed items should be included in the CAT NMS Plan rather than in the Technical Specifications? Do Commenters believe that the list addresses all of the areas that should be included in the Technical Specifications? Are there other aspects of the CAT that require Technical Specifications? If so, please identify and explain why the additional Technical Specifications are needed.

52. Do Commenters believe the Plan Processor should have sole discretion to amend and publish interpretations regarding the Technical Specifications, except for Material Amendments? Why or why not? What discretion or input, if any, should the Operating Committee or other parties, including the Advisory Committee, have in amending and publishing Technical Specifications interpretations?

53. How should Technical Specifications be communicated to the industry? Why?

54. What are the incentives for the Operating Committee to review the Plan Processor's interpretation of Technical Specifications and verify that the interpretation is consistent with the regulatory objectives of the Plan? What are the best practices to ensure sufficient review by the Operating Committee? What provisions of the Plan are in place to ensure that the Operating Committee follows these practices? What provisions, if any, could be strengthened? Please explain and provide supporting examples and evidence, if available.

55. The CAT NMS Plan provides that non-Material Amendments and published interpretations will be deemed approved ten days following provision to the Operating Committee, unless two unaffiliated Participants call for a vote to be taken on the proposed amendment or interpretation. Do Commenters have any views on this process? If so, please explain.

56. Do Commenters have any views regarding the definition of Material Amendments? Is the definition too broad? Too narrow? Please explain. Do Commenters have any views on who should be responsible for determining whether an amendment to the Technical Specifications is a Material Amendment? Do Commenters believe the CAT NMS Plan clearly states who shall have the responsibility to make the determination? Do Commenters have any views on how the determination should be made? Please explain.

57. The CAT NMS Plan requires that Material Amendments be approved by the Operating Committee by Supermajority Vote and allows the Operating Committee to amend the Technical Specifications on its own motion by Supermajority Vote. Do Commenters have any views on these processes? If so, please explain.

58. The CAT NMS Plan provides that the Plan Processor's business continuity planning must include a secondary site for critical staff, capable of recovery and restoration of services within 48 hours, with the goal of next day recovery. Should the CAT NMS Plan provide additional details regarding “the goal of next day recovery”? Do Commenters believe a 48-hour recovery and restoration period is too long? Too short? Please explain. Should the CAT NMS Plan impose any other requirements on the Plan Processor to better assure the Plan Processor is able to transition to the secondary site within the specified time frames? If so, what?

H. Surveillance

Surveillance issues are described in Section 6.10. Using the tools provided for in Appendix D, Functionality of the CAT System, each Participant will develop and implement a surveillance system, or enhance existing surveillance systems, reasonably designed to make use of the consolidated information contained in the Central Repository. Unless otherwise ordered by the SEC, within fourteen months after the Effective Date, each Participant must initially implement a new or enhanced surveillance system(s) as required by Rule 613 and Section 6.10(a) of the Plan. Participants may, but are not required to, coordinate surveillance efforts through the use of regulatory services agreements and agreements adopted pursuant to Rule 17d-2 under the Exchange Act.

Consistent with Appendix D, Functionality of the CAT System, the Plan Processor will provide Participants and the SEC with access to all CAT Data stored in the Central Repository. Regulators will have access to processed CAT Data through two different methods: (1) An online targeted query tool; and (2) user-defined direct queries and bulk extracts. The online targeted query tool will provide authorized users with the ability to retrieve CAT Data via an online query screen that includes the ability to choose from a variety of pre-defined selection criteria. Targeted queries must include date(s) and/or time range(s), as well as one or more of a variety of fields. The user-defined direct queries and bulk extracts will provide authorized users with the ability to retrieve CAT Data via a query tool or language that allows users to query all available attributes and data sources.

Extraction of CAT Data will be consistent with all permission rights granted by the Plan Processor. All CAT Data returned will be encrypted, and PII data will be masked unless users have permission to view the PII contained in the CAT Data that has been requested.

The Plan Processor will implement an automated mechanism to monitor direct query usage. Such monitoring will include automated alerts to notify the Plan Processor of potential issues with bottlenecks or excessively long queues for queries or CAT Data extractions. The Plan Processor will provide the Operating Committee or its designee(s) details as to how the monitoring will be accomplished and the metrics that will be used to trigger alerts.

The Plan Processor will reasonably assist regulatory Staff (including those of Participants) with creating queries. Without limiting the manner in which regulatory Staff (including those of Participants) may submit queries, the Plan Processor will submit queries on behalf of regulatory Staff (including those of Participants) as reasonably requested. The Plan Processor will staff a CAT help desk, as described in Appendix D, CAT Help Desk, to provide technical expertise to assist regulatory Staff (including those of Participants) with questions about the content and structure of the CAT Data.

Request for Comment

59. What features of the CAT NMS Plan will facilitate the creation of enhanced surveillance systems? Are the minimum functional and technical requirements for the Plan Processor set forth in Appendix D consistent with the creation of enhanced surveillance systems? What, if any, additional requirements or details should be provided in the CAT NMS Plan to ensure that the Plan facilitates the creation of enhanced surveillance systems?

60. Under the CAT NMS Plan, will regulatory Staff have appropriate access to the Central Repository? Specifically, do Commenters believe that the online targeted query tool and user-defined direct queries and bulk extracts described in Sections 8.1 and 8.2 of Appendix D will enable regulatory Staff to use the data in the Central Repository to carry out their surveillance, analysis, and other regulatory functions? If not, why not and what should be added? Does the CAT NMS Plan provide sufficient detail to determine if regulators will have appropriate access? If not, what additional details should be provided?

61. Do Commenters believe that the provisions in Section 6.10(c)(ii) of the CAT NMS Plan regarding permission rights granted by the Plan Processor, encryption, and masking of PII are appropriate and reasonable? Would these provisions affect the ability of Commission or SRO regulatory Staff to access and use the data in the Central Repository? If so, what additional or different provisions would mitigate the impact on regulatory access to and use of the data?

62. Do Commenters believe that the query monitoring mechanism to be implemented by the Plan Processor, as described in Section 6.10(c)(iii) of the CAT NMS Plan, is appropriately designed to help enable regulators to carry out their regulatory functions? If not, what additional details or functionality should be provided? Will the provisions regarding Plan Processor assistance of regulatory Staff and submission of regulatory Staff queries (Sections 6.10(c)(iv)-(v) of the CAT NMS Plan) and the CAT user support functionality (as described in Section 10.2 of Appendix D) provide sufficient assistance to regulators in carrying out their regulatory functions?

I. Information Security Program

As set forth in Section 6.12, the Plan Processor is required to develop and maintain a comprehensive information security program for the Central Repository that contains, at a minimum, the specific requirements detailed in Appendix D, Data Security. The information security program must be approved and reviewed at least annually by the Operating Committee.

Request for Comment

63. Do Commenters believe the CAT NMS Plan should include a discussion of policies and procedures applicable to members of the Advisory Committee to ensure the security and confidentiality of the operation of the CAT (for example, requiring members of the Advisory Committee to enter into a non-disclosure agreement with the Company)? If so, what additional measures should be considered?

64. Do Commenters believe the CAT NMS Plan should detail the policies and procedures applicable to regulatory users of the CAT that would ensure the security and confidentiality of the CAT Data and the operation of the CAT? If so, what measures should be considered? Do Commenters have any views on how such policies and procedures should be enforced? Please explain.

(6) Financial Matters

Articles VII and VIII of the Plan address certain financial matters related to the Company. In particular, the Plan states that, subject to certain special allocations provided for in Section 8.2, any net profit or net loss will be allocated among the Participants equally. In addition, subject to Section 10.2, cash and property of the Company will not be distributed to the Participants unless the Operating Committee approves by Supermajority Vote a distribution after fully considering the reason that such distribution must or should be made to the Participants, including the circumstances contemplated under Section 8.3, Section 8.6, and Section 9.3. To the extent a distribution is made, all Participants will participate equally in any such distribution except as otherwise provided in Section 10.2.

Article XI addresses the funding of the Company. On an annual basis the Operating Committee will approve an operating budget for the Company. The budget will include the projected costs of the Company, including the costs of developing and operating the CAT System for the upcoming year, and the sources of all revenues to cover such costs, as well as the funding of any reserve that the Operating Committee reasonably deems appropriate for prudent operation of the Company.

Subject to certain funding principles set forth in Article XI, the Operating Committee will have discretion to establish funding for the Company, including: (1) Establishing fees that the Participants will pay; and (2) establishing fees for Industry Members that will be implemented by Participants. In establishing the funding of the Company, the Operating Committee will seek to: (1) Create transparent, predictable revenue streams for the Company that are aligned with the anticipated costs to build, operate and administer the CAT and the other costs of the Company; (2) establish an allocation of the Company's related costs among Participants and Industry Members that is consistent with the Exchange Act, taking into account the timeline for implementation of the CAT and distinctions in the securities trading operations of Participants and Industry Members and their relative impact upon Company resources and operations; (3) establish a tiered fee structure in which the fees charged to: (a) CAT Reporters that are Execution Venues, including ATSs, are based upon the level of market share, (b) Industry Members' non-ATS activities are based upon message traffic, and (c) the CAT Reporters with the most CAT-related activity (measured by market share and/or message traffic, as applicable) are generally comparable (where, for these comparability purposes, the tiered fee structure takes into consideration affiliations between or among CAT Reporters, whether Execution Venues and/or Industry Members); (4) provide for ease of billing and other administrative functions; (5) avoid any disincentives such as placing an inappropriate burden on competition and a reduction in market quality; and (6) build financial stability to support the Company as a going concern. The Participants will file with the SEC under Section 19(b) of the Exchange Act any such fees on Industry Members that the Operating Committee approves, and such fees will be labeled as “Consolidated Audit Trail Funding Fees.”

To fund the development and implementation of the CAT, the Company will time the imposition and collection of all fees on Participants and Industry Members in a manner reasonably related to the timing when the Company expects to incur such development and implementation costs. In determining fees for Participants and Industry Members, the Operating Committee shall take into account fees, costs and expenses (including legal and consulting fees and expenses) incurred by the Participants on behalf of the Company prior to the Effective Date in connection with the creation and implementation of the CAT, and such fees, costs and expenses shall be fairly and reasonably shared among the Participants and Industry Members. Consistent with Article XI, the Operating Committee will adopt policies, procedures, and practices regarding the budget and budgeting process, assignment of tiers, resolution of disputes, billing and collection of fees, and other related matters. As a part of its regular review of fees for the CAT, the Operating Committee will have the right to change the tier assigned to any particular Person pursuant to this Article XI. Any such changes will be effective upon reasonable notice to such Person.

The Commission notes that Section 11.1(b) of the CAT NMS Plan states that the Participants would file fees for Industry Members approved by the Operating Committee with the Commission. The Operating Committee may only change the tier to which a Person is assigned in accordance with a fee schedule filed with the Commission.

The Operating Committee will establish fixed fees to be payable by Execution Venues as follows. Each Execution Venue that executes transactions, or, in the case of a national securities association, has trades reported by its members to its trade reporting facility or facilities for reporting transactions effected otherwise than on an exchange, in NMS Stocks or OTC Equity Securities will pay a fixed fee depending on the market share of that Execution Venue in NMS Stocks and OTC Equity Securities. The Operating Committee will establish at least two and no more than five tiers of fixed fees, based on an Execution Venue's NMS Stocks and OTC Equity Securities market share. For these purposes, market share will be calculated by share volume. In addition, each Execution Venue that executes transactions in Listed Options will pay a fixed fee depending on the Listed Options market share of that Execution Venue. The Operating Committee will establish at least two and no more than five tiers of fixed fees, based on an Execution Venue's Listed Options market share, with market share calculated by contract volume. Changes to the number of tiers after approval of the Plan would require a Supermajority Vote of the Operating Committee and Commission approval under Section 19(b) of the Exchange Act, as would the establishment of the initial fee schedule and any changes to the fee schedule within the tier structure.

The Commission notes that the Participants could choose to submit the proposed fee schedule to the Commission as individual SROs pursuant to Rule 19b-4 or jointly as Participants to an NMS plan pursuant to Rule 608 of Regulation NMS. Because the proposed fee schedule would establish fees, whether the Participants individually file it pursuant to Section 19(b)(3)(A)(ii) of the Act, or jointly file it pursuant to Rule 608(b)(3)(i) of Regulation NMS, the proposed fee schedule could take effect upon filing with the Commission. See 15 U.S.C. 78s(b)(3)(A)(ii); 17 CFR 242.608(b)(3)(i).

The Operating Committee also will establish fixed fees payable by Industry Members, based on the message traffic generated by such Industry Member. The Operating Committee will establish at least five and no more than nine tiers of fixed fees, based on message traffic. For the avoidance of doubt, the fixed fees payable by Industry Members pursuant to this paragraph will, in addition to any other applicable message traffic, include message traffic generated by: (1) An ATS that does not execute orders that is sponsored by such Industry Member; and (2) routing orders to and from any ATS system sponsored by such Industry Member.

Furthermore, the Operating Committee may establish any other fees ancillary to the operation of the CAT that it reasonably determines appropriate, including: Fees for the late or inaccurate reporting of information to the CAT; fees for correcting submitted information; and fees based on access and use of the CAT for regulatory and oversight purposes (and not including any reporting obligations).

As it relates to any fees that the Operating Committee may impose for access and use of the CAT for regulatory and oversight purposes, the Commission interprets the provisions in the Plan relating to the collection of fees as applying only to Participants and Industry Members, and thus the Commission would not be subject to such fees.

The Company will make publicly available a schedule of effective fees and charges adopted pursuant to the Plan as in effect from time to time. Such schedule will be developed after the Plan Processor is selected. The Operating Committee will review the fee schedule on at least an annual basis and will make any changes to such fee schedule that it deems appropriate. The Operating Committee is authorized to review the fee schedule on a more regular basis, but will not make any changes on more than a semi-annual basis unless, pursuant to a Supermajority Vote, the Operating Committee concludes that such change is necessary for the adequate funding of the Company.

The Operating Committee will establish a system for the collection of fees authorized under the Plan. The Operating Committee may include such collection responsibility as a function of the Plan Processor or another administrator. Alternatively, the Operating Committee may use the facilities of a clearing agency registered under Section 17A of the Exchange Act to provide for the collection of such fees.

Each Participant will require each Industry Member to pay all applicable fees authorized under Article XI within thirty days after receipt of an invoice or other notice indicating payment is due (unless a longer payment period is otherwise indicated). If an Industry Member fails to pay any such fee when due, such Industry Member will pay interest on the outstanding balance from such due date until such fee is paid at a per annum rate equal to the lesser of: (1) The Prime Rate plus 300 basis points; or (2) the maximum rate permitted by applicable law. Each Participant will pay all applicable fees authorized under Article XI as required by Section 3.7(b).

Disputes with respect to fees the Company charges Participants pursuant to Article XI will be determined by the Operating Committee or a Subcommittee designated by the Operating Committee. Decisions by the Operating Committee on such matters shall be binding on Participants, without prejudice to the rights of any Participant to seek redress from the SEC pursuant to SEC Rule 608 or in any other appropriate forum. The Participants will adopt rules requiring that disputes with respect to fees charged to Industry Members pursuant to Article XI be determined by the Operating Committee or a Subcommittee. Decisions by the Operating Committee or Subcommittee on such matters will be binding on Industry Members, without prejudice to the rights of any Industry Member to seek redress from the SEC pursuant to SEC Rule 608 or in any other appropriate forum.

Request for Comment

65. Do Commenters believe that the provisions in the CAT NMS Plan regarding the funding and budget of the Company to operate the CAT (as described in Article XI) are appropriate and reasonable? Specifically, do Commenters believe that the tiered funding model described in Section 11.2(c) of the CAT NMS Plan and the fixed-tier funding model described in Section 11.3 of the CAT NMS Plan are appropriate and reasonable?

66. What are Commenters' views regarding the methodology in the CAT NMS Plan to establish and impose fees on Participants and the industry? Do Commenters believe that the fee system described in Sections 11.2 and 11.3 of the CAT NMS Plan will result in an equitable and fair allocation of CAT-related fees between Participants, other types of Execution Venues, and Industry Members? Will the fee system in the Plan, including consideration of the distinctions in securities trading operations, impose higher costs upon or result in any competitive advantage to some types of Execution Venues or Industry Members as opposed to others? If yes, are those differences in fees appropriate and reasonable? Will this proposed fee system create incentives to execute orders in certain Execution Venues over others? What alternative fee systems, if any, would be more appropriate?

67. Do Commenters believe that assessing fees based on market share and message traffic, as described in Sections 11.2 and 11.3 of the CAT NMS Plan, is appropriate and reasonable? Specifically, is it appropriate and reasonable to base Industry Member fees on message traffic and Execution Venue fees on market share? Will this method of calculating fees impose higher costs upon or result in any competitive advantage to some types of Execution Venues or Industry Members as opposed to others? What fee calculation method, if any, would be more appropriate?

68. Are the tier levels appropriate and reasonable? Why or why not? Is the number of tiers contemplated (2-5 for Execution Venues and 5-9 for Industry Members) appropriate and reasonable? Why or why not?

69. Do Commenters believe that giving the right to the Operating Committee to change the fee tier assigned to any particular Person as set forth in Section 11.1(d) of the CAT NMS Plan is appropriate and reasonable? If not, why not? What alternative process, if any, would be more appropriate?

70. Do Commenters believe that giving the right to the Operating Committee to change the fee tier assigned to any particular Person as set forth in Section 11.1(d) of the CAT NMS Plan conflicts with the tier structure of fees as set forth in Section 11.2(c) of the CAT NMS Plan, which will be based on the market share for Execution Venues, and message traffic for Industry Members? Why or why not?

71. Section 11.1(d) of the CAT NMS Plan also provides that any change to a Person's fee tier will be effective upon reasonable notice to such Person. Do Commenters believe that a notice to any such Person is necessary, given that the CAT NMS Plan provides that a Person will change fee tiers based on market share or message traffic, as applicable? Why or why not? What should constitute reasonable notice?

72. Do Commenters believe the Operating Committee's ability to establish additional fees for “access and use of the CAT for regulatory and oversight purposes” (as described in Section 11.3(c) of the CAT NMS Plan) is appropriate and reasonable? Would this provision affect the ability of regulatory Staff to access and use the data in the Central Repository? If so, what additional or different provisions would mitigate the impact upon regulatory access to and use of the data?

73. Do Commenters believe that the funding provisions in Section 11.1 of the CAT NMS Plan provide sufficient authority and guidance to the Operating Committee to establish and maintain such reserves as are reasonably deemed appropriate by the Operating Committee for the prudent operation of the Company? If not, why not?

74. Do Commenters believe that the provisions in the CAT NMS Plan regarding the collection of fees (Section 11.4 of the CAT NMS Plan) and fee disputes (Section 11.5 of the CAT NMS Plan) are appropriate and reasonable? If not, what alternatives do Commenters suggest?

75. Do Commenters believe the CAT NMS Plan provides sufficient detail regarding the proposed cost allocation among the Plan Processor and regulators with respect to hardware and software costs that may be required in order to use CAT Data? If not, what are the risks of not providing sufficient detail and what requirements should be set forth in the CAT NMS Plan? For example, since there will only be one Plan Processor, what are the risks of significant costs for regulators to the extent regulators will need to contract with the Plan Processor for additional computing resources, storage costs and data transfer costs?

76. Should the Operating Committee be required to consult the Advisory Committee when setting fees and performing regular reviews of fees? Please explain.

(7) Amendments

Section 12.3 of the CAT NMS Plan, which governs amendments to the Plan, states that, except with respect to the addition of new Participants (Section 3.3), the transfer of Company Interest (Section 3.4), the termination of a Participant's participation in the Plan (Section 3.7), amendments to the Selection Plan (Section 5.3 [sic]) and special allocations (Section 8.2), any change to the Plan requires a written amendment authorized by the affirmative vote of not less than two-thirds of all of the Participants, or with respect to Section 3.8 by the affirmative vote of all the Participants. Such proposed amendment must be approved by the Commission pursuant to Rule 608 or otherwise becomes effective under Rule 608. Notwithstanding the foregoing, to the extent that the SEC grants exemptive relief applicable to any provision of this Agreement, Participants and Industry Members will be entitled to comply with such provision pursuant to the terms of the exemptive relief so granted at the time such relief is granted irrespective of whether the LLC Agreement has been amended.

(8) Compliance Rule Applicable to Industry Members

Under Article III, each Participant agrees to comply with and enforce compliance by its Industry Members with the provisions of Rule 613 and the Plan, as applicable, to the Participant and its Industry Members. Accordingly, the Participants will endeavor to promulgate consistent rules (after taking into account circumstances and considerations that may impact Participants differently) requiring compliance by their respective Industry Members with the provisions of Rule 613 and the Plan.

(9) Plan Appendices

The Plan includes three appendices. Appendix A provides the Consolidated Audit Trail National Market System Plan Request for Proposal, as issued February 26, 2013 and subsequently updated. In addition, Rule 613(a)(1) requires that the Plan discuss twelve considerations that explain the choices made by the Participants to meet the requirements specified in Rule 613 for the CAT. In accordance with this requirement, the Participants have addressed each of the twelve considerations in Appendix C. Finally, Appendix D describes the technical requirements for the Plan Processor.

Appendix B is reserved for future use.

b. Governing or Constituent Documents

Rule 608 requires copies of all governing or constituent documents relating to any person (other than a self-regulatory organization) authorized to implement or administer such plan on behalf of its sponsors. The Participants will submit to the Commission such documents related to the Plan Processor when the Plan Processor is selected.

c. Development and Implementation Phases

The terms of the Plan will be effective immediately upon approval of the Plan by the Commission (the “Effective Date”). The Plan sets forth each of the significant phases of development and implementation contemplated by the Plan, together with the projected date of completion of each phase. These include the following, each of which is subject to orders otherwise by the Commission:

Within two months after the Effective Date, the Participants will jointly select the winning Shortlisted Bid and the Plan Processor pursuant to the process set forth in Article V. Following the selection of the Initial Plan Processor, the Participants will file with the Commission a statement identifying the Plan Processor and including the information required by Rule 608;

Within four months after the Effective Date, each Participant will, and, through its Compliance Rule, will require its Industry Members to, synchronize its or their Business Clocks and certify to the Chief Compliance Officer (in the case of Participants) or the applicable Participant (in the case of Industry Members) that it has met this requirement;

Within six months after the Effective Date, the Participants must jointly provide to the SEC a document outlining how the Participants could incorporate into the CAT information with respect to equity securities that are not NMS Securities, including Primary Market Transactions in securities that are not NMS Securities, which document will include details for each order and Reportable Event that may be required to be provided, which market participants may be required to provide the data, the implementation timeline, and a cost estimate;

In the Amendment to the CAT NMS Plan, Section 6.11 excludes OTC Equity Securities from the document the Participants would submit to the Commission, since the Participants plan to include OTC Equity Securities as well as NMS Securities in the initial phase in of CAT.

Within one year after the Effective Date, each Participant must report Participant Data to the Central Repository;

Within fourteen months after the Effective Date, each Participant must implement a new or enhanced surveillance system(s);

Within two years after the Effective Date, each Participant must, through its Compliance Rule, require its Industry Members (other than Small Industry Members) to report Industry Member Data to the Central Repository; and

Within three years after the Effective Date, each Participant must, through its Compliance Rule, require its Small Industry Members to provide Industry Member Data to the Central Repository.

In addition, Industry Members and Participants will be required to participate in industry testing with the Central Repository on a schedule to be determined by the Operating Committee. Furthermore, Appendix C, A Plan to Eliminate Existing Rules and Systems (SEC Rule 613(a)(1)(ix)), and Appendix D, Data Types and Sources, set forth additional implementation details concerning the elimination of rules and systems.

The Chief Compliance Officer will appropriately document objective milestones to assess progress toward the implementation of this Agreement.

Request for Comment

77. Under the CAT NMS Plan, the SROs' rules would require that their members become CAT Reporters. What mechanism should there be to ensure that all CAT Reporters would participate in all pre-implementation activities, including connectivity and testing? Please explain.

78. Do Commenters believe that the CAT NMS Plan allows for sufficient pre-implementation testing support for CAT Reporters, including providing CAT Reporter feedback and accuracy reports? If not, what requirements should be added to the CAT NMS Plan?

79. Do Commenters believe that full implementation of the CAT would allow for the retirement of OATS? Please explain. Are any identified gaps with respect to OATS' data elements not addressed in the CAT NMS Plan? If yes, what are they?

80. The CAT NMS Plan provides for a single Plan Processor. As such, do Commenters believe there are adequate and appropriate incentives for continuous CAT innovation and cost reductions by the Plan Processor and the Participants? If not, explain and describe what additional incentives may be implemented in the CAT NMS Plan or related documentation. What competition might be encouraged to lead to further innovations and reduced costs for future CAT technologies?

81. Do Commenters believe that the proposed CAT NMS Plan sets forth acceptable milestones to measure the progress of developing and implementing the CAT? Why or why not?

82. The CAT NMS Plan sets forth significant phases of development and implementation and a projected timetable for each stage. Are these projections appropriate and reasonable? If not, why not, and what is a more appropriate and reasonable timeline?

83. The CAT NMS Plan's “Access to the Central Repository for Regulators” Section sets forth a milestone requiring the publication of the finalized document detailing methods of access to the Central Repository one (1) month before Participants are required to begin reporting. Do Commenters believe this allows sufficient time for Participants to build applications to access the Central Repository when CAT goes live? If not, please explain and describe any related modifications to this Section.

See CAT NMS Plan, supra note 3, at Appendix C, Section C.10(d).

d. Analysis of Impact on Competition

The Commission reiterates that Section III.A of this Notice, including this subsection III.A.3.d, is substantially as prepared and submitted by the SROs to the Commission. The Commission's Economic Analysis in respect of the Plan's impact on competition is set forth in Section IV of this Notice.

The Plan states that it does not impose any burden on competition that is not necessary or appropriate in furtherance of the purposes of the Exchange Act. Section 8 of Appendix C, An Analysis of the Impact on Competition, Efficiency and Capital Formation, discusses the competition impact of the Plan in detail. In addition, the Participants do not believe that the Plan introduces terms that are unreasonably discriminatory for the purposes of Section 11A(c)(1)(D) of the Exchange Act. As noted in Section III.A.3.a, supra, the Participants are aware that potential conflicts of interest are raised because a Participant, or an Affiliate of a Participant, may be both submitting a Bid (or participating in a Bid (e.g., as a subcontractor)) and participating in the evaluation of Bids to select the Plan Processor. As described in Section III.A.3.a, the Selection Plan previously approved by the Commission and incorporated in the Plan includes multiple provisions designed to mitigate the potential impact of these conflicts by imposing restrictions on the Voting Senior Officers and by requiring the recusal of Bidding Participants for certain votes taken by the Selection Committee.

The Commission notes that as required under Rule 613(a)(1)(viii), the SROs set forth in the CAT NMS Plan a discussion of their analysis of the impact on competition, efficiency and capital formation of creating, implementing, and maintaining the CAT NMS Plan. See 17 CFR 242.613(a)(1)(viii) and CAT NMS Plan, supra note 3, at Appendix C, Section B.8. The SROs' analysis in Section B.8 of Appendix C to the CAT NMS Plan, which is more detailed than as set forth in this Section III of this Notice, is organized as follows: (a) Impact on Competition—both for Participants and Broker-Dealers, (b) Impact on Efficiency, (c) Impact on Capital Formation, and (d) Impacts of the CAT NMS Plan Governance on Efficiency, Competition, and Capital Formation. See CAT NMS Plan, supra note 3, at Appendix C, Section B.8. The Commission's analysis in respect of the Plan's impact on competition, efficiency and capital formation includes discussions of the SROs' analysis regarding the same and is in Section IV of this Notice. See Section IV.G, infra.

e. Written Understanding or Agreements Relating to Interpretation of, or Participation in, the Plan

The Participants have no written understandings or agreements relating to interpretations of, or participation in, the Plan other than those set forth in the Plan itself. For example, Section 4.3(a)(iii) states that the Operating Committee only may authorize the interpretation of the Plan by Majority Vote, Section 6.9(c)(i) addresses interpretations of the Technical Specifications, and Section 8.2 addresses the interpretation of Sections 8.1 and 8.2. In addition, Section 3.3 sets forth how any entity registered as a national securities exchange or national securities association under the Exchange Act may become a Participant.

f. Dispute Resolution

The Plan does not include a general provision addressing the method by which disputes arising in connection with the operation of the Plan will be resolved. The Plan does, however, provide the means for resolving disputes regarding the Participation Fee. Specifically, Article III states that, in the event that the Company and a prospective Participant do not agree on the amount of the Participation Fee, such amount will be subject to the review by the SEC pursuant to Section 11A(b)(5) of the Exchange Act. In addition, the Plan addresses disputes with respect to fees charged to Participants and Industry Members pursuant to Article XI. Specifically, such disputes will be determined by the Operating Committee or a Subcommittee designated by the Operating Committee. Decisions by the Operating Committee or such designated Subcommittee on such matters will be binding on Participants and Industry Members, without prejudice to the rights of any Participant or Industry Member to seek redress from the SEC pursuant to Rule 608 or in any other appropriate forum.

This marks the end of the statement of purpose as set forth above and as substantially prepared and submitted by the SROs.

B. Summary of Additional CAT NMS Plan Provisions and Request for Comment

The Commission requests and encourages any interested person to comment generally on the proposed CAT NMS Plan. In addition to the specific requests for comment throughout the release, the Commission requests general comment on all aspects of the proposed CAT NMS Plan. The Commission encourages Commenters to provide information regarding the advantages and disadvantages of each aspect of the proposed CAT NMS Plan. The Commission invites Commenters to provide views and data as to the costs and benefits associated with the proposed CAT NMS Plan. The Commission also seeks comment regarding other matters that may have an effect on the proposed CAT NMS Plan.

1. Reporting Procedures

The CAT NMS Plan requires CAT Reporters to comply with specific reporting procedures when reporting CAT Data to the Central Repository. Specifically, CAT Reporters must format CAT Data to comply with the format specifications approved by the Operating Committee. CAT Reporters must record CAT Data contemporaneously with the applicable Reportable Event and report such data to the Central Repository by 8:00 a.m. Eastern Time on the next Trading Day. The obligation to report CAT Data applies to “each NMS Security registered or listed for trading on [a national securities] exchange or admitted to unlisted trading privileges on such exchange,” and “each Eligible Security for which transaction reports are required to be submitted to such [national securities] association.” Further, the Participants are required to adopt Compliance Rules that require Industry Members, subject to their SRO jurisdiction, to report CAT Data.

See CAT NMS Plan, supra note 3, at Sections 6.3-6.4; Appendix D, at Section 2.1.

See id. at Sections 6.3(a), 6.4(a). The CAT NMS Plan also requires that the Operating Committee-approved format must be a format specified by the Plan Processor and Rule 613 compliant.

See id. at Section 6.3(b)(i) and Section 6.4(b)(i).

See id. at Section 6.3(b)(ii), Section 6.4(b)(ii), and Appendix C, Section A.1(a)(ii). Participants may voluntarily report CAT Data prior to the 8:00 a.m. Eastern Time deadline. Id. The CAT NMS Plan defines “Trading Day” as the date “as is determined by the Operating Committee.” The CAT NMS Plan also provides that “the Operating Committee may establish different Trading Days for NMS Stocks (as defined in SEC Rule 600(b)(47), Listed Options, OTC Equity Securities, and any other securities that are included as Eligible Securities from time to time.” Id. at Section 1.1.

See id. at Section 6.3(c)(i)-(ii) and Section 6.4(c)(i)-(ii).

The CAT NMS Plan defines the “Compliance Rule” to mean “with respect to a Participant, the rules promulgated by such Participant as contemplated by Section 3.11.” Id. at Section 1.1. Section 3.11 of the CAT NMS Plan provides that “each Participant shall comply with and enforce compliance, as required by SEC Rule 608(c), by its Industry Members with the provisions of SEC Rule 613 and of [the LLC Agreement], as applicable, to the Participant and its Industry Members. The Participants shall endeavor to promulgate consistent rules (after taking into account circumstances and considerations that may impact Participants differently) requiring compliance by their respective Industry Members with the provisions of SEC Rule 613 and [the LLC Agreement].” Id. at Section 3.11.

See id. at Section 6.4(c)(i)-(ii).

The CAT NMS Plan requires specific data elements of CAT Data that must be recorded and reported to the Central Repository upon: (i) “original receipt or origination of an order,” (ii) “routing of an order,” and (iii) “receipt of an order that has been routed.” Additionally, the CAT NMS Plan requires that a CAT Reporter must record and report data related to an “order [that] is modified or cancelled,” and an “order [that] is executed, in whole or in part,” as well as “other information or additional events as may be prescribed in Appendix D, Reporting and Linkage Requirements.” The CAT NMS Plan also requires Industry Member CAT Reporters to report additional data elements for (i) an “order [that] is executed, in whole or in part,” (ii) a “trade [that] is cancelled,” or (iii) “original receipt or origination of an order.” Further, each Participant shall, through Compliance Rules, require Industry Members to record and report to the Central Repository information or additional events as may be prescribed to accurately reflect the complete lifecycle of each Reportable Event.

For “original receipt or origination of an order,” the CAT NMS Plan specifies the following data elements: (i) Firm Designated ID(s) for each Customer; (ii) CAT-Order-ID; (iii) SRO-Assigned Market Participant Identifier of the Industry Member receiving or originating the order; (iv) date of order receipt or origination; (v) time of order receipt or origination (using time stamps pursuant to Section 6.8 of the CAT NMS Plan); and (vi) Material Terms of the Order. Id. at Section 6.3(d)(i).

For “routing of an order,” the CAT NMS Plan specifies the following data elements: (i) CAT-Order-ID; (ii) date on which the order is routed; (iii) time at which the order is routed (using time stamps pursuant to Section 6.8 of the CAT NMS Plan); (iv) SRO-Assigned Market Participant Identifier of the Industry Member or Participant routing the order; (v) SRO-Assigned Market Participant Identifier of the Industry Member or Participant to which the order is being routed; (vi) if routed internally at the Industry Member, the identity and nature of the department or desk to which the order is routed; and (vii) Material Terms of the Order. Id. at Section 6.3(d)(ii).

For “receipt of an order that has been routed,” the CAT NMS Plan specifies the following data elements: (i) CAT-Order-ID; (ii) date on which the order is received; (iii) time at which the order is received (using time stamps pursuant to Section 6.8); (iv) SRO-Assigned Market Participant Identifier of the Industry Member or Participant receiving the order; (v) SRO-Assigned Market Participant Identifier of the Industry Member or Participant routing the order; and (vi) Material Terms of the Order. Id. at Section 6.3(d)(iii).

For an “order [that] is modified or cancelled,” the CAT NMS Plan specifies the following data elements: (i) CAT-Order-ID; (ii) date the modification or cancellation is received or originated; (iii) time at which the modification or cancellation is received or originated (using time stamps pursuant to Section 6.8 of the CAT NMS Plan); (iv) price and remaining size of the order, if modified; (v) other changes in the Material Terms of the Order, if modified; and (vi) whether the modification or cancellation instruction was given by the Customer or was initiated by the Industry Member or Participant. Id. at Section 6.3(d)(iv).

For an “order [that] is executed, in whole or in part,” the CAT NMS Plan specifies the following data elements: (i) CAT-Order-ID; (ii) date of execution; (iii) time of execution (using time stamps pursuant to Section 6.8 of the CAT NMS Plan); (iv) execution capacity (principal, agency or riskless principal); (v) execution price and size; (vi) SRO-Assigned Market Participant Identifier of the Participant or Industry Member executing the order; and (vii) whether the execution was reported pursuant to an effective transaction reporting plan or the Plan for Reporting of Consolidated Options Last Sale Reports and Quotation Information. Id. at Section 6.3(d)(v).

See id. at Section 6.3(d)(vi).

For an “order [that] is executed, in whole or in part,” the CAT NMS Plan specifies the following additional data elements: (i) An Allocation Report; (ii) SRO-Assigned Market Participant Identifier of the clearing broker or prime broker, if applicable; and (iii) CAT-Order-ID of any contra-side order(s). Id. at Section 6.4(d)(ii)(A).

For a “trade [that] is cancelled,” the CAT NMS Plan specifies the following additional data element: A cancelled trade indicator. Id. at Section 6.4(d)(ii)(B).

For “original receipt or origination of an order,” the CAT NMS Plan specifies the following additional data element(s): The Firm Designated ID, Customer Account Information, and Customer Identifying Information for the relevant Customer. Id. at Section 6.4(d)(ii)(C).

Id. at Appendix D, Section 3.

Request for Comment

84. Do Commenters believe that the data recording, reporting, and formatting procedures described in the CAT NMS Plan are appropriate and reasonable? Would providing additional details or requirements on these procedures enhance the quality of CAT Data reported to the Central Repository or the efficiency and cost-effectiveness of the CAT?

85. Do Commenters believe that the CAT NMS Plan, including Appendix D thereto, requires sufficient outreach, support, training, guidance and/or documentation to ensure that CAT Reporters are able to make data transmissions to the Central Repository that are complete and timely? If not, please explain. Describe what, if any, further requirements may be needed.

86. Do Commenters believe that the CAT NMS Plan should have a formal communications plan, other than the public Web site, to provide CAT Reporters the information they would need in order to set-up or configure their systems to record and report CAT Data to the Central Repository? If so, how, when, and by whom should such information be disseminated to CAT Reporters?

87. Do Commenters believe the Plan should require a specific method for entering CAT Data upon each CAT Reportable Event or upon updates and corrections to CAT Reportable Events? If so, what method? Please explain.

88. Do Commenters believe that the CAT NMS Plan should include a requirement that the Participants and the Plan Processor set forth a more detailed schedule, with milestones, for CAT Reporters to adhere to in setting-up or configuring their systems to become CAT Data reporting compliant? If so, please explain and describe what details and milestones should be included in the schedule (e.g., publication of Technical Specifications and announcements of CAT Reporter-facing technology changes).

2. Timeliness of Data Reporting

Section 6.3(b)(ii) of the CAT NMS Plan requires each Participant to report Participant Data to the Central Repository by 8:00 a.m. Eastern Time on the Trading Day following the day the Participant records such data. Additionally, a Participant may voluntarily report such data prior to this deadline. Section 6.4(b)(ii) states that each Participant shall, through its Compliance Rule, require its Industry Members to report Recorded Industry Member Data to the Central Repository by 8:00 a.m. Eastern Time on the Trading Day following the day the Industry member records such data, and Received Industry Member Data to the Central Repository by 8:00 a.m. Eastern Time on the Trading Day following the day the Industry Member receives such data. Section 6.4(b)(ii) of the CAT NMS Plan also states that each Participant shall, through its Compliance Rule, permit its Industry Members to voluntarily report such data prior to the applicable 8:00 a.m. Eastern Time deadline.

See CAT NMS Plan, supra note 3, at Section 6.3(b)(ii); see also id. at Appendix C, Section A.1(a)(ii); Appendix D, Sections 3.1, 6.1.

Id. at Section 6.3(b)(ii).

Id. at Section 6.4(b)(ii).

Id.

Request for Comment

89. The CAT NMS Plan requires that all Participants report Participant Data to the Central Repository by 8:00 a.m. Eastern Time on the Trading Day following the day the Participant records such data, and that Industry Members report Recorded Industry Member Data to the Central Repository by 8:00 a.m. Eastern Time on the Trading Day following the day the Industry Member records such data and Received Industry Member Data to the Central Repository by 8:00 a.m. Eastern Time on the Trading Day following the day the Industry Member receives such data. Do Commenters believe that the CAT NMS Plan provides sufficient detail and information to determine whether the applicable 8:00 a.m. Eastern Time data reporting deadlines provided in the CAT NMS Plan are achievable? If not, why not?

Id. at Section 6.3(b)(ii).

Id. at Section 6.4(b)(ii).

Id.

90. Do Commenters believe that CAT Reporters will submit their reports at or about the same time? If all or most of the CAT Reporters would report at or just before 8:00 a.m. Eastern Time, what, if any, impact would there be on the necessary CAT infrastructure? Would this place an excessive burden on the Plan Processor? Do Commenters believe this would increase operational risk and/or increase costs? If so, please explain. Are there alternative reporting mechanisms that could reduce such risks?

91. The CAT NMS Plan provides that the Plan Processor must be able to handle two times the historical peak data to ensure that, if a significant number of CAT Reporters choose to submit data at or around the same time, the Plan Processor could handle the influx of data. Do Commenters believe that the SROs' estimate of capacity is sufficient? If not, why not and what capacity should be required?

Id. at Appendix C, Section A.1(a)(ii); see also id. at Section IV.H.2.g., infra.

92. Do Commenters believe that the CAT NMS Plan allocates, or requires the Plan Processor to have, sufficient resources to work with the approximately 1,800 CAT Reporters that would, under the CAT NMS Plan, have to establish secure connections over which CAT Data will flow from their systems to the Central Repository? Do Commenters believe that the Plan Processor could implement the CAT Reporters' Central Repository connections nearly simultaneously without compromising testing periods and implementation timelines?

3. Uniform Format

The CAT NMS Plan does not mandate the format in which data must be reported to the Central Repository. Appendix D states that the Plan Processor will determine the electronic format in which data must be reported, and that the format will be described in the Technical Specifications. Appendix C specifies that CAT Reporters could be required to report data either in a uniform electronic format, or in a manner that would allow the Central Repository to convert the data to a uniform electronic format, for consolidation and storage. Similarly, Sections 6.3(a) and 6.4(a) of the CAT NMS Plan require that CAT Reporters report data to the Central Repository in a format or formats specified by the Plan Processor, approved by the Operating Committee, and compliant with Rule 613.

See CAT NMS Plan, supra note 3, at Appendix C, Section D.12(f); see also id. at Appendix C, Section A.1(a).

Id. at Appendix D, Section 2.1. Appendix D states that more than one format may be allowed to support the various market participants that would report information to the Central Repository. Id.; see also id. at Section 6.9.

Id. at Appendix C, Section A.1(b).

Id. at Section 6.3(a) and Section 6.4(a).

The CAT NMS Plan requires that data reported to the Central Repository be stored in an electronic standard format. Specifically, Section 6.5(b)(i) of the CAT NMS Plan requires the Central Repository to retain the information collected pursuant to Rule 613(c)(7) and (e)(7) in a convenient and usable standard electronic data format that is directly available and searchable electronically without any manual intervention by the Plan Processor for a period of not less than six (6) years. Such data must be linked when it is made available to the Participant's regulatory Staff and the Commission.

Pursuant to the Plan, for data consolidation and storage, as noted above, such data must be reported in a uniform electronic format or in a manner that would allow the Central Repository to convert the data to a uniform electronic format. Id. at Appendix C, Section A.1(b).

Id. at Section 6.5(b)(i).

Id.

Request for Comment

93. The CAT NMS Plan provides that CAT Reporters could be required to report data either in a uniform electronic format, or in a manner that would allow the Central Repository to convert the data to a uniform electronic format, for consolidation and storage. Do Commenters believe that if data is reported to the Central Repository in a non-uniform format, the proposed CAT NMS Plan includes sufficient requirements or details to determine whether the Central Repository could reliably and accurately convert such data to a uniform electronic format, for consolidation and storage, without affecting the quality of the data? If not, what additional requirements or details should be provided in the CAT NMS Plan prior to the Commission's approval of such plan?

94. If Commenters believe that it is not necessary to provide additional requirements or details, if any, in the CAT NMS Plan, what additional requirements or details should be included in the Technical Specifications to determine whether the Central Repository could reliably and accurately convert such data to a uniform electronic format, for consolidation and storage?

95. Do Commenters believe the CAT NMS Plan's lack of a mandated uniform format in which data must be reported to the Central Repository would affect the accuracy of CAT Data collected and maintained under the CAT? If so, how? Would reporting data in a uniform format result in greater accuracy? If so, please explain.

96. Do Commenters believe the CAT NMS Plan's lack of a mandated uniform format in which data must be reported to the Central Repository would affect the completeness of CAT Data collected and maintained under the CAT? If so, how? Would reporting data in a uniform format result in more complete CAT Data? If so, please explain.

97. Do Commenters believe the CAT NMS Plan's lack of a mandated uniform format in which data must be reported to the Central Repository would affect the accessibility of CAT Data collected and maintained under the CAT? If so, how? Would reporting data in a uniform format result in a different level of accessibility? If so, please explain.

98. Do Commenters believe allowing CAT Reporters to report data to the Central Repository in a non-uniform format would affect the timeliness of data collected and maintained under the CAT? How would the requirement that the Central Repository convert non-uniform data to a uniform format affect the timeliness of the data collected and maintained under the CAT? Would reporting data in a uniform format result in a different level of timeliness of data reporting? If so, please explain.

99. Do Commenters believe that allowing CAT Reporters to report data to the Central Repository in a non-uniform format is more efficient and cost-effective than requiring data to be reported in a uniform format? Would allowing CAT Reporters to report data to the Central Repository in a non-uniform format merely transfer the costs from individual CAT Reporters to the Central Repository? Would centralization of the costs of converting data to a uniform format reduce costs? Please explain.

100. Do Commenters believe that allowing CAT Reporters to report data to the Central Repository in a non-uniform format would affect the security and confidentiality of CAT Data? If so, how? Would reporting data in a uniform format create different security or confidentiality concerns? If so, please explain.

4. Clock Synchronization

Pursuant to Section 6.8(a) of the CAT NMS Plan, each Participant and Industry Member, (through the Compliance Rule adopted by every Participant), must synchronize its Business Clocks, at a minimum, to within 50 milliseconds of the time maintained by the NIST, consistent with industry standards. The Participants believe that a 50-millisecond clock offset tolerance represents the current industry clock synchronization standard. Industry Members must maintain such a clock synchronization standard; certify periodically (according to a schedule to be defined by the Operating Committee) that their Business Clocks meet the requirements of the Compliance Rule; and report to the Plan Processor and the Participant any violation of the Compliance Rule pursuant to the thresholds set by the Operating Committee. Pursuant to Section 6.8(c) of the CAT NMS Plan, the Chief Compliance Officer, in conjunction with the Participants and other appropriate Industry Member advisory groups, annually must evaluate and make a recommendation to the Operating Committee as to whether the industry standard has evolved such that the clock synchronization standard should be tightened.

The CAT NMS Plan defines a “Business Clock” to mean “a clock used to record the date and time of any Reportable Event required to be reported under SEC Rule 613.” Id. at Section 1.1.

Id. at Section 6.8(a)(i)-(ii).

See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(c).

Id. at Section 6.8(a)(ii).

Id. at Section 6.8(c).

Appendix C describes the process by which Participants determined that a 50-millisecond clock offset tolerance was consistent with industry standards. To that end, the Participants and Industry Members reviewed their respective internal clock synchronization technology practices, and reviewed the results of The Financial Information Forum (“FIF”) Clock Offset Survey, a clock synchronization survey conducted by FIF. In light of their internal reviews and the FIF Clock Offset Survey, the Participants concluded that a clock offset tolerance of 50 milliseconds represented an aggressive but achievable standard.

Id. at Appendix C, Section D.12(p).

Id.

Id. at Appendix C, n.236. See Financial Information Forum, FIF Clock Offset Survey Preliminary Report (February 17, 2015), available at http://www.catnmsplan.com/industryfeedback/p602479.pdf and http://catnmsplan.com/web/groups/catnms/@catnms/documents/appsupportdocs/p602479.pdf . (“FIF Clock Offset Study”).

Id. The Participants note in Appendix C that according to the FIF Clock Offset Survey, annual maintenance costs would escalate to 102%, 123% and 242% if clock synchronization standards moved to 5 milliseconds, 1 millisecond and 100 microseconds, respectively, indicating that maintenance costs rapidly escalate as clock synchronization standards increase beyond 50 milliseconds. Id.

Appendix C discusses mechanisms to ensure compliance with the 50-millisecond clock offset tolerance. The Participants anticipate that they and Industry Members will adopt policies and procedures to verify the required clock synchronization each trading day before the market opens, as well as periodically throughout the trading day. The Participants also anticipate that they and Industry Members will document their clock synchronization procedures and maintain a log recording the time of each clock synchronization performed, and the result of such synchronization, specifically identifying any synchronization revealing any clock offset between the Participant's or Industry Member's Business Clock and the time maintained by the NIST exceeding 50 milliseconds. The CAT NMS Plan states that once both large and small broker-dealers begin reporting to the Central Repository, and as clock synchronization technology matures further, the Participants will assess, in accordance with Rule 613, tightening CAT's clock synchronization standards to reflect changes in industry standards.

See id. at Appendix C, Section A.3(c).

See id.

See id. It was noted that such a log would include results for a period of not less than five years ending on the then current date. Id.

See id. at Appendix C, Section D.12(p).

Request for Comment

See Sections IV.D.3, IV.E.4 and IV.H.5, infra, for further clock synchronization related requests for comment.

101. Do Commenters believe that a clock offset tolerance of 50 milliseconds is appropriate and reasonable, in light of the increase in the speed of trading over the last several years? If not, what would an appropriate and reasonable standard be?

102. What are current clock synchronization practices? Do Commenters believe that current industry clock synchronization practices are sufficiently rigorous in light of current trading speeds? If not, please explain.

103. Would a smaller clock offset tolerance be reasonably achievable? If so, please identify such tolerance and any incremental additional costs that achieving that smaller clock offset tolerance might entail.

104. If Commenters believe that, in light of the current speed of trading, the clock offset tolerance should be more rigorous, what, if any transition period would be reasonable and appropriate for reducing the clock offset tolerance standards of CAT?

105. What is the range of clock synchronization practices across the industry?

106. Do Commenters believe the range of clock synchronization practices should be considered when considering the appropriate clock synchronization standard?

107. If an SRO or broker-dealer can or does synchronize its clocks to an offset tolerance more rigorous than 50 milliseconds, do Commenters believe that that SRO or broker-dealer should be required to synchronize its clocks to that standard? Why or why not? If so, how, if at all, would that affect sequencing of Reportable Events in CAT?

108. Do Commenters believe that certain categories of market participants should be held to a smaller or larger clock offset tolerance? If so, what category of market participant and why? How, if at all, would that affect sequencing of Reportable Events in CAT?

109. Do Commenters believe a 50-millisecond clock offset tolerance would materially impair the quality and accuracy of CAT Data? If so, please explain. Would such a standard undermine the ability of the Central Repository to accurately and reliably link order and sequence event data across venues, or combine it with other sources of trade and order data? If so, please explain. Is there a benefit from applying the same uniform clock offset tolerance to all market participants, or would a variable clock offset tolerance approach be preferable? For example, should a high-volume market participant trading on multiple exchanges and ATSs have the same clock offset tolerance as a small retail-focused regional office? Would the benefits of a smaller clock offset tolerance for service bureaus that report but do not record order events be lower than for other types of CAT Reporters? Would the benefits of a smaller clock offset tolerance for clearing brokers that record and report information available only after an execution be lower than for other types of CAT Reporters? Please explain.

110. The CAT NMS Plan provides that as time synchronization standards evolve, the Participants would assess, on an annual basis, the ability to tighten the clock synchronization standards for CAT to reflect changes in industry standards. Do Commenters believe that this would establish an appropriately rigorous process and schedule for the Participants to evaluate whether the clock synchronization standard should be tightened? Are there any other factors that should affect when and how to tighten the clock synchronization standard?

111. Do Commenters believe the CAT NMS Plan provides adequate enforcement provisions to ensure CAT Reporters synchronize Business Clocks within the proposed 50-millisecond clock offset tolerance? If not, what additional enforcement provisions should the CAT NMS Plan provide?

112. Do Commenters believe that sufficient detail has been provided in the CAT NMS Plan concerning the reasonable justification or exceptional circumstances that would permit a pattern or practice of reporting events outside of the specified clock synchronization standard?

113. The CAT NMS Plan generally requires CAT Reporters to record and report Reportable Events with a time stamp of at least to the millisecond but provides for a 50 millisecond clock offset tolerance. Do Commenters believe the time stamp granularity requirement and the clock offset tolerance should correspond more closely or even identically? If so, please explain, including what such time stamp granularity requirement and clock offset tolerance should be.

5. Time Stamp Granularity

The CAT NMS Plan requires CAT Reporters to record and report the time of each Reportable Event using time stamps reflecting current industry standards, which should be at least to the millisecond, except with respect to events that involve non-electronic communication of information (“Manual Order Events”). Furthermore, the Plan requires Participants to adopt rules requiring that CAT Reporters that use time stamps in increments finer than milliseconds use those finer increments when reporting to the Central Repository. For Manual Order Events, the Participants determined that time stamp granularity at the level of a millisecond is not practical. Accordingly, the CAT NMS Plan provides that each Participant and Industry Member shall be permitted to record and report Manual Order Events to the Central Repository in increments up to and including one second, provided that Participants and Industry Members shall be required to record and report the time when a Manual Order Event has been captured electronically in an order handling and execution system of such Participant or Industry Member (“Electronic Capture Time”) in milliseconds.

See CAT NMS Plan, supra note 3, at Section 1.1. The SROs requested exemptive relief from Rule 613 so that the CAT NMS Plan may permit CAT Reporters to report Manual Order Events with a time stamp granularity of one second, in lieu of a time stamp granularity of one millisecond. See Exemptive Request Letter, supra note 16, at 34. The Commission granted exemptive relief on March 1, 2016 in order to allow this alternative to be included in the CAT NMS Plan and subject to notice and comment. See Exemption Order, supra note 18.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(c).

See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(c). The Participants state that they received industry feedback through the DAG that suggests that the established business practice with respect to Manual Order Events is to manually capture time stamps with granularity at the level of a second because finer increments cannot be accurately captured when dealing with manual processes which, by their nature, take longer to perform than a time increment of under one second. Id. The Participants agree that, due to the nature of transactions originated over the phone, it is not practical to attempt granularity finer than one second, as any such finer increment would be inherently unreliable. Id.

See CAT NMS Plan, supra note 3, at Section 6.8(b).

Request for Comment

See Section IV.D.3, infra, for further time stamp granularity related requests for comment.

114. Are the time stamp granularity standards for both electronic and non-electronic reportable events appropriate and reasonable? If not, why not and what would be a better alternative?

115. Do Commenters believe the CAT NMS Plan's time stamp granularity requirement is precise enough to reliably and accurately sequence Reportable Events? If not, why not? Is there a better time stamp approach and what should the requirement(s) be?

116. To what degree does the millisecond or less time stamp granularity requirement enable or prevent regulators' ability to sequence events that occur in different execution venues? Please explain.

117. Are certain CAT Reportable Events more time-sensitive than other CAT Reportable Events? If so, what events are more time-sensitive and why? What systems are more likely to process these more sensitive events and to what level of time stamp granularity are such events processed? Where are those systems located (i.e., within broker-dealers, service bureaus, execution venues)? Please explain.

118. What market participant systems, if any, should have less granular time stamp requirements? Why? What time stamp granularity standard should these systems have? Why?

119. What market participant systems, if any, should have more granular time stamp requirements? Why? What time stamp granularity standard should these systems have? Why?

120. The Commission granted an exemption from Rule 613 in order to allow the alternative of permitting CAT Reporters to report Manual Order Events with a time stamp granularity of one second, in lieu of the Rule 613 requirement that the CAT NMS Plan require CAT Reporters to report with a time stamp granularity of one millisecond, to be included in the CAT NMS Plan and subject to notice and comment. Do Commenters believe that the CAT NMS Plan's one-second time stamp granularity standard for Manual Order Events is appropriate and reasonable? If not, why not? Would a more granular time stamp requirement for Manual Order Events be feasible?

See Exemption Order, supra note 18.

121. What alternative approach with respect to Manual Order Events may be preferable? Could the provisions in the CAT NMS Plan related to Manual Order Events be more narrowly tailored to, for example, only apply to CAT Reporters who are unable to record and report Manual Order Events with a time stamp granularity of one millisecond?

122. The SROs note in the Exemption Request that recording and reporting Manual Order Events with a time stamp granularity of at least one second would result in little additional benefit, and, in fact, could result in adverse consequences such as creating a false sense of precision for data that is inherently imprecise, while imposing additional costs on CAT Reporters. Do Commenters agree? Why or why not?

123. If Manual Order Events are recorded and reported with a time stamp granularity of one second, what, if any, challenges do Commenters believe would arise with respect to the sequencing of order events (for the same order) and orders (for a series of orders)? Would the one millisecond standard originally provided for in Rule 613 be preferable? Please explain.

124. Do Commenters believe the CAT NMS Plan's requirement that time stamp granularity (other than for Manual Order Events) should be to at least the millisecond is granular enough in light of current practices? If not, why not?

125. The CAT NMS Plan provides that as time stamp standards evolve, the Participants would assess, on an annual basis, the ability to require more precise time stamp granularity standards for CAT to reflect changes in industry standards. Do Commenters believe that this establishes an appropriately rigorous schedule for the Participants to evaluate whether time stamp granularity requirements could potentially be set to finer increments? Are there any other factors that should affect when and how the requirements for time stamp granularity increments could be made more precise?

126. Do Commenters believe the CAT NMS Plan provides adequate enforcement provisions to ensure CAT Reporters time stamp Reportable Events to a granularity of one millisecond (and for Manual Order Events to a granularity of one second)? If not, what additional enforcement provisions should the CAT NMS Plan provide?

127. Do Commenters believe that the CAT NMS Plan's requirement that Participants and Industry Members synchronize Business Clocks used solely for Manual Order Events to within one second of the time maintained by the NIST is appropriate and reasonable? Would a tighter clock synchronization standard for Business Clocks used solely for Manual Order Events be feasible?

6. CAT-Reporter-ID

Sections 6.3 and 6.4 of the CAT NMS Plan require CAT Reporters to record and report to the Central Repository an SRO-Assigned Market Participant Identifier for orders and certain Reportable Events to be used by the Central Repository to assign a unique CAT-Reporter-ID for purposes of identifying each CAT Reporter associated with an order or Reportable Event (the “Existing Identifier Approach”). The CAT NMS Plan requires the reporting of SRO-Assigned Market Participant Identifiers of: The Industry Member receiving or originating an order; the Industry Member or Participant from which (and to which) an order is being routed; the Industry Member or Participant receiving (and routing) a routed order; the Industry Member or Participant executing an order, if an order is executed; and the clearing broker or prime broker, if applicable, if an order is executed. An Industry Member would report to the Central Repository its existing SRO-Assigned Market Participant Identifier used by the relevant SRO specifically for transactions occurring at that SRO. Similarly, an exchange reporting CAT Reporter information would report data using the SRO-Assigned Market Participant Identifier used by the Industry Member on that exchange or its systems. Over-the-counter (“OTC”) orders and Reportable Events would be reported with an Industry Member's FINRA SRO-Assigned Market Participant Identifier.

The CAT NMS Plan defines an “SRO-Assigned Market Participant Identifier” as “an identifier assigned to an Industry Member by an SRO or an identifier used by a Participant.” See CAT NMS Plan, supra note 3, at Section 1.1.

Rule 613 defines a CAT-Reporter-ID as “a code that uniquely and consistently identifies [a CAT Reporter] for purposes of providing data to the central repository.” 17 CFR 242.613(j)(2).

The SROs requested exemptive relief from Rule 613 so that the CAT NMS Plan may permit the Existing Identifier Approach, which would allow a CAT Reporter to report an existing SRO-Assigned Market Participant Identifier in lieu of requiring the reporting of a universal CAT-Reporter-ID. See Exemptive Request Letter, supra note 16, at 19. The Commission granted exemptive relief on March 1, 2016 in order to allow this alternative to be included in the CAT NMS Plan and subject to notice and comment. See Exemption Order, supra note 18.

See CAT NMS Plan, supra note 3, at Section 6.3(d)(i) and Section 6.4(d)(i).

Id. at Section 6.3(d)(ii) and Section 6.4(d)(i).

Id. at Section 6.3(d)(iii) and Section 6.4(d)(i).

Id. at Section 6.3(d)(v) and Section 6.4(d)(i).

Id. at Section 6.4(d)(ii)(A)(2). Industry Members are required by the CAT NMS Plan to record and report this information. See CAT NMS Plan, supra note 3, at Section 6.4(d)(ii).

See Exemption Order, supra note 18, at 31-41.

See id. at 20.

Id.

The CAT NMS Plan requires the Plan Processor to develop and maintain the mechanism to assign (and to change, if necessary) CAT-Reporter-IDs. For the Central Repository to link the SRO-Assigned Participant Identifier to the CAT-Reporter-ID, each SRO must submit, on a daily basis, all SRO-Assigned Market Participant Identifiers used by its Industry Members (or itself), as well as information to identify the corresponding market participant (for example, a CRD number or Legal Entity Identifier (“LEI”)) to the Central Repository. Additionally, each Industry Member shall be required to submit to the Central Repository information sufficient to identify such Industry Member (e.g., CRD number or LEI, as noted above). The Plan Processor would use the SRO-Assigned Market Participant Identifiers and identifying information (i.e., CRD number or LEI) to assign a CAT-Reporter-ID to each Industry Member and SRO for internal use across all data within the Central Repository. The Plan Processor would create and maintain a database in the Central Repository that would map the SRO-Assigned Market Participant Identifiers to the appropriate CAT-Reporter-ID.

See CAT NMS Plan, supra note 3, at Appendix D, Section 10.1. Changes to CAT-Reporter-IDs must be reviewed and approved by the Plan Processor. Id. The CAT NMS Plan also requires the Central Repository to generate and assign a unique CAT-Reporter-ID to all reports submitted to the system based on sub-identifiers that are currently used by CAT Reporters in their order handling and trading processes (described in the Exemption Request as SRO-assigned market participant identifiers). See CAT NMS Plan, supra note 3, at Appendix D, Section 3; see also Exemption Order, supra note 18, at 31-41.

See CAT NMS Plan, supra note 3, at Section 6.3(e)(i).

Id. at Section 6.4(d)(vi).

See Exemption Order, supra note 18, at 31-41.

Id. at 20.

The consolidated audit trail must be able to capture, store, and maintain current and historical SRO-Assigned Market Participant Identifiers. The SRO-Assigned Market Participant Identifier must also be included on the Plan Processor's acknowledgment of its receipt of data files from a CAT Reporter or Data Submitter, on daily statistics provided by the Plan Processor after the Central Repository has processed data, and on a secure Web site that the Plan Processor would maintain that would contain each CAT Reporter's daily reporting statistics. In addition, data validations by the Plan Processor must include confirmation of a valid SRO-Assigned Market Participant Identifier.

See CAT NMS Plan, supra note 3, at Appendix D, Section 2.

See id. at Appendix D, Section 7.1.

See id. at Appendix D, Section 7.2.

See id. at Appendix D, Section 10.1.

See id. at Appendix D, Section 7.2. The CAT NMS Plan also notes that both the CAT-Reporter-ID and the SRO-Assigned Market Participant Identifier would be data fields for the online targeted query tool described in the CAT NMS Plan as providing authorized users with the ability to retrieve processed and/or validated (unlinked) data via an online query screen. See id. at Appendix D, Section 8.1.1.

Request for Comment

128. The Commission granted an exemption from Rule 613 in order to allow the Existing Identifier Approach to be included in the CAT NMS Plan and subject to notice and comment. The Existing Identifier Approach would allow a CAT Reporter to report an existing SRO-Assigned Market Participant Identifier in lieu of Rule 613's requirement that a CAT Reporter must report a universal CAT-Reporter-ID. Do Commenters believe that allowing the Existing Identifier Approach would be more efficient and cost-effective than the Rule 613 approach of requiring a CAT-Reporter-ID to be reported for each order and reportable event in accordance with Rule 613(c)(7)? Why or why not? Or do Commenters believe that the Rule 613 approach is preferable? Why or why not? Would implementation of the Existing Identifier Approach merely transfer costs from CAT Reporters to the Central Repository?

See Exemption Order, supra note 18.

See supra note 142.

129. Do Commenters believe that the Existing Identifier Approach would affect the accuracy of CAT Data? Would the Rule 613 approach result in greater accuracy? If so, please explain.

130. Do Commenters believe that the CAT NMS Plan's proposed Existing Identifier Approach would affect the accessibility of CAT Data? If so, how? Would the Rule 613 approach result in a different level of accessibility? If so, please explain.

131. Do Commenters believe that the CAT NMS Plan's proposed Existing Identifier Approach would affect the timeliness of CAT Data? If so, how? Would the Rule 613 approach result in greater timeliness? If so, please explain.

132. Do Commenters believe the Existing Identifier Approach would affect the security and confidentiality of CAT Data? If so, how? Would the Rule 613 approach result in a different level of security and confidentiality? If so, please explain.

133. What challenges or risks do Commenters believe the Plan Processor would face in linking all SRO-Assigned Market Participant Identifiers to the appropriate CAT-Reporter-IDs? What, if anything, could be done to mitigate those challenges and risks?

134. The CAT NMS Plan does not require that an Industry Member provide its LEI to the Plan Processor as part of the identifying information used to assign a CAT-Reporter-ID. The CAT NMS Plan permits an Industry Member to report its CRD number in lieu of its LEI for this purpose. Do Commenters believe that the CAT NMS Plan should mandate that Industry Members provide their LEIs, along with their SRO-Assigned Market Participant Identifiers, to the Plan Processor for purposes of developing a unique CAT-Reporter-ID? Why or why not?

7. Customer-ID

a. Customer Information Approach

Rule 613(c)(7)(i)(A) requires that for the original receipt or origination of an order, a CAT Reporter report the “Customer-ID(s) for each Customer.” “Customer-ID” is defined in Rule 613(j)(5) to mean “with respect to a customer, a code that uniquely and consistently identifies such customer for purposes of providing data to the Central Repository.” Rule 613(c)(8) requires that “[a]ll plan sponsors and their members shall use the same Customer-ID and CAT-Reporter-ID for each customer and broker-dealer.”

In Appendix C, the Participants describe the “Customer Information Approach,” an alternative approach to the requirement that a broker-dealer report a Customer-ID for every Customer upon original receipt or origination of an order. Under the Customer Information Approach, the CAT NMS Plan would require each broker-dealer to assign a unique Firm Designated ID to each Customer. As the Firm Designated ID, broker-dealers would be permitted to use an account number or any other identifier defined by the firm, provided each identifier is unique across the firm for each business date (i.e., a single firm may not have multiple separate customers with the same identifier on any given date). According to the CAT NMS Plan, broker-dealers would submit an initial set of Customer information to the Central Repository, including, as applicable, the Firm Designated ID, the Customer's name, address, date of birth, individual tax payer identifier number (“ITIN”)/social security number (“SSN”), individual's role in the account (e.g., primary holder, joint holder, guardian, trustee, person with power of attorney) and LEI, and/or Large Trader ID (“LTID”), if applicable, which would be updated as set forth in the CAT NMS Plan.

The SROs requested exemptive relief from Rule 613 so that the CAT NMS Plan may permit the Customer Information Approach, which would require each broker-dealer to assign a unique Firm Designated ID to each trading account and to submit an initial set of information identifying the Customer to the Central Repository, in lieu of requiring each broker-dealer to report a Customer-ID for each Customer upon the original receipt or origination of an order. See Exemptive Request Letter, supra note 16, at 12. The Commission granted exemptive relief on March 1, 2016 in order to allow this alternative to be included in the CAT NMS Plan and subject to notice and comment. See Exemption Order, supra note 18.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.1(a)(iii).

Id. at Appendix C, Section A.1(a)(iii). The CAT NMS Plan defines a “Firm Designated ID” as “a unique identifier for each trading account designated by Industry Members for purposes of providing data to the Central Repository, where each such identifier is unique among all identifiers from any given Industry Member for each business date.” See id. at Section 1.1.

Id. at Appendix C, Section A.1(a)(iii).

The CAT NMS Plan provides that where a validated LEI is available for a Customer or entity, this may obviate a need to report other identifier information (e.g., Customer name, address, EIN). Id. at Appendix C, Section A.1(a)(iii) n.31.

The CAT NMS Plan states that the Participants anticipate that Customer information that is initially reported to the CAT could be limited to Customer accounts that have, or are expected to have, CAT Reportable Event activity. For example, the CAT NMS Plan notes accounts that are considered open, but have not traded Eligible Securities in a given time frame, may not need to be pre-established in the CAT, but rather could be reported as part of daily updates after they have CAT Reportable Event activity. Id. at Appendix C, Section A.1(a)(iii) n.32.

Under the Customer Information Approach, broker-dealers would be required to report only the Firm Designated ID for each new order submitted to the Central Repository, rather than the “Customer-ID” as defined by Rule 613(c)(j)(5) and as required by Rule 613(c)(7)(i)(A), and the Plan Processor would associate specific Customers and their Customer-IDs with individual order events based on the reported Firm Designated IDs. Within the Central Repository, each Customer would be uniquely identified by identifiers or a combination of identifiers such as an ITIN/SSN, date of birth, and, as applicable, LEI and LTID. The Plan Processor would be required to use these unique identifiers to map orders to specific Customers across all broker-dealers. To ensure information identifying a Customer is updated, broker-dealers would be required to submit to the Central Repository daily updates for reactivated accounts, newly established or revised Firm Designated IDs, or associated reportable Customer information.

See id. at Appendix C, Section A.1(a)(iii). The CAT NMS Plan also requires broker-dealers to report “Customer Account Information” upon the original receipt of origination of an order. See CAT NMS Plan, supra note 3, at Section 1.1, Section 6.4(d)(ii)(C).

See CAT NMS Plan, supra note 3, at Appendix C, Section A.1(a)(iii).

Id.

The CAT NMS Plan notes that because reporting to the CAT is on an end-of-day basis, intra-day changes to information could be captured as part of the daily updates to the information. To ensure the completeness and accuracy of Customer information and associations, in addition to daily updates, broker-dealers would be required to submit periodic full refreshes of Customer information to the CAT. The scope of the “full” Customer information refresh would need to be further defined, with the assistance of the Plan Processor, to determine the extent to which inactive or otherwise terminated accounts would need to be reported. Id. at Appendix C, Section A.1(a)(iii) n.33.

Appendix C provides additional requirements that the Plan Processor must meet under the Customer Information Approach. The Plan Processor must maintain information of sufficient detail to uniquely and consistently identify each Customer across all CAT Reporters, and associated accounts from each CAT Reporter, and must document and publish, with the approval of the Operating Committee, the minimum list of attributes to be captured to maintain this association. In addition, the Plan Processor must maintain valid Customer and Customer Account Information for each trading day and provide a method for Participants and the Commission to easily obtain historical changes to that information (e.g., name changes, address changes). The Plan Processor also must design and implement a robust data validation process for submitted Firm Designated IDs, Customer Account Information and Customer Identifying Information, and be able to link accounts that move from one CAT Reporter to another due to mergers and acquisitions, divestitures, and other events. Under the Customer Information Approach, broker-dealers will initially submit full account lists for all active accounts to the Plan Processor and subsequently submit updates and changes on a daily basis. Finally, the Plan Processor must have a process to periodically receive full account lists to ensure the completeness and accuracy of the account database.

See id. at Appendix C, Section A.1(a)(iii).

Id. Section 9.1 of Appendix D also addresses, among other things, the minimum attributes that CAT must capture for Customers and the validation process for such attributes. Id. at Appendix D, Section 9.1.

Id. at Appendix D, Section 9.1. In relevant part, “Customer Account Information” is defined in the Plan to include, but not be limited to, account number, account type, customer type, date account opened, and large trader identifier (if applicable). See id. at Section 1.1.

See id. at Appendix C, Section A.1(a)(iii).

Id. at Appendix C, Section A.1(a)(iii). The CAT NMS Plan defines “Customer Identifying Information” to mean “information of sufficient detail to identify a Customer, including, but not limited to, (a) with respect to individuals: Name, address, date of birth, individual tax payer identification number (“ITIN”)/social security number (“SSN”), individual's role in the account (e.g., primary holder, joint holder, guardian, trustee, person with the power of attorney); and (b) with respect to legal entities: name, address, Employer Identification Number (“EIN”)/LEI) or other comparable common entity identifier, if applicable; provided, however, where the LEI or other common entity identifier is provided, information covered by such common entity identifier (e.g., name, address) would not need to be separately submitted to the Central Repository.” See id. at Section 1.1.

Id. at Appendix C, Section A.1(a)(iii).

Id.

b. Account Effective Date vs. Account Open Date

Rule 613(c)(7)(viii)(B) requires broker-dealers to report to the Central Repository “Customer Account Information” upon the original receipt or origination of an order. The CAT NMS Plan defines “Customer Account Information” to include, in part, the Customer's account number, account type, customer type, date account opened and LTID (if applicable). The Plan, however, provides that in two limited circumstances, a broker-dealer could report the “Account Effective Date” in lieu of the date an account was opened. The first circumstance is where a relationship identifier—rather than an actual parent account—has been established for an institutional Customer relationship. In this case, no account open date is available for the institutional Customer parent relationship because there is no parent account, and for the same reason, there is no account number or account type available. Thus, the Plan provides that in this circumstance, a broker-dealer could report the “Account Effective Date” of the relationship in lieu of an account open date. Further, the Plan provides that where such an institutional Customer relationship was established before the broker-dealer's obligation to report audit trail data is required, the “Account Effective Date” would be either (i) the date the broker-dealer established the relationship identifier, or (ii) the date when trading began (i.e., the date the first order is received) using the relevant relationship identifier, and if both dates are available and differ, the earlier date. Where such relationships are established after the broker-dealer's obligation to report audit trail data is required, the “Account Effective Date” would be the date the broker-dealer established the relationship identifier and would be no later than the date the first order was received. Regardless of when the relationship was established for such institutional Customers, the Plan provides that broker-dealers may report the relationship identifier in place of Rule 613(c)(7)(viii)(B)'s requirement to report the “account number,” and report “relationship” in place of “account type.”

17 CFR 242.613(c)(7)(viii)(B). “Customer Account Information” is defined in Rule 613(j)(4) to “include, but not be limited to, account number, account type, customer type, date account opened, and large trader identifier (if applicable).” 17 CFR 242.613(j)(4).

See CAT NMS Plan, supra note 3, at Section 1.1.

Id. The SROs requested exemptive relief from Rule 613 so that the CAT NMS Plan may permit broker-dealers to report to the Central Repository the “effective date” of an account in lieu of requiring each broker-dealer to report the date the account was opened in certain limited circumstances. See Exemptive Request Letter, supra note 16, at 13. The Commission granted exemptive relief on March 1, 2016 in order to allow this alternative to be included in the CAT NMS Plan and subject to notice and comment. See Exemption Order, supra note 18.

See Exemption Order, supra note 18; see also September 2015 Supplement, supra note 16, at 4-5.

See September 2015 Supplement, supra note 16, at 6.

See CAT NMS Plan, supra note 3, at Section 1.1.

See id.

See id.

See id.

The second circumstance where a broker-dealer may report the “Account Effective Date” rather than the date an account was opened as required in Rule 613(c)(7)(viii)(B) is when particular legacy system data issues prevent a broker-dealer from providing an account open date for any type of account (i.e., institutional, proprietary or retail) that was established before CAT's implementation. According to the Plan, these legacy system data issues may arise because:

See id.; see also September 2015 Supplement, supra note 16, at 7-9.

(1) A broker-dealer has switched back office providers or clearing firms and the new back office/clearing firm system identifies the account open date as the date the account was opened on the new system;

(2) A broker-dealer is acquired and the account open date becomes the date that an account was opened on the post-merger back office/clearing firm system;

(3) Certain broker-dealers maintain multiple dates associated with accounts in their systems and do not designate in a consistent manner which date constitutes the account open date, as the parameters of each date are determined by the individual broker-dealer; or

(4) No account open date exists for a proprietary account of a broker-dealer.

See CAT NMS Plan, supra note 3, at Section 1.1.

Thus, when legacy systems data issues arise due to one of the four reasons above and no account open date is available, the Plan provides that broker-dealers would be permitted to report an “Account Effective Date” in lieu of an account open date. When the legacy systems data issues and lack of account open date are attributable to above reasons (1) or (2), the “Account Effective Date” would be the date the account was established, either directly or via a system transfer, at the relevant broker-dealer. When the legacy systems data issues and lack of account open date are attributable to above reason (3), the “Account Effective Date” would be the earliest available date. When the legacy systems data issues and lack of account open date are attributable to above reason (4), the “Account Effective Date” would be (i) the date established for the proprietary account in the broker-dealer or its system(s), or (ii) the date when proprietary trading began in the account, i.e., the date on which the first orders were submitted from the account.

Id.

Id.

Id.

Id.

c. Modification/Cancellation

Rule 613(c)(7)(iv)(F) requires that “[t]he CAT-Reporter-ID of the broker-dealer or Customer-ID of the person giving the modification or cancellation instruction” be reported to the Central Repository. Because the Customer Information Approach no longer requires that a Customer-ID be reported upon original receipt or origination of an order, and because reporting the Customer-ID of the specific person that gave the modification or cancellation instruction would result in an inconsistent level of information regarding the identity of the person giving the modification or cancellation instruction versus the identity of the Customer that originally received or originated an order, Section 6.3(d)(iv)(F) of the CAT NMS Plan modifies the requirement in Rule 613 and instead requires CAT Reporters to report whether the modification or cancellation instruction was “given by the Customer or was initiated by the Industry Member or Participant.”

17 CFR 242.613(c)(7)(iv)(F) (emphasis added).

See CAT NMS Plan, supra note 3, at Section 6.3(d)(iv)(F). The SROs requested exemptive relief from Rule 613 so that the CAT NMS Plan may permit CAT Reporters to report whether a modification or cancellation instruction was given by the Customer associated with the order, or was initiated by the broker-dealer or exchange associated with the order, in lieu of requiring CAT Reporters to report the Customer-ID of the person giving the modification or cancellation instruction. See Exemptive Request Letter, supra note 16, at 12-13. The Commission granted exemptive relief on March 1, 2016 in order to allow this alternative to be included in the CAT NMS Plan and subject to notice and comment. See Exemption Order, supra note 18.

Request for Comment

135. The Commission granted an exemption from Rule 613 in order to allow the Customer Information Approach to be included in the CAT NMS Plan and subject to notice and comment. The Customer Information Approach would require each broker-dealer to assign a unique Firm Designated ID to each trading account and to submit an initial set of information identifying the Customer to the Central Repository, in lieu of Rule 613's requirement that a CAT Reporter must report a Customer-ID for each Customer upon the original receipt or origination of an order. Do Commenters believe that allowing broker-dealers to report a Firm Designated ID to the Central Repository is more efficient and cost-effective than the Rule 613 approach of requiring broker-dealers to report a unique Customer-ID upon original receipt or origination of an order? Would allowing CAT Reporters to report a Firm Designated ID to the Central Repository merely transfer the costs from individual broker-dealers to the Central Repository? Or do Commenters believe that the Rule 613 approach is preferable? Why or why not?

136. If broker-dealers are permitted to report a Firm Designated ID, do Commenters believe the proposed CAT NMS Plan includes sufficiently detailed requirements to determine whether the Plan Processor could use the Firm Designated ID to identify a Customer?

137. Do Commenters believe the CAT NMS Plan's proposal to permit reporting a Firm Designated ID would affect the accuracy of CAT Data collected and maintained under the CAT compared to the Rule 613 approach that requires a unique Customer-ID? If so, how? Would permitting reporting a Firm Designated ID result in more complete CAT Data? If so, please explain.

138. Do Commenters believe the CAT NMS Plan's proposal to permit reporting a Firm Designated ID would affect the accessibility of CAT Data collected and maintained under the CAT compared to the Rule 613 approach? If so, how? Would permitting reporting a Firm Designated ID result in CAT Data being more accessible? If so, please explain.

139. Do Commenters believe allowing broker-dealers to report a Firm Designated ID to the Central Repository would affect the timeliness of data collected and maintained under the CAT compared to the Rule 613 approach? Would permitting reporting a Firm Designated ID result in more timely CAT Data? If so, please explain.

140. Do Commenters believe there are any increased risks related to allowing a broker-dealer to report a Firm Designated ID rather than a unique Customer-ID to the Central Repository? How difficult would it be for the Central Repository to utilize a Firm Designated ID for each account?

141. Do Commenters believe that the CAT NMS Plan has provided sufficient information to determine whether the Central Repository could use a Firm Designated ID to efficiently, reliably and accurately link orders and Reportable Events to a Customer?

142. Do Commenters believe that the CAT NMS Plan includes sufficient safeguards or policies to assure that the same Firm Designated ID would not be used for multiple Customers?

143. The CAT NMS Plan does not require that a broker-dealer provide an LEI to the Plan Processor as part of the identifying information used to assign a Customer-ID at the Central Repository. The CAT NMS Plan provides that a broker-dealer must report its LEI, if available, but allows a broker-dealer to report another comparable common entity identifier, if an LEI is not available. Do Commenters believe that the CAT NMS Plan should mandate that broker-dealers provide an LEI as part of the information used by the Plan Processor to uniquely identify Customers? Why or why not?

144. Do Commenters believe that reporting the Firm Designated ID, rather than a unique Customer-ID, would affect the security and confidentiality of CAT Data? If so, how? Would permitting reporting a Firm Designated ID result in a different level of security and confidentiality of CAT Data? If so, please explain.

145. The CAT NMS Plan provides that an initial set of Customer Account Information and Customer Identifying Information would be reported to the Central Repository by broker-dealers upon the commencement of reporting audit trail data to the Central Repository by that broker-dealer, and that such Customer Identifying Information would be updated as set forth in the CAT NMS Plan. Do Commenters believe that the approach for reporting an initial set of Customer Account Information and Customer Identifying Information and updates to such information thereafter as set forth in the CAT NMS Plan would affect the quality, accuracy, completeness, accessibility or timeliness of the data? If so, what additional requirements or details should be provided in the CAT NMS Plan?

146. Do Commenters believe that allowing broker-dealers to report an initial set of Customer Account Information and Customer Identifying Information and updates to such information thereafter is more efficient and cost-effective than the Rule 613 approach for identifying Customers under Rule 613? Or do Commenters believe that the Rule 613 approach is preferable? Why or why not?

147. Do Commenters believe there are any increased risks as a result of allowing a broker-dealer to report an initial set of Customer Account Information and Customer Identifying Information and updates to such information thereafter to be reported to the Central Repository? How difficult would it be for the Central Repository to ingest the Customer Account Information and Customer Identifying information, and any updates thereafter?

148. Do Commenters believe that the CAT NMS Plan provides sufficient information to determine whether the Central Repository could use the initial set of Customer Account Information and Customer Identifying Information and updates to such information thereafter to efficiently, reliably and accurately link orders and Reportable Events to a Customer?

149. Do Commenters believe that reporting an initial set of Customer Account Information and Customer Identifying Information and updates to such information thereafter would affect the security and confidentiality of CAT Data? If so, how? Would reporting an initial set of Customer Account Information and Customer Identifying Information and updates to such information result in a different level of security and confidentiality? If so, please explain.

150. As part of the Customer Identifying Information reported to the Central Repository, the CAT NMS Plan requires a broker-dealer to report PII such as the Customer's name, address, date of birth, and ITIN/SSN. Do Commenters believe there is data that could be reported by broker-dealers and used by the Central Repository to identify Customers that is not PII? What types of data would this be? If data other than PII is used to identify a Customer, do Commenters believe that such data would be sufficiently unique to ensure that Customers can be accurately identified by the Central Repository?

151. If data other than PII is used by the Central Repository to identify a Customer, would the use of such data affect the quality or completeness of the CAT audit trail, as compared to the use of PII to identify a Customer?

152. Do Commenters believe that if broker-dealers reported data other than PII to identify Customers, the accessibility and timeliness of the data collected and maintained under the CAT would be affected? If the data would be affected, in what way(s)?

153. Would relying on data other than PII to identify a Customer be a more efficient and cost-effective way to identify Customers, as compared to relying on PII to identify a Customer?

154. Do Commenters believe that there would be increased risks to the reliability of the CAT audit trail data if broker-dealers were required to identify a Customer with data that does not include PII?

155. If broker-dealers report data other than PII to identify Customers, do Commenters believe that the Central Repository could efficiently, reliably and accurately link orders and Reportable Events to a Customer?

156. Do Commenters believe that the proposed CAT NMS Plan provides sufficient information to determine when broker-dealers would report the “Account Effective Date”, rather than the date the Customer's account was opened as required by Rule 613? Is there any ambiguity in the circumstances under which a broker-dealer would report an “Account Effective Date” rather than the date a Customer's account was opened?

157. Do Commenters believe reporting of the “Account Effective Date” rather than the account open date for a Customer's account under the Rule 613 approach would affect the quality, accuracy, completeness, accessibility or timeliness of the CAT data? If it does, what additional requirements or details should be provided in the CAT NMS Plan prior to the Commission's approval of such Plan? Or do Commenters believe that the Rule 613 approach is preferable? Why or why not?

158. Do Commenters believe that reporting the “Account Effective Date” would provide sufficient information to the Central Repository to facilitate the ability of the Plan Processor to link a Customer's account with the Customer?

159. Do Commenters believe that allowing the reporting of the “Account Effective Date” would be more efficient and cost-effective than requiring the Rule 613 approach of reporting of a Customer's account open date? Or do Commenters believe that the Rule 613 approach is preferable? Why or why not? Would allowing CAT Reporters to report the “Account Effective Date” rather than the date a Customer's account was opened merely transfer the costs from individual CAT Reporters to the Central Repository?

160. Do Commenters agree that the proposed approach for reporting the “Account Effective Date,” which differs depending on whether the account was established before or after the commencement of reporting audit trail data to the Central Repository as set forth in the CAT NMS Plan, is a reasonable approach? Why or why not?

161. The Commission granted an exemption from Rule 613 to permit the alternative of allowing CAT Reporters to report whether the modification or cancellation of an order was given by a Customer, or initiated by a broker-dealer or exchange, in lieu of requiring the reporting of the Customer-ID of the person giving the modification or cancellation instruction, to be included in the CAT NMS Plan and subject to notice and comment. To what extent does the approach permitted by the exemption affect the completeness of the CAT? Would the information lost under the approach permitted by the exemption affect investigations or surveillances? If so, how?

8. Order Allocation Information

Section 6.4(d)(ii)(A)(1) of the CAT NMS Plan provides that each Participant through its Compliance Rule must require that Industry Members record and report to the Central Repository an Allocation Report that includes the Firm Designated ID when an execution is allocated in whole or part. The CAT NMS Plan defines an Allocation Report as “a report made to the Central Repository by an Industry Member that identifies the Firm Designated ID for any account(s), including subaccount(s), to which executed shares are allocated and provides the security that has been allocated, the identifier of the firm reporting the allocation, the price per share of shares allocated, the side of shares allocated, the number of shares allocated to each account, and the time of the allocation.” The CAT NMS Plan explains, for the avoidance of doubt, that an Allocation Report shall not be required to be linked to particular orders or executions.

See CAT NMS Plan, supra note 3, at Section 6.4(d)(ii)(A)(1); see also April 2015 Supplement, supra note 16. The SROs requested exemptive relief from Rule 613 so that the CAT NMS Plan may permit Industry Members to record and report to the Central Repository an Allocation Report that includes the Firm Designated ID when an execution is allocated in whole or part in lieu of requiring the reporting of the account number for any subaccount to which an execution is allocated, as is required by Rule 613. See Exemptive Request Letter, supra note 16, at 26-27. The Commission granted exemptive relief on March 1, 2016 in order to allow this alternative to be included in the CAT NMS Plan and subject to notice and comment. See Exemption Order, supra note 18.

See CAT NMS Plan, supra note 3, at Section 1.1; see also April 2015 Supplement, supra note 16.

See CAT NMS Plan, supra note 3, at Section 1.1.

Request for Comment

162. The Commission granted an exemption from Rule 613 in order to allow the alternative of permitting the CAT NMS Plan to provide that Industry Members record and report to the Central Repository an Allocation Report that includes the Firm Designated ID when an execution is allocated in whole or part. This alternative is in lieu of the requirement in Rule 613 that Industry Members must report the account number for any subaccount to which an execution is allocated. Do Commenters believe that providing the information required in an Allocation Report as a means to identify order events and information related to the subaccount allocation information (the “Allocation Report Approach”) would be more efficient and cost-effective than the Rule 613 approach requiring the reporting of the account number for any subaccount to which an execution is allocated? Or do Commenters believe that the Rule 613 approach is preferable? Why or why not?

See Exemption Order, supra note 18.

163. Do Commenters believe that the Allocation Report Approach would affect the completeness of CAT Data? If so, how? Would the Allocation Report Approach result in more complete CAT Data? If so, please explain.

164. Do Commenters believe that the Allocation Report Approach would affect the accessibility of allocation information? If so, how? Would the Allocation Report Approach result in more accessible CAT Data? If so, please explain.

165. Do Commenters believe that the Allocation Report Approach would affect the timeliness of allocation information? If so, how? Would the Allocation Report Approach result in more timely CAT Data? If so, please explain.

166. Do Commenters believe the Allocation Report Approach would affect the security and confidentiality of CAT Data? If so, how? Would the Allocation Report Approach result in a different level of security or confidentiality? If so, please explain.

167. Do Commenters believe that the Allocation Report Approach described by the SROs is feasible? What challenges or risks would CAT Reporters face in providing such information? What challenges or risks would the Plan Processor face when ingesting such information and linking it to the appropriate Customers' accounts?

9. Options Market Maker Quotes

Section 6.4(d)(iii) of the CAT NMS Plan states that, with respect to the reporting obligations of an Options Market Maker under Sections 6.3(d)(ii) and (iv) regarding its quotes in Listed Options, such quotes shall be reported to the Central Repository by the relevant Options Exchange in lieu of reporting by the Options Market Maker. Section 6.4(d)(iii) further states that each Participant that is an Options Exchange shall, through its Compliance Rule, require its Industry Members that are Options Market Makers to report to the Options Exchange the time at which a quote in a Listed Option is sent to the Options Exchange (and, if applicable, the time of any subsequent quote modification and/or cancellation where such modification or cancellation is originated by the Options Market Maker). Such time information also shall be reported to the Central Repository by the Options Exchange in lieu of reporting by the Options Market Maker.

Rule 613(c)(7) provides that the CAT NMS Plan must require reporting of the details for each order and each Reportable Event, including the routing and modification or cancellation of an order. 17 CFR 242.613(c)(7). Rule 613(j)(8) defines “order” to include “any bid or offer.” 17 CFR 242.613(j)(8).

See CAT NMS Plan, supra note 3, at Section 6.4(d)(iii). The SROs requested exemptive relief from Rule 613 so that the CAT NMS Plan may permit Options Market Maker quotes to be reported to the Central Repository by the relevant Options Exchange in lieu of requiring that such reporting be done by both the Options Exchange and the Options Market Maker, as is required by Rule 613. See Exemptive Request Letter, supra note 16, at 2. In accord with the exemptive relief requested, the SROs committed to require Options Market Makers to report to the Exchange the time at which a quote in a Listed Option is sent to the Options Exchange. Id. at 3. The Commission granted exemptive relief on March 1, 2016 in order to allow this alternative to be included in the CAT NMS Plan and subject to notice and comment. See Exemption Order, supra note 18.

See CAT NMS Plan, supra note 3, at Section 6.4(d)(iii).

Id.

Request for Comment

168. The Commission granted an exemption from Rule 613 in order to allow the alternative of permitting Options Exchanges to report Options Market Maker quotes to the Central Repository in lieu of requiring such reporting by both the Options Exchange and the Options Market Maker as is required by Rule 613, to be included in the CAT NMS Plan and subject to notice and comment. Do Commenters believe that permitting exchanges to report quote information sent to them by Options Market Makers, including the Quote Sent Time, to the Central Repository would affect the completeness or quality of CAT Data? If so, what information would be missing?

See Exemption Order, supra note 18.

169. Under Rule 613, Options Market Makers would report their quotes to the Central Repository and time stamps would be attached to such quotes. Under the exemption, Options Market Makers would include the Quote Sent Time when sending quote information to the Options Exchanges. What, if any, are the risks of permitting the Options Exchanges to report information Options Market Makers otherwise would be required to report?

170. Do Commenters believe that the cost savings from permitting Options Exchanges to report information Options Market Makers would otherwise have to report makes this a preferable approach than Rule 613?

10. Error Rates

The CAT NMS Plan defines Error Rate as “the percentage of [R]eportable [E]vents collected by the [C]entral [R]epository in which the data reported does not fully and accurately reflect the order event that occurred in the market.” Under the CAT NMS Plan, the Operating Committee sets the maximum Error Rate that the Central Repository would tolerate from a CAT Reporter reporting data to the Central Repository. The Operating Committee reviews and resets the maximum Error Rate, at least annually. If a CAT Reporter reports CAT Data to the Central Repository with errors such that their error percentage exceeds the maximum Error Rate, then such CAT Reporter would not be in compliance with the CAT NMS Plan or Rule 613. As such, “the Participants as Participants or the SEC may take appropriate action for failing to comply with the reporting obligations under the CAT NMS Plan and SEC Rule 613.” The CAT NMS Plan, however, does not detail what specific compliance enforcement provisions would apply if a CAT Reporter exceeds the maximum Error Rate.

See CAT NMS Plan, supra note 3, at Section 1.1; see also Rule 613(j)(6).

See id. at Section 6.5(d)(i).

See id. at Appendix C, Section A.3(b).

See id. at Appendix C, Section A.3(b) and Rule 613(g) and (h).

See id. at Appendix C, Section A.3(b).

The CAT NMS Plan sets the initial maximum Error Rate at 5% for any data reported pursuant to subparagraphs (3) and (4) of Rule 613(c). The SROs highlight that “the Central Repository will require new reporting elements and methods for CAT Reporters and there will be a learning curve when CAT Reporters begin to submit data to the Central Repository” in support of a 5% initial rate. Further, the SROs state that “many CAT Reporters may have never been obligated to report data to an audit trail.” The SROs believe an initial maximum Error Rate of 5% “strikes the balance of making allowances for adapting to a new reporting regime, while ensuring that the data provided to regulators will be capable of being used to conduct surveillance and market reconstruction.” In the CAT NMS Plan, the Participants compared the contemplated Error Rates of CAT Reporters to the error rates of OATS reporters in the time periods immediately following three significant OATS releases in the last ten years. The Participants state that for the three comparative OATS releases: An average of 2.42% of order events did not pass systemic validations; an average of 0.36% of order events were not submitted in a timely manner; an average of 0.86% of orders were unsuccessfully matched to a trade reporting facility trade report; an average of 3.12% of OATS Route Reports were unsuccessfully matched to an exchange order; and an average of 2.44% of OATS Route Reports were unsuccessfully matched to a report by another reporting entity.

See id. at Section 6.5(d)(i).

See id. at Appendix C, Section A.3(b).

See id.

See id.

See id. The SROs note that the three comparative releases are known as “(1) OATS Phase III, which required manual orders to be reported to OATS; (2) OATS for OTC Securities which required OTC equity securities to be reported to OATS; and (3) OATS for NMS which required all NMS stocks to be reported to OATS.” Id.

See id. The SROs note that the calculated “combined average error rates for the time periods immediately following [the OATS] release across five significant categories for these three releases” was used in setting in the initial maximum Error Rate. Id.

See id.

The Participants, moreover, anticipate reviewing and resetting the maximum Error Rate once Industry Members (excluding Small Industry Members) begin to report to the Central Repository and again once Small Industry Members report to the Central Repository.

See id.

The Participants thus propose a phased approach to lowering the maximum Error Rates among CAT Reporters based on the period of time reporting to the Central Repository and whether the CAT Reporters are Participants, large broker-dealers or small broker-dealers. The Plan sets forth a goal of the following maximum Error Rates where “Year(s)” refers to year(s) after the CAT NMS Plan's date of effectiveness:

See id.

See id.

Table 1—Maximum Error Rates Schedule

One year % Two years % Three years % Four years %
Participants 5 1 1 1
Large Industry Members N/A 5 1 1
Small Industry Members N/A N/A 5 1

The CAT NMS Plan requires that the Plan Processor to: (i) Measure and report errors every business day; (ii) provide CAT Reporters daily statistics and error reports as they become available, including a description of such errors; (iii) provide monthly reports to CAT Reporters that detail a CAT Reporter's performance and comparison statistics; (iv) define educational and support programs for CAT Reporters to minimize Error Rates; and (v) identify, daily, all CAT Reporters exceeding the maximum allowable Error Rate. To timely correct data-submitted errors to the Central Repository, the Participants require that the Central Repository receive and process error corrections at all times. Further, the CAT NMS Plan requires that CAT Reporters be able to submit error corrections to the Central Repository through a web-interface or via bulk uploads or file submissions, and that the Plan Processor, subject to the Operating Committee's approval, support the bulk replacement of records and the reprocessing of such records. The Participants, furthermore, require that the Plan Processor identify CAT Reporter data submission errors based on the Plan Processor's validation processes.

See id. The CAT NMS Plan sets forth that the Plan Processor shall provide the Operating Committee with regular Error Rate reports. Id. at Section 6.1(o)(v). The Error Rate reports shall include each of the following—if the Operating Committee deems them necessary or advisable—“Error Rates by day and by delta over time, and Compliance Thresholds by CAT Reporter, by Reportable Event, by age before resolution, by symbol, by symbol type (e.g., ETF and Index) and by event time (by hour and cumulative on the hour)[.]” Id.

See id. at Appendix C, Section A.3(b).

See id.

See id. at Appendix D, Section 10.1. The CAT NMS Plan sets forth support programs that shall include educational programs, including FAQs, a dedicated help desk, industry-wide trainings, certifications, industry-wide testing, maintaining Technical Specifications with defined intervals for new releases/updates, emailing CAT Reporter data outliers, conducting annual assessments, using test environments prior to releasing new code to production, and imposing CAT Reporter attendance requirements for testing sessions and educational and industry-wide trainings. Id.

See id. at Appendix D, Section 10.4.

See id. at Appendix C, Section A.3(b).

See id.

See id. At a minimum, the processes would include validating the data's file format, CAT Data format, type, consistency, range, logic, validity, completeness, timeliness and linkage. See id. at Appendix D, Section 7.2.

Request for Comment

See Section IV.E.4, infra, for further Error Rate related requests for comment.

171. Do Commenters believe the CAT NMS Plan's initial maximum Error Rate of 5% for CAT Data reported to the Central Repository is appropriate in light of OATS' current error rate of less than 1%? Why or why not?

See Section IV.E.1.b(1), infra.

172. Please provide examples of error rates that are generally accepted with respect to other regulatory data reporting systems. At what error rate should data be considered materially unreliable? Please explain.

173. Do Commenters believe the CAT NMS Plan's initial maximum Error Rate of 5% would negatively affect the quality of CAT Data? Why or why not? In explaining why or why not, please address each quality (accuracy, completeness, timeliness and accessibility) separately.

174. Do Commenters believe that it was reasonable for the Participants to compare the contemplated Error Rates of CAT Reporters to the error rates of OATS reporters in the time periods immediately following three significant OATS releases in the last ten years? Why or why not?

175. If not 5%, what initial maximum Error Rate do Commenters believe Participants and Industry Members should be subject to and why?

176. What impact, if any, do Commenters believe a 5% initial maximum Error Rate would have on Industry Members' costs of compliance? Please describe the costs of correcting audit trail data. Given the costs of correcting audit trail data, do Commenters believe that establishing a lower maximum Error Rate could be less costly to Industry Members? Why or why not? How much less costly?

177. What impact, if any, do Commenters believe a 5% initial maximum Error Rate would have on the timing of the retirement of any redundant audit trail systems and any related costs? Please explain. Should the actual Error Rate for CAT Data affect the timing of the retirement of any redundant audit trail systems? If so, why? If not, why not?

178. Do Commenters believe the CAT NMS Plan's target maximum Error Rate of 1% for CAT Data reported to the Central Repository pursuant to the CAT NMS Plan's phased approach is the appropriate target maximum Error Rate in light of current industry standards? If not, why not? If not 1%, what target maximum Error Rate do Commenters believe Participants and Industry Members should be subject to and why?

179. Do Commenters believe there are any increased risks as a result of allowing CAT Data subject to an initial maximum Error Rate of 5% to be reported to the CAT? How difficult would it be for the Central Repository to process and analyze CAT Data based on data reported subject to an initial maximum Error Rate of 5%? Specifically, what are the increased risks, if any, of CAT Data reported subject to an Error Rate of 5% in respect of combining or linking data within the Central Repository or across other sources of trade and order data currently available to regulators?

180. Do Commenters believe there are any increased risks as a result of allowing CAT Data subject to a target maximum Error Rate of 1% to be reported to the CAT? How difficult would it be for the Central Repository to process and analyze CAT Data based on data reported subject to a target maximum Error Rate of 1%? Specifically, what are the increased risks, if any, of CAT Data reported subject to an Error Rate of 1% in respect of combining or linking data within the Central Repository or across other sources of trade and order data currently available to regulators?

181. The CAT NMS Plan provides that the Participants would review and reset, at least on an annual basis, the maximum Error Rate. Do Commenters believe that this establishes an appropriately rigorous schedule for the Participants to evaluate whether the maximum Error Rate could potentially be set to a lower rate? Are there any other factors that should affect when and how the maximum Error Rate is set?

182. The CAT NMS Plan provides as a goal a four-year phased approach schedule to lower the maximum Error Rate segmented by Participants, large broker-dealers and small broker-dealers. Do Commenters believe a phased schedule is appropriate and reasonable? Do Commenters believe establishing segments is appropriate and reasonable, and if so are these the appropriate Error Rate groupings? What alternative groupings, if any, do Commenters believe are the appropriate Error Rate groupings?

183. Do Commenters believe that the CAT NMS Plan is clear whether the four-year phased approach is a goal? Should it be more than a goal? Please explain.

184. Do Commenters believe the phased approach for CAT implementation, whereby SROs would begin reporting CAT Data one year prior to other CAT Reporters and two years prior to small CAT Reporters, would affect the quality of the CAT Data and the number of available CAT Data items in the audit trail?

185. Do Commenters believe the CAT NMS Plan provides adequate enforcement provisions to ensure CAT Reporters submit data to the Central Repository no higher than the maximum Error Rate? If not, what additional enforcement provisions should the CAT NMS Plan provide?

186. Do Commenters believe that there should be a lower initial maximum Error Rate and/or a more accelerated or slower reduction of the target maximum Error Rate? Would an accelerated reduction of the target maximum Error Rate facilitate the earlier retirement of any redundant audit trail system? What should the initial maximum Error Rate and/or what should be the schedule for reducing the target maximum Error Rate?

187. What framework and criteria should regulators adopt when determining whether to retire potentially redundant regulatory data reporting systems? Please explain when and how such retirement should take place.

188. Do Commenters believe the CAT NMS Plan sets forth sufficient consequences for a CAT Reporter exceeding the maximum Error Rates? If not, what should be those consequences?

189. Do Commenters believe that some errors are of greater concern than others? If so, what types of errors are more or less problematic? Should the type of error be considered when calculating Error Rates? If so, how should the Plan Processor take into account different types of errors when calculating Error Rates? How should the Participants take into account different types of errors when setting Error Rates?

11. Regulatory Access

Under Section 6.5(c) of the CAT NMS Plan, the Plan Processor must provide regulators access to the Central Repository for regulatory and oversight purposes and create a method of accessing CAT Data that includes the ability to run complex searches and generate reports. Section 6.10(c) requires regulator access by two different methods: (1) An online targeted query tool with predefined selection criteria to choose from; and (2) user-defined direct queries and bulk extractions of data via a query tool or language allowing querying of all available attributes and data sources. Additional requirements concerning regulator access appear in Section 8 of Appendix D.

See CAT NMS Plan, supra note 3, at Section 6.5(c). Appendix C provides objective milestones to assess progress concerning regulator access to the Central Repository. See id. at Appendix C, Section C.10(d).

Id. at Section 6.10(c). Section 6.10(c) also requires the Plan Processor to reasonably assist regulatory staff with queries, submit queries on behalf of regulatory staff as requested, and maintain a help desk to assist regulatory staff with questions concerning CAT Data. Id.

See id. at Appendix D, Section 8.

The CAT NMS Plan requires that CAT must support a minimum of 3,000 regulatory users and at least 600 such users accessing CAT concurrently without an unacceptable decline in performance. Moreover, CAT must support an arbitrary number of user roles and, at a minimum, include defined roles for both basic and advanced regulatory users.

Id. at Appendix D, Section 8.1.

Id.

a. Online Targeted Query Tool

Sections 8.1.1, 8.1.2, and 8.1.3 of Appendix D contain further specifications for the online targeted query tool. The tool must allow for retrieval of processed and/or validated (unlinked) data via an online query screen that includes a choice of a variety of pre-defined selection criteria. Targeted queries must include date(s) and/or time range(s), as well as one or more of a variety of fields listed in Section 8.1.1 (e.g., product type, CAT-Reporter-ID, and Customer-ID). Targeted queries would be logged such that the Plan Processor could provide monthly reports to the SROs concerning metrics on performance and data usage of the search tool. The CAT NMS Plan further requires that acceptable response times for the targeted search be in increments of less than one minute; for complex queries scanning large volumes of data or large result sets (over one million records) response times must be available within 24 hours of the request; and queries for data within one business date of a 12-month period must return results within three hours regardless of the complexity of criteria. Under the CAT NMS Plan, regulators may access all CAT Data except for PII data (access to which would be limited to an authorized subset of Participant and Commission employees) and the Plan Processor must work with regulators to implement a process for providing them with access and routinely verifying a list of active users.

Id. at Appendix D, Sections 8.1.1-8.1.3.

Id. at Appendix D, Section 8.1.1.

Id.

Id.

Id. at Appendix D, Section 8.1.2. Appendix D, Section 8.1.2 contains further performance requirements applicable to data and the architecture of the online query tool. Id.

Id. at Appendix D, Section 8.1.3.

b. User-Defined Direct Queries and Bulk Extraction of Data

Section 8.2 of Appendix D outlines the requirements for user-defined direct queries and bulk extraction of data, which regulators would use to obtain large data sets for internal surveillance or market analysis. Under the CAT NMS Plan, regulators must be able to create, save, and schedule dynamic queries that would run directly against processed and/or unlinked CAT Data. Additionally, CAT must provide an open application program interface (“API”) that allows use of analytic tools and database drivers to access CAT Data. Queries submitted through the open API must be auditable and the CAT System must contain the same level of control, monitoring, logging, and reporting as the online targeted query tool. The Plan Processor must also provide procedures and training to regulators that would use the direct query feature. Sections 8.2.1 and 8.2.2 of Appendix D contain additional specifications for user-defined direct queries and bulk data extraction, respectively.

Id. at Appendix D, Section 8.2.

Id.

Id.

Id. Direct queries must not return or display PII data but rather display non-PII unique identifiers (e.g., Customer-ID or Firm Designated ID). The PII corresponding to these identifiers could be gathered using the PII workflow described in Appendix D, Data Security, PII Data Requirements. See id. at Appendix D, Section 4.1.6.

Id. at Appendix D, Section 8.2.

Id. at Appendix D, Sections 8.2.1 and 8.2.2.

c. Regulatory Access Schedule

Section A.2 of Appendix C addresses the time and method by which CAT Data would be available to regulators. Section A.2(a) requires that data be available to regulators any point after the data enters the Central Repository and passes basic format validations. After errors are communicated to CAT Reporters on T+1, CAT Reporters would be required to report corrected data back to the Central Repository by 8 a.m. Eastern Time on T+3. Regulators must then have access to corrected and linked Order and Customer data by 8:00 a.m. Eastern Time on T+5. Section A.2(b) generally describes Bidders' approaches regarding regulator access and use of CAT Data and notes that although the SROs set forth the standards the Plan Processor must meet, they do not endorse any particular approach. Section A.2(c) outlines requirements the Plan Processor must meet for report building and analysis regarding data usage by regulators, consistent with, and in addition to, the specifications outlined in Section 8 of Appendix D.

Id. at Appendix C, Section A.2.

Id. at Appendix C, Section A.2(a). Appendix C, Section A.3(e) indicates this would be no later than noon EST on T+1. Id. at Appendix C, Section A.3(e).

Id. at Appendix C, Section A.1(a)(iv); Appendix D, Section 6.1.

Id. at Appendix C, Section A.2(a).

Id. at Appendix C, Section A.2(b).

Id. at Appendix C, Section A.2(c). Appendix C, Section A.2(d) addresses system service level agreements that the SROs and Plan Processor would enter into. Id. at Appendix C, Section A.2(d).

Request for Comment

See Section IV.H.5, infra, for further regulatory access related requests for comment.

190. Do Commenters believe the CAT NMS Plan's “Functionality of the CAT System” Section (Section 8 of Appendix D) describes with sufficient detail how a regulator would access, use and analyze CAT Data? If not, describe what, if any, additional requirements and details should be provided and how.

191. Do Commenters believe the CAT NMS Plan's “Functionality of the CAT System” Section sufficiently addresses all regulators' end-user requirements? If not, please explain. Describe what, if any, additional requirements and details should be provided and how.

192. If Commenters believe that the CAT NMS Plan's “Functionality of the CAT System” Section does not cover all regulators' end-user requirements, please describe how regulators would integrate their applications in a timely and reasonable manner.

193. The CAT NMS Plan permits the CAT to be implemented in a way that would (1) require regulators to download entire data sets and analyze such data within the regulator or the regulators' cloud or (2) permit regulators to analyze sets of data within the CAT using applications or programs selected by the Commission. What do Commenters believe are the advantages and disadvantages to each approach?

194. Do Commenters believe the CAT NMS Plan's T+5 schedule for regulatory access to corrected and linked Order and Customer data is the appropriate schedule in light of current industry standards? If not, why not? Do Commenters believe that the SROs' determination of current industry standards is reasonable or appropriate? Do Commenters believe that it is appropriate to base the timing for regulatory access on industry standards? Why or why not?

195. If the T+5 schedule is not appropriate, when do Commenters believe regulatory access to corrected and linked Order and Customer data should be provided and why? Do Commenters believe the SROs' should include in the CAT NMS Plan detailed provisions with milestones in achieving a more accelerated regulatory access schedule to corrected and linked Order and Customer data?

196. Do Commenters believe the Plan's proposed error correction timeframe—i.e., communication of errors on T+1, corrected data resubmitted by CAT Reporters by T+3, and corrected data available to regulators by T+5—is feasible and appropriate in light of current industry standards? If not, why not, and how long do Commenters believe these error correction timeframes should be and why? Are shorter timeframes feasible and appropriate in light of current industry standards? Why or why not?

197. To what extent do Commenters believe the CAT NMS Plan's T+5 regulatory access schedule to corrected and linked Order and Customer data would affect the accuracy, completeness, accessibility and/or timeliness of CAT Data collected and maintained under the CAT? How?

198. To what extent do Commenters believe the Plan's three-day window of error correction would affect the accuracy, completeness, accessibility and/or timeliness of CAT Data collected and maintained under the CAT? How?

199. Regulators' technology teams would be required to work with the Plan Processor to integrate their applications under the CAT NMS Plan. What, if any, are the risks to this approach? Should the Plan Processor be required to enter into support contracts with regulators? If so, please explain. Describe what, if any, service contract terms should be set forth in the CAT NMS Plan or set forth in any related documents. Do Commenters have any concerns about the security or confidentiality of CAT Data resulting from a service contract between the Plan Processor and the regulators? If so, please explain. If Commenters have any security or confidentiality concerns resulting from a service contract between the Plan Processor and the regulators, please specify any appropriate service contract terms that would address the concerns.

200. How do Commenters believe the Plan Processor should set pricing for a regulator seeking additional functionality from the Plan Processor under the CAT? What, if anything, do Commenters believe should govern pricing for additional functionality by the Plan Processor? For example, should pricing or contract standards (e.g., reasonable, commercially reasonable, etc.), agreed-upon profit margins—or minimums and maximums, etc.—be included under the CAT NMS Plan or any related documentation? If so, please explain.

201. Do Commenters believe the CAT NMS Plan appropriately encourages or incentivizes the Participants and the Plan Processor to incorporate new technology and to innovate? Does the CAT NMS Plan appropriately encourage or incentivize the Plan Processor to have a flexible and scalable solution? Do Commenters believe that the CAT NMS Plan would result in a CAT that has adequate system flexibility and scalability to incorporate improvements in technology and future regulatory, analytic and data capture needs? Why or why not?

202. Does the regulatory access approach set forth in the CAT NMS Plan provide regulators with sufficient tools to maximize their regulatory activities, actions, and improve their surveillances? If not, why not and what should be added?

203. The CAT NMS Plan provides that targeted queries and data extractions would be logged so that the Plan Processor can provide the Operating Committee, the Participants, and the Commission with monthly performance and usage reports including data such as the user ID of the person submitting the query and the parameters of the query. Do Commenters believe that the data to be recorded in these logs and provided in these reports to each Participant and to the SEC would be appropriate and useful? Should any data elements be added or removed from these reports?

204. Do Commenters believe it is appropriate for the Plan Processor and the Operating Committee to also have access to these logs and monthly performance and usage reports? How should the Plan Processor and Operating Committee be permitted to use these logs and reports? To the extent that these logs and reports are accessible by the Plan Processor and the Operating Committee, should any data elements be added or removed? Should additional details or requirements be added to the CAT NMS Plan to clarify what the content of these logs and reports would be and which parties would have access to them?

12. Security, Confidentiality, and Use of Data

The CAT NMS Plan provides that the Plan Processor is responsible for the security and confidentiality of all CAT Data received and reported to the Central Repository, including during all communications between CAT Reporters and the Plan Processor, data extraction, data manipulation and transformation, loading to and from the Central Repository, and data maintenance by the Central Repository. The Plan Processor must, among other things, require that individuals with access to the Central Repository agree to use CAT Data only for appropriate surveillance and regulatory activities and to employ safeguards to protect the confidentiality of CAT Data.

See CAT NMS Plan, supra note 3, at Section 6.5(f)(i), (iv).

Id. at Section 6.5(f)(i).

In addition, the Plan Processor must develop a comprehensive information security program as well as a training program that addresses the security and confidentiality of all information accessible from the CAT and the operational risks associated with accessing the Central Repository. The Plan Processor must also designate one of its employees as Chief Information Security Officer; among other things, the Chief Information Security Officer is responsible for creating and enforcing appropriate policies, procedures, and control structures regarding data security. The Technical Specifications, which the Plan Processor must publish, must include a detailed description of the data security standards for CAT.

Id. at Sections 6.1(m), 6.12.

Id. at Section 6.2(b).

Id. at Section 6.9.

Appendix D of the CAT NMS Plan sets forth minimum data security requirements for CAT that the Plan Processor must meet. For example, Appendix D enumerates various connectivity, data transfer, and encryption requirements such as that the CAT System must have encrypted internet connectivity, CAT Reporters must connect to CAT infrastructure using secure methods such as private lines or virtual private network connections over public lines, CAT Data must be encrypted in flight using industry standard best practices, PII data must be encrypted both at rest and in flight, and CAT Data stored in a public cloud must be encrypted at rest. Additional requirements regarding data storage, data access, breach management, and PII data are also specified in Appendix D.

Id. at Appendix D, Section 4.

Id. at Appendix D, Section 4.1.1, 4.1.2.

Id. at Appendix D, Section 4.1.3-4.1.6.

In addition, the Participants must establish and enforce policies and procedures that ensure the confidentiality of the CAT Data obtained from the Central Repository, limit the use of CAT Data obtained from the Central Repository solely for surveillance and regulatory purposes, implement effective information barriers between each Participant's regulatory and non-regulatory Staff with regard to CAT Data, and limit access to CAT Data to designated persons. However, a Participant may use the Raw Data it reports to the Central Repository for “commercial or other” purposes if not prohibited by applicable law, rule or regulation.

The Commission notes that regulatory purposes includes, among other things, analysis and reconstruction of market events, market analysis and research to inform policy decisions, market surveillance, examinations, investigations, and other enforcement functions.

Id. at Section 6.5(f)(ii), (g).

Raw data is defined as “Participant Data and Industry Member Data that has not been through any validation or otherwise checked by the CAT System.” Id. at Section 1.1.

Id. at Section 6.5(f)(i).

Request for Comment

205. Do Commenters believe that the CAT NMS Plan appropriately allocates responsibility for the security and confidentiality of CAT Data among the Participants, the Plan Processor, and other parties? If not, how should these responsibilities be allocated?

206. Do Commenters believe that the data security requirements set out in Appendix D are appropriate and reasonable? Should any additional details or requirements be provided?

207. What, if any, specific details or requirements regarding data security and confidentiality do Commenters believe should be included in the information security program, training program, and Technical Specifications to be developed by the Plan Processor? Should additional details on the content of these programs and specifications be provided?

208. What, if any, specific details or requirements regarding data confidentiality do Commenters believe should be included in the policies and procedures to be developed by the Participants? Should additional details on the content of these policies and procedures be provided?

209. Do Commenters believe that the CAT NMS Plan includes sufficient safeguards to prevent the misuse of CAT Data by employees or agents of the Participants or other persons with access to the Central Repository? For example, do Commenters believe that requiring information barriers between regulatory and non-regulatory staff and permitting the use of CAT Data only for regulatory, surveillance, and commercial or other purposes as permitted by law are effective measures to prevent the misuse of CAT Data? Should the CAT NMS Plan set forth additional detail regarding the distinction between regulatory and non-regulatory staff and between the appropriate and inappropriate use of CAT Data for commercial or other purposes? Should the CAT NMS Plan prescribe any specific information barriers? If so, what should be prescribed in the CAT NMS Plan?

See id. at Section 6.5(f)(ii)(A).

See id. at Section 6.5(f)(i)(A).

210. Do Commenters believe the data access and breach management provisions described in Appendix D of the CAT NMS Plan are effective mechanisms for monitoring and preventing the misuse of CAT Data? Why or why not? Would any additional details or requirements make these provisions more effective?

See id. at Appendix D, Sections 4.1.4, 4.1.5.

211. Which persons or entities should have the responsibility to monitor for and prevent the misuse of CAT Data? For example, should the Chief Compliance Officer or the Chief Information Security Officer have this responsibility? Why or why not? Should additional details be provided to clarify where this responsibility lies?

212. Do Commenters believe it is appropriate for Participants to be permitted to use all Raw Data reported to the Central Repository for commercial purposes? If not, what particular types of Raw Data would be inappropriate to use for commercial purposes?

213. Do Commenters believe that the CAT NMS Plan adequately addresses the protection and security of PII in CAT? If not, why not and what should be added to the CAT NMS Plan? For example, should the CAT NMS Plan provide that PII is accessible only when required, that PII be properly masked, and/or that it be safeguarded such that it would not be improperly accessible?

214. Do Commenters believe that there are alternative methods or information that could be used in lieu of requiring the reporting of Customer PII to the Central Repository that, without diminishing the quality of CAT Data available to regulators or impairing regulators' ability to use CAT Data to carry out their functions, would create less risk of a breach of the security or confidentiality of the personal information of Customers? If so, what methods or information, specifically, could serve as such an alternative to PII?

See Section III.B.7, supra, for additional PII related requests for comment.

215. Do Commenters believe that the CAT NMS Plan includes adequate requirements regarding the operational security of the CAT System? What, if any, additional details or requirements should be provided? Should the CAT NMS Plan require the Plan Processor to have the ability to monitor for threats, attacks, and anomalous activity on a 24/7 basis through a Security Operations Center (“SOC”) or a similar capability? What would be the costs and benefits of such a requirement?

216. Appendix C of the CAT NMS Plan discusses solutions for encrypting data at rest and in motion. Appendix D of the CAT NMS Plan states that all CAT Data must be encrypted in flight, and PII Data must encrypted in flight and at rest. Do Commenters believe that the Plan's data encryption requirements are adequate for CAT Data and PII Data? Why or why not? Do Commenters believe that the CAT NMS Plan provides sufficient information and clarity regarding data encryption requirements? Do Commenters believe that there is a particular method for data encryption, in motion and/or at rest, that should be used?

217. Appendix D, Section 4.1.1 of the CAT NMS Plan states that the CAT System must have “encrypted internet connectivity.” What are the risks, if any, of allowing Internet access from the Central Repository, even if encrypted? Please explain. Do Commenters believe that the encrypted connection requirement in the CAT NMS Plan should apply to communication paths from the Central Repository to the Internet and/or connections from CAT to/from trusted parties? What challenges would the Plan Processor face in implementing either option? Does one option provide more robust security than the other? Why or why not?

218. To the extent the requirement for “encrypted internet connectivity” applies to connectivity between the Central Repository and trusted parties such as the Commission and the Participants, do Commenters believe that the CAT NMS Plan should require that these parties and the Plan Processor enter into formal Memoranda of Understanding or Interconnection Security Agreements that document the technical, operational, and management details regarding the interface between the CAT System and these parties? Why or why not?

219. With respect to industry standards, do Commenters believe that the CAT NMS Plan should be updated to include standards and requirements of other NIST Special Publications (“SPs”) that were not mentioned in Appendix D (e.g., NIST SP 800-86 for incident handling, 800-44 for securing public-facing web servers, 800-146 for cloud security)? Why or why not?

220. Do Commenters believe that the Plan should be updated more broadly to include the NIST family of guidance documents? Why or why not?

221. Throughout the Plan, there are numerous references to leveraging “industry best practices” pertaining to compliance subjects such as system assessments and disaster recovery/business continuity planning. How do “industry best practices” compare to NIST guidance in these areas? Do Commenters believe that the Plan Processor should implement NIST guidance for the Plan rather than industry best practices? Why or why not?

222. The CAT NMS Plan states that the Plan Processor must conduct third party risk assessments at regular intervals to verify that security controls implemented are in accordance with NIST SP 800-53. Do Commenters believe that the CAT NMS Plan should adopt the meaning and terminology of Security Assessment and Authorization as defined by the NIST and/or other NIST guidance in the CAT NMS Plan, particularly within the requirements set forth in Appendix D to the CAT NMS Plan? Why or why not?

See CAT NMS Plan, supra note 3, at Appendix D, Section 5.3.

223. Do Commenters believe that the CAT NMS Plan should include requirements regarding how the Plan Processor should categorize data from a security perspective? For example, should the Plan Processor be required to implement data categorization standards consistent with Federal Information Processing Standard (“FIPS”) 199 or NIST SP 800-60? Why or why not? Would including data categorization requirements in the CAT NMS Plan improve data integrity, availability, segmentation, auditing, and incident response? Why or why not?

224. The CAT NMS Plan provides that CAT must follow NIST SP 800-137—Information Security Continuous Monitoring for Federal Information Systems and Organizations in addition to a limited number of related monitoring provisions. Do Commenters believe that the CAT NMS Plan provides sufficient and robust information related to continuous monitoring program requirements? Why or why not?

See id. at Sections 6.1(g), 6.10(c), Appendix C, Section A.4, Appendix D, Sections 2.2, 4.1.2, 4.1.4, 4.2, 8.3, 8.4.

225. Do Commenters believe the CAT NMS Plan adequately sets forth the roles and responsibilities of independent third party risk assessment functions, including the consistent description of their specific functions and performance frequency? For example, are the CAT NMS Plan independent third party risk assessment provisions consistent with “industry best practices”? Or should the CAT require a greater or lesser performance frequency than as described in the CAT NMS Plan? As another example, do the technical assessments described in Section 6.2, Appendix C, Section A.5, and the NIST SP 800-53 requirements noted in Appendix D, Section 4.2, adequately and clearly establish the roles and responsibilities of the parties assessing the technical aspects of the CAT?

226. Do Commenters believe the CAT NMS Plan should specify the general audit and independent assessment requirements and the proper timeframes for when those assessments should occur? For instance, are there assessments that may need to occur on an annual basis? If so, what are those assessments? Are there assessments that may need to occur more frequently? If so, what are those assessments and why do they need to occur more frequently?

227. Do Commenters believe that the CAT NMS Plan requirements for conducting ad hoc penetration testing and an application security code audit by a reputable third-party in Appendix D, Section 4.1.3 “prior to launch” and periodically as defined by SLAs are consistent with industry best practices? Should additional testing or audits be required? Why or why not? Should testing or audits be required to occur more frequently than required by the CAT NMS Plan and SLAs? Why or why not?

228. Do Commenters believe that the third party risk assessments and penetration tests required by the CAT NMS Plan could themselves compromise the security or confidentiality of CAT Data? Please explain.

229. In Section 6.2(b)(vi) of the CAT NMS Plan, the Chief Information Security Officer is required to report to the Operating Committee the activities of the Financial Services Information Sharing and Analysis Center (“FS-ISAC”) or other comparable body. Do Commenters believe there are other cyber and threat intelligence bodies, in addition to FS-ISAC, that the Plan Processor should join? Why or why not?

230. Do Commenters believe the CAT NMS Plan effectively describes the verification process when CAT Reporters connect to the Central Repository network? For example, which specific individual(s) at a CAT Reporter would be allowed access to CAT for reporting and verification purposes? Should there be a public key exchange process?

231. Do Commenters believe the CAT NMS Plan provides sufficient detail regarding the ability of CAT to determine whether a regulator's queries are shielded from the Plan Processor (including its staff, officers, and administrators) as well as other regulators and users of CAT? If not, what specifically should be added to the CAT NMS Plan?

232. Do Commenters believe that the CAT NMS Plan should require an audit of all CAT Reporters' data security? If so, which person or entity should have responsibility for such an audit, and what should the scope and elements of the audit be? Please estimate the cost of such audits. What other changes, if any, should be made to the CAT NMS Plan to provide for the allocation of sufficient resources whereby such an audit could be carried out?

233. Do Commenters believe the CAT NMS Plan should require the Plan Processor to provide a “blanket” security authorization to operate (“ATO”) document (or its equivalent) prior to CAT Reporters sending CAT Data?

IV. Economic Analysis

A. Introduction

When adopting Rule 613, the Commission noted that the adopted Rule permitted the SROs to consider a wider array of solutions than did the proposed Rule. The Commission stated its belief that, as a result, “the economic consequences of the consolidated audit trail now will become apparent only over the course of the multi-step process for developing and approving an NMS plan that will govern the creation, implementation, and maintenance of a consolidated audit trail.” In particular, the Commission noted its belief that “the costs and benefits of creating a consolidated audit trail, and the consideration of specific costs as related to specific benefits, is more appropriately analyzed once the SROs narrow the expanded array of choices they have under the adopted Rule and develop a detailed NMS plan.” The Commission also noted that a “robust economic analysis of . . . the actual creation and implementation of a consolidated audit trail itself . . . requires information on the plan's detailed features (and their associated cost estimates) that will not be known until the SROs submit their NMS plan to the Commission for its consideration.” Accordingly, the Commission deferred its economic analysis of the actual creation, implementation, and maintenance of the CAT until after submission of an NMS plan.

See Adopting Release, supra note 9, at 45725-6.

Id.

Id. at 45726.

To assist in that analysis, Rule 613, as adopted, requires that the SROs: (1) Provide an estimate of the costs associated with creating, implementing, and maintaining the consolidated audit trail under the terms of the NMS plan submitted to the Commission for its consideration; (2) discuss the costs, benefits, and rationale for the choices made in developing the NMS plan submitted; and (3) provide their own analysis of the submitted NMS plan's potential impact on competition, efficiency and capital formation. The Commission stated that it believed that these estimates and analyses would help inform public comment regarding the CAT NMS Plan and would help inform the Commission as it evaluates whether to approve the CAT NMS Plan.

Id.; see also 17 CFR 242.613(a)(1)(vii), (viii), (xi), (xii).

See Adopting Release, supra note 9, at 45726. Rule 613(a)(5) requires that “[i]n determining whether to approve the national market system plan, or any amendment thereto, and whether the national market system plan or any amendment thereto is in the public interest under [Rule] 608(b)(2), the Commission shall consider the impact of the national market system plan or amendment, as applicable, on efficiency, competition, and capital formation.” 17 CFR 242.613(a)(5).

The Commission is sensitive to the economic effects of the CAT NMS Plan, including its costs and benefits and its impact on efficiency, competition and capital formation. In the Adopting Release for Rule 613, the Commission considered the economic effects of the actions the SROs were required to take upon approval of the adopted Rule, specifically the requirement that the SROs develop an NMS plan, utilizing their own resources and undertaking their own research, that addresses the specific details, cost estimates, considerations, and other requirements of the Rule. As noted in the Adopting Release, however, Rule 613 provided the SROs with “flexibility in how they [chose] to meet the requirements of the adopted Rule,” allowing the SROs to consider a number of different approaches in developing the CAT NMS Plan.

See CAT NMS Plan, supra note 3.

See Adopting Release, supra note 9, at 45726.

Id. at 45725.

In accordance with the approach articulated by the Commission in the Adopting Release, the Commission is hereby publishing its economic analysis of the CAT NMS Plan and is soliciting comment thereon. This Section reflects the Commission's preliminary analysis and conclusions regarding the economic effects of the creation, implementation and maintenance of the CAT pursuant to the details proposed in the NMS plan submitted to the Commission for its consideration. The analysis is divided into eight topics: (1) A summary of the expected economic effects of approving the CAT NMS Plan; (2) a description of the economic framework for analyzing the economic effects of approving the CAT NMS Plan; (3) a discussion of the current, or “Baseline,” audit trail data available to regulators, and the sources of such data; (4) a discussion of the potential benefits of the CAT NMS Plan; (5) a discussion of the potential costs of the CAT NMS Plan; (6) an economic analysis of the CAT NMS Plan's impact on efficiency, competition, and capital formation; (7) a discussion of alternatives to various features of the CAT NMS Plan and to the CAT NMS Plan itself; and (8) a request for comment on the Commission's preliminary economic analysis.

B. Summary of Expected Economic Effects

As the Commission explained in the Adopting Release, the Commission believes that the regulatory data infrastructure on which the SROs and the Commission currently must rely is outdated for effective oversight of a complex, dispersed, and highly automated national market system. In performing their oversight responsibilities, regulators today must attempt to cobble together disparate data from a variety of existing information systems, each lacking in completeness, accuracy, accessibility, and/or timeliness—a model that neither supports the efficient aggregation of data from multiple trading venues nor yields the type of complete and accurate market activity data needed for robust market oversight. The Commission has analyzed the expected economic effects of the CAT NMS Plan in light of these existing shortcomings and the goal of improving the ability of SROs and the Commission to perform their regulatory activities to the benefit of investors.

See id. at 45723.

See id.

The Commission noted current SRO audit trail limitations in the Proposing Release and the Adopting Release. See Proposing Release, supra note 9, at 32563-68; Adopting Release, supra note 9, at 45726-30. Rule 613 is designed to address these limitations.

In general, the Commission preliminarily believes that, if approved, the CAT NMS Plan would result in benefits by improving the quality of the data available to regulators in four areas that affect the ultimate effectiveness of core regulatory efforts—completeness, accuracy, accessibility and timeliness. The Commission preliminarily believes that the improvements in these data qualities that would be realized from approval of the CAT NMS Plan would substantially improve regulators' ability to perform analysis and reconstruction of market events, and market analysis and research to inform policy decisions, as well as perform other regulatory activities, in particular market surveillance, examinations, investigations, and other enforcement functions. Regulators depend on data for many of these activities and the improvements in the data qualities would thus improve the efficiency and effectiveness of such regulatory activities. As explained further below, these improvements could benefit investors by giving regulators more and better regulatory tools to provide investors with a more effectively regulated trading environment, which could increase capital formation, liquidity, and price efficiency. Data improvements could enhance regulators' ability to provide investors and the public with more timely and accurate analysis and reconstruction of market events, and to develop more effective responses to such events. Improved understanding of emerging market issues resulting from enhanced market analysis and research could inform regulatory policies that improve investor protection through better market quality, more transparency, and more efficient prices.

See Adopting Release, supra note 9, at 45727 (discussing four “qualities” of trade and order data that impact the effectiveness of core SRO and Commission regulatory efforts: Accuracy, completeness, accessibility, and timeliness); see also Section IV.E, infra, for a detailed discussion of the expected benefits of the CAT NMS Plan.

See Section IV.E.2, infra.

See Section IV.E.2.a, infra.

In terms of completeness, the Plan requires the reporting of certain additional data fields, events, and products. More importantly, the CAT NMS Plan requires certain data elements useful for regulatory analysis to be available from a single data source. Having relevant data elements available from a single source would simplify regulators' data collection process and facilitate more efficient analyses and surveillances that incorporate cross-market and cross-product data.

See CAT NMS Plan, supra note 3, at Sections 6.3, 6.4; see also 17 CFR 242.613(c)(7).

With respect to the accuracy of available data, the Commission preliminarily believes that the requirements in the Plan would improve data accuracy significantly. For example, the Commission expects that the requirements to store the CAT Data in a uniform linked format and the use of consistent identifiers for customers and market participants would result in fewer inaccuracies as compared to current data sources. These accuracy improvements should significantly reduce the time regulators spend processing the data and finding solutions when faced with inaccurate data. The Commission preliminarily believes that the requirements in the Plan for clock synchronization and time stamp granularity would improve the accuracy of data with respect to the timing of market events, but the improvements would be modest. The Commission preliminarily believes that the Plan would improve regulators' ability to determine the sequence of a small percentage of market events relative to all surrounding events.

The CAT NMS Plan would also require that CAT Reporters' business clocks be synchronized to within 50 milliseconds of the time maintained by the NIST, which would increase the precision of the time stamps provided by the 39% of broker-dealers who currently synchronize their clocks with less precision than what is called for by the Plan. See supra note 125. Independent of the potential time clock synchronization benefits, the order linking data that would be captured in CAT should increase the proportion of events that could be sequenced accurately. This reflects the fact that some records pertaining to the same order could be sequenced by their placement in an order lifecycle (e.g., an order submission must have occurred before its execution) without relying on time stamps. This information may also be used to partially sequence surrounding events.

The Commission also preliminarily believes that the Plan would increase the accessibility of data for SROs and the Commission, because regulators would be able to access the CAT Data directly. This, coupled with the improvements in completeness, would vastly increase the scope of information readily available to regulators and significantly reduce the number of data requests from the several hundred thousand requests regulators make each year. The increased scope of readily available information should facilitate more data-driven regulatory policy decisions, broaden the potential surveillances, expand the opportunities for SRO and Commission analysis to help target broker-dealers and investment advisers for examinations and help to perform those examinations.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.2, Appendix D, Section 8.1; see also 17 CFR 242.613(e)(2).

Finally, the Commission preliminarily believes that the CAT NMS Plan would improve the timeliness of available data. Because regulators would be able to access uncorrected data the day after an order event and would be able to access corrected and linked data five days after an order event, many data elements would be available to regulators more quickly than they are currently and the amount of time regulators would need to acquire and process data before running analyses would be reduced. For example, the corrected and linked data available on T+5 would identify the customer account associated with all order events, information that currently takes ten days or longer for regulators to obtain and then need to link to other data sources for use. These improvements in timeliness, combined with improvements in completeness, accessibility, and accuracy discussed above, would improve the efficiency of regulatory analysis and reconstruction of market events, as well as market analysis and research that informs policy decisions, and make market surveillance, examinations, investigations, and other enforcement functions more efficient, allowing, for example, the SROs and the Commission to review tips and complaints more effectively.

CAT Data would be reported by 8:00 a.m. Eastern Time on day T+1 and made available to regulators in raw form after it is received and passes basic formatting validations with an error correction process completed by 8:00 a.m. Eastern Time on day T+5. While the Plan does not specify exactly when these validations would be complete, the requirement to link records by 12:00 p.m. Eastern Time on day T+1 gives a practical upper bound on this timeline. See CAT NMS Plan, supra note 3, at Appendix C, Sections A.2(a), A.3(a), Appendix D, Section 6.2.

The Commission notes that the Plan lacks information regarding the details of certain elements of the Plan likely to affect the costs and benefits associated with it, primarily because those details have not yet been determined, and this lack of information creates some uncertainty about the expected economic effects. As discussed further below, lack of specificity surrounding the processes for converting data formats and linking related order events creates uncertainty as to the anticipated improvements in accuracy because such processes have the potential to create new data inaccuracies. Lack of specificity surrounding the process for regulators to access the CAT Data also creates uncertainty around the expected improvements in accessibility. For example, while the Plan indicates that regulators would have an on-line targeted query tool and a tool for user-defined direct queries or bulk extraction, the Plan itself does not provide an indication for how user-friendly the tools would be or the particular skill set needed to use the tools for user-defined direct queries. However, the Commission has analyzed the expected economic effects of the Plan to the extent possible with the information available, noting areas of uncertainty in its analysis where applicable. The Commission has also considered whether certain provisions related to the operation and administration of the Plan could mitigate some of the uncertainties.

See CAT NMS Plan, supra note 3, at Appendix D, Sections 8.1.1, 8.1.2.

See Section IV.E.3.d, infra.

The Commission also preliminarily believes that more effective and efficient regulation of securities markets and market participants resulting from approval of the CAT NMS Plan could significantly benefit investors and the integrity of the market. For example, the Commission preliminarily believes that more effective and efficient surveillance and enforcement would detect a higher proportion of violative market activity. This additional detection could not only reduce violative behavior through potential enforcement actions, but through deterrence if market participants believe violative activities are more likely to be detected. Because violative activity degrades market quality and imposes costs on investors and market participants, reductions in violative activity would benefit investors and market integrity. Likewise, more effective and efficient risk assessment and risk-based examinations should more effectively facilitate the selection of market participants for examination who have characteristics that elevate their risk of violating the rules. Decreasing the amount of violative activity by targeting exams in this way would provide investors with a more effectively regulated trading environment and hence better market quality. Further, access to audit trail data that is comprehensive, accurate, and timely could improve regulatory reconstruction of market events, market analysis, and research resulting in an improved understanding of emerging market issues and regulatory policies that better encourage industry competition, thus improving investor protection through better transparency and more efficient prices.

See Section IV.E.2.a, IV.E.2.b, infra.

Further, regulatory initiatives that are based on a more thorough understanding of underlying events and their causes, and that are narrowly tailored to address any market deficiency, could improve market quality and thus benefit investors. Moreover, access to more complete and linked audit trail data would improve regulators' ability to analyze and reconstruct market events, allowing regulators to provide investors and the public with more accurate explanations of market events, to develop more effective responses to such events, and to use the information to assist in retrospective analyses of their rules and pilots.

The Commission has also evaluated the potential costs that would result from approval of the CAT NMS Plan. In particular, using information included in the Plan, information gathered from market participants through discussions, surveys of market participants, and other relevant information, the Commission has preliminarily estimated the potential costs associated with building and maintaining the Central Repository as well as the costs to report data to the Central Repository. Currently, the 20 Participants spend $154.1 million annually on reporting regulatory data and performing surveillance, while the approximately 1,800 broker-dealers anticipated to have CAT reporting responsibilities spend $1.6 billion annually on regulatory data reporting, for total current industry costs of $1.7 billion annually for regulatory data reporting and surveillance by SROs. The Commission preliminarily estimates the cost of the Plan as approximately $2.4 billion in initial aggregate implementation costs and recurring annual costs of $1.7 billion. The primary driver of the annual costs is the data reporting costs for broker-dealers, which are estimated to be $1.5 billion per year. For both large and small broker-dealers, the primary driver of both current $1.6 billion reporting costs and projected $1.5 billion CAT reporting costs is costs associated with staffing. Estimates of the costs to build the Central Repository are based on Bids that vary in a range as high as $92 million. Current estimates of annual operating costs are based on Bids that vary in a range up to $135 million. The eventual magnitude of Central Repository costs is dependent on the Participants' selection of the Plan Processor, and may ultimately differ from estimates discussed above if Bids are revised as the bidding process progresses. Furthermore, the Plan anticipates a period of duplicative reporting responsibilities preceding the retirement of potentially duplicative regulatory data reporting systems; these duplicative reporting costs are likely to be significant.

See Section IV.F.2, Table 9, infra.

The economic analysis discusses duplicative reporting costs in Section IV.F.2, infra.

Drawing from the discussion in the CAT NMS Plan, the Commission expects that, if approved, the Plan would have a number of additional economic effects, including effects on efficiency, competition, and capital formation. The Commission preliminarily believes that the Plan generally promotes competition. However, the Commission recognizes that the Plan could increase barriers to entry because of the costs to comply with the Plan. Further, the Commission's analysis identifies several limiting factors to competition but Plan provisions and Commission oversight could address such limiting factors. The Commission preliminarily believes that the Plan would improve regulatory analysis and reconstruction of market events, as well as market analysis and research that informs policy decisions. In addition, the Plan would improve enforcement related activities, including the efficiency of regulatory activities such as market surveillance, examinations, investigations, and other enforcement functions that could enhance market efficiency by reducing violative activity that harms market efficiency. Finally, the Commission preliminarily believes that the Plan could have positive effects on capital formation and allocative efficiency and that the threat of a security breach at the Central Repository is unlikely to significantly harm capital formation. The Commission recognizes that the Plan's likely effects on competition, efficiency and capital formation are dependent to some extent on the performance and decisions of the Plan Processor and the Operating Committee in implementing the Plan, and thus there is necessarily some uncertainty in the Commission's analysis. Nonetheless, the Commission believes that the Plan contains certain governance provisions, as well as provisions relating to the selection and removal of the Plan Processor, that mitigate this uncertainty by promoting decision-making that could, on balance, have positive effects on competition, efficiency, and capital formation.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8; see also Section IV.G, infra.

The Commission notes that while the Participants developed the Plan in compliance with Rule 613 by considering information from industry representatives, the Commission has discretion to approve the Plan subject to changes or conditions that the Commission deems necessary or appropriate. Therefore, as a part of this economic analysis, the Commission analyzed numerous alternatives to provisions of the CAT NMS Plan and to the CAT NMS Plan itself. The Commission analyzes alternatives to the approaches the Exemption Order permitted the Participants to include in the Plan; alternatives to certain specific approaches in the Plan; alternatives to the scope of certain specific elements of the Plan; and the broad alternative of modifying OATS or another existing system to meet the requirements of Rule 613 instead of approving the Plan. Finally, the Commission requests comment on alternatives discussed in this economic analysis, alternatives considered in the Plan, and on whether the Commission should consider any additional alternatives.

See 17 CFR 242.608(b)(1) (“No national market system plan . . . shall become effective unless approved by the Commission . . .”); 17 CFR 242.608(b)(2) (“Within 120 days of the date of publication of notice of filing of a national market system plan . . . the Commission shall approve such plan . . . with such changes or subject to such conditions as the Commission may deem necessary or appropriate, if it finds that such plan or amendment is necessary or appropriate in the public interest, for the protection of investors and the maintenance of fair and orderly markets, to remove impediments to, and perfect the mechanisms of, a national market system, or otherwise in furtherance of the purposes of the Act.”).

See Exemption Order, supra note 18.

C. Framework for Economic Analysis

As discussed above, the Commission is conducting an economic analysis of the CAT NMS Plan filed by the SROs on February 27, 2015, as amended, as anticipated in the Adopting Release for Rule 613. In particular, the Commission has carefully evaluated the information in the CAT NMS Plan, including the twelve considerations required by Rule 613 and the details of the decisions left to the discretion of the SROs. The Commission has also considered information drawn from outside the Plan in order to assess potential economic effects not addressed therein. To provide context for this analysis, this Section describes the economic framework for the analysis and seeks to identify uncertainties within that framework.

See Adopting Release, supra note 9, at 45789.

1. Economic Framework

a. Benefits

The CAT NMS Plan would create a new data source that could replace the use of some current data sources for many regulatory activities. As such, the economic benefits of the CAT NMS Plan would come from any expanded and more efficient regulatory activities facilitated by improvements to the data regulators use. Therefore, the framework for examining benefits in this economic analysis involves first considering whether and to what degree the CAT Data would improve on the Baseline of current trading and order data in terms of the four qualities of accuracy, completeness, accessibility, and timeliness.

See Adopting Release, supra note 9, at 45727.

Through these improvements in the data, the economic analysis then considers the degree to which the Plan would result in improvements to regulatory activities such as the analysis and reconstruction of market events, in addition to market analysis and research conducted by SROs and Commission Staff, as well as market surveillance, examinations, investigations, and other enforcement functions. These potential improvements, based on the regulatory objectives of the CAT NMS Plan described in the Adopting Release, relate to the overall goal of substantially enhancing the ability of the SROs and the Commission to oversee securities markets and fulfill their regulatory responsibilities under the securities laws. The economic analysis explores how the improvements to these regulatory activities provide economic benefits to investors and the market. Among other things, potential benefits that could result from the CAT NMS Plan include benefits rooted in changes in the behavior of market participants. For example, requirements to report certain data elements or events to the CAT could have the beneficial effect of deterring rule violations because the inclusion of certain data fields and improvements in the ability to surveil for violations could increase the perceived costs of violating rules and regulations. Potential benefits could also stem from improved investor protection, such as from more effective surveillance and more informed, data-driven rulemaking.

See id. at 45730.

(1) Data Qualities

In the Adopting Release, the Commission identified four qualities of trade and order data that impact the effectiveness of core SRO and Commission regulatory efforts: Accuracy, completeness, accessibility, and timeliness. In assessing the potential benefits of the CAT NMS Plan, the Commission's economic analysis compares the data that would be available under the Plan to the trading and order data currently available to regulators to determine whether and to what degree the Plan would improve the available data with respect to those four qualities.

See id. at 45727. Accuracy refers to whether the data about a particular order or trade is correct and reliable. Completeness refers to whether a data source represents all market activity of interest to regulators, and whether the data is sufficiently detailed to provide the information regulators require. While current data sources provide the trade and order data required by existing rules and regulations, those sources generally do not provide all of the information of interest to regulators in one consolidated audit trail. Accessibility refers to how the data is stored, how practical it is to assemble, aggregate, and process the data, and whether all appropriate regulators could acquire the data they need. Timeliness refers to when the data is available to regulators and how long it would take to process before it could be used for regulatory analysis. As explained in the Baseline, Section IV.D, infra, the trading and order data currently available to regulators suffers from deficiencies in all four dimensions.

(2) Regulatory Activities

Any economic benefits would derive from how such improved data would affect regulatory activities. Therefore, to analyze the potential benefits of the CAT NMS Plan, the economic analysis also evaluates the potential of the CAT NMS Plan to meet the regulatory objectives set out in the Adopting Release for Rule 613. The objectives are: Improvements in the analysis and reconstruction of broad-based market events; improvements in market analysis in support of regulatory decisions; and improvements in market surveillance, investigations, and other enforcement activities.

See Adopting Release, supra note 9, at 45730.

A. Analysis and Reconstruction of Broad-Based Market Events

The economic analysis considers whether and to what extent the CAT NMS Plan would facilitate regulators' performance of analysis and reconstruction of market events, potentially helping to better inform both regulators and investors about such market events and speeding the regulatory response following market events. Regulators perform reconstructions of market events so that they and the public can be informed by an accurate accounting of what happened (and, possibly, why it happened). As discussed in the Benefits Section, market reconstructions can take a significant amount of time, in large measure due to various deficiencies in the currently available trading and order data in terms of the four qualities described above. The sooner regulators complete a reconstruction and analysis of a market event, the sooner investors can be informed and the sooner regulators can begin reviewing the event to determine what happened, who was affected and how, and whether the analysis supports potential regulatory responses. In addition, the improved ability for regulators to generate prompt and complete market reconstructions could provide improved market knowledge, which could assist regulators in conducting retrospective analysis of their rules and pilots.

See Section IV.E.2.a, infra.

See Section IV.C.1.a(1), supra.

See Adopting Release, supra note 9, at 45732.

B. Market Analysis in Support of Regulatory Decisions

The economic analysis considers whether and to what extent the CAT NMS Plan would enhance the ability of the SROs and the Commission to conduct market analysis and research, including analysis of market structure, and the degree to which it would improve regulators' market knowledge and facilitate consideration of policy questions of interest. The SROs and Commission Staff conduct data-driven analysis on market structure, in direct support of both rulemaking and other regulatory decisions such as SRO rule approvals. The Commission also relies on such analysis to improve understanding of market structure in ways that could inform policy. Finally, SROs conduct market analysis and research on their own regulatory initiatives. Improvements in the ability to conduct market analysis could further improve analysis related to regulatory decisions and potentially influence those regulatory decisions to the benefit of investors and the markets more generally.

C. Market Surveillance and Investigations

The economic analysis examines whether the CAT NMS Plan would improve market surveillance and investigations, potentially resulting in more effective oversight of trading, better investor protection, and deterrence of violative behavior. As described in more detail in the Baseline Section, both SROs and the Commission conduct market surveillance, examinations, investigations, and other enforcement functions targeting illegal activities such as insider trading, wash sales, or manipulative practices. Improvements in market surveillance and investigations could come in the form of “facilitating risk-based examinations, allowing more accurate and faster surveillance for manipulation, improving the process for evaluating tips, complaints, and referrals . . ., and promoting innovation in cross-market and principal order surveillance.”

See Section 0, infra.

See Adopting Release, supra note 9, at 45730.

b. Costs

The economic analysis evaluates the costs of building and operating the Central Repository; the costs of CAT reporting for Participants, broker-dealers, and service bureaus; and other CAT-related costs. Where the CAT NMS Plan provides estimates of these costs, the economic analysis evaluates those estimates and re-estimates them when necessary. The economic analysis also discusses the drivers of these costs, and whether broker-dealers may or may not pass these costs down to their customers. In addition, the economic analysis assesses whether the CAT NMS Plan has the potential to result in cost savings. Rule 613 requires the Plan to discuss “[a] plan to eliminate existing rules and systems (or components thereof) that would be rendered duplicative by the consolidated audit trail.” As a part of its consideration of the costs of the CAT NMS Plan, the economic analysis considers costs from duplicative reporting for some period of time as well as potential cost savings from the retirement of duplicative regulatory reporting systems.

The economic analysis also considers whether the CAT NMS Plan could result in second order effects, such as changes to the behavior of market participants, that impose certain costs. For example, the CAT NMS Plan's tiered funding model could lead to costly efforts by market participants to try to control their tiers in order to affect their fee payments, such as reducing activity levels near the end of an activity level measuring period to avoid being classified as a higher activity level firm. In addition, Participants, their members, and investors could incur costs if their private information were accessed in the event of a security breach of the Central Repository. The economic analysis considers these and other elements of the Plan that could lead to distortions in behavior by market participants.

2. Existing Uncertainties

The Commission has carefully analyzed the information in the CAT NMS Plan, as well as other relevant data, in order to assess the economic effects of the Plan. As discussed throughout the analysis, in certain cases the Commission lacks information needed to evaluate all of the potential economic effects of the CAT NMS Plan, creating uncertainty in some potential benefits and costs. The primary drivers of uncertainty include the fee schedule applicable to funding the Central Repository (the “Funding Model”), which has not yet been finalized, the deferral of decisions on certain discretionary elements including the Technical Specifications applicable to the CAT, and a lack of detailed information that would enable the Commission to assess certain economic effects with greater precision. The implications of each primary area of uncertainty for the Commission's economic analysis are discussed below.

First, as noted above, the economic analysis evaluates information provided in the CAT NMS Plan on the economic effects of the Plan, as well as information drawn from outside of the Plan. However, the Commission lacks detailed information regarding some of the individual costs and discretionary decisions in the Plan, including the Funding Model. Specifically, the Plan does not outline the proportion of CAT costs that would be allocated to Participants versus broker-dealers. This uncertainty limits the Commission's ability to evaluate the economic effects of the Plan in some cases. However, the Commission has analyzed the expected economic effects of the Plan to the extent possible with the information available, and where the Commission can identify such areas of uncertainty, the economic analysis addresses this uncertainty. In addition, the Commission requests comments to help resolve such uncertainties during the consideration of the CAT NMS Plan.

Second, certain elements of the CAT NMS Plan would not be finalized until after the selection of a “Plan Processor.” Among these are the security and confidentiality procedures of the Central Repository, the precise methods by which regulators would access data in the Central Repository, and the complete Technical Specifications. The Plan also provides the Plan Processor the “sole discretion” to publish interpretations of the Technical Specifications, including interpretations of permitted values in data elements.

See CAT NMS Plan, supra note 3, at Article VI. The Plan Participants have engaged in a bidding process to select a Plan Processor, and the leading candidate bidders have proposed different solutions. In certain instances, the Plan Participants have decided to adopt the solutions proposed by whichever bidder they select.

See Section 0, infra, for additional discussion of risks and uncertainties related to data security.

Rule 613(e)(1) requires the CAT NMS Plan to create a Central Repository to collect, link, and store CAT Data and to make that data available to regulators. See 17 CFR 242.613(e)(1).

The CAT NMS Plan contains minimum standards and principles for setting many of Technical Specifications, see CAT NMS Plan, supra note 3, at Section 6.9, and the Commission's economic analysis reflects those minimum standards and principles. However, because the detailed Technical Specifications are not yet finalized by the Participants, the Commission cannot fully assess any corresponding costs and benefits.

See id. at Section 6.9.

Because these and other elements of the Plan have not yet been finalized, the Commission cannot assess how and to what extent they could affect the overall economic effects of the Plan. The Commission's economic analysis is therefore limited to the extent that the economic effects of the Plan depend on decisions that would be made after approval of the Plan. However, the Commission has identified these areas of uncertainty and has assessed the economic effects of the Plan to the best of its ability in light of these existing uncertainties.

Given the range of possible outcomes with respect to both the costs and benefits of the CAT NMS Plan that depend on future decisions, the Commission also recognizes the importance of provisions of the Plan related to the operation and administration of the CAT. In particular, governance provisions of the Plan related to voting by the Operating Committee and the involvement of the Advisory Committee may help promote better decision-making by the relevant parties. Such provisions could mitigate concerns about potential uncertainty in the economic effects of the Plan by giving the Commission greater confidence that its expected benefits would be achieved in an efficient manner and that costs resulting from inefficiencies would be avoided. As part of this economic analysis, the Commission therefore considers these features of the Plan.

See Section 0, infra.

3. Request for Comment on the Framework

The Commission requests comment on all aspects of the Framework for the Economic Analysis on the CAT NMS Plan. In particular, the Commission seeks responses to the following questions:

234. Do Commenters believe that the general economic framework applied in this analysis is appropriate? If not, which considerations should be added or removed?

235. Do Commenters agree with the approach to identifying benefits of the CAT NMS Plan? Are there important sources of benefits that are not discussed here? Are the data qualities important for regulatory uses? Are there additional data qualities that the Commission should consider? Are the regulatory objectives important and beneficial for investors? Are there additional regulatory objectives that the Commission should consider?

236. Do Commenters agree with the approach taken in this analysis for examining the costs of CAT? Please explain.

237. Do the Commenters agree with the approach for analyzing second order effects? Are there other sources of economic effects that the Commission should consider?

238. Do Commenters agree with the Commission's characterization of uncertainties in the economic analysis? How important are these uncertainties to the Commission's consideration of the CAT NMS Plan? Are there other sources of uncertainty that the Commission should consider?

239. Do Commenters agree with the Commission's preliminary assessment that governance provisions of the Plan related to voting by the Operating Committee and the involvement of the Advisory Committee may help promote better decision-making by the relevant parties and thus mitigate concerns associated with uncertainties in the economic effects of the Plan? Please explain.

D. Baseline

The CAT NMS Plan would create a new regulatory dataset that SROs and the Commission would use to supplement or replace their current data sources. The Adopting Release states that “improvements [in the quality of audit trail data] should have the potential to result in the following: (1) [I]mproved market surveillance and investigations; (2) improved analysis and reconstructions of broad-based market events; and (3) improved market analysis.” To assess the overall economic impact of the CAT NMS Plan, the economic analysis uses as the Baseline the current state of trade and order data and the current state of regulatory activity that relies on that data. The Baseline discusses the currently available sources of data, limits in available data that could impact regulatory activity, how regulators currently use the available data, and the burden that producing that data imposes on SROs and broker-dealers.

See Adopting Release, supra note 9, at 45730.

1. Current State of Regulatory Activities

The SROs and the Commission use data to analyze and reconstruct market events, conduct market analysis and research in support of regulatory decision-making, and conduct market surveillance, examinations, investigations, and other enforcement functions. The trend in this area is to use more automated and data-intensive methods as regulators' activities adjust to the data and technology available. The following Sections describe these regulatory activities and how regulators currently use data.

a. Analysis and Reconstruction of Market Events

In the Adopting Release, the Commission described how it expected CAT Data to significantly improve the ability of regulators to reconstruct market events so that the public might be informed by an accurate and timely accounting of the events in question. In a market reconstruction, regulators seek to provide an accurate and factual accounting of what transpired during a market event of interest by conducting a thorough analysis of the available market data. These events often encompass activity in many securities across multiple trading venues, requiring the linking and analysis of data from multiple sources. Examples of recent market reconstructions include the Commodity Futures Trading Commission (“CFTC”) and SEC's analysis of the May 6, 2010 “Flash Crash,” analysis of equity market volatility on August 24, 2015, and the multi-agency report on the U.S. Treasuries market on October 15, 2014.

See id. at 45732-33.

See Findings Regarding the Market Events of May 6, 2010: Report of the Staffs of the CFTC and SEC to the Joint Advisory Committee on Emerging Regulatory Issues (September 30, 2010) (“Flash Crash Analysis”), available at http://www.sec.gov/news/studies/2010/marketevents-report.pdf .

See Staff of the Office of Analytics and Research, Division of Trading and Markets, Research Note: Equity Market Volatility on August 24, 2015 (Dec. 2015) available at http://www.sec.gov/marketstructure/research/equity_market_volatility.pdf ; see also Austin Gerig and Keegan Murphy, The Determinants of ETF Trading Pauses on August 24th, 2015, White Paper (February 2016) available at http://www.sec.gov/marketstructure/research/determinants_eft_trading_pauses.pdf .

See U.S. Department of the Treasury, Board of Governors of the Federal Reserve System, Federal Reserve Bank of New York, U.S. Securities and Exchange Commission, and U.S. Commodity Futures Trading Commission, Joint Staff Report: The U.S. Treasury Market on October 15, 2014 (July 13, 2015), available at http://www.sec.gov/reportspubs/special-studies/treasury-market-volatility-10-14-2014-joint-report.pdf .

b. Market Analysis and Research

In the Adopting Release, the Commission described how it expected CAT Data to improve the ability of regulators to monitor overall market structure and better understand its relationship with market behavior, so that the Commission and the SROs could be better informed in their policy decisions. The Commission and SRO Staffs conduct data-driven analysis on market structure, in direct support of both rulemaking and other regulatory decisions such as SRO rule approvals as well as retrospective analyses of rules and pilots. The Commission also relies on data analysis to inform its market structure policy. SROs also conduct market analysis and research on their own regulatory initiatives. Examples of data-driven market analysis include reports on OTC trading, small capitalization stock trading, the Limit Up-Limit Down Pilot, short selling, and high frequency trading.

See Adopting Release, supra note 9, at 45733.

See Laura Tuttle, Alternative Trading Systems: Description of ATS Trading in National Market System Stocks (October 2013) available at http://www.sec.gov/divisions/riskfin/whitepapers/alternative-trading-systems-10-2013.pdf ; Laura Tuttle, OTC Trading: Description of Non-ATS OTC Trading in National Market System Stocks (March 2014), available at http://www.sec.gov/dera/staff-papers/white-papers/otc-trading-white-paper-03-2014.pdf .

See Securities Exchange Act Release No. 74892, Order Approving the National Market System Plan to Implement a Tick Size Pilot Program (May 6, 2015), 80 FR 27514, 27534, 27541 (May 13, 2015); see also Charles Collver, A Characterization of Market Quality for Small Capitalization US Equities (September 2014), available at http://www.sec.gov/marketstructure/research/small_cap_liquidity.pdf .

See SRO Supplemental Joint Assessment, available at http://www.sec.gov/comments/4-631/4-631.shtml ; Memo to File from the Division of Economic and Risk Analysis regarding the Cornerstone Analysis of the Impact of Straddle States on Options Market Quality (February 8, 2016), available at http://www.sec.gov/comments/4-631/4631-42.pdf ; see also Gerig and Murphy, supra note 323.

See Memo to Chairman Christopher Cox from Daniel Aromi and Cecilia Caglio regarding an Analysis of Short Selling Activity during the First Weeks of September 2008, (December 16, 2008) available at http://www.sec.gov/comments/s7-08-09/s70809-369.pdf ; Memo to Chairman Christopher Cox from Daniel Aromi and Cecilia Caglio regarding an Analysis of a Short Sale Price Test Using Intraday Quote and Trade Data (December 17, 2008) available at http://www.sec.gov/comments/s7-08-09/s70809-368.pdf ; Memo from the Office of Economic Analysis regarding an Analysis of the July Emergency Order Requiring a Pre-borrow on Short Sales (January 14, 2009) available at http://www.sec.gov/spotlight/shortsales/oeamemo011409.pdf .

See Austin Gerig, High-Frequency Trading Synchronizes Prices in Financial Markets, available at http://www.sec.gov/dera/staff-papers/working-papers/dera-wp-hft-synchronizes.pdf ; see also Staff of the Office of Analytics and Research, Division of Trading and Markets, Research Note: Equity Market Volatility on August 24, 2015 (December 2015) available at http://www.sec.gov/marketstructure/research/equity_market_volatility.pdf .

c. Market Surveillance and Investigations

Regulators perform market surveillance and investigation functions that rely on access to multiple types of market data. In the Adopting Release, the Commission discussed how data limitations impact surveillance and investigations, including risk-based examinations, market manipulation investigations, tips and complaints, and cross-market and principal order surveillance. The following Sections update and broaden the discussion from the Adopting Release to describe the current state of SRO surveillance and SRO and Commission examinations and enforcement investigations.

See Adopting Release, supra note 9, at 45730-32.

(1) Current SRO Surveillance

Rule 613(f) requires the SROs to develop and implement a surveillance system, or enhance existing surveillance systems, reasonably designed to make use of the CAT Data. For the purposes of this economic analysis, the Commission considers surveillance to involve SROs running automated processes on routinely collected or in-house data to identify potential violations of rules or regulations. As such, surveillance does not include processes run on data that the SROs request only when needed. SRO surveillance can help protect investors by having systems in place that can be used to detect fraudulent behavior and anomalous trading. For instance, SROs use surveillance systems, developed internally or by a third party, to detect violations of trading rules, market abuse, or unusual behavior, in real time, within one day, or within a few weeks of the activity in question. The exchanges are responsible for surveillance of their own exchanges, and FINRA is responsible for off-exchange and cross-market surveillance. FINRA also provides surveillance services to U.S. equity and options exchanges through regulatory services agreements with nearly every equity market and all options exchanges. FINRA also currently conducts several cross-market surveillance patterns, such as surveillance focused on wash sales, front running, relationship trading, and high frequency trading.

See Richard G. Ketchum, FINRA Chairman and CEO, Testimony Before the Subcommittee on Capital Markets and Government Sponsored Enterprises Committee on Financial Services (May 1, 2015), available at https://www.finra.org/newsroom/speeches/050115-testimony-subcommittee-capital-markets-and-government-sponsored-enterprises ; Richard G. Ketchum, FINRA Chairman and CEO, Testimony Before the Subcommittee on Securities, Insurance and Investment, United States Senate (March 3, 2016), available at http://www.finra.org/newsroom/speeches/030316-testimony-subcommittee-securities-insurance-and-investment-united-states .

FINRA has responsibility to oversee and regulate OTC trading of exchange-listed and non-exchange-listed securities, as well as trading in corporate and municipal debt instruments and other fixed income instruments. Also, FINRA conducts cross-market surveillance for approximately 99% of the listed equity market and approximately 70% of the listed options market. To conduct cross-market surveillance, FINRA uses a variety of online and offline surveillance techniques and programs to reconstruct market activity, using trading data and quote information that is captured throughout the trading day, as well as order audit trail data reported daily. FINRA's cross-market surveillance is able to identify a single broker-dealer's manipulative activity across multiple markets, as well as manipulative activity of multiple market participants acting in concert across multiple markets.

See Richard G. Ketchum, FINRA Chairman and CEO, Testimony Before the Subcommittee on Securities, Insurance and Investment, United States Senate (March 3, 2016), available at http://www.finra.org/newsroom/speeches/030316-testimony-subcommittee-securities-insurance-and-investment-united-states .

See FINRA 2015 Regulatory and Examinations Priorities Letter, at 14, available at https://www.finra.org/sites/default/files/p602239.pdf ; see also FINRA 2016 Regulatory and Examinations Priorities Letter, at 12, available at https://www.finra.org/sites/default/files/2016-regulatory-and-examination-priorities-letter.pdf .

Additional surveillance is conducted by exchange-operating SROs, some of it conducted as trading activity occurs. This surveillance can include detection of market manipulation, violations of trading rules, and other unusual behavior.

(2) Examinations

In the Adopting Release, the Commission explained how it expected CAT Data to facilitate risk-based examinations. SROs currently conduct exams of broker-dealers for violations of trading-related federal laws, rules, and regulations and for violations of SRO rules and regulations. In 2015, FINRA's Member Regulation Department conducted approximately 2,400 broker-dealer examinations. The Commission currently conducts exams of broker-dealers, transfer agents, investment advisers, investment companies, municipal advisers, clearing agencies, the national securities exchanges, other SROs such as FINRA and the Municipal Securities Rulemaking Board, and the Public Company Accounting Oversight Board (“PCAOB”). The Commission conducted 493 broker-dealer examinations in 2014 and 484 in 2015, 70 exams of the national securities exchanges and FINRA in 2014 and 21 in 2015. In addition, the Commission conducted 1,237 investment adviser and investment company examinations in 2014 and 1,358 in 2015. Virtually all investment adviser examinations and a significant proportion of the Commission's other examinations involve analysis of trading and order data.

See Adopting Release, supra note 9, at 45730-31.

SEC Rule 17d-2 permits SROs to propose joint plans among two or more SROs for the allocation of regulatory responsibility. Where 17d-2 agreements are in place, SROs have joint plans with respect to their common members (i.e., members of both/all the SROs party to an agreement under Rule 17d-2) for common rules (i.e., rules that are identical or substantially identical). Commission approval of a plan filed pursuant to Rule 17d-2 relieves an SRO of those regulatory responsibilities allocated by the plan to another SRO. See 17 CFR 240.17d-2. Exchanges also enter into Regulatory Services Agreements (“RSAs”) whereby one SRO contractually agrees to perform regulatory services for another. However, RSAs do not relieve the contracting SRO from regulatory responsibility for the performance of any regulatory services allocated pursuant to the RSA and are not filed with the Commission for approval.

This estimate is based on Staff discussions with FINRA. See also FINRA overview of Member Regulation available at http://www.finra.org/industry/member-regulation.

Examinations of broker-dealers and investment advisers involve intensive analysis of trading data. Examinations seek to determine whether the entity being examined is: Conducting its activities in accordance with the federal securities laws, rules adopted under these laws, and SRO rules; adhering to the disclosures it has made to its clients, customers, the general public, SROs and/or the Commission; and implementing supervisory systems and/or compliance policies and procedures that are reasonably designed to ensure that the entity's operations are in compliance with the applicable legal requirements.

See SEC, Examination Information for Entities Subject to Examination or Inspection by the Commission (June, 2014), available at http://www.sec.gov/about/offices/ocie/ocie_exambrochure.pdf.

The Commission and certain SROs, such as FINRA, use a risk-based approach to select candidates and to determine exam scope and focus. “Risk-based examinations” seek to increase regulatory efficiency by using preliminary data analysis to direct examination resources towards entities and activities where risks of violative or illegal activity are the highest. The Commission uses risk and data analysis before opening an exam to identify broker-dealers and investment advisers for areas of focus such as suspicious trading, as well as during an exam to identify the particular activities of a broker-dealer or investment adviser that could trigger certain compliance and supervisory risks.

FINRA conducts regulatory examinations by contract on behalf of all the options and equities exchanges, except for the Chicago Stock Exchange, Inc. (“CHX”) and the National Stock Exchange, Inc. (“NSX”). Accordingly most exchanges also employ a risk-based approach to examination selection and scope. CHX examines members on a cycle basis. NSX recently resumed operations in December, 2015. See Securities Exchange Act Release No. 76640 (December 14, 2015), 80 FR 79122 (December 18, 2015).

Because of the data-intensive nature of examinations, the Commission and SROs have systems, such as the Commission's National Exam Analytics Tool (“NEAT”), to combine, standardize, and analyze exam data. The NEAT system allows examiners to import trade blotter data to conduct commission analysis, cross trades analysis, bunch price analysis, trading pattern analysis, and restricted trade analysis. However, as discussed further below, there are limitations on the trade blotter data imported by the NEAT system.

See Section IV.D.2.b, infra.

(3) Enforcement Investigations

The Adopting Release details how the Commission expects the CAT Data to aid in the analysis of potential manipulation. The Commission and SROs undertake numerous investigations to enforce the securities laws and related rules and regulations, including investigations of market manipulations (e.g., marking the close, order layering, spoofing, wash sales, trading ahead), insider trading, and issuer repurchase violations. As noted below, the Commission estimates that 30-50% of enforcement investigations use trade and order data, and any of these types of investigations, in addition to numerous other investigations, could potentially utilize CAT Data.

See Adopting Release, supra note 9, at 45731.

Layering and spoofing are manipulations where orders are placed close to the best buy or sell price with no intention to trade in an effort to falsely overstate the liquidity in a security.

See infra note 345 and accompanying text. The percentage of enforcement investigations that could be expected to utilize CAT Data depends on the percentage of investigations that involve broker-dealers, investment advisers and investment companies.

SROs rely primarily on surveillance to initiate investigations based on anomalies in the trading of securities. The Commission initiates enforcement investigations when SROs or others submit reliable tips, complaints, or referrals, or when the Commission becomes aware of anomalies indicative of manipulation. After the detection of potential anomalies, a tremendous amount of time and resources are expended in gathering and interpreting trade and order data to construct an accurate picture of when trades were actually executed, what market conditions were in effect at the time of the trade, which traders participated in the trade, and which beneficial owners were affected by the trade. In 2015, the Commission filed 807 enforcement actions, including 39 related to insider trading, 43 related to market manipulation, 124 related to broker-dealers, 126 related to investment advisers/investment companies, and one related to exchange or SRO duties. In 2014, the Commission filed 755 enforcement actions, including 52 related to insider trading, 63 related to market manipulation, 166 related to broker-dealers, and 130 related to investment advisers/investment companies, many of which involved trade and order data. Similarly, FINRA brought 1,397 disciplinary actions in 2014 and 1,512 in 2015.

See Year-by-Year SEC Enforcement Statistics, available at https://www.sec.gov/news/newsroom/images/enfstats.pdf. The total number of actions filed is not necessarily the same as the number of investigations. An investigation may result in no filings, one filing, or multiple filings. Additionally, trade and order data may be utilized in enforcement investigations that do not lead to any filings.

See FINRA statistics available at http://www.finra.org/newsroom/statistics .

(4) Tips and Complaints

The Adopting Release discussed how the Commission expected CAT Data to improve the processes used by the SROs and the Commission for evaluating tips and complaints. Market participants or those with experience in analyzing market data sometimes notice atypical trading or quoting patterns in publicly available market data, and these observations sometimes result in a tip or complaint to a regulator. Regulators investigate thousands of tips and complaints each year. In fiscal years 2014 and 2015, the Commission received around 15,000 entries in its Tips, Complaints and Referrals (“TCR”) system, approximately one third of which related to manipulation, insider trading, market events, or other trading and pricing issues.

See Adopting Release, supra note 9, at 45731-32.

Analysis of tips and complaints follows three general stages. First, regulators ensure that the tip or complaint contains sufficient information to facilitate analysis. The second stage involves a triaging effort in which regulators may use directly accessible data or make phone calls and other informal queries to determine if the tip or complaint is credible. For tips and complaints that seem credible, the third stage involves a more in-depth investigation or examination, which follows the processes described above for examinations and enforcement investigations.

2. Current State of Trade and Order Data

To assess how and to what degree the CAT NMS Plan would affect the trade and order data available to regulators, the economic analysis considers what data regulators use currently and the limitations in that data.

a. Current Sources of Trade and Order Data

The SROs and the Commission currently use a range of trading and order data sources for the regulatory activities discussed above. The types of data and ease of use can vary widely from one source to the next. Some data sources provide access to in-depth information on a narrow slice of the market, while others reveal more limited information but with broader market coverage. This Section reviews the primary sources of data currently available to regulators, describing the content of the data provided and examples of their specialized uses. There are limitations on each of the data sources discussed below that reduce their usefulness for regulatory purposes. These limitations and their impact on the ability of the SROs and the Commission to use the data sources for regulatory purposes are explained in Section IV.D.2.b below.

(1) SRO Data

Most SROs maintain audit trails that contain the trade and order data that they obtain from members. Regulators have access to at least three sources of audit trail data. First, the National Association of Securities Dealers (“NASD”) established its Order Audit Trail System (“OATS”) in 1998, which required NASD (n/k/a FINRA) members to report certain trade and order data regarding NASDAQ-listed equity securities. OATS was later expanded to include OTC equity securities and all NMS stocks. Second, beginning in 2000, several of the current options exchanges implemented the Consolidated Options Audit Trail System (“COATS”). Finally, each equity and options exchange keeps an audit trail of orders and trades that occur on its market.

In 2007, NASD and the member-related functions of NYSE Regulation, Inc., the regulatory subsidiary of New York Stock Exchange LLC (“NYSE”), were consolidated. As part of this regulatory consolidation, the NASD changed its name to FINRA. See Securities Exchange Act Release No. 56146 (July 26, 2007), 72 FR 42190 (August 1, 2007). FINRA and the National Futures Association (“NFA”) are currently the only national securities associations registered with the Commission; however, the NFA has a limited purpose registration with the Commission under Section 15A(k) of the Exchange Act. 15 U.S.C. 78o-3(k); see also Securities Exchange Act Release No. 44823 (September 20, 2001), 66 FR 49439 (September 27, 2001).

See Securities Exchange Act Release No. 39729 (March 6, 1998), 63 FR 12559 (March 13, 1998) (order approving proposed rules comprising OATS) (“OATS Approval Order”).

The FINRA Web site states: “FINRA has established the Order Audit Trail System (OATS), as an integrated audit trail of order, quote, and trade information for all NMS stocks and OTC equity securities. FINRA uses this audit trail system to recreate events in the life cycle of orders and more completely monitor the trading practices of member firms.” FINRA, OATS, available at http://www.finra.org/industry/oats (listing further information on OATS).

See Securities Exchange Act Release No. 63311 (November 12, 2010), 75 FR 70757 (November 18, 2010) (order approving proposed rule change by FINRA relating to the expansion of OATS to all NMS stocks).

See, e.g., In the Matter of Certain Activities of Options Exchanges, Order Instituting Public Administrative Proceedings Pursuant to Section 19(h)(1) of the Securities Exchange Act of 1934, Making Findings and Imposing Remedial Sanctions, Securities Exchange Act Release No. 43268 (September 11, 2000) (“Options Settlement Order”); Securities Exchange Act Release No. 50996 (January 7, 2005), 70 FR 2436 (January 13, 2005) (order approving proposed rule change by Chicago Board Options Exchange, Incorporated (“CBOE”) relating to Phase V of COATS).

See, e.g., infra notes 358-364 and accompanying text. For example, the NYSE tracks counterparties on every trade in its Consolidated Equity Audit Trail Data (“CAUD”) system, and records electronic order events in a System Order Data (“SOD”) database. See Proposing Release, supra note 9, at 32564-68 (proposing Consolidated Audit Trail and discussing equity exchange audit trails). The SROs provided data in various proprietary formats to the Commission in support of the investigation of the May 6th, 2010 “Flash Crash.” These data sources are briefly discussed in the Flash Crash Analysis, supra note 322.

Specifically, for each of these stages in the life of an order, FINRA Rule 7440 requires the recording and reporting of the following information, as applicable, including but not limited to: For the receipt or origination of the order, the date and time the order was first originated or received by the reporting member, a unique order identifier, the market participant symbol of the receiving reporting member, and the material terms of the order; for the internal or external routing of an order, the unique order identifier, the market participant symbol of the member to which the order was transmitted, the identification and nature of the department to which the order was transmitted if transmitted internally, the date and time the order was received by the market participant or department to which the order was transmitted, the material terms of the order as transmitted, the date and time the order was transmitted, and the market participant symbol of the member who transmitted the order; for the modification or cancellation of an order, a new unique order identifier, original unique order identifier, the date and time a modification or cancellation was originated or received, and the date and time the order was first received or originated; and for the execution of an order, in whole or in part, the unique order identifier, the designation of the order as fully or partially executed, the number of shares to which a partial execution applies and the number of unexecuted shares remaining, the date and time of execution, the execution price, the capacity in which the member executed the transaction, the identification of the market where the trade was reported, and the date and time the order was originally received. FINRA Rule 7440 also requires reporting of the account type, the identification of the department or terminal where an order is received from a customer, the identification of the department or terminal where an order is originated by a reporting member, and the identification of a reporting agent if the agent has agreed to take on the responsibilities of a reporting member under Rule 7450.

The specific information required to be reported includes: The number of shares; designation as a buy or sell or short sale; designation of the order as market, limit, stop, or stop limit; limit or stop price; date on which the order expires and if the time in force is less than one day, the time when the order expires; the time limit during which the order is in force; any request by a customer that a limit order not be displayed, or that a block size limit order be displayed, pursuant to Rule 604(b) of Regulation NMS; any special handling requests; and identification of the order as related to a program trade or index arbitrage trade. See FINRA Rule 7440(b).

The specific information required includes the number of shares to which the transmission applies, and whether the order is an intermarket sweep order. See FINRA Rule 7440(c).

For cancellations or modifications, the following information also is required: If the open balance of an order is canceled after a partial execution, the number of shares canceled; and whether the order was canceled on the instruction of a customer or the reporting member. See FINRA Rule 7440(d).

“Account type” refers to the type of beneficial owner of the account for which the order was received or originated. Examples include institutional customer, individual customer, employee account, market making, and proprietary. See FINRA, OATS Reporting Technical Specifications, at 4-2, available at http://www.finra.org/sites/default/files/OATSTechSpec_01112016.pdf.

A majority of options exchanges require their members to provide the following information with respect to orders entered onto their exchange: (1) The material terms of the order; (2) order receipt time; (3) account type; (4) the time a modification is received; (5) the time a cancellation is received; (6) execution time; and (7) the clearing member identifier of the parties to the transaction.

The specific information required includes option symbol; underlying security; expiration month; exercise price; contract volume; call/put; buy/sell; opening/closing transaction; price or price limit; and special instructions. See, e.g., BATS Exchange, Inc. (“BATS”) Rule 20.7; BOX Options Exchange LLC (“BOX”) Chapter V, Section 15; CBOE Chapter VI, Rules 6.24 and 6.51; NASDAQ Options Market (“NOM”) Rule Chapter V, Section 7; NYSE Amex Rules 153, Commentary .01, and 962; NYSE Arca Rules 6.67, 6.68, and 6.69; and NASDAQ OMX PHLX LLC (“Phlx”) Rules 1063 and 1080.

The required information also includes identification of the terminal or individual completing the order ticket. See id.

See id.

Although SROs that operate exchanges collect much of their audit trail information directly from their internal systems, broker-dealers also have the responsibility to report regulatory data to SRO audit trails. Some broker-dealers perform nearly all of these data reporting requirements in-house, whereas others contract with service bureaus to accomplish this data reporting. This reporting can represent a significant burden on broker-dealers.

See Section IV.F.1.c(2), infra, for a discussion of how broker-dealers decide whether or not to outsource their regulatory reporting.

Audit trail data have become more useful to regulators over time. As noted above, FINRA expanded OATS from covering only NASDAQ listed securities to include OTC equity securities and all NMS stocks. Commission Staff understands that FINRA has also begun collecting additional SRO audit trail data, provided voluntarily from most exchanges, to supplement OATS data. In addition, NYSE, NYSE Amex LLC (n/k/a “NYSE MKT LLC”) (“NYSE Amex”), and NYSE ARCA, Inc. (“NYSE Arca”) eliminated their OTS audit trail requirements and replaced them to coordinate with the OATS requirements, so that members who are also members of either FINRA or NASDAQ (and therefore subject to OATS requirements) are able to satisfy their reporting obligations by meeting the OATS requirements. As a result of all of these changes, the combined data from these different audit trails now cover most order events in equities.

See supra note 351.

See Securities Exchange Act Release No. 65523 (October 7, 2011), 76 FR 64154 (October 17, 2011) (concerning NYSE); Securities Exchange Act Release No. 65524 (October 7, 2011), 76 FR 64151 (October 17, 2011) (concerning NYSE Amex); Securities Exchange Act Release No. 65544 (October 12, 2011), 76 FR 64406 (October 18, 2011) (concerning NYSE Arca).

Other SRO audit trails have varied reporting requirements. Some exchanges have detailed audit trail data submission requirements for their members covering order entry, transmittal, and execution. See CHX Article 11, Rule 3(b); NASDAQ Rules 6950-6958 (substantially similar to the OATS rules); NASDAQ OMX BX Rules 6950-6958 (substantially similar to OATS rules). The audit trail rules of the other exchanges incorporate only standard books and records requirements in accordance with Section 17 of the Exchange Act, 15 U.S.C. 78q. See, e.g., NSX Chapter VI, Rule 4.1.; BATS Chapter IV, Rule 4.1; CBOE Rule 15.1 (applicable to CBOE Stock Exchange (“CBSX”)); International Securities Exchange, LLC (“ISE”) Rule 1400; NYSE Arca Equities Rule 2.24. One exchange only requires its members to make and keep books and records and other correspondence in conformity with Section 17 of the Exchange Act and the rules thereunder, with all other applicable laws and the rules, regulations and statements of policy promulgated thereunder, and with the exchange's rules. See NSX Chapter VI, Rule 4.1. Though not an audit trail, the Large Options Position Report (“LOPR”) is also a source of SRO data that is used for surveillance, examination, and enforcement purposes by SRO and Commission staff. The data is collected pursuant to FINRA Rule 2360(b)(5), Reporting of Options Positions, under which each member must file a report for each account in which they have an interest in a position of 200 or more options contracts, on the same side of the market. Any increases or decreases in this position must also be reported. The Options Clearing Corporation (“OCC”) is the service provider for the processing of these reports, which are used at will by the SROs for surveillance purposes. The Commission also frequently uses LOPR for enforcement investigations of insider trading and market manipulation cases.

SRO audit trail data is used for market reconstructions and market analyses, and to inform policy decisions, both by the Commission and by SROs. Regulators also use SRO audit trail data extensively for surveillance, examinations, investigations, and other enforcement functions. Current SRO market surveillance relies primarily on data from the SRO audit trails, generated directly from the exchange servers and from OATS. Likewise, SRO examinations and investigations pull information from their own audit trails before seeking data from others. Commission examinations and investigations also rely heavily on SRO audit trails to start the process of tracing a particular trade from its execution to the order initiations and customer information, and the audit trails can be useful for manipulation investigations or other regulatory activities that require analyses of microcap securities trading activity. There are, however, limitations on SRO audit trail data that reduce their usefulness to regulators. For example, for the examinations mentioned above, Commission examination Staff may undertake a laborious process of linking SRO audit trail data with EBS data, because SRO audit trail data does not contain customer information. These and other limitations are discussed in Section IV.D.2.b, infra.

See Section IV.D.2.b, infra.

(2) Equity and Option Cleared Reports

The SROs and Commission also have access to equity and option cleared reports. Clearing broker-dealers report their equity and option cleared data on a daily basis and the NSCC and the OCC aggregate the data across the market and generate the reports. The reports show the number of trades and daily cleared trade and share volume, by clearing member, for each equity and listed option security in which transactions took place. Regulators can query these reports directly through an internal online system that interfaces with the Depository Trust and Clearing Corporation (“DTCC”) data by security name and CUSIP number. The originating source of the DTCC cleared equity data is the Securities Information Automation Corporation (“SIAC”) and the originating source of the cleared options data is the OCC.

NSCC provides clearing, settlement, risk management, central counterparty services and a guarantee of completion for certain transactions for virtually all broker-to-broker trades involving equities, corporate and municipal debt, American depositary receipts, exchange-traded funds, and unit investment trusts. See DTCC, About DTCC, NSCC, available at http://www.dtcc.com/about/businesses-and-subsidiaries/nscc.aspx. The OCC is an equity derivatives clearing organization that is registered as a clearing agency under Section 17A of the Act, 15 U.S.C. 78q-1, and operates under the jurisdiction of both the Commission and the CFTC. See OCC, About OCC, available at http://www.optionsclearing.com/about/corporate-information/what-is-occ.jsp.

A CUSIP number is a unique alphanumeric identifier assigned to a security and facilitates the clearance and settlement of trades in the security. See SEC, Fast Answers, CUSIP Number, available at www.sec.gov/answers/cusip.htm.

Equity and option cleared reports provide a way for regulators to directly access a dataset to see how much trading volume is accounted for by a particular clearing broker. As such, these data are often used at the beginning of an examination or investigation to start identifying the market participants that may have additional data needed to pinpoint a particular activity. But there are limitations on these reports that reduce their usefulness to regulators. For example, the information available on the reports is limited to the date, the clearing firm, and the number of transactions cleared by each clearing firm on each SRO. These and other limitations are discussed in Section IV.D.2.b, infra.

(3) Electronic Blue Sheets

Broker-dealers provide detailed data to regulators in the form of EBS. The EBS data, provided pursuant to Rule 17a-25 under the Act, facilitate investigations by the SROs and Commission Staff, particularly in the areas of insider trading and market manipulations. The EBS system provides certain detailed execution information in its electronic format upon request by SRO or Commission Staff. This information often includes the employer of the beneficial owner of an account, which can be important to insider trading investigations, and in some cases, a tax identification number.

17 CFR 240.17a-25. Rule 17a-25 codified the requirement that broker-dealers submit to the Commission, upon request, information on their customer and proprietary securities transactions in an electronic format. The Rule requires submission of the same standard customer and proprietary transaction information that SROs request through the EBS system in connection with their market surveillance and enforcement inquiries.

For a proprietary transaction, Rule 17a-25 requires a broker-dealer to provide the following information electronically upon request: (1) Clearing house number or alpha symbol used by the broker-dealer submitting the information; (2) clearing house number(s) or alpha symbol(s) of the broker-dealer(s) on the opposite side to the trade; (3) security identifier; (4) execution date; (5) quantity executed; (6) transaction price; (7) account number; (8) identity of the exchange or market where the transaction was executed; (9) prime broker identifier; (10) average price account identifier; and (11) the identifier assigned to the account by a depository institution. See Rule 17a-25(a)(1), (b)(1)-(3), 17 CFR 240.17a-25(a)(1), (b)(1)-(3). For customer transactions, the broker-dealer also is required to include the customer's name, customer's address, the customer's tax identification number, and other related account information. See Rule 17a-25(a)(2), 17 CFR 240.17a-25(a)(2); see also infra note 372 and accompanying text (discussing additional information on “large traders” reported through EBS).

Employer information is required by some SRO EBS rules. See, e.g., NYSE and FINRA Rule 8211. While employer information is not required under Rule 17a-25, Commission staff sometimes request and receive this information.

Tax identification numbers are not required to be reported in EBS for average price, allocation, riskless principal, foreign accounts, and subaccounts.

The EBS system also provides additional information on market participants who meet the definition of “large traders” and have self-identified to the Commission as required by Rule 13h-1. Large traders who file Form 13H with the Commission are assigned a “large trader identification number” by the Commission and must provide that number to their brokers for inclusion in the EBS records that are maintained by the clearing brokers. Rule 13h-1, subject to relief granted by the Commission, requires that execution time be captured (to the second) for certain categories of large traders. Large trader data provide the Commission with a way to acquire information about the activities of large traders and allow the activities of large traders to be more readily aggregated across or partitioned by multiple broker-dealers. Regulators generally use data from the EBS system extensively in enforcement investigations, for which EBS data are vital, particularly insider trading investigations. But again, there are limitations on EBS data. For example, EBS data are cumbersome to use for broad analyses, such as analysis and reconstruction of market events, market analysis and research, and some examinations, because of the fragmentation of the data. These and other limitations are discussed in Section IV.D.2.b, infra.

See Securities Exchange Act Release No. 64976 (July 27, 2011), 76 FR 46960 (August 3, 2011). A “large trader” is defined as a person whose transactions in NMS securities equal or exceed 2 million shares or $20 million during any calendar day, or 20 million shares or $200 million during any calendar month. SEC Rule 13h-1, 17 CFR 240.13h-1, requires those market participants who meet the definition of “large traders” to comply with a number of requirements, including filing Form 13H with the Commission to receive a large trader identification number. Id.

See Securities Exchange Act Release No. 76322 (October 30, 2015), 80 FR 68590 (November 5, 2015).

(4) Trade Blotters and Order Tickets

Investment advisers and broker-dealers maintain data in the form of order tickets and trade blotters that regulators can obtain on request. Order tickets are in-house records maintained by investment advisers and broker-dealers that provide order details, including time stamps of order initiation and placement, special order types, any special instructions for the order, and plans for the allocation of shares and prices across accounts and subaccounts. Order tickets also identify account owners. Commission Staff collects order tickets regularly for examinations, and occasionally also for market manipulation investigations.

Rule 204-2 requires investment advisers to maintain a memorandum of each order given by the investment adviser for the purchase or sale of any security. 17 CFR 275.204-2(a)(3). Rule 17a-3(a)(1) requires broker-dealers to maintain a trade blotter. 17 CFR 240.17a-3(a)(1).

Broker-dealers maintain data in trade blotters that are similar to EBS. However, the trade blotters also contain more information, including the commissions paid in executing each order, time stamps of when an order is received and when it is executed (and the number of fills), and the pricing information for all executions in the order. SROs use trade blotters in examinations of their members. Commission Staff uses trade blotters frequently for examinations, including in almost every broker-dealer, investment adviser, and hedge fund examination, as well as for insider trading and market manipulation investigations. Regulators use trade blotter data to determine the order entry time and execution time for trades by a particular customer in examinations and enforcement investigations. Trade blotters are also the primary data source used in regulatory investigations for which subaccount allocation information is important for determining violative behavior, such as cherry-picking and front-running cases. There are limitations on trade blotter and order ticket data that reduce their usefulness to regulators, however. For example, regulators lack direct access to these data; in order to acquire trade blotter and order ticket data, regulators need to send a request to each individual broker-dealer to obtain its data, which can be a lengthy and cumbersome process. These and other limitations are discussed in Section IV.D.2.b, infra.

Regulators could also request a trade confirmation instead of a trade blotter. A trade confirmation shows the customer, the symbol, execution price, trade date, settlement date and commission. A trade blotter is more detailed than a trade confirmation. A trade blotter is what a firm itself records and the exact information recorded varies by firm. Typically, regulators look to the trade confirmation when they have questions about the veracity of a firm's blotter, but generally prefer to request the trade blotter due to its greater detail.

(5) Trading and Order Handling System Data

Broker-dealers and exchanges also collect and maintain records of activity in their order handling systems and internal matching systems. This data may include order receipt, modification or routing information not otherwise reported to SROs. Some elements of these data exceed the scope of information captured in EBS, SRO audit trail, trade blotter, or order ticket data; for example, SRO audit trail data sometimes excludes market-making activity. But certain market making activity is included in the data that broker-dealers and exchanges are required to maintain pursuant to Section 17(a) of the Act and Rule 17a-3 thereunder. Regulators use these trading and order handling system data in investigations and examinations to further analyze issues discovered during their analysis of data from other sources. Like other current sources of data, there are limitations on trading and order handling system data that reduce their usefulness to regulators. For example, a lack of standardization results in variations in trading and order handling system data across broker-dealers. These and other limitations are discussed in Section IV.D.2.b, infra.

Internal matching systems of broker-dealers may include Alternative Trading Systems (“ATSs”) or automated trading systems that provide liquidity to received orders without interacting on a registered exchange. The Commission understands that some broker-dealers rely on their clearing firms to collect and maintain records relating to routed orders on their behalf. Broker-dealers that operate their own internal matching systems are more likely to collect and maintain their own records.

17 CFR 240.17a-3. For example, market makers are only required to report information on orders that are executed.

(6) Public Data

Exchanges and SROs also make data available to the public, in some cases on a commercially-available basis, that regulators could access for their regulatory activities. One type of public data is “consolidated” data feeds that are disseminated by registered Securities Information Processors (“SIPs”) pursuant to joint SRO plans. For a fee, the SIPs distribute consolidated market data on recent equity and option transactions and the prevailing best quotes at each exchange to market data subscribers. In addition, all exchanges also make data available through direct data feeds. These feeds contain all data included in the SIP feed, but also include depth of book information and, depending on the exchange, may include additional data, such as the submission, cancellation and execution of all displayed orders and auction imbalance information on the exchange, among other things.

In other words, the exchanges and SROs sell the data publicly and regulators can purchase it.

ICE serves as the operator for the Consolidated Tape Association (“CTA”) Plan SIP and the Consolidated Quote System (“CQS”) Plan SIP. These SIPs collect and disseminate information on quotes and trades in listed securities, other than NASDAQ listed securities. The NASDAQ Stock Market LLC serves as the operator for the Unlisted Trading Privileges (“UTP”) Plan SIP, which collects and disseminates quote and trade information in NASDAQ listed securities.

An exchange's order book consists of all unexecuted orders at each price. Order book data typically includes the depth (aggregated number of shares) of the displayed orders at each price and might include all prices in the order book or the depth at each price over a range of prices. Displayed orders consist of any order in which the submitter did not instruct that some or all of the order be hidden from display.

The SEC's Market Information Data Analytics System (“MIDAS”) uses information disseminated by the SIP feeds, as well as exchange direct feeds consisting of data that individual exchanges choose to sell to subscribers. In addition, at the request of Commission Staff, most equities exchanges produce and make public two datasets with information on short sales: A file of short selling volume by stock, which contains the short selling and total volume on that exchange by symbol, and a file of short selling transactions, which contains trade information such as time, volume, and price for each transaction involving a short sale.

See Short Sale Reporting Study, infra note 413, for more information on available short selling data and the demands for additional short selling data. This study also describes information regarding data from Form SH filings. For ten months starting during the financial crisis, the Commission required certain institutional investors to submit weekly reports of their short selling activity and positions.

The Commission and SROs use these publicly available trade and order data to conduct market analyses, market reconstructions, examinations, and investigations. Because of the accessibility and ease of use of the public data, regulators often use it as a starting point or a basis of comparison to other data sources. For example, real-time surveillance can rely on SIP data, and some insider trading surveillance relies on information from other publicly available sources such as news sources. Further, investigations into short sale market manipulation sometimes start with an analysis of the short selling data. Some market analyses by regulators rely on public data alone. However, there are limitations on these data that reduce their usefulness to regulators. For example, they do not provide customer information, order entry time, information about special order handling codes, counterparties, or member identifiers. These and other limitations are discussed in Section IV.D.2.b, infra.

See Collver, supra note 327.

See also Staff of the Office of Analytics and Research, Division of Trading and Markets, Research Note: Equity Market Volatility on August 24, 2015 (December 2015) available at http://www.sec.gov/marketstructure/research/equity_market_volatility.pdf .

b. Current Limitations of Trade and Order Data

Although regulators have access to trade and order data from the sources described above, the available data are, for various reasons, limited in terms of the four qualities discussed above. In terms of completeness, current sources do not represent all of the market activity of interest in sufficient detail in one consolidated audit trail. In terms of accuracy, current sources may reflect data errors, insufficiently granular clock synchronization and time stamps, errors introduced in the process of combining data from different sources, a lack of consistent customer and broker-dealer identifiers, and data that is too aggregated at the record level to provide the information regulators need. With respect to accessibility, the SROs and Commission lack direct access to most of the data sources described above, and with respect to timeliness, obtaining trade and order data from current sources and converting the data into a form in which they can be analyzed can involve a significant delay from the time of a particular event of interest. The qualities of market data are important to the Commission's ability to fulfill its statutory mission in an efficient and effective manner. As a result of the limitations on current data sources, regulators are limited in their ability to perform the activities outlined in Section IV.D.1, above. Table 2: Currently Available Data Sources summarizes the key characteristics of the currently available data sources, which are discussed in more detail below.

See Section IV.D.2.a, supra.

As discussed above and in the Adopting Release, accuracy refers to whether the data about a particular order or trade is correct and reliable; completeness refers to whether the data represents all market activity of interest or just a subset, and whether the data is sufficiently detailed to provide the required information; accessibility refers to how the data is stored, how practical it is to assemble, aggregate, and process the data, and whether all appropriate regulators could acquire the data they need; and timeliness refers to when the data is available to regulators and how long it would take to process before it could be used for regulatory analysis. See supra note 306.

(1) Completeness

“Completeness” refers to whether the data represents all market activity of interest or just a subset, and whether the data is sufficiently detailed to provide the required information. While current data sources provide trade and order data specified by existing rules and regulations, those sources do not contain all market activity that might be required for certain market inquiries, in sufficient detail, within one consolidated audit trail. To obtain information regarding a particular market event, regulators may have to piece together information from different data sources. Further, some data is not required to be reported at all under existing regulations. Therefore, current data sources either cover only a limited number of events and products, or lack some data fields that would be useful to regulators, each of which impedes effective market surveillance.

See supra note 306.

See, e.g., Adopting Release, supra note 9, at 45726-30, 45741, 45750 n.286, 45756 n.361 (discussing the incompleteness of the data recorded by existing audit trail systems such as OATS, acknowledging that “certain elements are not collected by existing audit trails,” and noting that “existing SRO audit trails do not require customer information to be reported”); see also Proposing Release, supra note 9, at 32564-66, 32603 (discussing gaps in current required audit trail information and stating that the proposed rule would require “national securities exchanges, national securities associations, and their members to capture . . . information that is not currently captured under the existing audit trail or other regulatory requirements”).

A. Events and Products

There is currently no single data source that covers all market activities. EBS data contains executed trades but does not contain information on orders or quotes (and thus does not provide information on routes, modifications, or cancellations). Similarly, trade blotters and order tickets contain only information recorded by that particular broker-dealer or investment adviser and may contain limited information about full order lifecycles. SRO audit trail data are limited to identifying the activity of their members, can have incomplete information concerning their members, lack order lifecycle information occurring prior to receipt by an exchange, and may not contain information regarding principal trading. Furthermore, public consolidated and direct data feeds provide data about the entire market, but lack information regarding non-displayed orders and do not provide sufficient information to identify the different lifecycle events of a single order.

Individual SRO audit trails are extensive but still incomplete in their coverage of the activities of the market participants they cover; they contain only activity of their own members and many do not necessarily contain all activity by their members. For example, FINRA's OATS data does not include proprietary orders originated by a trading desk in the ordinary course of a member's market making activities, or options data. And while OATS collects data from FINRA members with respect to orders and trades involving NMS and OTC stocks, OATS does not include trade or order activity that occurs on exchanges or at broker-dealers that are not FINRA members. In addition, while broker-dealers who are not members of FINRA must be members of an exchange SRO, an individual exchange SRO's audit trail data is generally limited to activity taking place on that exchange. Because broker-dealers who are not members of FINRA may engage in trading activity in off-exchange markets, a substantial portion of the trading activity that an exchange SRO supervises is not reported to the supervising SRO.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(ii)(A). OATS includes records showing the routing of an order to an exchange, but not the outcome of that routing. In performing its regulatory oversight of the markets, FINRA has created an internal process in which it augments the data it collects via OATS with trade execution data from other exchanges with which it has regulatory service agreements. This process provides FINRA with a wider view of the markets than OATS previously provided, but linking data across these sources does not yield fully accurate results. See Section IV.D.2.b(2), infra for a discussion of the accuracy of linking across data sources. See infra note 1060 for a discussion of FINRA's RSAs.

Currently, Rule 15b9-1 offers an exemption from FINRA membership that applies if the firm is a member of a national securities exchange, carries no customer accounts, and has annual gross income of no more than $1,000 that is derived from securities transactions effected otherwise than on a national securities exchange of which it is a member (the `de minimis allowance'). Income derived from transactions for that dealer's own account with or through another registered broker-dealer do not count toward the $1,000 de minimis allowance. However, the national securities exchanges have not generally supervised their members' activity outside of the markets they operate. The Commission has proposed modifications to Rule 15b9-1 that would require a dealer to be a member of a registered national securities association to conduct most off-exchange activity. See Securities Exchange Act Release No. 74581 (March 25, 2015), 80 FR 18035, 18042 (April 2, 2015) (“Exemption for Certain Exchange Members”) (proposing to amend rule 15b9-1 and noting that “[n]on-Member Firms are not subject to oversight by [FINRA] and their off-exchange transactions typically are not overseen by the exchanges of which they may be members,” and that “[e]xchanges traditionally have not assumed the role of regulating the totality of the trading of their member-broker-dealers . . .”).

Id. at 18043 n.85. Broker-dealers that are not FINRA members accounted for 48% of orders sent directly to ATSs in 2014. Therefore, OATS includes incomplete information on a substantial portion of off-exchange trading. As of March 2015, 125 of the approximately 4,209 registered broker-dealers were not members of FINRA. Id. at 18052. Orders from non-FINRA members accounted for 40% of orders sent directly to ATSs in 2013, and 32% in 2012. Id. at 18038 n.21.

Further, not all FINRA members are obligated to report to OATS. FINRA's rules exempt from reporting certain members that engage in a non-discretionary order routing process. Additionally, FINRA has the authority to exempt other members who meet specific criteria from the OATS recording and reporting requirements, and has granted approximately 50 such exemptions.

See FINRA Rule 7410 (Definitions). The Rule specifically excludes from the definition of “Reporting Member” members that (1) engage in a non-discretionary order routing process and route all of their orders either to a single receiving Reporting Member or two Reporting Members, provided orders are routed to each receiving Reporting Member on a pre-determined schedule and the time period for the schedule does not exceed one year; (2) do not direct or maintain control over subsequent routing or execution by the receiving Reporting Member; and (3) have a written agreement with the receiving Reporting Member that specifies the respective functions and responsibilities of each party to effect full compliance with the OATS recording and reporting rules. Finally, the receiving Reporting Member must record and report all required information pertaining to the order.

See FINRA Rule 7470 (Exemption to the Order Recording and Data Transmission Requirements). The Rule provides that, for good cause shown, FINRA may exempt a member from its recording and reporting requirements if: (1) The member and current control affiliates and associated persons of the member have not been subject within the last five years to any final disciplinary action, and within the last ten years to any disciplinary action involving fraud; (2) the member has annual revenues of less than $2 million; (3) the member does not conduct any market making activities in NMS stock or OTC securities; (4) the member does not execute principal transactions with its customers; and (5) the member does not conduct clearing or carrying activities for other firms. This authority sunsets on July 10, 2019. Approximately 799 firms that are excluded or exempt from OATS would incur CAT reporting obligations if the Plan were approved; see also infra note 931, Section IV.F.1.c(2)B.i, infra.

Exchange audit trails also lack information on the order lifecycle events that occur prior to receipt at the exchange. SRO audit trail data available from the Intermarket Surveillance Group (“ISG”) does not capture quotes/orders away from a market's inside market (i.e., those quotes/orders below the best bid or above the best offer); currently identify market participants in a trade only to the clearing broker level; do not provide information on the executing broker; and contain certain data fields that are not mandatory.

The Commission understands that exchange routing broker-dealers, which route orders from exchanges to other Execution Venues, do substantial business, but it is very hard in current data sources to track orders sent to one exchange that are then sent to another exchange or off-exchange venue by the exchange routing broker-dealer.

The ISG was established in the early 1980s and is comprised of over 50 international exchanges, market centers, and market regulators that perform market surveillance in their respective jurisdictions. The purpose of the ISG is to provide a framework for the sharing of information and the coordination of regulatory efforts among exchanges trading securities, options on securities, security futures products, and futures and options on broad-based security indexes, to address potential inter-market manipulations and trading abuses. In effect, the ISG is an information-sharing cooperative governed by a written agreement. ISG also provides a forum for ISG members to discuss common regulatory concerns, thus enhancing members' ability to efficiently fulfill their regulatory responsibilities. As a condition to membership, every ISG member must represent that it has the ability to obtain and freely share regulatory information and documents with other ISG members, generally unencumbered by rules, nationally imposed blocking statutes or bank secrecy laws. Regulatory information is only shared on an as-needed basis and only upon request, and any information shared through ISG must be kept strictly confidential and used only for regulatory purposes. The SEC is not a member of ISG, nor is ISG subject to regulatory oversight by the SEC.

See Comment Letter from FINRA and NYSE Euronext regarding Proposing Release (August 9, 2010), available at https://www.sec.gov/comments/s7-11-10/s71110-46.pdf .

Additionally, some SRO audit trails do not include and are not required to include activity associated with principal trading, such as market-making activity. This may result in the exclusion of a significant amount of activity, particularly for firms with substantial market-making business activities. Principal trading activity represents a significant portion of market activity and there are aspects of the current market regime that may result in the underreporting of this trading activity. Indeed, an analysis by Commission Staff estimates that principal trading accounted for 40.5% of all reported transactions and principal activity accounted for 67% of all exchange message traffic. And, because these figures do not capture principal activity done by trading on-exchange through other broker-dealers, these estimates are likely to be biased downwards.

The analysis used audit trail data (where orders are identified at the broker-dealer level), from 10 exchanges, excluding CHX, and OATS reported off-exchange activity. Message traffic was defined as order placement, cancellation, or amendment.

The fact that off-exchange principal trading of non-FINRA member broker-dealers is not fully reported in OATS, may also bias these estimates downwards.

Finally, no single current data source integrates both equities and options. The lack of any combined equity and options audit trail data is a significant impediment to regulators performing cross-product surveillance.

Likewise no single audit trail combines futures with NMS Securities either. See Adopting Release, supra note 9, at 45744 for a discussion of the potential inclusion of futures in CAT Data.

B. Data Fields

Each of the available data sources discussed above is missing certain data fields that are useful for conducting a variety of regulatory activities. Furthermore, certain valuable data fields are not contained in any of the data sources discussed above. For example, the lack of completeness in the data sources makes it impossible to use certain key information, such as customer identifiers and allocation information, in market surveillance. Further, even for single-security events within a single trading venue, regulators may need to seek data from multiple sources such as an SRO audit trail and EBS.

See Section IV.D.2.a, supra.

See Section IV.D.2.a, supra, and Section IV.D.2.b(3) infra, for a discussion of how regulators access such data.

Most notably, the identity of the customer is unavailable from all current data sources that are reported to regulators on a routine basis. A unique customer identifier could be useful for many types of investigations and examinations such as market manipulation investigations and examinations of investment advisers. As noted above, some data sources—specifically Large Trader, EBS, trade blotters, and order tickets—identify customers. But these data sources are not reported on a routine basis, provide only one part of the order lifecycle, and have other inherent limitations.

Trade confirmation data also identifies customers, but trade confirmation data are much more basic than a trade blotter. See supra note 375.

Because there is currently no data source that includes customer identities across multiple parts an order lifecycle, regulators must engage in a process of linking EBS, trade blotters and order tickets with SRO audit trails, which can be a burdensome and imperfect process. For example, trade blotter and order ticket data that identifies customers from one broker-dealer may only include customer names and thus may not be readily matched to similar data from another broker-dealer, or may require substantial effort and uncertainty to reconcile across firms. Further, EBS data's limited coverage of trading activity and lack of some detailed trade information creates inefficiencies in insider trading investigations. These investigations often begin with a request for EBS data of trades before a significant corporate news event that affected a company's stock price. After identifying accounts that made suspicious trades, investigators often request additional EBS data of all trades by the accounts during the same period. If the additional data reveal suspicious trades by the accounts of the securities of other companies, investigators often must make a third round of EBS requests for data of trades by all accounts in those securities. If trading is done in an omnibus account, Commission Staff must ask firms to provide the identity of the account holder, and then request account information. To investigate for manipulation (e.g., marking the close, order layering, spoofing, wash sales, trading ahead), Commission Staff may also link data from multiple sources. First, Commission Staff obtains equity and option cleared reports from an internal online system that interfaces with data provided by the DTCC. Because the equity and option cleared reports do not have trade details, Commission Staff may also request trade information through EBS submissions from one or multiple firms. If a trade was executed on behalf of another firm, Commission Staff may then contact the other firm, until Staff can find out who placed the trade and the account holder. The Commission may then obtain granular trade information that contains order entry time and order execution time from firms or brokers via request or subpoena.

The Commission approved a FINRA rule that would require broker-dealers to report to OATS the identity of U.S. registered broker-dealers that are not FINRA members and broker-dealers that are not registered in the U.S. but have received an SRO-assigned identifier in order to access certain FINRA trade reporting facilities, from whom they receive or route an order. See Securities Exchange Act Release No. 77523 (April 5, 2016), 81 FR 21427 (April 11, 2016) (Order Approving FINRA Rule to Report Identity of Certain Broker-Dealers to OATS). CAT would similarly capture this information upon full implementation.

For further discussion of the problems associated with linking, see Section IV.D.2.b(2)C, infra.

See supra note 343.

The process to obtain detailed trade information from firms and brokers via requests or subpoenas generally takes anywhere from two to four weeks depending on the size of the request.

The methods for obtaining such information significantly reduce its utility, particularly for surveillance and market reconstruction purposes. Market reconstructions, for example, cannot take advantage of the detail in the EBS and trade blotter data because of the resources required to link so many data sources, lack of necessary elements (such as time stamps in milliseconds) needed to link data sources (for example, matching large trader reports to activity on a particular exchange), or the absence of standardized format. To examine a tip or complaint, regulators may consolidate data from each affected market participant to determine the identities of those responsible for the atypical activity in question. To the extent that the activity originates from several market participants, regulators must request data from each of those market participants, and possibly other market participants, to obtain information that could identify the customer(s) originating the orders that created the atypical activity.

For many regulatory activities, lack of completeness results in regulators initially relying upon the most accessible data sources, with significant information contained only in data sources made available by request. Starting regulatory functions with incomplete data sources requires regulators to later make data requests and link such data request responses. More importantly, however, incomplete or unconsolidated data interferes with effective surveillance. Access to data through non-routine means makes investigations and examinations less efficient, and makes automated surveillance less accurate and less effective. For example, the publicly available data discussed above identify exchanges but lack most of the fields found in some SRO audit trails or EBS, such as customer information, order entry time, order execution time, information about special order handling codes, counterparties, and member identifiers. Similarly, equity cleared reports contain only the date, the clearing firm, and the volume cleared by each clearing firm and not the trade size, trade time, or trade location. Option cleared reports contain only the date, the clearing firm, number of customer contracts, and number of firm contracts for the options.

See Section IV.D.2.a(6), supra.

Some valuable data fields, such as modifications that make an order non-displayed and other special handling instructions are consistently available on only a few data sources or require linking different data sources. The lack of direct, consistent access to order display information and special handling instructions creates inefficiencies in surveillances, examinations, and investigations that examine hidden liquidity and the treatment of customer orders. Data that are not directly accessible by regulators at all include buy-to-cover information and subaccount allocation information, including the allocation time. For example, no current data source allows regulators to directly identify when someone is buying to cover a short sale. Regulators could use this information to better understand short selling and for investigations of short sale manipulation. Indeed, the absence of this information during the financial crisis in 2008 reduced the efficiency of the reconstruction of investor positions in financial companies.

Order display information (i.e., whether the size of the order is displayed or non-displayed) is indicated in the “Customer Instruction Flag” and special handling instructions are indicated in the “Special Handling Code” of an OATS report. The Customer Instruction Flag is mandatory if a limit or stop price is provided. A Special Handling Code is required for order modifications, reserve size orders, when the order is routed electronically to another member, or when the terms and conditions of the order were derived from a related options transaction. See FINRA, OATS Reporting Technical Specifications, at Appendix A (June 26, 2015), available at http://www.finra.org/sites/default/files/TechSpec_20150825.pdf . This data is not directly available to all regulators. The Commission must request this data from FINRA.

Having access to buy-to-cover information was also one of the subjects of a Dodd-Frank-mandated study on short sale reporting. See SEC, Short Sale Position and Transaction Reporting (June 5, 2014) (“Short Sale Reporting Study”), available at http://www.sec.gov/dera/reportspubs/special-studies/short-sale-position-and-transaction-reporting.pdf .

Subaccount allocation information needed for regulatory activities can be difficult for regulators to collect and compile. SRO audit trails currently do not require allocation reports and broker-dealers may not have records of the time of a subaccount allocation. When regulators require an understanding of subaccount allocations for a regulatory task, they generally request and sift through trade blotter or EBS data in an attempt to identify allocations and the details of those allocations. Current trade blotter data contains limited customer information on allocations and is not required to contain allocation time information at the subaccount level. While the Commission is sometimes able to acquire allocation time on trade blotters, not all broker-dealers keep records in a manner that facilitates efficient regulatory requests for allocation time information.

The difficulty in obtaining allocation information and the difficulty in reconstructing allocations with data from broker-dealers limits the efficiency of certain surveillances and examinations. Allocation time at the subaccount level is critical for determining whether some customers are systematically given more favorable allocation treatment than others. For example, when a broker-dealer places an order or series of orders for multiple customer accounts that generates multiple executions at multiple prices, it is possible that different customers receive different prices in the allocation process. However, if some customers systematically receive less favorable prices than others when they should be receiving the same prices for their executions, this could indicate that the broker-dealer is handling allocations improperly.

If a group of orders are bundled together for execution, when those same orders are allocated, they should receive the same (usually average price) allocations. However, if executions are for orders that are not bundled together, it might be appropriate that customers for those separate orders would receive differently-priced allocations.

(2) Accuracy

In the Adopting Release, the Commission noted that while “to some extent, errors in reporting audit trail data to the central repository will occur,” the CAT NMS Plan would improve the quality of data including improvements to accuracy. Therefore, the economic analysis carefully considers the Baseline of the accuracy of data regulators currently use in order to consider whether and to what degree the CAT NMS Plan would provide more accurate data.

See Adopting Release, supra note 9, at 45730.

The prospect of inaccurate data can result in regulators expending extra resources to run additional quality checks to ensure reliable data and conclusions in enforcement investigations, or being unable to draw reliable conclusions at all. In addition, risk-based analysis may not properly identify a potential risk that justifies further examination if the underlying data suffers from inaccuracies. Ultimately, inaccurate data results in less efficient investigations as well as less effective surveillance and risk analyses. This economic analysis considers several forms of data inaccuracy, including data errors, inaccurate event sequencing, the inability to link data accurately, inconsistent identifiers, and obfuscating levels of irreversible data aggregation.

A. Data Errors 416

Based on Staff experience, the Commission preliminarily believes that data errors affect most current data and can persist even after corrections. For example, Commission Staff has investigated instances where information was inaccurately reported by broker-dealers, most notably in EBS data given to the Commission. In addition, the Commission believes that data sources that depend on data translated from back-office systems can be less accurate than those that come from trading systems, such as trade blotters and data sourced from exchanges' electronic trading systems, because the data translation process creates an additional source of potential errors in code that may not work as intended. Data from trading systems can also contain errors resulting from a coding error in the query pulling the data. Such coding errors can affect any data including trade blotters. For example, trade blotters are stored using the ticker symbol in effect at the time of the trade. If the ticker symbol changes between the trade and the data request, the coding may fail to take the ticker symbol change into account and fail to retrieve the correct data. The Commission has found that trade blotter data can often be inaccurate due to improper inclusion of cancelled orders or corrections, making accurate reconciliation difficult. Furthermore, trade blotter data can lack security information including CUSIP, symbol, or description at the subaccount level, which are important features for helping regulators determine potential violations.

For example, Commission staff have experienced frequent errors in EBS data such as omitted variables, decimals in the wrong places, blank account information, and data for the wrong securities. The Commission has instituted actions against entities in connection with inaccurate EBS data. See, e.g., Securities Exchange Act Release No. 75445 (July 14, 2015), In the Matter of OZ Management, LP, Administrative Proceeding File No. 3-16686 (OZ Management, LP admitted submitting inaccurate data to four of its prime brokers); see also Section IV.D.2.b(4), infra, for a discussion of one impact of inaccurate data.

In cases where Commission staff has used these data, it has found that the frequent omission of these important fields in trade blotter data is generally due to the manner in which the data is queried by broker-dealers. There are a variety of reasons why these fields may be excluded from a query. For example, over time firms make changes to their software systems; records stored by previous versions, particularly when the records are archived in a secondary location, may not be fully compatible with software that is written to access more current versions of this data. Additionally, sometimes when a broker-dealer or clearing firm merges or is acquired, its trade data may be compromised due to incompatible systems or inadequate data storage issues. This problem was particularly relevant following the financial crisis. Consequently, staff does not currently believe that this missing information is caused by a failure of broker-dealers to collect and retain these variables, but rather that over time this data becomes less accessible by software tools and may require hand processing by broker-dealers providing this information.

Audit trail data contain errors, as well. The CAT NMS Plan reports that 2.42% of order events submitted to OATS fail validation checks, resulting in the rejection of almost 425,000 reports per day, on average. While FINRA sends these records back to its members to correct, not all data errors are identified because OATS limits error correction requests to records with internal inconsistencies within a given member's submission. In particular, significant error rates in event linking are common because there is no cross-participant error resolution process; FINRA estimates that 0.5% of OATS routing reports directed to another FINRA member broker-dealer cannot currently be linked. The CAT NMS Plan reports that, following the rollouts of three major updates to OATS, 0.86% of Trade Reporting Facility (“TRF”) reported trades could not be matched to OATS execution reports, 3.12% of OATS route reports could not be matched to exchange orders, and 2.44% of inter-firm routes could not be matched to a record of the receiving firm's receipt of a routed order.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(b). When FINRA receives an end-of-day OATS file from a member, it performs over 152 validation checks on each order event reported to OATS. Each of these checks can result in rejecting an OATS data submission and generating an error message. In addition to validation checks, FINRA determines whether a file that is syntactically correct nevertheless contains errors in content related to internally inconsistent information about processing, linking, and routing orders. For some errors, FINRA requires the member to provide corrections within five business days after rejections are available. See OATS Reporting Technical Specifications, supra note 357, at 6-1—6-10. Duplicate records and records with symbols that are not reportable to OATS may result in rejections that do not require repair. Id. at 6-4. Validation checks refer to tests of whether data is consistent with a set of rules that specify conditions that should be met by valid data. Validation checks are typically limited to detecting errors that can be discovered by a concise set of logical rules using data within scope at the time the validation test is run. An incorrect price that is negative would likely be detected by a validation check, while a price that was a few cents too low may not. Validation checks that apply across multiple records may be difficult to apply across data that is submitted at different times.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(b); see also Adopting Release, supra note 9 at 45729.

See Section IV.D.2.b(2)C, infra.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(b).

Other audit trail data may also contain errors. For example, the Commission notes that exchange SROs populate most of the information with data from their in-house order and trading records, but a few of these exchange SROs also rely on members to complete their audit trails.

B. Event Sequencing

The ability to sequence market events is crucial to the efficacy of detecting and investigating some types of manipulation, particularly those involving high frequency trading, those in liquid stocks in which many order events can occur within microseconds, and those involving orders spread across various markets. In today's market, high frequency and algorithmic traders can react to changes in the market in a few milliseconds or less. Investigations involving algorithmic trading, therefore, can require the ability to sequence the order and trade events to within a few milliseconds; however, regulators relying on currently available data may have difficulty sequencing events that occur within a second on different trading venues or broker-dealer systems. In addition, in one type of trade-based manipulation, a manipulator might build a short position in a stock, submit sell orders designed to decrease the stock price, and finally buy at an artificially low price. To analyze this activity, except when cover orders precede the sell activity, it would be necessary to determine whether the orders intending to create an artificial price came before the orders intending to profit from the artificial price, which becomes difficult when the systems on which order events occurring close in time to each other have clocks that are not synchronized. Further, insufficiently granular time stamps can make sequencing events across venues impossible.

See, e.g., Joel Hasbrouck and Gideon Saar, Low-Latency Trading, 16 Journal of Financial Markets 646 (2013) in which the authors report apparent HFT response times to market events of 2-3 milliseconds. Given technology advances, it is likely that response times have decreased since their sample period, which ends in June 2008.

Regulators can sequence events occurring on the same venue or on the same systems at broker-dealers, but sequencing across venues or broker-dealer systems that could have clocks that are not synchronized with each other is more difficult.

Thus, the sequencing of order events requires both sufficient clock synchronization across market participants and time stamps that are granular enough for accurate sequencing. As discussed below, current clock synchronization standards make this process difficult.

For example, if two market participants report that two non-simultaneous events happened at 10:15:45, then the time stamps are not granular enough to sequence the events and regulators would need sub-second time stamp granularity to distinguish them. If the two market participants each have up to one-second clock drift from the actual time, the 10:15:45 time stamps only show that the event happened between 10:15:44 and 10:15:46. Only when regulators have both adequate time stamp granularity and sufficient clock offset tolerances can events be sequenced using time stamps.

i. Clock Synchronization

Clock synchronization refers to the synchronization of the business clocks used by market participants for the purposes of recording the date and time of market events to a centralized benchmark clock, often that maintained by the NIST. Clock synchronization helps to ensure that the time stamps used by various participants are consistent, thereby allowing regulators to compare time stamps across participants and to use multiple time stamps to determine the sequence of market events. The ability of regulators to accurately sequence events can be limited by the permitted “offset” between the clocks—i.e., the length of the gap that is permitted between a participant's clock and the time maintained by a centralized benchmark clock. For example, if the offset between the clocks is one second, regulators cannot accurately determine the correct sequence of events in the market occurring within a two-second period, because each clock may be up to one second fast or slow.

For example, if a participant's clock records a point in time as 11:00:00 and the NIST clock records the same point in time as 11:00:01, then the offset between the clocks is one second.

Current rules require most broker-dealers to synchronize their system clocks to within one second. In particular, FINRA specifies a clock offset tolerance of one second, and the NASDAQ Stock Market and NASDAQ OMX BX require members to comply with FINRA clock synchronization rules. CHX specifies a clock offset tolerance of 500 milliseconds. NYSE MKT and NASDAQ OMX PSX require members to synchronize their clocks relative to a time source designated by the Exchange, but do not specify the standard. NYSE Arca allows options traders to use any time provider source for clock synchronization as long as the business clocks it uses on the Exchange are accurate to within three seconds of the NIST clock or the United States Naval Observatory Master Clock in Washington DC.

See FINRA Rule 7430 (requiring each member to “synchronize its business clocks that are used for purposes of recording the date and time of any event that must be recorded pursuant to the FINRA By-Laws or other FINRA rules, with reference to a time source as designated by FINRA, and shall maintain the synchronization of such business clocks in conformity with such procedures as are prescribed by FINRA.”). Section 2 of the OATS Technical Specifications states that all computer system clocks and mechanical time stamping devices must be synchronized to within one second of the NIST clock and must be synchronized every day. See OATS Reporting Technical Specifications, supra note 357, at 2-1. In November 2014, FINRA issued a Regulatory Notice seeking comment on a proposal to change the clock offset tolerance to be 50 milliseconds. This proposal also proposed to move the clock offset tolerance from the OATS Technical Specifications to FINRA's books and records rules so that the requirements apply to the recording of the date and time of any event that FINRA By-Laws or Rules require, not just OATS requirements. See FINRA, Equity Trading Initiatives: Synchronization of Business Clocks, Regulatory Notice 14-47, available at http://www.finra.org/sites/default/files/notice_doc_file_ref/Notice_Regulatory_14-47.pdf . On February 9, 2016, FINRA filed a proposed rule change with the Commission. The proposal would reduce the clock offset tolerance for members' computer clocks that are used to record events in NMS securities, including standardized options, and OTC Equity Securities, from within one second of the NIST atomic clock to within a 50-millisecond tolerance of the NIST atomic clock. FINRA would require firms with systems that capture time in milliseconds to comply with the new 50-millisecond clock offset tolerance within six months of the effective date; remaining firms that do not have systems which capture time in milliseconds would have 18 months from the effective date to comply with the 50-millisecond standard. The proposal would not change the current one-second clock offset tolerance of the NIST clock requirement for mechanical clocks or time stamping devices. The proposal would consolidate and codify the clock synchronization requirements in new FINRA Rule 4590. The Commission has published notice of this proposed rule change. See Securities Exchange Act Release No. 77196 (February 19, 2016), 81 FR 9550 (February 25, 2016).

See NASDAQ Rule 7430A (“(a) Nasdaq members shall comply with FINRA Rule 7430 as if such Rule were part of Nasdaq's rules. (b) For purposes of this Rule, references to `the FINRA By-Laws or other FINRA rules' shall be construed as references to `the Nasdaq Rules'); NASDAQ OMX BX Rule 6953 (“(a) Exchange members shall comply with NASD Rule 6953 [superceded by FINRA Rule 7430] as if such Rule were part of the Exchange's rules. FINRA is in the process of consolidating certain NASD rules into a new FINRA rulebook. If the provisions of NASD Rule 6953 are transferred into the FINRA rulebook, then Equity Rule 6953 shall be construed to require Exchange members to comply with the FINRA rule corresponding to NASD Rule 6953 (regardless of whether such rule is renumbered or amended) as if such rule were part of the Rules of the Exchange. (b) For purposes of this Rule, references to `the By-Laws or other rules of the Association' shall be construed as references to `the Rules of the Exchange.' ”).

See CHX Rule 3, Interpretations and Policies .03 (“These rules shall not apply to orders sent or received through the Exchange's matching system or through any other electronic systems that the Exchange expressly recognizes as providing the required information in a format acceptable to the Exchange. The Exchange will not recognize a non-Exchange system as providing information in an acceptable format unless that system has synchronized its business clocks for recording data with reference to a time source designated by the Exchange and maintains that synchronization in conformity with procedures prescribed by the Exchange.”); Rule 4, Interpretations and Policies .02 (“Each Participant or layoff service provider shall synchronize its business clocks that are used for purposes of recording the date and time of any event that must be recorded pursuant to this provision with reference to a time source as designated by the Exchange, and shall maintain the synchronization of such business clocks in conformity with such procedures as are prescribed by the Exchange.”); Rule 5, Interpretations and Policies .01(a) (“Clock synchronization and timing of the determination of improper trade-throughs. The Exchange's systems shall routinely, throughout the trading day, use processes that capture the time reflected on the atomic clock operated by the National Institute of Standards and Technology and shall automatically make adjustments to the time recorded in the Exchange's Matching System to ensure that the period between the two times will not exceed 500 milliseconds. The Exchange shall determine whether a trade would create an improper trade-through based on the most recent NBBO that has been received and processed by the Exchange's systems.”).

See NYSE Rule 123, Supplementary Material .23 (“Any vendor or proprietary system used by a member or member organization on the Floor to record the details of an order or report for purposes of this rule must be synchronized with reference to a time source as designated by the Exchange.”); NYSE MKT Rule 7430 (“Each member organization shall synchronize its business clocks that are used for purposes of recording the date and time of any event that must be recorded pursuant to the Rules of the Exchange, with reference to a time source as designated by the Exchange, and shall maintain the synchronization of such business clocks in conformity with such procedures as are prescribed by the Exchange.”); NASDAQ OMX PSX Rule 3403 (“Each member organization shall synchronize its business clocks that are used for purposes of recording the date and time of any event that must be recorded pursuant to the rules of the Exchange, with reference to a time source as designated by the Exchange, and shall maintain the synchronization of such business clocks in conformity with such procedures as are prescribed by the Exchange.”).

See NYSE Arca Options Rule 6.20 (“(a) Each OTP Holder and OTP Firm must synchronize, within a time frame established by the Exchange, the business clocks that it uses for the purpose of recording the date and time of any event that must be recorded pursuant to the Rules of the Exchange. OTP Holders and OTP Firms may use any time provider source. Each OTP Holder and OTP Firm must, however, ensure that the business clocks it uses on the Exchange are accurate to within a three-second [sic] of the National Institute of Standards and Technology Atomic Clock in Boulder Colorado (`NIST Clock') or the United States Naval Observatory Master Clock in Washington DC (`USNO Master Clock'). This tolerance includes all of the following: (1) The difference between the NIST/USNO standard and a time provider's clock; (2) transmission delay from the source; and (3) the amount of drift of the OTP Holder or OTP Firm's business clock. For purposes of this Rule, `business clocks' mean an OTP Holder or OTP Firm's proprietary system clocks. OTP Holders and OTP Firms must set forth in their written supervisory procedures, required by Rule 11.18, the manner in which synchronization of business clocks will be conducted, documented and maintained.”).

In practice, some broker-dealers currently synchronize their clocks to smaller clock offset tolerances. FIF surveyed market participants to gather information on current broker-dealer clock synchronization practices. The survey found that 29% of respondents currently synchronize their clocks to permit a maximum clock offset of one second from NIST time. The survey further found that 10% of market participants permit a maximum offset from NIST time that is between 50 milliseconds and one second, 21% of respondents permit a 50-millisecond maximum offset, and 18% of respondents permit a maximum offset that is less than 50 milliseconds. The remaining 22% of survey respondents utilize multiple clock offset tolerances across their systems, ranging from five microseconds to one second. FIF noted that 69% of firms that achieve a maximum clock offset of 50 milliseconds or less are large firms reporting more than three million OATS records per month.

See FIF Clock Offset Survey, supra note 127. The Commission notes limitations to the survey that could result in downward bias and imprecision. Specifically, the broker-dealers represented by the survey are primarily complex and large broker-dealers in terms of market activity levels; consequently, smaller broker-dealers are underrepresented. But, as discussed below, the exclusion of small broker-dealers is unlikely to materially affect industry costs because smaller broker-dealers are unlikely to incur significant clock-synchronization costs because the majority of broker-dealers rely on service bureau clocks to time stamp their CAT Reportable Events.

Id.

Certain exchanges, the SIPs, and FINRA synchronize their clocks for their trading, recordkeeping, and other systems. According to FIF, all exchange matching engines meet a clock offset tolerance of 50 milliseconds. However, NASDAQ recently stated that all exchanges trading NASDAQ securities synchronize their matching engines and quotation systems to within 100 microseconds. The Commission understands that the NYSE, the options exchanges, and the SIAC SIP have comparable clock synchronization standards. In conversations with Commission Staff, the Participants stated that absolute clock offset on exchanges averages 36 microseconds.

Id.

See NASDAQ, UTP Vendor Alert #2015-7 (April 24, 2015), available at https://www.nasdaqtrader.com/TraderNews.aspx?id=UTP2015-07 (describing additional time stamps to be reported to the SIP, including information on exchange clock synchronization, and stating that “[e]xchanges use a clock sync methodology ensuring that timestamps are accurate within tolerances of 100 microseconds or less.”).

In response to questions from Commission Staff, the Participants surveyed the exchanges to establish their current average clock offset. All exchanges that currently operate matching engines responded to the survey, which measured the offset from the exchange clock to NIST. The Participants noted that the frequency with which exchanges measure their clock offset ranges from once per second to once per fifteen minutes, and the procedures to correct for clock offset vary. Some exchanges correct by slewing, in which the offset is gradually corrected, while others use stepping, in which the offset is immediately corrected. The process by which clock offset is corrected can impact the ability to order events time stamped by a single clock because stepping could result in a backwards adjustment in recorded time.

Because multiple order events can occur within timeframes of less than one second, current clock synchronization requirements and practices greatly limit the ability of regulators to accurately sequence order events. To examine, among other things, how many events can be synchronized with current clock offset tolerances, Commission Staff conducted an analysis of the frequency of events using MIDAS data. In the analysis, events are all real-time messages, consisting of trades, orders, modifications, cancellations and updates from exchange direct feeds and trades from the FINRA TRFs. The analysis focused on identifying whether, for each order event, an event at another venue occurred within a given time range. For the purposes of the analysis, events at another venue were called an “unrelated event.” The Commission recognized that order events occurring on the same venue have sequence numbers that allow sequencing even if orders have the same time stamp. Therefore, the analysis considered only whether any unrelated orders existed within a given time range that could complicate the sequencing across the market.

The MIDAS system does not contain all of the events in a given security that would be in CAT. Therefore, the analysis is limited, but still provides useful insights.

The methodology to calculate these percentages starts with sorting all event messages for every day chronologically by exchange time stamp. (MIDAS does not report the exchange time stamp; but it provides the difference between the MIDAS time stamp and the exchange or TRF time stamp; the analysis uses this value to derive the exchange time stamp.) For each event, it calculates the difference (Delta) between the current time stamp (t0) and the last time stamp (t-1) in the same security on a different venue.

Deltanearest last = t0,venue A − maximum(t-1,venue B, t-1,venue C, t-1,venue D, t-1,venue E)

This is the shortest time difference (Deltanearest last) between an event on venue A and a preceding event on any venue, except for venue A. Next, the analysis calculates the time difference (Deltanearest next) between the current time stamp (t0) and the next time stamp (t1) in the same security on a different venue.

Deltanearest next = minimum(t1,venue B, t1,venue C, t1,venue D, t1,venue E) − t0,venue A

Finally, the analysis uses the shorter of the time differences to evaluate whether an event occurs within a particular time period of another event in the same security on a different venue.

Deltanearest = minimum(Deltanearest last, Deltanearest next)

Values are aggregated over one week (June 15, 2015 through June 19, 2015) for the equities analysis; and the options analysis data is from one day (June 15, 2015).

Within the analysis, events reported to the TRF are treated as occurring on a different trading venue than other recent events because TRF data comprises many separate venues (such as ATSs and off-exchange market makers). While events within a single exchange with identical time stamps can potentially be sequenced through record identifiers recorded by the exchange, for TRF trades this is often untrue because many venues with independent clocks contribute to the aggregate TRF data.

Table 3—Percentage of Events Close to Unrelated Events

Nearest event time stamped within Percent of unrelated events
Equities Options
2 seconds 98.69 93.03
1 second 97.95 90.99
100 milliseconds 92.16 81.17
50 milliseconds 89.12 76.59
10 milliseconds 83.49 64.46
5 milliseconds 81.28 58.26
2 milliseconds 77.92 49.30
1 millisecond 74.31 41.13
200 microseconds 57.53 21.58
100 microseconds 48.09 14.51
10 microseconds 21.42 3.13
5 microseconds 14.44 3.12

Table 3 shows that 97.95% of the order events for listed equities and 91% of order events for listed options in the samples occurred within one second of another unrelated order event in the same security. At the other extreme in Table 3, 14.44% of the unrelated order events for listed equities and 3.12% of the unrelated order events for listed options in the same security occurred within 5 microseconds of another order event in the same security. The Commission notes that Table 3 underestimates the true frequency of unrelated events within the given time frames because it includes only order events that are included in the MIDAS data. As such, the analysis is unable to include events such as the placing of hidden orders on exchanges, the placing of orders on an ATS, order originations, order routes, order receipts, and order cancellations and modifications for any order not displayed on an exchange order book. Despite this limitation, Table 3 illustrates how the current frequency of order events makes sequencing unrelated order events difficult.

ii. Time Stamps

Given the frequency with which order events can occur, regulators need sufficiently granular time stamps to sequence events across orders and within order lifecycles. As noted above, even if the clocks recording time stamps have no clock offset, the granularity of the time stamp can limit regulators' ability to sequence events accurately.

In addition, Craig W. Holden and Stacey Jacobsen, Liquidity Measurement Problems in Fast, Competitive Markets: Expensive and Cheap Solutions, 69 Journal of Finance 1747 (2014), shows that using time stamps in seconds instead of milliseconds can yield liquidity measurement problems.

Current data sources have different time stamp granularity standards. Many public data sources report time in seconds or milliseconds and some, including direct data feeds, report time in microseconds or nanoseconds. For example, the Options Price Reporting Authority (“OPRA”) allows for time stamps in nanoseconds and the other SIPs require time stamps in microseconds for equity trades and quotes, whereas the short sale transactional data released by exchanges contains time stamps in seconds. Currently, OATS requires time stamps in milliseconds for firms that capture time in milliseconds, but does not require members to capture time in milliseconds. EBS trade times are recorded only to the second; other EBS records must contain time stamps containing only the transaction date. The lack of uniform and granular time stamps can limit the ability of regulators to sequence events accurately and to link data with information from other data sources.

See OPRA Option Price Reporting Authority Binary Participant Interface Specification Version 1.7 (January 2015), available at http://www.opradata.com/specs/opra_binary_part_spec.pdf ; see also NYSE, Modified Timestamps and Additional Timestamp Information for Daily TAQ (June 22, 2015), available at http://www.nyxdata.com/nysedata/default.aspx?tabid=993&id=2784 ; UTP Vendor Alert #2015-7, supra note 435, regarding additional time stamps to be reported to the SIP.

See FINRA Rule 7440 (providing that “[e]ach required record of the time of an event shall be expressed in terms of hours, minutes, and seconds; provided that the time of an event shall be expressed in hours, minutes, seconds, and milliseconds if the member's system captures time in milliseconds.”). The Commission approved the requirement that time be expressed in milliseconds if the member's system captures time in milliseconds on February 27, 2014. See Securities Exchange Act Release No. 71623 (February 27, 2014), 79 FR 12558 (March 5, 2014); see also, FINRA, Equity Trade Reporting and OATS, Regulatory Notice 14-21 (May 2014), available at http://www.finra.org/sites/default/files/NoticeDocument/p506337.pdf.

C. Linking and Combining Data

Sometimes one order or market activity event may be reflected in information contained in various data sources or in different fields within the same data source, and fully understanding that activity requires linking information across the different data sources. Therefore, regulators analyzing an event or running a surveillance pattern often need to link data. For example, cross-market examinations require the cumbersome and time-consuming task of linking many different data sources. Regulators combine trading data from sources such as public feeds, SRO audit trails, EBS data, and trade blotters when reviewing surveillance alerts to determine whether violations of rules such as Rule 611 of Regulation NMS occurred or to examine, for example, whether an entity availing itself of a market maker exemption is engaging in bona fide market making. In fact, the data needed for an examination often consist of many audit trails and are stored in non-uniform formats. In addition, the analysis and reconstruction of market events could require linking many different data sources, such as a dozen SRO audit trails.

Such linking is typically conducted electronically with an algorithm unless the size of the data set is small. This requires the person attempting to combine and link the data to write computer code to identify and match the records that need to be linked. This task involves extensive testing and debugging the first time that person tries to combine and link those specific data sources. Further, given the variation in formats across broker-dealers and other data sources, the code may need to change for each investigation, requiring a repeat of the extensive testing and debugging process.

17 CFR 242.611.

In the context of the CAT NMS Plan, the Commission does not distinguish data format from data taxonomy. See Section III.B.3, supra. In discussing data format, the Commission combines data format with data taxonomy. Id. The distinction between format and taxonomy is not significant in the context of the CAT NMS Plan because the Plan does not specify either for incoming data and the Plan effectively requires uniformity in both for regulator access. Id. SRO audit trails currently differ in both format and taxonomy as do many other trading and order data sources.

Regardless of whether order lifecycle reports are reflected in the same or different data sources, the process of linking lifecycle events is complex and can create inaccuracies. Merging different data sources often involves translating the data sources into the same format, which can be a complex process that is prone to error. Linking records within or across data sources also requires the sources to share “key fields” that facilitate linkage, along with a successful linking algorithm. Regulators may be unable to link some data source combinations accurately because the data sources do not have key fields in common or the key fields are not sufficiently granular. For example, regulators cannot always link trade records accurately to EBS records. The EBS records contain a symbol and date, but the price and size on the records may reflect multiple trades spread over a period of time. Sometimes, different data sources may have key fields in common but the relationship between the fields is not straightforward. In these cases, the algorithm to link them may be necessarily complex and not entirely successful. Further, within a single order lifecycle, the order number may change when a broker-dealer routes the order to another broker-dealer or exchange or even to another desk at the same broker-dealer. The inability to link all records affects the accuracy of the resulting data and can force an inefficient manual linkage process that would delay the completion of the data collection and analysis portion of the examination, investigation, or reconstruction.

For example, different data sources can format dates and times differently or may use different notations to signify that the field contains no value.

D. Customer and Broker-Dealer Identifiers

The data sources described in Section IV.D.2.a also lack consistent customer and broker-dealer identifiers, which limit regulators' ability to track the activity of one client or broker-dealer across the market. There is no standard convention for how broker-dealers identify customers.

Regulators face challenges in tracking broker-dealers' activities across markets due to inconsistent identifiers and a lack of a centralized database. These challenges occur primarily in the context of regulatory activities that require manual or ad hoc data analysis, as is often the case in particular investigations, examinations, and market studies. In the case of broker-dealers, SROs generally identify their members within their data using market participant identifiers (“MPIDs”). However, the MPIDs that identify broker-dealers on Execution Venues are not standardized across venues; consequently, a broker-dealer identified as “ABCD” on one venue may be identified differently on another venue, where “ABCD” may refer to a different broker-dealer entirely. Therefore, aggregating a broker-dealer's activity across venues requires verifying the MPIDs assigned to a broker-dealer on each venue, usually referencing the broker-dealer by its Central Registration Depository (“CRD”) number. In the course of manual data analysis, the Commission notes that its Staff have experienced challenges in identifying broker-dealers using CRD numbers. These challenges can be due to the fact that, although every broker-dealer has a CRD number, a broker-dealer that routes an order seldom, if ever, provides a CRD number to the broker-dealer that accepts the order.

The CRD is an automated database operated by FINRA that stores and maintains information on broker-dealers and their registered persons relating to their licensing, registration, complaints, professional background, and disciplinary history. Each broker-dealer and their registered persons are assigned a CRD number for identification.

The Commission and the SROs have generally overcome these challenges in the context of automated regulatory data analysis, and found ways to reduce these challenges in some manual data analysis and can efficiently track broker-dealers across venues. The Commission understands that FINRA can track broker-dealers across venues pursuant to its responsibilities under a plan for allocating regulatory responsibilities pursuant to Rule 17d-2. On September 12, 2008, the Commission declared effective a plan for allocating regulatory responsibilities pursuant to Rule 17d-2 filed by the American Stock Exchange, LLC, Boston Stock Exchange, Inc., CBOE Stock Exchange, LLC, CHX, FINRA, ISE, NASDAQ, NSX, NYSE, NYSE Arca, NYSE Regulation, Inc., and Philadelphia Stock Exchange, Inc. (the “Participating Organizations,” which have since been updated to be the following SROs: BATS, BYX, CBOE, CHX, EDGA, EDGX, FINRA, NASDAQ OMX BX, NASDAQ OMX PHLX, NASDAQ, NSX, NYSE, NYSE MKT [f/k/a NYSE Amex], and NYSE Arca) (“Insider Trading Rule 17d-2 Plan”). The Insider Trading Rule 17d-2 Plan allocates regulatory responsibility over common FINRA members (members of FINRA and at least one of the Participating Organizations) (collectively “Common FINRA Members”) for the surveillance, investigation, and enforcement of (i) Federal securities laws and rules promulgated by the Commission pertaining to insider trading, and (ii) the rules of the Participating Organizations that are related to insider trading (“common insider trading rules”). Under that Plan, the Participating Organizations, other than FINRA, have been relieved of regulatory responsibility over Common FINRA Members (i.e., the broker-dealer and its associated persons) for surveillance, investigation, and enforcement of the common insider trading rules over such persons with respect to “Listed Stocks” (as defined in that Plan). Accordingly, FINRA retains regulatory responsibility for Common FINRA Members with respect to the common insider trading rules—irrespective of the market(s) on which the relevant trading may occur. Separately, FINRA performs investigations and enforcement with respect to non-Common FINRA Members pursuant to a regulatory services agreement between FINRA and several of the other Participating Organizations. See Securities Exchange Act Release No. 58536 (September 12, 2008), 73 FR 54646 (September 22, 2008); see also Securities Exchange Act Release Nos. 58806 (October 17, 2008), 73 FR 63216 (October 23, 2008); 61919 (April 15, 2010), 75 FR 21051 (April 22, 2010); 63103 (October 14, 2010), 75 FR 64755 (October 20, 2010); 63750 (January 21, 2011), 76 FR 4948 (January 27, 2011); and 65991 (December 16, 2011), 76 FR 79714 (December 22, 2011).

Regulators sometimes find it necessary to analyze trading activity at the customer level instead of the broker-dealer level. Consistently identifying customer account owners across the multiple broker-dealers with whom they transact is difficult and prone to error. Although, for example, the EBS system provides the names associated with each account traded, these names are drawn from the separate records of each broker-dealer providing data to the EBS system, and the same party may be identified by a different name across multiple broker-dealers. Further, the lack of tax identification numbers in many EBS records limits the ability for regulators to trace the trading activity of customers across broker-dealers. Tax identification numbers are not required to be reported in EBS for average price, allocation, riskless principal, foreign accounts, and subaccounts. In fact, when one broker-dealer executes for a second broker-dealer, the tax identification number is that of the second broker-dealer regardless of whether the second broker-dealer is trading for a customer.

E. Aggregation

The practice used in some data records of bundling together data from different orders and trades also can make it difficult to distinguish the different orders and trades in a given bundle. As an example, brokers frequently utilize average-price accounts to execute and aggregate multiple trades for one or more customers. In these cases, for example with EBS data, the system does not reflect the details of each individual trade execution, because it reports only the average aggregate prices and volumes of the various trades within a series that have been bundled together for reporting purposes. Further, information on trade allocations aggregate the trade information to such an extent that it is difficult for regulators to identify when particular clients may be afforded preferential treatment because it is challenging to link subaccount allocations to orders and trades.

Equity and options cleared reports provide valuable data to regulators, but aggregation reduces their usefulness, because the reports do not have detailed trade information and do not include activity that does not require clearing. The volume in these reports cannot be fully disaggregated and reconciled with the equity trade execution volume from other data sources used by the Commission, e.g., TAQ and MIDAS, because the volume in the cleared reports is not necessarily a summation of all trades. For example, the same trade can be reported two or more times, by both the buy and the sell sides, for some OTC transactions and for all trades in NASDAQ exchanges. Similarly, option cleared reports bundle together multiple executions by compressing or netting them to facilitate clearing. This aggregation limits regulators' ability to link records across data sources, as well as limiting the accuracy with which the data source reflects market events, which is particularly problematic in applications that require market reconstruction.

The option cleared volume from the OCC contains the clearing firm, number of customer contracts, and number of firm contracts for the options.

This scenario of a trade being reported several times is generally the result of agreements that permit a broker-dealer to clear trades on behalf of another broker-dealer and send trades directly to the NSCC. Broker-dealers often enter into these agreements to simplify their clearing processes, achieve lower transaction costs, and take advantage of extended hours of service.

Finally, issuer repurchase information is aggregated at the monthly and quarterly level. This aggregation limits the use of such data in investigations of the timing of issuer repurchases and issuer stock price manipulation and in analysis of the use of the Rule 10b-18 issuer repurchase safe harbor.

Issuers report quarterly and monthly repurchases pursuant to Item 703 of Regulation S-K. This data includes all issuer repurchases, including tender offers and open market repurchases, but does not distinguish the type of repurchase. The Commission notes that Item 703 provides, in part, that issuers must disclose “the number of shares purchased other than through a publicly announced plan or program and the nature of the transaction (e.g., whether the purchases were made in open-market transactions, tender offers, in satisfaction of the company's obligations upon exercise of outstanding put options issued by the company, or other transactions.” See 17 CFR 229.703.

Rule 10b-18 provides issuers with a “safe harbor” from liability for manipulation under Section 9(a)(2) of the Act, 15 U.S.C. 78i(2), and Rule 10b-5 thereunder, 17 CFR 240.10b-5, solely by reason of the manner, timing, price, and volume of their repurchases when they repurchase common stock in the market in accordance with the Section's manner, timing price, and volume conditions. See 17 CFR 240.10b-18.

(3) Accessibility

The SROs and Commission also lack direct access—meaning the ability to log into a system in a manner that would allow them to gather and analyze the data they need—to many of the data sources described above. SROs generally have direct access only to their own audit trails and the public data feeds. While SROs control the manner in which they access their own data, their investigations in some cases require access to the data of other SROs because firms could trade across multiple SROs. To access another SRO's data, SROs must send requests to the other SROs or to the ISG. SROs needing information not included in their audit trails or the audit trail of another SRO must request such information from their members. The SROs might not be able to acquire data from entities that are not members of that SRO; non-members are not obligated to provide SROs with data, any data provided by the regulator of the non-member firm would be on a voluntary basis, or pursuant to the terms of the ISG Agreement.

FINRA does receive data from certain SROs on a daily basis and subsequently has direct access to that data.

Commission staff understands that SROs receiving information requests from other SROs will typically provide the information, although they are not required to do so.

See supra note 399.

See, e.g., NYSE Rule 2.A.xvi.—Jurisdiction (noting that the exchange has jurisdiction over matters related to non-member broker-dealers that choose to be regulated by the exchange). The Commission may, by rule or order, subject non-members to the rules of national securities exchanges if it deems it necessary or appropriate in the public interest and for the protection of investors, to maintain fair and orderly markets, or to assure equal regulation. Section 6(f)(2) of the Act, 15 U.S.C. 78f(f)(2); see also Sections 6(b)(1), 15A(b)(2) of the Act, 15 U.S.C. 78f(b)(1), 78o-3(b)(2) (requiring national securities exchanges and securities associations, respectively, to have the capability to enforce compliance by their members with applicable Exchange Act requirements and exchange or association rules).

The Commission has direct access only to the public data feeds and the equity and option cleared data; it lacks direct access to information provided in EBS or contained in trade blotters, order tickets, order handling data, SRO audit trails, and OATS data. Unlike the SROs, the Commission can subpoena data from entities that are not registered with the Commission, such as professional traders that are neither broker-dealers nor investment advisers.

If a regulator does not have direct access to data it needs, the regulator would request it. This can result in many data requests to broker-dealers, SROs, and others, which are burdensome to fill. The Commission recognizes that data requests could impose burdens on the entities responding to the request, in addition to the burden on the regulators to put the request together. Broker-dealers, investment advisers, and SROs responding to a data request must incur costs in order to produce, store, and transmit the data for the Commission or SRO. Further, as indicated above, regulators may need to request the data needed from many different data providers because of fragmentation in the data, and thus one analysis, such as an investigation, can generate many data requests.

In the context of an investigation or a court, in litigation, the Commission can request or subpoena information from entities, including those not registered with the Commission. See SEC Rule of Practice 232. Pursuant to their rules, SROs can request information from their registered entities; see also supra notes 454-456 and accompanying text (discussing how SROs request information from other parties, including other SROs).

See, e.g., CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(ii)(B) (discussing the current process for broker-dealers and SROs to respond to data requests, and stating that broker-dealers must commit staff to respond to requests for EBS or large trader data and may take varied approaches to fulfilling their regulatory reporting obligations).

Fragmentation in trade and order data can take many forms. First, an analysis may require the same type of data from many market participants. Second, the required data fields for an analysis may be reflected in different types of data. Finally, an analysis may require data on different products covered in separate data sources. The fragmentation in the data across market participants is a function of the fragmentation of trading and broker-dealer services. In today's equity markets, trades execute across 12 exchanges, more than 40 ATSs, and around 250 dealers. With its RSAs, FINRA can consolidate much of the SRO audit trails in equities. In the options markets, 14 different exchanges trade listed options with no off-exchange trading of standardized options and no entity aggregating each audit trail into one dataset. The vast majority of stocks trade in more than one location and most options trade on multiple exchanges.

See Securities Exchange Act Release No. 76474 at 81008, 81112, “Regulation of NMS Stocks” (November 2015), available at http://www.sec.gov/rules/proposed/2015/34-76474.pdf ; see also Laura Tuttle, OTC Trading: Description of Non-ATS OTC Trading in National Market System Stocks (March 2014), available at http://www.sec.gov/dera/staff-papers/white-papers/otc-trading-white-paper-03-2014.pdf .

FINRA has access to data from OATS and each equities exchange except CHX. See supra note 333 and accompanying text. This reduces the data fragmentation as it relates to the number of data requests for equities.

Exchange SROs generally limit their data collection to securities traded on their own exchanges, and limit the scope of their audit trails to transactions occurring on their exchanges. While ATSs and dealers report order events in equities to OATS, each of the 12 equities exchanges has its own audit trail. As a result of this structure, a market reconstruction for a single security may involve data requests to multiple exchanges. Likewise, a project involving options data may require data from each of the 14 options exchanges.

To acquire broker-dealer order records, EBS, trade blotters, and order tickets, regulators need to send a request to each broker-dealer to obtain its data. In the Commission's experience requested data can suffer from missing variables, truncations, and formatting problems due to the way that the data is queried by the broker-dealer. These problems can lead to substantial delays in using data and loss in regulatory productivity. Many different broker-dealers could have trading records in a given security on a given day of interest, so one narrow investigation could generate many data requests. As a result, in 2014 the Commission made 3,722 EBS requests that generated 194,696 letters to broker-dealers for EBS data. Likewise, the Commission understands that FINRA requests further generate about half this number of letters. In addition, for examinations of investment advisers and investment companies, the Commission makes approximately 1,200 data requests per year. Further, an investigation that requires tracing a single trade or a set of trades back to an investor or investors can generate many data requests. For such investigations, regulators would first need to request data from the exchanges or market participants executing the trades. This data would tell the regulators which members, subscribers, or broker-dealers sent the orders that led to the executions. Regulators would then need to go to the members, subscribers, and broker-dealers to get information on the orders and repeat until they get to the broker-dealer who initiated the order to see the customer behind the order.

Finally, some regulatory activities require data on both equities and options. Because current data sources do not contain information regarding both equities and options, regulators needing data on both types of securities would need to make several data requests. Closely related securities are sometimes traded on entirely different exchanges, complicating cross-product analyses. For example, COATS data covers options trades but excludes the trading of the underlying assets. Often investigations or analyses require examining both options and their underlying assets, creating the need for regulators to request data from multiple sources.

This data fragmentation also results in disparate requirements for industry members to record and report the same information in multiple formats. Because each SRO has its own data requirements, a market participant that is a member of multiple SROs may be required to report audit trail data in numerous formats and interact with multiple regulators in response to normal data queries. That said, the Commission understands that the number of disparate formats faced by each member may have reduced over the past several years.

For example, some exchange audit trails require floor brokers who operated on their own systems to submit order records to the exchange. These same floor brokers could be members of other SROs that require different formats for submitting order reports. The Commission understands that the volume of trading conducted on an exchange but not on the exchange's systems has declined sharply. Therefore, the activity generating the disparate reporting requirements has declined.

(4) Timeliness

In order to respond promptly to market events, regulators must be able to obtain and analyze relevant data in a timely fashion. Currently, obtaining trade and order data and converting the data into a form in which they can be analyzed can involve a significant delay from the time of a particular event of interest. Indeed, in some cases the length of time from when an event occurs until regulators can use relevant data in an investigation or analysis can be weeks or months. This is especially true for trading data that includes customer information.

Some of the data sources described above can be accessed by SROs and the Commission without significant delay. For example, SROs and the Commission have some real-time direct access to public data and, through MIDAS, the Commission has next-day direct access to analytics that are based on public data, such as volumes over various time horizons. Regulators can also sometimes request and receive trade blotter data on the same day as the trade(s) of interest because trade blotters are generally stored in systems immediately. Further, the Commission understands that FINRA receives audit trail data from exchanges pursuant to RSAs at the end of each trading day. However, it has been the Commission's experience that trade blotter data requests can take weeks or in excess of a month depending on the scope of the request and how accustomed the broker-dealer is with fulfilling such requests.

The regulated entities that respond to data requests need to query data to respond to the request while still maintaining normal operations. Large data requests can take significant computing time and thus, may require the respondents to time the queries to minimize disruptions. Further, respondents need to write code to execute the query. More experienced respondents would have existing code that they could modify without significant debugging whereas less experienced respondents would need to take time to code and debug their queries.

Corrected FINRA OATS data may be available less than two weeks after an event and uncorrected data on day T+1. In particular, FINRA members submit OATS data on a daily basis, submitting end-of-day files by 8:00 a.m. Eastern Time the following day or they are marked late by FINRA. FINRA acknowledges receipt of the data an hour after the member submits it, before running its validation process. FINRA then takes approximately four hours after acknowledging receipt of OATS data to determine if the data contain any syntax errors. In addition to the four hours needed to identify errors within a report, it takes another 24 hours for context checking, which identifies duplicates or secondary events without an originating event. Once a context rejection is available, the member has up to five business days to repair the rejection. Reports for files that contain internally inconsistent information about processing, linking, and routing orders may be available within two business days. FINRA attempts to match the inconsistent information against any additional data received up to day T+2 for linking errors and day T+30 for routing errors. The timing for surveillance programs varies depending on the type of surveillance being performed; data is assumed to be completely processed and corrected at day T+8.

FINRA currently receives exchange data from most SROs at the end of the trading day. Information on broker-dealer data reporting timeframes is available at OATS Reporting Technical Specifications, supra note 357, at 8-1; see also Adopting Release, supra note 9, at 45768 n.504.

See Section IV.D.2.b(2)A, supra (providing more detail on the validation and error checking process for OATS and other data sources).

See OATS Reporting Technical Specifications, supra note 357, at 6-3. Other types of errors and corrections adhere to slightly different time-lines. See, e.g., id. at 6-12.

FINRA has the capability to query data that is not fully corrected, processed and linked to investigate market activity at T+1.

Because market participants generally do not report or compile datasets immediately after an order event, there is a delay before regulators may access some data sources. For example, the compilation of equity and option cleared reports occurs on day T+1 for options and day T+3 for equities (i.e., the clearing day) and the electronic query access for equities is available from SIAC on day T+3. Additionally, when broker-dealers receive a request for EBS, the firm must first fill in the EBS report and then, if it does not self-clear, pass the reports on to its clearing firm to compile and send to SIAC. The EBS submission process can take up to ten business days. More immediate requests for cleared options data can be submitted to FINRA, but even this process takes up to two days. Because EBS data do not contain order entry time and order execution time, regulators must obtain this information from firms and brokers using either data requests or subpoenas, and this process generally can take from two to four weeks depending on the size of the request.

As discussed above, the lack of direct access to most data sources may further delay the ability of regulators to use data in certain cases. When regulators have direct access to a data source, the time needed to receive data is only the time it takes for a query to run. For example, depending on the scope of the search, it can take just a few minutes to return the results of a query of equity and option clearing data. As a result of direct access to their own audit trails, some SRO surveillance occurs on the same day as the trading activity. FINRA, however, typically gets direct access to exchange data, uncorrected OATS data, and corrected OATS data at the time it receives it, unlike the exchanges and broker-dealers that have some access to the data as it is generated. However, when regulators lack direct access, their data requests can consume significant time, including both the time required to put the request together and response times from the SROs, broker-dealers, and others producing the data. For example, obtaining complete responses from each broker-dealer for an EBS request can take days or weeks depending on the scope of the request. Likewise, responses from the ISG for SRO audit trail data can take days or weeks.

See Section IV.D.2.b(3), supra.

MIDAS, one example of a direct access data source, queries return data in seconds for single ticker, intraday queries and within hours for complex multi-ticker, multi-day queries. The data response times from MIDAS vary depending on the format of the resulting data and the number of other users on the system. A query that pulls all message traffic in an equity on a single day would take around thirty minutes.

FINRA typically collects exchange data at the end of the trading day and, as noted above, OATS on T+1. FINRA can begin to access each data source, but, as discussed below, FINRA has direct access to combined data only after the completion of the OATS error process and the processing necessary to reformat and merge the data sources.

As discussed above, because analysis of some events requires the collection of data from numerous sources, the time to request and receive data may be significant. The more fragmented the necessary data is, the longer it would take regulators to put together the data request. Putting together an EBS request, for example, could involve first identifying to which broker-dealers to send the requests and then manually creating a request letter for each broker-dealer. The Commission does recognize, however, that regulators can request and receive trade blotter data on the same day as the trade event if the request is for a small amount of data from an experienced provider. In fact, two years of trade blotter data from an experienced investment adviser can take several days while two years of data from clearing firms can take six weeks to several months.

Once regulators receive requested data, the data often have to be processed into a form in which they can be analyzed. As discussed above, it can take considerable time for regulators to combine data from different sources and link records from within or across data sources. Furthermore, the lack of consistency in format adds complexity to projects involving data from multiple data sources, even when the project does not involve linking of these different data. For example, the Commission understands that FINRA takes approximately three days to process exchange data to transform it into a common format and prepare it for surveillance. Therefore, FINRA cross-market surveillances and surveillance of the off-exchange market typically assumes data is fully corrected and processed on T+8. Any processing that requires linking order life-cycle events or other types of data can be time consuming to perform, even if all of the data comes from the same data source. In some cases, the laborious process of assembling the data delays other critical investigative or analytical steps.

See Section IV.D.2.b(2)C, supra.

Because no single data source is complete, regulators often need to combine data across sources to get a full picture. For example, regulators may need to compile SRO audit trail records from multiple SROs. Not all SROs collect data using the OATS format. The different data formats implemented by SROs thus involve a significant investment of staff time to reconcile. In addition, each options exchange maintains its own COATS audit trail in a different format and includes different supplemental data items in its audit trail. These differences make it difficult and labor intensive for regulators to view options trading activity across multiple markets.

FINRA can access data as soon as T+1 when necessary.

The first step in linking involves finding a key to link the records. The key can be one field or a series of fields in the data. The second step involves designing an algorithm to use the key to link records. If each data source formats or stores the fields in the key differently, the algorithm can be complex. Even within a single data source, the creation of the algorithm may be complicated because the fields needed to build the key can change with each market participant. For example, each member can report a different order ID for the same order, and this order ID may even change within the same member. The algorithm for linking needs to recognize how order IDs change and use additional information in the order records to piece an order lifecycle together. As noted above in Section IV.D.2.b(2)C, linking algorithms have varying rates of success and significant error rates in event linking are common. The lack of success could be due to the lack of a cross-participant error resolution process, the complexity in the linkage, or otherwise missing key information needed for linkage. As a result, regulators may invest significant time and resources into linking data only to achieve a success rate significantly less than 100%. Linking across multiple data sources makes linking even more time consuming.

In addition, those who use regulatory data also typically take time to ensure the accuracy of the data. When regulators question the accuracy of data, they often check several alternative sources until they are comfortable that their data are accurate. This checking of data accuracy and augmentation process adds time to an investigation or analysis. In some cases, regulators may filter out unreliable data or refocus an investigation to avoid relying on data after spending time and resources unsuccessfully attempting to ensure accuracy.

As discussed in the Adopting Release, the timely accessibility of data to regulators also impacts the efficacy of detecting (and possibly mitigating the effects of) some types of market manipulation. For example, some pernicious trading schemes are designed to generate large “quick-hit” profits in which market participants attempt to transfer the proceeds from the activity to accounts outside of the reach of domestic law enforcement as soon as the offending transactions have settled in the brokerage account (typically three days after execution). The timeframes currently required to acquire data generally complicate the prevention of these asset transfers.

See Adopting Release, supra note 9, at 45731.

3. Request for Comment on the Baseline

The Commission requests comment on all aspects of the Baseline for the economic analysis of the CAT NMS Plan. In particular, the Commission seeks responses to the following questions:

240. Do Commenters agree with the Commission's assessment of the Baseline for the economic effects of the CAT NMS Plan? Why or why not?

241. Do Commenters believe that the Baseline appropriately describes current market surveillance, examination, and investigation activities by regulators? Why or why not?

242. Do Commenters believe that the Baseline appropriately describes current market event analysis and reconstruction activities by regulators? Why or why not?

243. Do Commenters believe that the Baseline appropriately describes market analysis activities by regulators? Why or why not?

244. Do Commenters believe that the Baseline appropriately describes the sources of trade and order data currently available to regulators? Why or why not?

245. Are there additional sources of trade and order data currently available to regulators? Please explain and describe those sources in detail, including any limitations.

246. Do Commenters agree with the Commission's assessment of the completeness of the trade and order data currently available to regulators? Why or why not? Does the fragmented nature of current data sources pose significant challenges to regulators seeking complete data?

247. Do Commenters agree with the Commission's assessment of the accuracy of the trade and order data currently available to regulators? Why or why not?

248. Do Commenters agree that the error rates in current data sources or in responses to ad hoc data requests pose significant challenges to regulators? Why or why not? Do Commenters have additional statistics on error rates in these data?

249. Do Commenters agree with the Commission's assessment of the Baseline of clock synchronization for broker-dealers, exchanges, and others in the securities industry? Please explain. Does the Commission's analysis appropriately describe the frequency of orders that regulators may need to sequence and the challenges to sequencing given current clock synchronization standards? If not, do Commenters have more appropriate analyses? How could the Commission improve the analysis? Please explain.

250. Do Commenters believe that the Baseline appropriately describes granularity of time stamps in the trade and order data currently available to regulators? Please explain.

251. Do Commenters agree with the Commission's assessment of regulators' ability to combine or link data across the sources of trade and order data currently available to regulators? Please explain.

252. Do Commenters believe that the Baseline appropriately describes customer and broker-dealer identifiers in the sources of trade and order data currently available to regulators? Please explain.

253. Do Commenters believe that the Baseline appropriately describes aggregation within the sources of trade and order data currently available to regulators? Please explain.

254. Do Commenters agree with the Commission's assessment of the current ability of regulators to access trade and order data? Why or why not?

255. Do Commenters agree with the Commission's assessment of the timeliness of the trade and order data currently available to regulators? Why or why not?

256. Is there any other information that the Commission should include in the Baseline? Please explain.

E. Benefits

As noted in the Framework Section above, the economic benefits of the CAT NMS Plan would come from any expanded or more efficient regulatory activities facilitated by improvements to the data regulators use because the Plan would create a new consolidated data source, CAT Data that could replace the use of some current data sources for many regulatory activities. Therefore, the Benefits Section first describes how CAT Data compares to data regulators currently use for regulatory activities. Then this Section describes how the CAT Data would improve regulatory activities and how these improvements benefit investors.

The Commission preliminarily believes that the CAT NMS Plan would produce data that would improve on current data sources, because CAT Data would result in regulators having direct access to consolidated audit trail data that would improve many of the regulatory activities discussed in the Baseline Section. As summarized in Table 4, if the Plan is approved, the Commission preliminarily believes that the Plan would generate improvements in the quality of data that regulators would have access to in the areas of completeness, accuracy, accessibility, and timeliness. The Commission preliminarily believes that the improvements in the quality of regulatory data within these categories would significantly improve the ability of regulators to perform a wide range of regulatory activities, which would lead to benefits for investors and markets. In addition, the Commission preliminarily believes that certain provisions in the Plan related to future upgrades of the Central Repository, the promotion of the accuracy of CAT Data, the promotion of the timeliness of CAT Data, and the inclusion of specific governance provisions identified by the Commission in the Adopting Release for Rule 613, increase the likelihood that the potential benefits of the CAT NMS Plan described below would be realized.

In the category of completeness, the ability for regulators to access more material data elements from a consolidated source would enable regulators to more efficiently carry out investigations, examinations, and analyses because regulators could acquire from a single source data that they would otherwise need to compile from many data sources. This data source would include data elements that regulators currently acquire with difficulty (if at all), including customer information, allocation records, open/close position information for equities, and certain other trade and order information not consistently available in SRO audit trails.

See CAT NMS Plan supra note 3, at Sections 1.1, 6.3 and 6.4; see also 17 CFR 242.613(c)(7).

In the category of accuracy, the Commission preliminarily believes that the Plan would substantially improve data accuracy by requiring CAT Data to be collected, compiled, and stored in a uniform linked format using consistent identifiers for customers and market participants. These requirements should over time result in fewer inaccuracies in the data as well as fewer inaccuracies introduced in combining data compared to the current data regime. The CAT NMS Plan would also require that CAT Reporters' business clocks be synchronized to within 50 milliseconds of the time maintained by the NIST, which would increase the precision of the time stamps provided by the 39% of broker-dealers who currently synchronize their clocks with less precision than what is called for by the Plan. This information may also be used to partially sequence surrounding events. However, while the Commission preliminarily believes that the requirements in the Plan for clock synchronization and time stamp granularity would improve the accuracy of data with respect to the sequencing of market events, the improvements would be modest, as regulators' would experience improvement for a small percentage of market events relative to all surrounding events. Independent of the potential time clock synchronization benefits, the order linking data that would be captured in CAT should increase the proportion of events that could be sequenced accurately. This reflects the fact that some records pertaining to the same order could be sequenced by their placement in an order lifecycle (e.g., an order submission must have occurred before its execution) without relying on time stamps.

The Commission recognizes that the high initial Error Rate tolerance of the CAT NMS Plan could reduce the accuracy of raw CAT Data relative to current data sources. However, as stated in the Plan “the Participants expect that error rates after reprocessing of error corrections will be de minimis.” See CAT NMS Plan supra note 3, at Appendix C, Section 3(b), n.102.

See FIF Clock Offset Survey, supra note 127.

In the category of accessibility, the Commission preliminarily believes that the Plan would substantially improve the access of data for regulators due to the Plan's requirement for regulators to have direct access to CAT Data. While some elements of CAT Data can currently be obtained from other sources, it can take regulators weeks or months to obtain this data. As opposed to the current state of fragmented data with indirect regulatory access, if the CAT NMS Plan is approved, regulators would have direct access to consolidated trade and order data from a single source. Therefore, instead of requesting data from multiple sources, the Plan would allow regulators to log into a single system and query data directly from the system. This direct access for regulators would dramatically reduce the hundreds of thousands of requests that regulators must make each year in order to obtain data, thus reducing the burden on the industry.

In the category of timeliness, the Commission preliminarily believes that the Plan would significantly improve the timeliness of data acquisition and use, which could improve the timeliness of regulatory actions that use data. CAT Data would be reported by 8:00 a.m. Eastern Time on day T+1 and made available to regulators in raw form after it is received and passes basic formatting validations, with an error correction and linkage process that would be completed by 8:00 a.m. Eastern Time on day T+5. These requirements would ensure that data is available to regulators faster than in the current system and should also reduce the amount of time regulators would need to process data prior to usage.

While the Plan does not specify exactly when these validations would be complete, the requirement to link records by 12:00 p.m. Eastern Time on day T+1 gives a practical upper bound on this timeline.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.2(a).

Regulatory activities expected to benefit from improved data quality would include surveillance, investigations, examinations, analysis and reconstruction of market events, and analysis in support of rulemaking initiatives. Data is essential to all of these regulatory activities and therefore substantial improvements in the quality of the regulatory data should result in substantial improvements in the efficiency and effectiveness of these regulatory activities, which should translate into benefits to investors and markets. For example, improved data could lead to more effective and efficient surveillance that better protects investors and markets from violative behavior and facilitates more efficient and effective risk-based investigations and examinations that more effectively protect investors. Together, these improved activities could better deter violative behavior of market participants, which could improve market efficiency. Furthermore, this increase in directly accessible data should improve regulators' understanding of the markets, leading to more informed public policy decisions that better address market deficiencies to the benefit of investors and markets.

The Commission notes that the Plan lacks information regarding the details of certain elements of the Plan, primarily because many details likely to affect the benefits of the Plan have not yet been determined, which creates some uncertainty about the expected economic effects. As discussed further below, lack of specificity surrounding the processes for converting data formats and linking related order events creates uncertainty in the anticipated improvements in accuracy because such processes have the potential to create new data inaccuracies. Lack of specificity surrounding the process for regulators to access the CAT Data also creates uncertainty around the expected improvements in accessibility. For example, while the Plan indicates that regulators would have an on-line targeted query tool and a tool for user-defined direct queries or bulk extraction, the Plan itself does not provide an indication for how user-friendly the tools would be or the particular skill set needed to use the tools for user-defined direct queries.

See CAT NMS Plan, supra note 3, at Appendix D, Sections 8.1, 8.2.

1. Improvements in Data Qualities

As explained above, in the Adopting Release the Commission identified four qualities of trade and order data that impact the effectiveness of core SRO and Commission regulatory efforts: Accuracy, completeness, accessibility, and timeliness. In assessing the potential benefits of the CAT NMS Plan, the Commission's economic analysis compares the data that would be available under the Plan to the trading and order data currently available to regulators. As explained in detail below, the Commission preliminarily believes that the Plan would improve data in terms of all four qualities noted above, although uncertainty remains as to the expected degree of improvement in some areas.

See Adopting Release, supra note 9, at 45727.

Changes in all four data qualities affect certain data-driven regulatory activities. The benefits of the Plan derive from the changes to these regulatory activities.

a. Completeness

The CAT NMS Plan, if approved, would result in regulators having direct access to a single data source that would be more complete than any current data source. The CAT Data would be more complete than other data sources because it would contain data from a greater number of broker-dealers on more event types, products, and data fields, when compared to existing SRO audit trails and other data sources. As discussed in more detail below, while some current data sources contain many of the elements that would be included in CAT Data, the CAT Data would consolidate that data into one source to produce a data source much more complete than any existing source. CAT Data would also include some elements that are not available from any current data source. Having this data consolidated in a single source would provide numerous benefits that are described below.

See Sections IV.C.1.a(1) and IV.D.2.b(1), supra for a definition of completeness.

(1) Events and Products

CAT Data would be more complete than any current data source because it combines currently fragmented information into one data source. In particular, the Plan states that the Central Repository, under the Plan Processor's oversight, shall receive, consolidate, and retain all CAT Data. “CAT Data” is defined as “data derived from Participant Data, Industry Member Data, SIP Data, and such other data as the Operating Committee may designate as CAT Data from time to time.” Section 6.3 of the Plan describes the data to be received from Participants that are national securities exchanges, which would include data for “each NMS Security registered or listed for trading on such exchange or admitted to unlisted trading privileges on such exchange.” Participants that are a national securities association (i.e., FINRA) must report data for each “Eligible Security for which transaction reports are required to be submitted to that association.” “Eligible Security” is defined in the Plan as all NMS Securities and all OTC Equity Securities, and “OTC Equity Security” is defined as “any equity security, other than an NMS Security, subject to prompt last sale reporting rules of a registered national securities association and reported to one of such association's equity trade reporting facilities.” “Industry Member Data” refers to audit trail data reported by members of the exchanges and national associations, which includes Options Market Makers. SIP Data is defined in the Plan as information, including size and quote condition, on quotes including the National Best Bid and National Best Offer (“NBBO”) for each NMS Security; Last Sale Reports and transaction reports reported pursuant to an effective transaction reporting plan filed with the Commission pursuant to, and meeting the requirements of Rule 601 and 608; trading halts, limit-up limit-down (“LULD”) price bands, and LULD indicators; and summary data or reports described in the specifications for each of the SIPs and disseminated by the respective SIP.

See CAT NMS Plan, supra note 3, at Section 6.5(a)(i).

See id. at Section 1.1.

An “NMS Security” is defined as “any security or class of securities for which transaction reports are collected, processed, and made available pursuant to an effective transaction reporting plan, or an effective national market system plan for reporting transactions in listed options.” See 17 CFR 242.600(b)(46).

See CAT NMS Plan, supra note 3, at Section 6.3(c)(ii).

See id. at Section 1.1. Audit trail data regarding OTC Equity Securities was not required under Rule 613, but the Participants, in consultation with the DAG, included OTC Equity Securities in the CAT NMS Plan so as to permit the retirement of OATS and thereby reduce costs to the industry. See CAT NMS Plan, supra note 3, at Appendix C, Section C.9, Section A.1(a) n.16. The determination to include OTC Equity Securities would also have a positive effect on further reducing fragmentation of data sooner.

See CAT NMS Plan, supra note 3, at Section 1.1.

See id. at Section 6.4(d).

See Plan to Address Extraordinary Volatility for information on LULD, available at http://www.finra.org/sites/default/files/regulation-NMS-plan-to-address-extraordinary-market-volatility.pdf;; see also Securities Industry Automation Corporation, Consolidated Tape System, CTS, Output Multicast Interface Specifications, available at https://www.nyse.com/publicdocs/ctaplan/notifications/trader-update/cts_output_spec.pdf Securities Industry Automation Corporation, Consolidated Tape System, CQS, Output Multicast Interface Specifications, available at https://www.nyse.com/publicdocs/ctaplan/notifications/trader-update/cqs_output_spec.pdf. The UTP Plan Trade Data Feed SM. (UTDFSM. ), Direct Subcriber Interface Specification, Version 14.4 available at http://www.utpplan.com/DOC/utdfspecification.pdf.

See id. at Section 1.1 and Section 6.5(a)(ii).

CAT Data would include data from all SRO audit trails, combined into a single data source. In addition, it would include some off-exchange activity not captured on current SRO audit trails. Section 6.4(d) of the Plan requires the Participants to require their Industry Members to record and report order events to the Central Repository. The Commission notes that SRO audit trails currently do not include the activity of firms that are not members of that SRO. And, currently only FINRA requires its members to report their off-exchange activity. While broker-dealers that trade off-exchange must be members of FINRA unless their activity fits the terms of the exemption in Rule 15b9-1, firms that qualify for the exemption in that rule and that are not FINRA members do not report their off-exchange activity to OATS. This exemption amounts to a large percentage of off-exchange activity. Broker-dealers that are not FINRA Members accounted for 48% of orders sent directly to ATSs in 2014, 40% of orders sent directly to ATSs in 2013, and 32% in 2012. Because all SROs are Participants in the Plan, under the Plan all broker-dealers with Reportable Events, including off-exchange, would be required to report the required CAT Data to the Central Repository. And, the inclusion of these additional Reportable Events would make CAT Data more complete than the combination of current SRO audit trails.

This information can sometimes be inferred through data reported by member firms. See Securities Exchange Act Release No. 74581 (March 25, 2015), 80 FR 18036 (April 2, 2015) (“Proposed Amendments to Rule 15b9-1”), Section V.B.2; see also CAT NMS Plan, supra note 3, at Appendix C Section B.7(a)(ii)(A).

See id. for details on the exemption to Rule 15b9-1 and the proposed modifications to the Exemption for Certain Exchange Members that would require a dealer to be a member of a registered national securities association to conduct most off-exchange activity. If these modifications are adopted, Section IV.F.1.c(2)B.i discusses counts of broker-dealers currently not represented in OATS; the 15b9-1 exclusion applies to approximately 125 firms, most of which are not expected to incur OATS reporting obligations if 15b9-1 modifications are approved.

Furthermore, not all FINRA members are obligated to report to OATS. FINRA's rules exempt from reporting certain members that engage in a non-discretionary order routing process; additionally, FINRA has the authority to exempt other members who meet specific criteria from the OATS recording and reporting requirements, and has granted many such exemptions. See supra notes 396 and 397, and accompanying text. Approximately 799 firms that are excluded or exempt from OATS would incur CAT reporting obligations if the Plan were approved; see also Section IV.F.1.c(2)B.i, infra.

See Proposed Amendments to Rule 15b9-1, supra note 498, at n.21. If the Commission adopts the proposed amendments to Rule 15b9-1 set out in the proposed modifications to the Exemption for Certain Exchange Members, the percentage of off-exchange activity captured by CAT Data that is not currently captured by another audit trail would be smaller, and fewer broker-dealers would be excluded from OATS, reducing the number of broker-dealers that would be added to regulatory data if the Plan were approved. Section IV.F.1.c(2)B.ii discusses counts of broker-dealers currently not represented in OATS; the 15b9-1 exclusion applies to approximately 125 firms, most of which are not expected to incur OATS reporting obligations if 15b9-1 modifications are approved. Specifically, the exemption from FINRA membership would be limited to dealers that effect transactions off the exchanges of which they are members solely for the purpose of hedging the risks of their floor-based activity, and brokers and dealers that effect transactions off the exchange resulting from orders that are routed by a national securities exchange of which they are members. Id. at Section II.

CAT Data would also include many Reportable Events such as order origination, order routing, receipt of a routed order, order modifications, cancellations, and executions, and trade cancellations. Currently, OATS data contains most of these Reportable Events but does not cover all participants and does not include options. For example, CAT Data would contain more events than EBS data, trade blotters, and public data. As previously noted, OATS data also do not include proprietary orders originated by a trading desk in the ordinary course of a member's market making activities (or “principal activity”). But, pursuant to Rule 613(j)(8), principal trading would be included in CAT reporting requirements, an improvement over OATS. This requirement significantly improves completeness because such events are not included in current SRO audit trails, and account for a significant portion of market activity (40.5% of all transactions and 67% of all exchange message traffic according to a Commission analysis). This would improve regulatory activities in which observation of pricing information, as it relates to market activity, is important for determining the legality and consequences of market activity of interest as well as regulatory analysis of market behavior in general.

See Section IV.D.2.b(1)A, supra.

Id.

See Section IV.D.2.b(1)A, supra for a description of this analysis.

CAT Data also would include the information described above for listed equities and options and OTC Equity Securities. Therefore, the inclusion in CAT Data of all these products adds an additional level of completeness relative to current data sources.

See supra note 494.

(2) Data Fields

The CAT NMS Plan also would improve completeness by consolidating in a single source fields that currently may only be available from some data sources, and by including some fields that are difficult for regulators to compile. Not every data field that would be in CAT Data is currently included in SRO audit trails, and very few fields are included in all data sources.

The inclusion of consistent unique customer information, in particular, in the CAT Data represents a significant improvement over current SRO audit trails in terms of completeness. Rule 613(c)(7)(i) requires that a CAT Reporter report information to the Central Repository that uniquely identifies a customer across all broker-dealers. As noted in the Baseline, very few current data sources contain customer information, and those that do are largely limited in the completeness and accuracy of this information, all of which significantly limits regulatory efficiency. The identification of customers underlies numerous enforcement activities and many examination and surveillance activities of regulators. This would also allow regulators to obtain information efficiently regarding customers, such as issuers repurchasing their stock and short sellers.

17 CFR 242.613(c)(7)(i). Specifically, Sections 9.1 and 9.2 of Appendix D of the Plan require the CAT Data to include the following Customer information, at minimum: social security number or individual taxpayer identification number, date of birth, current name, current address, previous name and previous address. For legal entities, the Plan requires the reporting of the LEI (if available), tax identifier, full legal name and address. The Plan also requires that the following information about a Customer be reported to the Central Repository, at a minimum: Account owner name, account owner mailing address, account tax identifier, market identifiers, type of account, firm identifier number, prime broker ID, bank repository ID, and clearing broker. See CAT NMS Plan supra note 3, at Sections 9.1 and 9.2. The CAT Data must also support account structures that have multiple account holders. See id. Relatedly, the unique Customer-ID also improves accuracy because Rule 613 requires that it be consistent and associated with all Reportable Events involving that Customer. Current data sources do not provide consistent customer identifiers. See Sections IV.D.2.b(2)D supra, and IV.E.1.b(4), infra.

See Sections IV.D.2.a(1) and IV.D.2.b(1)B, supra. As discussed above, the Commission notes that SRO audit trails typically do not provide customer information but a recent FINRA rule change would require its members to report to OATS non-FINRA member customers who are broker-dealers. See supra note 407.

See Short Sale Reporting Study, supra note 413, for a discussion of the benefits of being able to identify short sellers. Because CAT Data would include a short sale mark and identify customers, regulators could use CAT Data to identify short sellers.

In addition to data fields providing customer information, the Plan would improve completeness by including other data fields not found on current SRO audit trails. For example, CAT Data would include allocation information, open/close information, Quote Sent Time, and information on whether a Customer gave a modification or cancellation instruction.

The information in the Allocation Report required by the CAT NMS Plan represents a significant improvement in completeness over current sources for subaccount allocation data, such as trade blotter and EBS data. Under the Plan, an Allocation Report would include the Firm Designated ID for any account(s), including subaccount(s), to which executed shares are allocated, the security that has been allocated, the identifier of the firm reporting the allocation, the price per share of shares allocated, the side of shares allocated, the number of shares allocated to each account, and the time of the allocation. While most of the fields required on the Allocation Report are included on trade blotter or EBS data, their inclusion in CAT Data would significantly reduce the time and effort expended for regulators to acquire such information. Because it is not required on EBS or in broker-dealer recordkeeping rules, the allocation time field on the Allocation Report provides information that is currently even more difficult for regulators to acquire than the other information on the Allocation Report. These data improvements should facilitate the use of allocation data in regulatory investigations and should result in more effective and efficient investigative processes. Allocation data also serves an important role in many other regulatory activities that aim to protect investors. Indeed, allocation time is an extremely important data field because it is critical in investigations of violations like market manipulation and cherry-picking.

See CAT NMS Plan, supra note 3, at Section 1.1; see also Exemption Order, supra note 18, at 11867.

See Section IV.D.2.b(1)B, supra, for further information on Allocation Reports.

Id.

Id.

In addition, while many of the elements contained in the definition of “Material Terms of the Order” are collected in current SRO audit trails, the CAT NMS Plan's definition of Material Terms of the Order expands the CAT Data beyond the coverage of current SRO audit trails and other sources. The CAT NMS Plan requires that the Material Terms of the Order be reported for order origination, routing, and the receipt of a routed order. And Material Terms of the Order is defined to include the security symbol, security type, price (if applicable), size (displayed and non-displayed), side (buy/sell), order type, if a sell order, whether the order is long, short, or short exempt, open/close indicator, time in force (if applicable), and any special handling instructions. In addition, if the order is for a Listed Option, the Material Terms of the Order would be defined to include option type (put/call), option symbol or root symbol, underlying symbol, strike price, expiration date, and open/close.

See CAT NMS Plan, supra note 3, Section 1.1; see also 17 CFR 242.613(j)(7).

Id.

Because data on open/close indicators are not currently included in SRO audit trails, obtaining data on whether a trade opens or closes a position in equities is currently very difficult. Ready access to this information would facilitate regulators' ability to determine whether a purchase or sale increases or decreases equity exposure, such as when a buy covers a short position. This would help regulators reconstruct customer positions without requiring specific position data and would assist in analysis of rules such as Rule 105 of Regulation M, governing when short sellers can participate in a follow-on offering. This information is also useful in investigating short selling abuses and short squeezes. Among other things, a build-up of a large short position by one investor along with the spreading of rumors may be indicative of using short selling as a tool to potentially manipulate prices. Information on when the position decreases is also useful for indicating potential manipulation, insider trading, or other rule violations. The ability to determine whether an order adds to a position, along with the timing of the order, is particularly important in detecting and investigating portfolio pumping or marking the close.

The open/close indicator would help to identify buy to cover orders because a buy order that closes a position would presumably be a buy-to-cover order. See Proposing Release supra note 9, at 32575. The Commission notes that the accuracy of this data field may depend on how the Plan Processor interprets when CAT Reporters should populate the field with particular permitted values. See infra note 537 and accompanying text.

17 CFR 242.105.

For a discussion of additional benefits of position information and buy to cover information, see Short Sale Reporting Study, supra note 413; see also Press Release: SEC Charges Six Firms for Short Selling Violations in Advance of Stock Offerings (October 14, 2015), available at http://www.sec.gov/news/pressrelease/2015-239.html .

See Proposing Release, supra note 9 at 32575.

Id.

Id.

The CAT Data would also include information regarding the sent time for Options Market Maker quotes and information about whether a modification or cancellation instruction for an order was given by a Customer associated with an order, or was initiated by a broker-dealer or exchange associated with the order. Neither of these data fields is currently readily available from existing SRO audit trails. Quote sent time is particularly informative for certain narrow market reconstructions for enforcement investigations, and knowing whether the member or Customer made a modification or cancellation helps regulators understand the decisions that broker-dealers and others make in the interest of best execution.

See Exemption Order, supra note 18 at 11857 and 11861.

The remaining data fields included in CAT Data are also included in some or all current SRO audit trails, although no single source contains all of them. For instance, Rule 613(c)(7)(vi)(C) requires the collection of audit trail data that links executions to contra-side orders and a CAT-Order-ID for the contra-side order. An order identifier for the contra-side order(s) would help regulators better reconstruct executions. Although some current exchange audit trails identify counterparties to trades, this identification is sometimes more difficult for off-exchange equity trading. Further, while all SRO audit trails contain time stamps, as CAT Data would, some sources of regulatory data do not currently include all the types of time stamps that would be in CAT Data.

17 CFR 613.242(c)(7)(vi)(C).

For off-exchange trading, OATS records sometimes do not directly identify counterparties. In the case of ATS trades, sometimes counterparty broker-dealers can only be identified through TRF records; sometimes ATS OATS records alone suffice. For internalized trades, the reporting broker-dealer is the counterparty. By combining OATS with TRF data, regulators can identify the broker-dealers representing the counterparties for over 99% of TRF reported trades, but identifying customer account information generally requires a data request to those broker-dealers. See Section IV.D.2.b(2)A, supra.

Additionally, the inclusion of order display information (i.e., whether the size of the order is displayed or non-displayed), and special handling instructions in CAT Data improve completeness because they are not always mandatory in SRO audit trail data, and therefore may not be consistently available without data requests to broker-dealers. Order display information is useful for examining how hidden liquidity affects markets or how regulatory changes affect hidden liquidity, and special order handling instructions could assist in examinations of best execution and could allow regulators to better understand the role and trends of these instructions in the market.

See supra note 412.

Other information required by the CAT NMS Plan includes the security symbol, date and time of the Reportable Event, the identity of each Industry Member or Participant accepting, routing, receiving, modifying, canceling, or executing each order, the identity and nature of the department or desk to which an order is routed, if an order is routed internally within the system of an Industry Member, a CAT-Order-ID, changes in any Material Term of the Order (if the order is modified), execution capacity, the CAT-Order-ID of any contra-side order(s), and the SRO-Assigned Market Participant Identifier of the clearing broker or prime broker. Of these fields, the security symbol and date are the only data found on all current data sources.

See CAT NMS Plan , supra note 3, Sections 6.3(d); 6.4(d).

The Commission preliminarily believes that the CAT Data would include all data elements that would be useful and efficient to include in a consolidated audit trail. The Commission previously considered which fields should be reported to CAT when proposing and adopting Rule 613. The set of data fields required by Rule 613 reflected the Commission's assessment, as informed by public comment, of the benefits and costs of including various data elements in CAT. While the costs and benefits of including particular fields can change due to technological advances and/or changes in the nature of markets, the Plan contains provisions regarding periodic reviews and upgrades to CAT that could lead to proposing additional data fields that are deemed important. In addition the Commission reviewed gap analyses that examine whether the CAT Data would contain all important data elements in current data sources. As a result of this review, the Commission is aware that one data gap involves OATS data fields that allow off-exchange transactions to be matched to their corresponding trade reports at trade reporting facilities, and recognizes that these fields are important to assure trade reporting requirements are being met for off-exchange trading. Similarly, the Commission notes that EBS includes 13 data elements that are not required by CAT or derivable through other CAT fields and would thus reflect some limitations of the Plan if EBS were retired before those missing data elements were incorporated into CAT. However, as discussed in Section 3 of Appendix D of the Plan, prior to the retirement of existing systems, the CAT Data must contain data elements sufficient to ensure the same regulatory coverage provided by existing systems that are anticipated to be retired. The Commission therefore expects that any missing elements that are material to regulators would be incorporated into CAT Data prior to the retirement of the systems that currently provide those data elements to regulators. And the Commission preliminarily believes that CAT Data would include the audit trail data elements that currently exist in audit trail data sources and that could be retired upon implementation of the CAT.

See Adopting Release, supra note 9, at 45751.

See Section IV.E.3a, infra for a discussion of adding new data fields and other requirements for upgrading the CAT Data after approval.

The Commission acknowledges that the Participants are continuing to study gaps between current regulatory data sources and the Plan as filed. CAT NMS Plan , supra note 3, at Appendix C, Section C.9; see also SEC Rule 613—Consolidated Audit Trail (CAT) OATS—CAT Gap Analysis and SEC Rule 613—Consolidated Audit Trail (CAT) Revised EBS—CAT GAP Analysis, available at http://www.catnmsplan.com/gapanalyses/index.html.

The Commission notes that Rule 613 does not require the inclusion of this information. This information did not exist at the time the Commission adopted Rule 613 and such information on exchange trades does not exist today. The Commission expects that the requirements discussed in Section 3 of Appendix D of the Plan would result in the inclusion of this information in the CAT Data.

See SEC Rule 613—Consolidated Audit Trail (CAT) Revised EBS—CAT GAP Analysis, available at http://www.catnmsplan.com/gapanalyses/p450537.pdf .

See CAT NMS Plan, supra note 3, at Appendix D, Section 3.

b. Accuracy

This Section analyzes the expected effect of the CAT NMS Plan, if approved, on the accuracy of the data available to regulators. In general, the Commission preliminarily believes that the requirements in the CAT NMS Plan for collecting, consolidating, and storing the CAT Data in a uniform linked format, the use of consistent identifiers for Customers, and the focus on sequencing would promote data accuracy.

As discussed above and in the Adopting Release, accuracy refers to whether the data about a particular order or trade is correct. See Adopting Release, supra note 9, at 45727.

The Commission notes that the full extent of improvement that would result from the Plan is currently unknown, because the Plan defers many decisions relevant to accuracy until the Plan Processor publishes the Technical Specifications and interpretations. In particular, the CAT NMS Plan specifies that the “[t]echnical Specifications shall include a detailed description of . . . each data element, including permitted values, in any type of report submitted to the Central Repository” and “the Plan Processor shall have sole discretion to amend and publish interpretations regarding the Technical Specifica- tions.” This leaves open precise definitions and parameters for the data fields to be included in CAT Data.

See CAT NMS Plan, supra note 3, at Section 6.9.

Id. at Section 6.9(b)(v).

The CAT NMS Plan provides details regarding how the responsibility for these decisions would be shared between the Operating Committee and the Plan Processor, with the Plan Processor having responsibility for data definitions and interpretations. See CAT NMS Plan, supra note 3, at Section 6.9(c)(i).

For example, the completeness Section notes that the open/close indicator for equities does not exist in current data sources (see Section IV.E.1.a(2)). The accuracy of the open/close indicator would be subject to Plan Processor discretion, because the Plan Processor would have responsibility for defining the permitted values and interpreting when CAT Reporters would use such permitted values and the Plan Processor would not have guidance from previous data sources on how to define or interpret such a field. While the Commission would ultimately be able to correct such misinterpretations, regulators may not detect such a misinterpretation until the misinterpretation harms an investigation, exam, or other analysis. Based on its experience with short sale indicators, the Commission believes that defining and interpreting the open/close indicator would be particularly complex. See SEC, Division of Market Regulation: Responses to Frequently Asked Questions Concerning Regulation SHO, Question 2.5, available at http://www.sec.gov/divisions/marketreg/mrfaqregsho1204.htm (“Regulation SHO FAQs”).

Nonetheless, the Commission preliminarily believes that the Plan provides some procedural protections to mitigate this uncertainty and help promote accuracy. For example, the Plan requires that, at a minimum, the Technical Specifications be “consistent with [considerations and minimum standards discussed in] Appendices C and D,” and that the initial Technical Specifications and any Material Amendments thereto must be provided to the Operating Committee for approval by Supermajority Vote. Further, all non-Material Amendments and all published interpretations must be provided to the Operating Committee in writing at least ten days before publication, and shall be deemed approved unless two or more unaffiliated Participants call the matter for a vote of the full Operating Committee.

Id. at Section 6.9(a). The Commission notes that the standards in Appendices C and D do not cover all decisions that would affect the accuracy of the data.

See CAT NMS Plan, supra note 3, at Section 6.9(c)(i).

(1) Data Errors

The CAT NMS Plan specifies a high-level process for handling errors that includes target Error Rates for data initially submitted by CAT Reporters and a correction process and timeline. In particular, the Plan specifies an initial maximum Error Rate, which measures errors by CAT Reporters and linkage validation errors, of 5% for reports received by the Central Repository before the error correction process and contemplates the reduction of this Error Rate over time. It is difficult to conclude whether the Error Rates and processes in the CAT NMS Plan would constitute an accuracy improvement as compared to current data sources.

The Commission notes that there is some uncertainty on whether the Error Rate definition includes any additional errors attributable to the Plan Processor because the Plan does not explicitly state whether Plan Processor errors are included in the Error Rate or not; it is also not clear whether Plan Processor errors are included in linking errors. See id. at Article VI, 6.1(n)(v) n.1; Appendix C, Section A.3(b), n.102. Additional uncertainty exists because the Operating Committee would determine the details regarding error definitions in the Technical Specifications after the Plan is approved.

The Plan states that 5% is an appropriate initial Error Rate, to allow CAT Reporters the opportunity to get used to a new reporting regime, and that the Error Rate should be reduced over time, with goal of a 1% Error Rate to be achieved one year after each new category of Reporters is required to begin reporting. This was determined based on Participants' experience with OATS. The initial rejection rates for OATS when it was initially implemented was 23%, although more recent experience with OATS reporting indicates error rates below 3% following the implementation of additional OATS upgrades over the past 10 years and a current error rate of less than 1%.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(b).

See id. at Appendix C, Section A.3(b), n.106.

See Memorandum to File No. S7-11-10 regarding Telephone Conferences with FINRA (April 17, 2012) available at http://www.sec.gov/comments/s7-11-10/s71110-116.pdf .

But, because the current OATS error rate is below 1%, the Commission preliminarily believes that the initial percentage of errors in CAT would be higher than current percentage of errors in OATS, though the OATS error rate may not be directly comparable to the Error Rate in the Plan. Given the magnitude of CAT, the fact that many CAT Reporters would be new to audit trail reporting, and that options would be covered for the first time, the Participants believe that 5% is an appropriate initial Error Rate. And the Plan injects some uncertainty by asserting that this initial 5% rate is subject to the quality assurance testing period to be performed prior to launch, and then again before each new batch of CAT Reporters are brought online. In time, the rate could be lowered, but it also could be raised.

See Section IV.D.2.b(2)A, supra, for discussion of current regulatory data error rates. It is important to note that both the 1% OATS error rate and the 5% proposed CAT Error Rate represent error rates measured at initial data submission. Furthermore, some situations that do not qualify as an error in OATS (i.e., a route that cannot be linked because the routing destination is not required to report OATS) would qualify as an error under CAT. Furthermore, error rates after data correction are not known for OATS, and are anticipated to be “de minimis” under CAT, as discussed in note 547, infra. Finally, definitions of “error” for both OATS and CAT Data are dependent on proscribed data validation checks; if data is reported and passes validation checks, it is assumed to be correct. When validation checks are exhaustive and stringent, error rates are expected to be higher than when validation checks are minimal. Consequently, the Commission is cautious in directly comparing OATS reported and proposed CAT Error Rates.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(b). See also Section IV.H.2.b, infra for a discussion and solicitation of comment on alternative Error Rates.

See id. at Appendix C, Section A.3(b).

The Plan specifies an error correction process after initial reports are received and indicates that practically all errors identifiable by the validations used in the error correction process would be corrected by 8:00 a.m. Eastern Time on day T+5, stating that errors are expected to be “de minimis” after the error correction period. Specifically, the Plan Processor must run initial validation checks on the data by noon eastern time on day T+1 (four hours after the submission deadline for the data). Those validation checks must be published in the Technical Specifications (as discussed further below) and have the objective to ensure that data is accurate, timely, and complete as near as possible to the time of submission. Once errors are identified, the Plan Processor must accept corrections via manual web-based entry and via batch uploads. Although there is a specific timeframe for performing these corrections, the Plan Processor must accept error corrections at any time.

See id. at Appendix C, Section A.3(b) n.102. “De minimis” is not defined and no numerical Error Rate is given. The Plan also includes a compliance program intended to help achieve this goal.

See Section IV.E.1.d, infra. The RFP requested that Bidders provide information on how data format and context validations for order and quote events would be performed and how errors would be communicated to CAT Reporters; a system flow diagram showing how and when different types of validations would be completed; and how Customer information would be validated. Bidders noted that the validations would be performed via rules engines (using standard data validation techniques like format checks, data type checks, consistency checks, limit and logic checks, or data validity checks), and processing would be done in real time during data ingestion. The Plan Processor would be required to perform validations within three specified categories, which must be set out in the Technical Specifications document: File Validations (confirmation that the file is received in the correct format); Validation of CAT Data (checks of format, data type, consistency, range/logic, data validity, completeness, and timeliness); and Linkage Validation (checking the “daisy chain”). See CAT NMS Plan, supra note 3, at Appendix D, Section 7.2. If errors are found, the data would be stored in an error database and notification sent to the CAT Reporter.

Rather than providing details on the validations that would occur, however, the Plan provides high-level requirements for the validations and delegates the detailed design of the specific validations to the Plan Processor (with the involvement of the Operating Committee and the Advisory Committee). Additionally, the Plan does not provide the level of detail necessary to verify whether the CAT validation process would run the same validations as OATS, whether current validations would be relevant, and what validations, if any, would be added.

See CAT NMS Plan, supra note 3, at Appendix D, Section 7.2 (discussing validation requirements); see also id. Appendix C at Section A.3(b) (delegating responsibility regarding measurement of Error Rates to the Plan Processor).

As noted above, it is therefore difficult to conclude whether the Error Rates and processes in the CAT NMS Plan would constitute an accuracy improvement as compared to current data sources. With respect to OATS, FINRA currently performs over 152 validation checks on each order event reported. After corrections, approximately 1-2% of each day's recorded events remain unmatched (i.e., multi-firm events, such as order routing that cannot be reconciled). However, the Commission is not certain that those error rates are directly comparable to the Error Rates permitted for CAT Data in the Plan given the increased scope and level of linkages specified in the Plan, and the new, large, and untested system. The Commission is not aware of other systems that track and record similar error rates, although the Commission does experience issues with errors contained in other sources of data when the Commission attempts to use that data. Accordingly, the Commission is unable to conclude whether the Error Rates and processes in the Plan would constitute an accuracy improvement compared to current data.

See Adopting Release, supra note 9, at 45729.

Id. at 45778.

(2) Event Sequencing

A. Clock Synchronization

Rule 613(d)(1) and (2) requires that the CAT NMS Plan require that the business clocks of Participants and their members be synchronized to a specified standard of precision and for protocols to be in place for that standard to be maintained over time. Complying with this clock synchronization standard will require that, for the purpose of recording the date and time of Reportable Events, the business systems of Participants and their members be synchronized consistently with “industry standards.” The Commission did not define the term “industry standard” in Rule 613, though it noted that it expected the Plan to “specify the time increment within which clock synchronization must be maintained, and the reasons the plan sponsors believe this represents the industry standard.”

See Adopting Release, supra note 9, at 45774.

The CAT NMS Plan describes the “industry standard” in this context in terms of the technology adopted by the majority of the industry. The Plan therefore bases its clock synchronization standard on current practices of the broker-dealer industry generally and provides that one standard would apply to all CAT Reporters. Specifically, Section 6.8(a) of the CAT NMS Plan requires CAT Reporters to synchronize their time clocks to the time maintained by the NIST with an allowable clock offset of 50 milliseconds, which the Plan determines is consistent with the current industry standards, as defined in the Plan. The Plan further requires annual review of the clock synchronization standard to evaluate its achievement of the Plan's goals related to clock synchronization. Section 6.8(c) of the Plan requires the Chief Compliance Officer to annually evaluate the clock offset tolerance and to make recommendations to the Operating Committee regarding whether industry standards have evolved such that the standard in Section 6.8(a) should be shortened.

See CAT NMS Plan, supra note 3, at Appendix C, Section 12(p).

See id. at Section 6.8.(c) and Appendix C, Section A.3.(c)

The Commission preliminarily believes that the clock synchronization standards in the CAT NMS Plan are reasonably designed to improve the accuracy of market activity sequencing by increasing the percentage of order events that could be chronologically sequenced relative to other order events, but notes that the improvements to the percentage of sequenceable order events by Plan standards are modest and the requirements of the Plan may not be sufficient to completely sequence the majority of market events relative to all other events.

Independent of the potential time clock synchronization benefits, the order linking data that would be captured in CAT should increase the proportion of events that could be sequenced accurately. This reflects the fact that some records pertaining to the same order could be sequenced by their placement in an order lifecycle (e.g., an order submission must have occurred before its execution) without relying on time stamps. This information may also be used to partially sequence surrounding events.

As discussed in the Baseline Section, 39% of the broker-dealers responding to the FIF Clock Offset Survey currently synchronize their clocks to a clock offset tolerance of greater than 50 milliseconds. Accordingly, the 50 millisecond requirement for all CAT Reporters (except on manual order handling systems) would result in the availability of more precise time stamps from many broker-dealers and would increase the number of order events that could be accurately sequenced relative to each other.

See Section IV.D.2.b(2)B.i, supra (reporting results of this survey); see also FIF Clock Offset Study, supra note 127.

As noted above, FINRA has indicated that it is considering proposing a rule change that would require a 50 millisecond clock offset tolerance. If this rule change is proposed and approved, more entities would record time stamps with data at a 50 millisecond clock offset tolerance regardless of whether the CAT NMS Plan is approved.

To evaluate the proportion of order events that could be sequenced with the clock offset tolerance specified in the CAT NMS Plan, the Commission has conducted an analysis of the frequency of market events occurring within 100 milliseconds of an event in a different trading venue in the same security. Table 5 (CAT and Current Clock Offset Tolerance) shows the percentage of events for listed equities and options that could be accurately sequenced with one-second and 50-millisecond clock offset tolerances.

The methodology to calculate these frequencies starts with the steps described in supra note 438 and then subtracts the result from one to get the percentage of unrelated orders that could be sequenced. This assumes that consecutive unrelated events within twice the clock offset tolerance cannot be sequenced. An unrelated event is an order event at a different venue.

Table 5—CAT and Current Clock Offset Tolerance

% of Unrelated order events
Minimum time between adjacent events Clock offset tolerance Equities (%) Options (%)
2 seconds 1 second 1.31 6.97
100 milliseconds 50 milliseconds 7.84 18.83

The analysis finds that the current FINRA one-second clock offset tolerance allows only 1.31% of unrelated order events for listed equities and 6.97% of unrelated order events for listed options to be sequenced. The proposed 50-millisecond clock offset tolerance could accurately sequence 7.84% for listed equities and 18.83% for listed options of such events included in the MIDAS data. This analysis overestimates the portion of unrelated events that the proposed clock synchronization standard could sequence because the analysis includes only trade and quote events observable in the MIDAS data. The data currently available to the Commission provides only a rough and upwardly-biased estimate of how many of these events could be sequenced by the order data that would be captured by the CAT. In sum, the results of the Commission's analysis suggest that the standards required by the Plan do represent an improvement over current standard but that the majority of market events would remain impossible to sequence based on the Plan's required clock synchronization standards.

This analysis does not consider events in OTC Equity Securities. The Commission believes that the proposed clock synchronization standard could accurately sequence a higher proportion of unrelated events in OTC Equity Securities because OTC Equity Securities trade less frequently than NMS equities and unrelated order events may be less frequent in OTC Equity Securities than in listed equities. The Commission therefore preliminarily believes that the proposed 50 millisecond clock offset tolerance in the CAT NMS Plan could improve accuracy by modestly increasing the number of events that could be sequenced in OTC Equity Securities.

The Plan acknowledges that the required clock offset tolerance, which is based on its determination of the current industry standard, would not be sufficient to accurately sequence all order events by their time stamps alone. In particular, the Plan states that “[f]or unrelated events, e.g., multiple unrelated orders from different broker-dealers, there would be no way to definitively sequence order events within the allowable clock drift as defined in Article 6.8.” This in turn limits the benefits of CAT in regulatory activities that require event sequencing, such as the analysis and reconstruction of market events, as well as market analysis and research in support of policy decisions, in addition to examinations, enforcement investigations, cross-market surveillance, and other enforcement functions.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(c). Order events occurring within a single system using the same time clock could be accurately sequenced by their time stamps, assuming that their time stamps are not identical. The CAT NMS Plan does not specify the approach that would be used to sequence events when time stamps are identical or indicate how this decision would be made.

Id. at n.110. Events involving the same order routed across systems could be logically sequenced using routing-related data, because a routed order must be sent before it can be received, and received before it can be executed. However, the Plan would not facilitate the accurate sequencing of events that occur in different systems within 100 milliseconds of each other (twice the clock offset tolerance) that are not linked using a parent-child order relationship. The CAT NMS Plan does not provide a solution that will sequence these events, but recognizes the issue and states that “the Participants plan to require that the Plan Processor develop a way to accurately track the sequence of order events without relying entirely on time stamps.” See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(c).

The Plan discusses its determination of the current industry standard and specifies implementation requirements for the clock synchronization standards in Appendix C. As noted above, the Plan bases industry standards on current practices of the broker-dealer industry, which are derived from a survey of broker-dealers, and on the assumption that a change in industry standards would be premised on “the extent existing technology that synchronizes . . . clocks with a lower tolerance . . . becomes widespread enough throughout the industry to constitute a new standard.”

See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(c).

Id.

The Commission notes however, that the current practices for exchanges and Execution Venues may differ from the industry standard for broker-dealers as defined in the Plan, and current practices for certain systems within broker-dealers may vary by the system within the broker-dealers. As noted in the Baseline Section, the Commission does not have precise information on the clock synchronization standards on exchange and ATS matching engines and quoting systems, but exchanges may currently synchronize their clocks to a 100 microsecond or less clock offset tolerance, and have an average clock offset of 36 microseconds. By defining industry standards based on practices of the broker-dealer industry generally, the Plan does not account for these differences. Further, defining industry standards by majority practices may have the unintended effect of setting a standard that delays adopting advances in technology.

See supra notes 435 and436.

Despite these limitations, it is worth noting that the Plan requires the CCO of the Plan Processor to develop and conduct an annual assessment of Business Clock synchronization. Moreover, Plan Participants must require Industry Members to certify periodically that their Business Clocks comply with the clock synchronization standard and that any violations thereof are reported to the Plan Processor and the Plan Participant. Thus, the Commission believes that these provisions would help ensure that the benefits of clock synchronization are maintained.

See CAT NMS Plan, supra note 3, Section 6.2(a)(v)(M).

See id. at Section 6.8(a)(ii) and (iii).

B. Time Stamp Granularity

The Commission preliminarily believes that the minimum time stamp granularity required by the Plan would result in some improvement in data accuracy, but that the level of improvement could be limited. Despite the modest level of direct improvements expected from the Plan's minimum time stamp granularity standards, the Commission preliminarily believes that the Plan should continue to have a time stamp granularity standard because the Plan provides a mechanism for making future improvements and monitoring whether more granular time stamps would provide better quality CAT Data and be feasible given technology improvements.

The level of precision or granularity with which time stamps are recorded has significant implications for the usability of audit trail data in terms of sequencing events, matching records, and linking the data to other data sources. In some current regulatory data, the relative lack of time stamp granularity standards for data reporters could lead to difficulties in accurately sequencing events or linking data with other data sources. Rule 613(d)(3) requires that CAT Reporters record time stamps to reflect current industry standards and be at least to the millisecond. Furthermore, the Plan requires Participants to adopt rules requiring that CAT Reporters that use time stamps in increments finer than milliseconds use those finer increments when reporting to the Central Repository. Consistent with Rule 613, Section 6.8(b) of the CAT NMS Plan requires millisecond or less time stamps. However, the Commission granted exemptive relief for manual orders to be recorded at the granularity of one second or better. Further, pursuant to Rule 613, if a CAT Reporter's system already utilizes time stamps in increments less than the minimum required by the Plan, the CAT Reporter must record time stamps in such finer increments.

17 CFR 242.613(d)(3). This requirement does not apply to certain Manual Order Events, which are exempted from the requirement and are captured at one-second increments. Time stamp granularity on manual order events is discussed separately in the Alternatives Section.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(c).

See CAT NMS Plan, supra note 3, at Section 6.8(b) and Appendix C, Section A.3(c) (explaining that recording Manual Order Events at the millisecond level would be costly and ultimately arbitrary or imprecise due to the human interaction); see also Exemption Order, supra note 18, at 11868-9.

Id.

The Plan asserts that the millisecond increment required for CAT Data reflects the industry standard level of granularity. As noted in the discussion of clock synchronization, the Commission did not define the term “industry standard” in Rule 613. The Plan therefore bases its standard for time stamp granularity on current practices of the broker-dealer industry generally, and provides that one standard would apply to all CAT Reporters. There appears to be a wide divergence of industry standards in practice, ranging from full seconds to microseconds for latency-sensitive applications, and the Plan describes the slower systems as mostly older ones that cannot support a finer time stamp granularity. Many of the systems from which regulators currently obtain data already capture time stamps in increments of milliseconds or less. For example, OPRA allows for time stamps in nanoseconds, and the other SIPs require time stamps in microseconds for equity trades and quotes. However, OATS and EBS do not. Current OATS rules require time stamps to be expressed to the nearest second, unless the member's system expresses time in finer increments; and as of September 2014, approximately 12% of OATS records contain time stamps greater than one millisecond. EBS records either do not contain times or express time stamps in seconds.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(c).

Id. Because older technology cannot support finer time stamp increments, members with older systems would incur significant effort and cost to upgrade those systems to support reporting data in milliseconds. The newest systems support finer increments, but include mostly the subset of systems dealing with low latency trading. Electronic Order Handling and Trading systems are commonly set at the millisecond level; see, e.g., FIF Letter.

See Section IV.D.2.b(2), supra.

Id.

Thus, to the extent that some current data sources report time stamps in increments coarser than a millisecond, which is the case for 12% of OATS records and all EBS records, the Commission expects the CAT millisecond time stamp requirement to improve data, and thereby allow regulators to more accurately determine the sequence of market events relative to surrounding events.

The Commission preliminarily believes, however, that benefits from the more granular time stamps could be limited by the level of clock synchronization required by the Plan. In particular, the Commission believes that time stamp granularity would not be the limiting factor in sequencing accuracy, because recording events with time stamps with resolutions of less than one millisecond cannot help to sequence events occurring on different venues with clocks that may be 100 milliseconds out of sync due to clock synchronization offsets. Therefore, the benefits of time stamping order events at increments finer than a millisecond would be limited without also improving the clock synchronization standards of the Plan.

For example, under the requirements in the Plan, an order event at Broker-Dealer A could have a time stamp that is 1 millisecond sooner than an order event at Broker-Dealer B even if the event at Broker-Dealer B actually occurred 99 milliseconds sooner. This could occur if Broker-Dealer A's systems are recording times 50 milliseconds ahead of NIST while Broker-Dealer B's systems are recording times 50 milliseconds behind NIST. Both broker-dealers' systems would be within the Plan's allowable clock synchronization tolerance.

(3) Linking and Combining Data

The Commission believes the requirements of Rule 613 and the Plan related to data linking would result in improvements to the accuracy of the data available to regulators, but the extent of the improvement would depend on the accuracy of the linking algorithm and the reformatting process that the Plan Processor would eventually develop.

As discussed in the Baseline, data is currently stored in multiple formats, is difficult to merge, and results in errors during the merging process. Moreover, in some cases, the data sources do not capture the information necessary to link records, while in other cases linking must be done with algorithms that accomplish the linking with some degree of error.

Rule 613(e)(1) generally requires the creation and maintenance of a Central Repository that would receive, consolidate, and retain information reported to the CAT. Further, the rule requires that the Central Repository store and make available to regulators data in a uniform electronic format and in a form in which all events pertaining to the same originating order are linked together in a manner that ensures timely and accurate retrieval of information reported to the CAT.

17 CFR 242.613(e)(1); see also CAT NMS Plan, supra note 3, at Section 6.5(a) and (b).

The Commission preliminarily believes that the requirement that data be stored in a uniform format would eliminate the need for regulators who are accessing the data to reformat the data. As noted in the Baseline Section above, regulators face delays and inaccuracies when attempting to reformat and link data from multiple sources, such as linking trade blotters from several broker dealers with SRO audit trails. Given that the reformatting of CAT Data would be accomplished by individuals that likely specialize in this activity and that repetitively do so in a prescriptive and formalized way, this requirement could reduce the errors that could be introduced in the current regime where reformatting data is often done on an ad hoc basis by regulatory Staff who need to work with the data. In other words, the Plan Processor would develop a reformatting process by working with CAT Reporters to build an expertise in harmonizing the various formats that it receives from Reporters. The Plan Processor could then build, test, and refine the reformatting process with the ability to go back to the CAT Reporters for further clarification. Even if only one Staff member at each SRO or Affiliated Participant developed the expertise necessary to reformat each of the various formats and ran a reformatting process on order data, this would result in a duplication of efforts compared to one centralized entity (the Plan Processor) developing the expertise and running the reformatting process. Storing data in a linked format removes the need for regulators to link information from multiple lifecycle events of an order or orders themselves, which could further reduce errors and increase the usability of the data. The Commission recognizes, however, that despite the potential improvements, the CAT Data could still contain errors introduced in the reformatting and linking processes.

Whether errors would decrease depends on the actual formatting process used.

The process for linking orders designated in the CAT NMS Plan is similar to the process FINRA currently uses to link OATS records across market participants. However, the Plan would significantly improve the ability of regulators to link order events compared to OATS, and would link this activity to specific customers unlike current audit trail data. CAT Reporters must report a series of unique identifiers that are designed to allow records of events that occur over the order's lifecycle to be linked together to determine how the order was handled and how the order interacted with other orders. The Plan Processor must then create the initial linkages in the submitted data; unlike in OATS, the Plan Processor would verify these linkages as part of its data validity checks. In general, the CAT NMS Plan would link orders using the “daisy chain approach,” where CAT Reporters assign their own identifiers to each order event that the Plan Processor later replaces with a single identifier (the CAT Order-ID) for all order events pertaining to the same order. The Central Repository at a minimum must be able to create linkages between all order events that are internalized, between the Customer execution and a proprietary order in the case of a riskless principal transaction, between two broker-dealers, between a broker-dealer and an exchange, and vice versa, between executed orders and trade reports, between various legs of option/equity complex orders, and between order events for all equity option order handling scenarios that currently are or could potentially be used by CAT Reporters.

As discussed above, the Commission notes that SRO audit trails typically do not provide customer information but a recent FINRA rule change requires its members to report to OATS non-FINRA member customers who are broker-dealers. See supra note 407.

See id. at Section 6.3(d)(i) through (vi).

These data validations are to be established in a Technical Specifications document by the Plan Processor. Consequently, it is as yet unclear precisely how that process would occur. See id. at Appendix D, Section 7.2; Appendix C, Section A.3(a) (validations ensure that data is submitted in required formats and that lifecycle events can be accurately linked).

See id. at Appendix D, Section 3.

See id.

Unlike OATS data, CAT Data would be less prone to breaking the order lifecycle chain when an order is sent across market participants because the order lifecycle linking procedure across reporters would be uniform and all industry participants would be reporters. Currently, linking procedures across SROs are not uniform, which complicates reconstructing order lifecycles. Furthermore, because some broker-dealers are not required to report to OATS, these broker-dealers' activity cannot be completely reconstructed from audit trail data, and therefore, orders that they handle cannot be traced through their lifecycle, effectively severing the links between the order being received and the order's final disposition. Furthermore, as covered elsewhere, unlike other data sources, CAT Data would link orders to Customers because the Plan requires the order lifecycle to be linked back to the original Customer, and the Plan Processor must be able to fix linkages when error correction files are submitted. While the success of such a matching process is dependent on the accurate reporting of order linkages by CAT Reporters, Appendix D directs the Plan Processor to ensure that breaks in certain lifecycle linkages must not cause the entire lifecycle to break or cause a CAT Reporter that correctly reports information to have its submission rejected.

See Section IV.D.2, supra.

See id.

For example, assume two broker-dealers handle an order that is ultimately executed on an exchange. Broker-Dealer A receives the order, and transmits it to Broker-Dealer B, that routes it to Exchange C where it is executed. In order for the Plan Processor to link these three order events, Broker-Dealer A would need to report the order and its routing to Broker-Dealer B; B would need to correctly echo A's order ID in its CAT reporting and its route to Exchange C, and C would need to correctly echo Broker-dealer B's order ID in its CAT reporting.

See CAT NMS Plan, supra note 3, at Appendix D, Section 7.3. The Commission also notes that, even if all CAT Reporters provide the required linking information, the success of the linking process would depend in part on the approach taken by the Plan Processor and whether or not that approach results in errors.

The CAT NMS Plan does not provide sufficiently detailed information for the Commission to estimate the likely Error Rates associated with the linking process required by the CAT NMS Plan. Indeed, the 5% Error Rate covers data from CAT Reporters, but the Plan Processor could create errors as well, for example, through the linking process. Further, the Plan does not include details on how the Plan Processor would perform the linking process, identify broken linkages, and seek corrected reports from CAT Reporters to correct broken linkages. Instead, the Plan defers key decisions regarding the validation process until the selection of a Plan Processor and the development of Technical Specifications. Accordingly, while the centralized linking should generally promote efficiencies and accuracies in linking, these uncertainties make it difficult for the Commission to gauge the degree to which the process for linking orders across market participants and SROs would improve accuracy compared to existing data, including OATS.

The CAT NMS Plan describes the Plan Processor's responsibility for creating the Technical Specifications. See CAT NMS Plan, supra note 3, at Section 6.9.

The Commission notes that the Plan Processor is required to create a quality assurance testing environment in which, during industry-wide testing, the Plan Processor provides linkage processing of data submitted, the results of which are reported back to Participants and to the Operating Committee for review. See CAT NMS Plan, supra note 3, at Appendix D, Section 1.2. This may help identify challenges in the linking process and allow for their early resolution.

Uncertainties also prevent the Commission from determining whether the process for converting data into a uniform format at the Central Repository would improve the accuracy of the data over existing audit trail accuracy rates. The Plan includes two alternative approaches to data conversion. In the first, called Approach 1, CAT Reporters would submit data to the Central Repository in an existing industry standard protocol of their choice such as the Financial Information eXchange (“FIX”) protocol. In Approach 2, CAT Reporters would submit data to the Central Repository in single mandatory specified format, such as an augmented version of the OATS protocol. Under Approach 1, the data must be converted into a uniform format at the Central Repository in a second step. Under Approach 2, the data is already in a uniform format at the time of submission. The Plan defers the decision regarding which approach to take until the selection of a Plan Processor and the development of Technical Specifications.

The Commission preliminarily believes that Approach 1 would likely result in a lower Error Rate than Approach 2. Under Approach 1, the CAT Reporters would presumably be submitting the actual data captured in real time without having to translate it into another format. In addition, under Approach 1, the conversion would be performed at the Central Repository by the Plan Processor, rather than the conversion being performed by each of the approximately 1,800 individual CAT Reporters or their vendors, which should reduce potential points where errors in formatting could be introduced, and provide for economies of scale. This would likely result in increased efficiency and accuracy due to specialization by the Plan Processor. However, while the Commission preliminarily believes that Approach 1 is likely to result in greater data accuracy than Approach 2, because of uncertainties regarding expected Error Rates and error rates in current data, the Commission is unable to evaluate the degree to which that approach would improve data accuracy relative to currently available data.

The Commission understands that a large proportion of reports that fail OATS validation checks do so because of errors in the translation of the data by the OATS reporter.

The Plan Processor is required to have policies and procedures, including standards, to ensure the accuracy of the consolidation by the Plan Processor of the data, per Rule 613(e)(4)(iii), which could mitigate errors as well. 17 CFR 242.613(e)(4)(iii).

Uniquely complex situations also pose a difficulty for assessing the ability of the Plan Processor to build a complete and accurate database of linked data that regulators could query for regulatory purposes. First, the Plan requires the Plan Processor, in consultation with industry, to develop a linking mechanism that would allow the option and equity legs of multi-leg trades to be linked within the Central Repository. Because the mechanism for this linkage is not yet determined, the Commission cannot assess the degree of the expected linkage error rate but, given that equities are not linked to options in current data sources, the Commission expects this feature to significantly improve the accuracy of linking equities to options.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.1(b).

Second, the Commission in the Proposing Release noted concern about the ability of the daisy chain approach to link a Customer order and a member's order from which the Customer is provided with an allocation. The Plan addresses this concern in the definition of an Allocation Report, which is a report that identifies accounts and subaccounts to which executed shares are allocated, but that is not required to be tied to a particular order or execution. The Report is required to be submitted to the Central Repository, but the lack of linkages in this case could make the resulting data less useful. Specifically, the content of the Allocation Report and the order lifecycles must contain content that permits regulators to draw certain conclusions about subaccount allocations even without a clean linkage.

See Proposing Release, supra note 9, at 32576.

See CAT NMS Plan, supra note 3, at Section 1.1.

See id. at Section 6.4(d)(ii).

While uncertainty about this issue remains, the Commission notes that the Plan's requirement for standardized Allocation Reports that consistently and uniquely identify Customers and reporters should improve the linkability of allocation information compared to current data, despite the limitation of direct linkage to order lifecycles, particularly in scenarios where potentially violative conduct is carried out by market participants operating through multiple broker dealers. This moderate improvement in the linkability of allocation data should improve regulators' ability to identify market participants who commit violations related to improper subaccount allocations.

(4) Customer and Reporter Identifiers

The Commission preliminarily believes that the inclusion of unique Customer and Reporter Identifiers described in the CAT NMS Plan would increase the accuracy of customer and broker-dealer information in data regulators use and provide benefits to a broad range of regulatory activities that involve audit trail data.

Currently, only a few data sources, which typically cover only a small portion of order lifecycles, include information regarding customers. Further, the customer information in these data sources is often incomplete and inconsistent and the data is currently only obtainable by regulators making requests to broker-dealers directly. Additionally, although broker-dealer identifiers, in the form of MPID numbers, CRD numbers, and clearing broker numbers, appear within the current sources of audit trail data, because of the lack of a centralized database and because these identifiers may vary across exchanges, the Commission faces challenges in relying on these identifiers to accurately identify broker-dealer activity across the market.

See Section IV.D.2.b(1)A, supra. As discussed above, the Commission notes that SRO audit trails typically do not provide customer information but a recent FINRA rule change would require its members to report to OATS non-FINRA member customers who are broker-dealers. See supra note 407.

See Section IV.D.2.b(1)D, supra.

Rule 613 requires the use of a unique Customer-ID that identifies the Customer involved in CAT Reportable Events. Based on a concern that requiring CAT Reporters to report a Customer-ID to the Central Repository with each order would disrupt existing business practices and that reporting on that basis could risk the leakage of order and Customer information into the market, the Plan requires the Plan Processor to translate a unique Customer identifier assigned by the firm to its Customer (the Firm Designated ID) into the Customer-ID to be used in CAT. Specifically, the Plan requires CAT Reporters to provide a Firm Designated ID for each Customer, which is defined as the unique identifier designated by the broker-dealer for each trading account for purposes of providing data to the Central Repository. Upon receipt of the Firm Designated ID, the Plan Processor would be required to generate and associate one or more Customer-IDs for orders received by the Customer of the CAT Reporter, which would also be linked to the relevant Reportable Events for that Customer's order. Pursuant to the Plan, therefore, the Customer-ID would be generated from the Firm Designated ID, and the Plan Processor would create a unique Customer-ID that would be consistent across that Customer's activity regardless of the originating broker-dealer.

Rule 613(c)(7) specifies the event records that would contain the Customer-ID. 17 CFR 242.613(c)(7). Event records that do not explicitly capture the Customer-ID could be linked to a record that does contain this information, typically using the Order-ID.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.1.(a)(iii).

Id. The Firm Designated ID could be anything, provided that it is unique across the firm for a given business date.

See id. at Section 6.3(d)(i)(A), n.2; see also id. at Section 1.1.

See CAT NMS Plan, supra note 3, at Appendix D, Section 3.

To facilitate the creation of Customer-IDs, certain information would be submitted to the Central Repository. Specifically, broker-dealers would be required to submit an initial set of information identifying a Customer to the Central Repository, including the Firm Designated ID and the other biographical information associated therewith including, for an individual, name, address, date of birth, ITIN/SSN, and individual's role in the account (e.g., primary holder, joint holder, guardian, trustee, person with power of attorney). With respect to legal entities, identifying information would include: name, address, EIN/LEI or other comparable common entity identifier. Broker-dealers must also submit to the Central Repository daily updates for reactivated accounts, newly-established or revised Firm Designated IDs, or other associated reportable Customer information. The Plan also calls for periodic refreshes of all Customer information from CAT Reporters. And the Plan Processor must have a way to periodically receive full account lists (i.e., not just the daily changes) to ensure the completeness and accuracy of the database.

See id. at Appendix C, Section A.1.(a)(iii); see also id. at Appendix D, Section 9.1. The CAT NMS Plan further provides, in the definition of Customer Identifying Information, that where the LEI or other comparable common identifier is provided, information covered by such common entity identifier (e.g., name, address) would not need to be separately submitted to the Central Repository. Id. at Section 1.1.

See id. at Appendix C, Section A.1.(a)(iii).

See id. at Appendix C, n.33 and Appendix D, Section 9.1.

See id. at Appendix D, Section 9.1.

Based on this information, the Plan Processor has to “maintain information of sufficient detail to uniquely and consistently identify each Customer across all CAT Reporters, and associated accounts from each CAT Reporter.” It is the Plan Processor's responsibility to document and publish, with the approval of the Operating Committee, the minimum list of data elements needed to maintain this association. Appendix D sets forth a list of minimum data elements needed to identify each Customer across all CAT Reporters, and associated accounts within a CAT Reporter, including SSN or ITIN, date of birth, current name, current address, previous name and address; and for legal entities, the LEI (if available), tax identifier, full legal name, and address. The Plan Processor must also support account structures that have multiple account owners and associated Customer information (e.g., joint accounts, managed accounts), and must be able to link accounts that move from one CAT Reporter to another, so it is possible that additional data fields would be necessary. Once a database is established, it must be maintained over time, and provide ready access to regulators to historical changes to that information.

See id. at Appendix C, Section A.1.(a)(iii).

See id. at Appendix D, Section 9.1.

See id.

See id. at Article VI, Section 6.5(b) and (c).

The Commission preliminarily believes that approval of the Plan would likely further remedy some of the inconsistencies and other limitations mentioned above. The Plan also contains provisions related to the accuracy of submitted Customer information. For example, a robust data validation process must be established for submitted Customer and Customer Account Information. There must also be a robust error resolution process for Customer information. The Central Repository must be able to accommodate minor data discrepancies (e.g., Road versus Rd in an address) on its own, while more substantial discrepancies (e.g., two different persons with the same SSN) would need to be transmitted to the CAT Reporter for resolution within the established error correction timeframe. While these elements should help increase the accuracy of Customer identification within CAT, there are some uncertainties, as the precise methods for submitting Customer data to the Central Repository, along with validations, are to be set out in Technical Specifications in the future.

See id. at Appendix C, Section A.1.(a)(iii); see also id. at Appendix D, Section 9.1.

See id. at Appendix D, Section 3.

See id. at Appendix C, Section A.1.(a)(iii).

In addition to Customer-IDs, the CAT NMS Plan calls for the use of CAT-Reporter-IDs. The data to be reported to the Central Repository includes the SRO-assigned Market Participant Identifier (MPID) of the Industry Member or Participant receiving, routing, or executing the order. Upon receipt of the data, the Plan Processor must map the SRO-assigned MPID to a CAT-Reporter-ID, which would be assigned by the Plan Processor in the CAT data. Specifically, the Plan Processor must be able to assign a CAT-Reporter-ID to all reports submitted to the Central Repository based on SRO-assigned MPIDs. To the extent that the different Participants assign the same MPID to different CAT Reporters, the Plan Processor must be able to properly associate the correct SRO-assigned MPIDs with the CAT Reporters. To do this, the Plan Processor must develop and maintain a mechanism for assigning CAT-Reporter-IDs based on the relevant SRO-assigned identifier (MPID, ETPID, or trading mnemonic) currently used by CAT Reporters in their order handling and trading processes, and also to change those identifiers should that be necessary (e.g., in the event of a merger), although changes are expected to be infrequent. Moreover, the SROs would have an obligation to provide all their SRO-assigned MPIDs to the Central Repository on a daily basis to ensure the accuracy of the information used to assign the CAT-Reporter-ID. The Plan Processor must capture, store, and maintain this information in a master/reference database, similar to how the Plan Processor would handle symbology changes. Finally, the validity of the SRO-assigned MPID is part of the initial file validation process upon receipt of a submission from a CAT Reporter, which should facilitate the accuracy of the Plan Processor's subsequent assignment of the CAT-Reporter-ID.

See Exemption Order, supra note 18, at 11863-11865; CAT NMS Plan, supra note 3, at Sections 6.3(d), 6.4(d).

See CAT NMS Plan, supra note 3, at Appendix D, Section 3.

See id.

See id. at Appendix D, Section 10.1.

See id. at Appendix D, Section 2 and Section IV.E.3.b, infra.

See id. at Appendix D, Section 7.2.

The Commission preliminarily believes that the Customer-ID approach in the CAT NMS Plan would significantly improve the accuracy of customer information available to regulators. As noted above, existing data does not consistently capture information about the customers involved in a trade or other market event, which negatively affects the ability of regulators to accurately track customers' activities across broker-dealers. Additionally, customer identities in many existing data sources use inconsistent definitions and mappings across market centers. Accordingly, it is difficult for regulators to identify the trading of a single customer across multiple market participants. The Customer-ID approach specified in the CAT NMS Plan constitutes a significant improvement because it would consistently identify the Customer responsible for market activity, obviating the need for regulators to collect and reconcile Customer identification information from multiple broker-dealers. This should reduce the risk of the introduction of errors into the data by regulators and save a significant amount of time.

See Adopting Release, supra note 9, at 45730; see also Section III.D.2.b(2)D, supra.

Furthermore, the Commission preliminarily believes that the Reporter ID approach specified in the CAT NMS Plan would improve the accuracy of tracking information regarding entities with reporting obligations, namely broker-dealers and SROs. Because the Commission currently face challenges in using MPIDs and CRD numbers, for example, to identify broker-dealers across the market, the Plan's requirement for consistent unique Reporter IDs would eliminate the need for the Commission to reconcile broker-dealer information from multiple data sources, which can be a costly task for regulatory Staff that is often limited in terms of accuracy by the inconsistencies and non-uniqueness of current identifiers, and facilitate more efficient and effective regulatory activities that protect investors from harm. Moreover, because CAT Data would include more Industry Members in the Reporter ID category than are currently in any current set of broker-dealer identifiers, the Commission preliminarily believes that approval of the Plan would likely further remedy some of the inconsistencies and other limitations mentioned above.

(5) Aggregation

Most CAT Data would be disaggregated data, meaning that CAT Data would not suffer from the limitations that characterize some of the aggregated data sources that regulators must currently use. As mentioned in the Baseline Section, subaccount allocation data and issuer repurchase data exist in forms that are aggregated and thus these data sources are limited for use in certain regulatory activities and interests. In particular, neither data type may necessarily indicate the individual executions. This data feature should promote more effective and efficient investigation by regulators of subaccount allocation issues and repurchase activity.

See Section IV.D.2.b.(2)E, supra. Item 703 of Regulation S-K requires issuers to report aggregated issuer repurchase data to the Commission on an annual and quarterly basis in Forms 10-K and 10-Q; see also 17 CFR 229.703 and supra note 451.

To meet the requirements of Rule 613, the CAT NMS Plan includes a required allocation reporting tool that would provide information on executions that are allocated to multiple subaccounts. The Allocation Reports required by the Plan would provide the Firm Designated ID for any account(s), including subaccount(s) to which executed shares are allocated, the security that has been allocated, the identifier of the firm reporting the allocation, the price per share of shares allocated, the side of shares allocated, the number of shares allocated to each account, and the time of the allocation. The Firm Designated IDs could facilitate linking back to the Customer-ID, so it may not be possible to perfectly link a Customer's aggregated orders, executions, and allocations for a day.

See CAT NMS Plan, supra note 3, at Section 6.4(d)(ii)(A)(1).

See Exemption Order, supra note 18, at 11867.

The Commission notes, however, that there may be allocations made by non-broker-dealers that are difficult to track if they involve multiple broker-dealers, or are not tracked if they involve non-CAT-reporters. See Exemptive Request Letter, supra note 16, at 26 n.61.

The Commission preliminarily believes that the CAT NMS Plan would improve the accuracy of allocation data compared to existing data available to regulators. It would provide disaggregated information on the identity of the security, the number of shares and price allocated to each subaccount, when the allocation took place, and how each Customer subaccount is associated with the master account. This would more accurately reflect which Customer ultimately received the shares that were purchased in a particular trade.

The Commission anticipates that regulators may use CAT Data for some purposes that they use cleared data for now because CAT is significantly less aggregated. As discussed above, regulators often used equity and option cleared reports to identify market participants involved in trading activity relevant to an investigation. Because these are aggregated, regulators can use them to identify clearing firms that may have higher volume in a particular stock on a particular day, but the data does not identify actual trades, and, therefore, regulators make data requests to access the underlying disaggregated data necessary to identify broker-dealers or customers that may be involved in the activity under investigation. If the CAT NMS Plan is approved, CAT Data could be used to identify individual trades and customers or other market participants who were involved in such activity with less delay and without requiring ad hoc data requests to clearing firms identified using equity or option cleared reports.

See Section IV.D.2.a(2), supra.

Likewise, the disaggregated issuer repurchase information that would be in the CAT data would be an improvement in the accuracy of information available to regulators about those issuer repurchases. In particular, the Plan would require that the Plan Processor link Customer information to the order lifecycle and the report would identify as Customers those issuers that are repurchasing their stock in the open market. This would provide much more granular data than what is available currently for open market issuer repurchases, which consists of monthly aggregations of those issuer repurchases.

See CAT NMS Plan, supra note 3, at Section 6.4(d)(iv).

See Section IV.D.2.b(2)E, supra for baseline information on current issuer repurchase data.

c. Accessibility

In general, the Commission believes that the Plan, if approved, would substantially improve the accessibility of regulatory data by providing regulators with direct access to the consolidated CAT Data, including some data elements that currently take weeks or months to obtain. However, there is some uncertainty regarding the process for regulatory access under the Plan, which creates uncertainty as to the degree of the expected improvement.

Accessibility refers to how the data is stored, how practical it is to assemble, aggregate, and process the data, and whether all appropriate regulators could acquire the data they need.

(1) Direct Access to Data

As discussed in the Baseline Section, one of the significant limitations of current regulatory data sources is lack of direct access. Rule 613(e)(1) requires the Central Repository to store and make available to regulators data in a uniform electronic format and in a form in which all events pertaining to the same originating order are linked together in a manner that ensures timely and accurate retrieval of the information for all Reportable Events for that order. Additionally, Rule 613(a)(1)(ii) requires that the CAT NMS Plan discuss the time and method of access by which the data would be made available to regulators. The CAT NMS Plan implements this requirement in Section 6.5(c) and further describes the direct access methods and functionality in the discussion of Consideration 2 and in Appendix D. Section 6.5(c) requires that the Participants and the Commission have access to the Central Repository, and access to and use of the CAT Data stored at the Central Repository, and further requires a method of access to the data that provides for the ability to run searches and generate reports, including complex queries. Specifically, the Central Repository must store 6 years of CAT data in a “convenient and usable standard electronic format” that is “directly available and searchable electronically without any manual intervention by the Plan Processor.” This access to the Central Repository is solely for the purpose of performing regulatory functions and must include the ability to run searches and generate reports; further, the Plan requires that the Central Repository shall allow the ability to return results of queries that are complex in nature, including market reconstructions and the status of order books at varying time intervals. The Central Repository must also maintain valid Customer and Customer Account Information and permit regulators access to “easily obtain historical changes to that information (e.g., name changes, address changes).”

See Section IV.D.2.b(3), supra.

17 CFR 242.613(a)(1)(ii).

See CAT NMS Plan, supra note 3, at Section 6.5(c).

See id. at Appendix C, Section A.2(b) and (c), Appendix D, Section 8.

See id. at Section 6.5(b)(i).

See id. at Section 6.5(c)(ii), Appendix D, Section 8.1.

See id. at Appendix C, Section A.1(a)(iii).

The Commission recognizes that improving accessibility relative to the Baseline requires ensuring that enough SRO and Commission Staff members are able to use the direct access system supplied by the Central Repository when they need it. The ability to use the direct access system depends, among other things, on how user-friendly the system is, whether it has enough capacity for the expected use of the system, and whether it contains the functionality that the SROs and Commission Staff require. The Commission preliminarily believes that the minimum requirements for the direct access system would ensure that the Plan would improve on the Baseline of access to current data, including the process of requesting data.

Appendix D provides minimum functional and technical requirements that must be met by the Technical Specifications to facilitate these methods of access, including the methods of selecting data that must be supported, query and bulk extract performance standards, and formats in which data could be retrieved. Specifically, CAT must be able to support a minimum of 3,000 regulatory users within the system, 600 of which might be accessing the system concurrently (which must be possible without an unacceptable decline in system performance) : 20% of the 3,000 users would be daily or weekly users, and 10% would require advanced regulatory-user access. Advanced user access includes the ability to run complex queries (versus basic users who may only run basic queries).

See id. at Appendix D, Section 8; see also Appendix C, Section A.2.

See id. at Appendix D, Section 8.1.

Id.

See id. at Appendix D, Section 8.1.1. Both Basic and Advanced Users may be established by an employee at the regulator designated to set up access to the system, if the Plan Processor chooses to do so versus processing it themselves. See id. at Appendix C, Section D.12(k). However, providing access to PII must always be done directly by the Plan Processor. Id.

Two types of query interfacing must be supported. The first, an online targeted query tool, must include a date or time range, or both, and allow users to choose from a broad menu of 26 pre-defined selection criteria (e.g., data type, listing market, size, price, CAT-Reporter-ID, Customer-ID, or CAT-Order-ID), with more to be defined at a later date. Results must be viewable in the tool or downloadable in a variety of formats and support at least a result size of 5,000 or 10,000 records, respectively, with a maximum result size to be determined by the Plan Processor. The other method for regulator access to the data is a user-defined direct query or bulk extraction. CAT must be able to support at least 3,000 daily queries, including 1,800 concurrently, and up to 300 simultaneous query requests with no performance degradation. Datasets generated by these direct queries could run from less than 1 GB to at least 10 TB or more of uncompressed data.

See id. at Appendix D, Section 8.1.1. This is a broad range of criteria from which to choose, although deferring additional selection fields to be defined at a later date makes the precise scope of this tool less certain.

See id.

See id. at Appendix D, Section 8.2.

See id. at Appendix D, Section 8.2.1.

See id.

The actual method of query support is to be determined by the Plan Processor, but must provide an open API that allows use of regulator-supplied common analytic tools (e.g., Python, Tableau) and ODBC/JDBC drivers. The Plan Processor is permitted to define a “limited set of basic required fields (e.g., date and at least one other field such as symbol, CAT-Reporter-ID, or CAT-Customer-ID)” that must be used by regulators in direct queries. Direct queries must be able to be created, saved, and run by regulators (either directly or at a prescheduled time), with automated delivery of scheduled query results. Finally, the Plan Processor must provide data models and data dictionaries for all processed and unlinked CAT Data, and the Plan Processor must provide procedures and training to regulators that would use the direct query feature (although it is up to the Plan Processor whether to require these training sessions). Consideration was given to requiring the Plan Processor to create an online Report Center that would provide pre-canned reports (i.e., recurring reports of interest to regulators), but due to the added complexity and lack of quantifiable use cases, the decision was made not to proceed. The Plan, however, provides that this decision would be reassessed when broker-dealers begin submitting data to the CAT.

See id. at Appendix D, Section 8.2. A discussion of the types of data tools that bidders proposed to support can be found in Appendix C, Section A.2(b).

See id. at Appendix D, Section 8.2.

See id. at Appendix D, Section 8.2.1.

See id. at Appendix D, Section 8.2.

See id. at Appendix D, Section 8.2.2.

All queries must be able to be run against raw (i.e., unlinked) or processed data, or both. A variety of minimum performance metrics apply to these queries. The Plan Processor must also provide certain support to regulatory users. Specifically, it must “develop a program to provide technical, operational and business support” to regulators, including creating and maintaining the CAT Help Desk to provide technical expertise to assist regulators with questions and/or functionality about the content and structure of the CAT query capability. The Help Desk must be available 24x7, support email and phone communication, and be staffed to handle 2,500 calls per month (although this resource would not be exclusive to regulators; CAT Reporters could use it as well). The Plan Processor must also develop tools, including an interface, to let users monitor the status of their queries and/or reports, including all in-progress queries/reports and estimated time to completion. In addition, the Plan Processor must develop communication protocols regarding system status, outages, and other issues affecting access, including access by regulators to a secure Web site to monitor CAT System status. Furthermore, the Plan Processor must develop and maintain documentation and other materials to train regulators, including training on building and running queries.

See Section IV.E.1.IV.E.1.d(3), infra, for additional for additional information.

See CAT NMS Plan, supra note 3, at Appendix D, Section 10.2.

See id. at Appendix D, Section 10.3.

See id. at Appendix D, Section 10.2.

See id.

See id.

The Commission preliminary believes that the direct access facilitated by provisions of the CAT NMS Plan described above is reasonably designed to substantially reduce the number of ad hoc data requests and provide access to substantial data without the delays and costly time and knowledge investments associated with the need to create and respond to data requests. For example, regulators do not have direct access to EBS or trade blotter data and therefore they must request such data when needed for regulatory tasks. As a result, in 2014 the Commission made 3,722 EBS requests that generated 194,696 letters to broker-dealers for EBS data. Likewise, the Commission understands that FINRA requests generate about half this number of letters. In addition, for examinations of investment advisers and investment companies, the Commission makes approximately 1,200 data requests per year. If the Plan is approved, the Commission preliminarily believes that the number of data requests would decline sharply. In addition to decreasing the amount of time currently required for regulators to access data sources, direct access to the CAT Data should decrease the costs that many regulators and market participants incur in either requesting data or fulfilling requests for data, such as the time and resources that regulators and data liaisons or back office IT staff at broker-dealers expend to understand and access broker-dealer data collected and provided in a particular way.

See Section IV.D.2.b(2), supra, for discussion of ad hoc data requests.

The Plan would also permit regulators to directly access customer information, which could improve the ability of SROs to conduct surveillance. Rule 613(e)(3) requires that the CAT provide the capability to run searches and generate reports. The CAT NMS Plan indicates that regulators would be able to run searches on many variables, including Customer-IDs. Appendix D further clarifies that both the online targeted query tool and the user-defined query/bulk extract process would produce records that provide Customer-IDs, but that do not themselves provide Customer PII data. Data containing PII, however, could be obtained by regulatory personnel specifically authorized to obtain PII access, through a process to be documented by the Plan Processor. Currently, most regulatory data sources do not directly link to specific customers. Instead, regulators can use an ad-hoc data request to identify the customer and follow up with an EBS request to identify the customer's other activity across market participants. In this regard, CAT would provide SROs with direct access to the data that is necessary to conduct surveillance of the trading behavior of individual market participants in a more timely fashion.

See CAT NMS Plan, supra note 3, at Appendix D, Section 8.2; See also supra note 632.

See id. at Appendix D, Section 4.1.6, Appendix D, Section 8.1.1-8.1.3.

See id. at Appendix D, Section 4.1.6.

The EBS system, trade blotters, order tickets, and trade confirmations are the existing data sources that contain customer information. See Section IV.D.2.b(1)A, supra; Adopting Release, supra note 9, at 45727. Also a recent FINRA rule change would require FINRA members to report to OATS non-FINRA member customers who are broker-dealers. See supra note 407.

Currently, FINRA receives exchange data from SROs at the end of the trading day. It takes approximately three days for FINRA to process and translate this data to a common format before surveillance programs can run. As noted in Section IV.D.1.c, this economic analysis considers surveillance to be SROs running automated processes on routinely collected or in-house data to identify potential violations of rules or regulations.

(2) Consolidation of Data

The Commission also preliminarily believes that, if approved, the Plan would improve accessibility by consolidating various data elements into one combined source, reducing data fragmentation. First, Rule 613 requires that the Central Repository collect data that includes the trading and routing of a given security from all CAT Reporters. Currently, audit trail data for securities that are traded on multiple venues (multiple exchanges or off-exchange venues) is fragmented across multiple data sources, with each regulator generally having direct access only to data generated on the trading venues it regulates. If approved, the Plan would bring audit trail data related to trading on all venues into the Central Repository where it could be accessed by all regulators. Second, Rule 613 requires that the Plan include both equity and options data. Currently no existing regulatory audit trail data source includes both options and equities data, so collecting this data and providing access would allow regulators to monitor and run surveillance on the activity of market participants in related instruments, such as when a market participant has activity in both options and the options' underlying assets.

The Commission recognizes that FINRA collects data from exchanges for which it provides regulatory services. However, this data is sent to FINRA by the exchanges with a delay, and the data formats are not standardized prior to receipt at FINRA.

See 17 CFR 242.613(c)(5), (c)(6).

The Plan would also marginally increase the accessibility of historical exchange data. In particular, Section 6.5(b)(i) of the Plan requires that the Central Repository make historical data available for not less than six years, in a manner that is directly accessible and searchable electronically without manual intervention by the Plan Processor.

See CAT NMS Plan, supra note 3, at Section 6.5(b)(i). Currently, broker-dealers retain data for six years, but exchanges are only required to retain data for five years. In practice, the Commission understands that most exchanges generally retain data for at least six years, but at least one exchange does not retain data for six or more years. Therefore, the CAT NMS Plan would improve the historical data available from at least one exchange.

In some dimensions of accessibility, the Commission notes that uncertainties exist that could affect the degree of expected improvement to accessibility. In particular, while the Plan provides detail on the method of access and the types of queries that regulators could run, many of the decisions regarding access have been deferred until after the Plan Processor is selected and finalizes the Technical Specifications; the Plan does not specify how regulators would access the data beyond providing for both an online query tool and user-defined direct queries that could do bulk extractions. For example, while the Plan indicates that regulators would have an on-line targeted query tool and a tool for user-defined direct queries or bulk extraction, the Plan itself does not provide an indication for how user-friendly the tools would be or the particular skill set needed to use the tools for user-defined direct queries.

See, e.g., CAT NMS Plan, supra note 3, at Appendix D, Section 8.2.

See CAT NMS Plan, supra note 3 at Appendix D, Sections 8.1.1, 8.1.2.

In addition, it is not known whether the Plan Processor would host a server workspace that regulators could use for more complex analyses, what software tools would be available to regulators within such a workspace, and whether complex analyses would be able to be performed without extracting significant data from the Central Repository's database.

While all Bidders included certain baseline functionality, such as some means for regulators to perform dynamic searches, data extraction, and “off-line analysis,” Bidders proposed using a variety of tools to provide regulators with access to and reports from the Central Repository, including direct access portals, web-based applications, and a number of different options for formatting the data provided to regulators in response to their queries. While all of these proposed solutions would presumably be compatible with achieving the accessibility benefits sought to be achieved through the Plan—i.e., they would all involve the aggregation of data from various sources and the provision of ready access to that data for regulators—the precise degree of functionality of the final system is still to be determined. Similarly, the details of system performance would depend on Service Level Agreements to be established between the Plan Participants and the eventual Plan Processor, which means that the details would not be known until after the Plan Processor is selected. These functionality and performance uncertainties create some uncertainty regarding the degree of improvement in regulatory access that would result from the Plan.

See id. at Appendix C, Section A.2(b). “Offline-analysis” refers to a regulator's analysis of data extracted from the Central Repository using the regulator's own analytical tools, software, and hardware to perform the analysis. See id. at Appendix C, Section A.2(b) n.77.

See id. at Appendix C, Section A.2(b).

See id. at Appendix D, Section 8.5.

Nonetheless, the requirements included in the Plan describe a system that, once implemented, would result in the ability to query consolidated data sources that represents a significant improvement over the currently available systems. This substantial reduction in data delays and costly data investments would permit regulators to complete market reconstructions, analyses, and research projects, as well as investigations and examinations, more effectively and efficiently and would lead to improved productivity in the array of regulatory matters that rely on data, which should lead to improved investor protection.

d. Timeliness

The Commission believes that, if approved, the CAT NMS Plan would significantly improve the timeliness of the reporting, compiling, and access of regulatory data, which would benefit a wide array of regulatory activities that use or could use audit trail data. The Commission preliminarily believes that the timeline for compiling and reporting data pursuant to the Plan constitutes an improvement over the processes currently in place for many existing data sources, and relative to some data sources the improvement is dramatic. Specifically, under the Plan, CAT Data would be compiled and made ready for access faster than is the case today for some data, both in raw and in corrected form; regulators would be able to query and manipulate the CAT Data without going through a lengthy data request process; and the data would be in a format to make it more immediately useful for regulatory purposes.

Timeliness refers to when the data is available to regulators and how long it would take to process before it could be used for regulatory analysis.

(1) Timing of Initial Access to Data

The Plan would require CAT Reporters to report data to the Central Repository at times that are on par with current audit trails that require reporting, but the Central Repository would compile the data for initial access sooner than some other such data. Sections 6.3(b)(ii) and 6.4(b)(ii) of the Plan require that the data required to be collected by CAT Reporters must be reported to the Central Repository by 8:00 a.m. Eastern Time on day T+1. These provisions also make clear that CAT Reporters could voluntarily report the required data prior to the deadline. As described in Table 4, the time at which data is reported often differs significantly from the time at which data is made available to various regulators. The CAT Data would be made available to regulators in raw form after it is received from reporters and passes basic formatting validations; the Plan does not specify exactly when these validations would be complete, but the requirement to link records by 12:00 p.m. (noon) Eastern Time on day T+1 gives a practical upper bound on this timeline for initial access to the data. Thus, to the extent that access to the raw (i.e., uncorrected and unlinked) data would be useful for regulatory purposes, the CAT NMS Plan provides a way for SROs and the Commission to access the uncorrected and unlinked data on day T+1 by 12:00 p.m. at the latest.

Compiling data refers to a process that aggregates individual data records into a data set. This could occur when regulators request data and when the regulators receive data from multiple providers. This is different from the act of reporting data.

See Rules 613(c)(3), (c)(4), 17 CFR 242.613(c)(3), (c)(4).

See CAT NMS Plan, supra note 3, at Appendix D, Section 3.1.

See Table 4, supra.

See CAT NMS Plan supra note 3, at Appendix C, Section A.2(a); Appendix C, Section A.3(e); Appendix D, Section 6.1.

As noted in the Baseline, some current data sources compile and report the data with delays. For example, equity and option clearing data are not compiled and reported to the NSCC and OCC until day T+3, and thus access to this data by the Commission cannot occur until day T+3 at the very soonest. Under the Plan, raw data would be available two days sooner to all regulators. In other cases such as EBS reports, the data are not compiled and reported to a centralized database until a request is received. OATS data is initially reported to FINRA by 8 a.m. on the calendar day following the reportable event, and it takes approximately 24 hours for FINRA to run validation checks on the file. However, SROs do not currently access OATS information for regulatory purposes until after the error correction process is complete, which imposes a further delay of several business days for non-FINRA SRO regulators' use. Uncorrected OATS data is, however, available at 8 a.m. on the calendar day following the reportable event to FINRA (several hours more timely than CAT Data would be)—and is available to other regulators upon request several weeks later. Uncorrected CAT Data would be available to all regulators at 12:00 p.m. on day T+1, which is at least several days sooner than OATS is available to non-FINRA regulators; however, the Commission notes that because OATS is reportable on the calendar day following the OATS-reportable event while CAT would be reported on T+1 following a Reportable Event, regulators' access to CAT Data from a day preceding a non-trading day (Fridays or days before market holidays) is likely to be less timely than it is currently, if that data would be covered by OATS. However, to the extent that the CAT would generally make CAT Data, which would include substantially more information than OATS data, available to all regulators, as opposed to just FINRA, in raw form by at least 12:00 p.m. Eastern Time on day T+1, the CAT would generally represent a significant improvement in timeliness for SROs other than FINRA compared to OATS.

The Commission notes, however, that broker-dealers could compile some data sources discussed in the baseline on the day of an event. For example, broker-dealers can compile trade blotters on the same day as the trade. Further, regulators can compile data received in real-time on the event day. For example, regulators can compile direct data feeds same day. The Commission does not believe the CAT NMS Plan would affect the timing of the compilation of such data, nor would it reduce the number of requests for data on the day of an event.

See Adopting Release, supra note 9, at 45729.

Id.

See OATS Reporting Technical Specifications Section 8.1, available at https://www.finra.org/sites/default/files/OATSTechSpec_01112016.pdf .

It is true that the Plan would not necessarily improve the timeliness of audit trail data in every case or for every regulator, as some kinds of audit trail data are currently timely for some regulators. For example, exchange SROs already have real-time access to their own audit trail data. However, regulators at other SROs or the Commission do not have real-time access to that exchange's audit trail, and therefore CAT Data could be more timely for these other regulators to access and use than obtaining that exchange's audit trail data through any means.

Under the Plan, SROs that are exchanges would still have the same real-time access to their own audit trail data as they currently do. The Commission does not expect that all SRO audit trails will be retired on implementation of the Plan because exchanges may use such audit trails to implement their CAT reporting responsibilities. CAT reporting requirements would require that exchanges collect and report audit trail information from their systems even if they elect to replace their current audit trails. However, CAT requirements may improve the completeness of real-time exchange audit trail data if the information that exchanges collect under the Plan is more complete than what they currently collect.

As noted, the SROs are generally currently able to access their own audit trail data on the same day of an event and the Commission is currently able to access some public data, like SIP and MIDAS, on the same day as an event. Further, OATS is available to FINRA at 8am on the day following an event. The Commission preliminarily does not expect the CAT NMS Plan would affect these regulators' access to most of these respective data sources.

(2) Timeliness of Access to Error-Corrected Data

Further, the Commission preliminarily believes that the error correction process required by the CAT NMS Plan is reasonably designed to provide additional improvements in timeliness for corrected data. The CAT NMS Plan specifies that the initial data validation and communication of errors to CAT Reporters must occur by noon on day T+1, corrections of these errors must be submitted by the CAT Reporters to the Central Repository by 8:00 a.m. Eastern Time on day T+3, and the corrected data made available to regulators by 8:00 a.m. Eastern Time on day T+5. During this interim time period between initial processing and corrected data availability, “all iterations” of processed data must be available for regulatory use. The Central Repository must be able to receive error corrections at any time, even if late; if corrections are received after day T+5, the Plan Processor must notify the SEC and SROs of this fact and how re-processing of the data (to be determined in conjunction with the Operating Committee) would be completed. Customer information (i.e., information containing PII) is processed along a slightly different timeline, but the outcome—corrected data available by 8:00 a.m. Eastern Time on day T+5—is the same. One exception to this timeline is if the Plan Processor has not received a significant portion of the data, as determined according to the Plan Processor's monitoring, in which case the Plan Processor could determine to halt processing pending submission of that data.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.2(a), Appendix D, Section 6.1.

Id. at Appendix D, Section 6.2.

See id. at Appendix C, Section A.3.(b), Appendix D, Section 7.4.

See id. at Appendix D, Section 6.2.

Id.

See id. at Appendix D, Section 6.1.

As discussed in the Baseline Section, the error resolution process for OATS is limited to five business days from the date a rejection becomes available. The CAT NMS Plan requires a three-day repair window for the Central Repository. Accordingly, if the Plan is approved, regulators would generally be able to access partially and fully corrected data earlier than they would for OATS.

See Section IV.D.2.b(4) and supra note 465.

Id. at Appendix C, Section A.2(a).

CAT Data being available on day T+5 may be later than for other current SRO audit trails.

(3) Timeliness of direct access

Improvements to timeliness would also result from the ability of regulators to directly access CAT Data. As noted in the Baseline Section and throughout this Section, most current data sources do not provide direct access to most regulators, and data requests can take as long as weeks or even months to process. Other data sources provide direct access with queries that can sometimes generate results in minutes—for example, running a search on all MIDAS message traffic in one day can take up to 30 minutes —but only for a limited subset of the data to be available in CAT, and generally only for a limited number of regulators. Accordingly, the Commission preliminarily believes that the ability of regulators to directly access and analyze the scope of audit trail data that would be stored in the Central Repository should reduce the delays that are currently associated with requesting and receiving data. For many purposes, therefore, CAT Data could be up to many weeks more timely than current data sources. Furthermore, direct access to CAT Data should reduce the costs of making ad hoc data requests, including extensive interactions with data liaisons and IT staff at broker-dealers, SROs, and vendors, developing specialized knowledge of varied formats, data structures, and systems, and reconciling data.

See CAT NMS Plan, supra note 3, Section 6.5(c).

See Section IV.D.2.b(4) and supra note 468.

As discussed above, Rule 613 generally requires that the Central Repository would receive, consolidate, and retain CAT Data in a linked uniform electronic format and the regulators would be able to directly access the data stored in the Central Repository. Queries take time to return data because they need to look up information across a range of data records, process that data, and compile it into an output dataset. Therefore, the improvements to timeliness depend on how long the queries take to return data. The CAT NMS Plan specifies that regulators would be able to query the Central Repository using an online targeted query tool with response times “measured in time increments of less than a minute” for targeted queries and within 24 hours for large or complex queries that either scan large amounts of data or return large result sets (i.e., sets of over 1 million records). That said, if the data request is limited to one business date, and that business date is within the last 12-month period, the query must not take more than 3 hours to run, regardless of complexity. Specifically, searches including only equities and options trade data must be returned within either 1 minute (events for a specific Customer or CAT Reporter with filterable other fields); 30 minutes (events for a specific Customer or CAT Reporter in a specified date range of less than 1 month); or 6 hours (events for a single Customer or CAT Reporter in a specified date range of up to 12 months within the last 24 months). Searches including equities and options trade data, along with NBBO data, must return within 5 minutes for all orders for a specific security from a specific Participant; and for all orders, cancellations, and NBBO (or the protected best bid and offer) for a specific security, and with several similar types of searches, within a specified window not to exceed 10 minutes for a single date.

See Section IV.E.1.c, supra.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.2(c); Appendix D, Section 8.1.2.

Id. at Appendix D, Section 8.1.2.

Id.

Id.

Furthermore, the search tool must include a resource management component, which could manage query requests to balance the workload, and categorize and prioritize query requests based on the input parameters, complexity of the query, and the volume of data to be parsed in the query, with the details on the prioritization plan to be provided at a later date. The database must support the estimated 600 concurrent users to ensure that there is not an unacceptable decline in system performance. The direct query and bulk extract features are also designed to ensure timely regulatory access to critical data. For example, the bulk extract of an entire day's worth of data should be able to be transferred in less than four hours (assuming the regulator's network could support the required data transfer speeds). The Plan Processor must have an automated mechanism to monitor user-defined direct queries and bulk data extracts, including automated alerts of issues with bottlenecks and excessively long queues for queries or data extractions. Monthly reporting on the delivery and timeliness of these tools to the Operating Committee and regulators is required.

See id. at Appendix D, Section 8.1.2.

See id. at Appendix D, Section 8.1.

See id. at Appendix D, Section 8.2.2.

Id.

Id.

(4) Timeliness of use of Data

The Commission also preliminarily expects the CAT NMS Plan to reduce the time required to process data before analysis. Currently regulators can spend days and up to months processing data they receive into a useful format. Part of this delay is due to the need to combine data across sources that could have non-uniform formats and to link data about the same event both within and across data sources. As discussed above, these kinds of linking processes can require sophisticated data techniques and substantial assumptions, and can result in imperfectly linked data. The Plan addresses this issue by stating that the Plan Processor must store the data in a linked uniform format. Specifically, the Central Repository will use a “daisy chain” approach to link and reconstruct the complete lifecycle of each Reportable Event, including all related order events from all CAT Reporters involved in that lifecycle. Therefore, regulators accessing the data in a linked uniform format would no longer need to take additional time to process the data into a uniform format or to link the data. Accordingly, the Commission preliminarily believes that the Plan would reduce or eliminate the delays associated with merging and linking order events within the same lifecycle. Further, the Plan would improve the timeliness of FINRA's access to the data it uses for much of its surveillance by several days because the corrected and linked CAT Data would be accessible on T+5 compared to FINRA's T+8 access to its corrected and linked data combining OATS with exchange audit trails.

See Section IV.D.2.b(4), supra.

See CAT NMS Plan, supra note 3, at Section 6.5(b)(i). The CAT NMS Plan does not link allocations to order events; see also 17 CFR 242.613(e)(1).

See CAT NMS Plan, supra note 3, at Appendix D, Section 3.

This does not apply if regulators choose to access raw data before the Central Repository processed them.

The expected improvements to data accuracy discussed above could also result in an increase in the timeliness of data that is ready for analysis, although uncertainty exists regarding the extent of this benefit. As noted in the Baseline, regulators currently take significant time to ensure data is accurate beyond the time that it takes data sources to validate data. In some cases, data users may engage in a lengthy iterative process involving a back and forth with the staff of a data provider in order to obtain accurate data necessary for a regulatory inquiry. Accordingly, to the extent that the Central Repository's validation process is sufficiently reliable and complete, the duration of the error resolution process regulators would perform with CAT Data may be shorter than for current data. Further, to the extent that the Central Repository's linking and reformatting processes are sufficiently successful, the SROs and Commission may not need a lengthy process to ensure the receipt of accurate data. However, as discussed above, the Commission lacks sufficient information on the validations, linking, and reformatting processes needed to draw a strong conclusion as to whether users would take less time to validate CAT Data than they take on current data. Nonetheless, the Commission preliminarily believes that the linking and reformatting processes at the Central Repository would be more accurate than the current decentralized processes such that it would reduce the time that regulators spend linking and reformatting data prior to use.

See Section IV.E.1.b, supra.

As discussed above, Rule 613 requires a validation process but leaves significant flexibility on the specific validations to be performed and the timeline for validation. The details regarding required validations do not appear in the CAT NMS Plan and instead would appear in the Technical Specifications, which would not be finalized until after approval of the CAT NMS Plan. See Section IV.E.1.b, supra.

2. Improvements to Regulatory Activities

The Commission preliminarily believes that improvements in the quality of available data have the potential to result in improvements in the analysis and reconstruction of market events; market analysis and research in support of regulatory decisions; and market surveillance, examinations, investigations, and other enforcement functions.

Regulators' abilities to perform analyses and reconstructions of market events would likely improve, allowing regulators to more quickly and thoroughly investigate these events. This would allow regulators to provide investors and other market participants with more timely and accurate explanations of market events, and to develop more effective responses to such events. The availability of the CAT Data would benefit market analysis and research in support of regulatory decisions, facilitating an improved understanding of markets and informing potential policy decisions. Regulatory initiatives that are based on an accurate understanding of underlying events and are narrowly tailored to address any market deficiency should improve market quality and benefit investors.

In the Commission's preliminary view, CAT Data would substantially improve both the efficiency and effectiveness of SRO broad market surveillance programs, which could benefit investors and market participants by allowing regulators to more quickly and precisely identify and address a higher proportion of market violations that occur, as well as prevent violative behavior through deterrence.

The Commission also preliminarily believes that CAT Data would enhance the SROs' and the Commission's abilities to effectively target risk-based examinations of market participants who are at elevated risk of violating market rules, as well as their abilities to conduct those examinations efficiently and effectively, which could also contribute to the identification and resolution of a higher proportion of violative behavior in the markets. The reduction of violative behaviors in the markets should benefit investors by providing investors with a safer environment for allocating their capital and making financial decisions. A reduction in violative behaviors could also benefit market participants whose business activities are harmed by the violative behavior of other market participants. The Commission further believes that more targeted examinations could also benefit market participants by resulting in proportionately fewer burdensome examinations of compliant market participants. A significant percentage of Commission enforcement actions involve trade and order data, and the Commission also preliminarily believes that CAT Data would significantly improve the efficiency and efficacy of enforcement investigations, including insider trading and manipulation investigations.

In 2015, the Commission filed 807 enforcement actions, including 39 related to insider trading, 43 related to market manipulation, 124 related to broker-dealers, 126 related to investment advisers/investment companies, and one related to exchange or SRO duties. In 2014, the Commission filed 755 enforcement actions, including 52 related to insider trading, 63 related to market manipulation, 166 related to broker-dealers, and 130 related to investment advisers/investment companies, many of which involved trade and order data. See Year-by-Year SEC Enforcement Statistics, available at https://www.sec.gov/news/newsroom/images/enfstats.pdf . The total number of actions filed is not necessarily the same as the number of investigations. An investigation may result in no filings, one filing, or multiple filings. Additionally, trade and order data may be utilized in enforcement investigations that do not lead to any filings. Based on these numbers, the Commission estimates that 30-50% of its enforcement actions incorporate trading or order data. A portion of FINRA's 1,397 disciplinary actions in 2014 and 1,512 in 2015 also involved trading or order data. See http://www.finra.org/newsroom/statistics .

The Commission further anticipates additional benefits associated with enhanced abilities to handle tips, complaints and referrals, and improvements in the speed with which they could be addressed, particularly in connection with the significant number of tips, complaints, and referrals that relate to manipulation, insider trading, or other trading and pricing issues. The benefits to investor protection of an improved tips, complaints, and referrals system would largely mirror the benefits to investor protection that would accrue through improved surveillance and examinations efficiency.

In fiscal years 2014 and 2015, the Commission received around 15,000 entries in its TCR system, approximately one third of which related to manipulation, insider trading, market events, or other trading and pricing issues.

a. Analysis and Reconstruction of Market Events

The Commission preliminarily believes that, if approved, the Plan would improve regulators' ability to perform analysis and reconstruction of market events. As noted in the Adopting Release, the sooner regulators can complete a market reconstruction, the sooner regulators can begin reviewing an event to determine what happened, who was affected and how, if any regulatory responses might be required to address the event, and what shape such responses should take. Furthermore, the improved ability for regulators to generate prompt and complete market reconstructions could provide improved market knowledge, which could assist regulators in conducting retrospective analysis of their rules and pilots.

See Adopting Release, supra note 9, at 45732.

The fragmented nature of current audit trail data and the lack of direct access to such data renders market reconstructions cumbersome and time-consuming. Currently, the information needed to perform these analyses is spread across multiple audit trails, with some residing in broker-dealer order systems and trade blotters. Requesting the data necessary for a reconstruction of a market event often takes weeks or months and, once received, regulators then need weeks to reconcile disparate data formats used in different data sources. For example, on the afternoon of May 6, 2010, the U.S. equity and equity futures markets experienced a sudden breakdown of orderly trading when indices, such as the Dow Jones Industrial Average Index and the S&P 500 Index, fell about 5% in five minutes, only to rebound soon after (the “Flash Crash”).

See CFTC and SEC, Findings Regarding the Market Events of May 6, 2010: Report of the Staffs of the CFTC and SEC to the Joint Advisory Committee on Emerging Regulatory Issues (September 30, 2010), available at http://www.sec.gov/news/studies/2010/marketevents-report.pdf .

The lack of readily available trade and order data resulted in delays and gaps in the Commission's analysis of the events of the Flash Crash. Ultimately, it took Commission Staff nearly five months to complete an accurate representation of the order books of the equity markets for May 6, 2010. Even then, the reconstruction only contained an estimated 90% of trade and order activity for that day.

For a further explanation of the limitations data deficiencies imposed on the Commission's investigation into the Flash Crash, see Adopting Release, supra note 9, at 45732-33.

Regulators, such as the Commission and SROs on whose exchanges events took place, faced similar challenges when reconstructing events around the May 2012 Facebook IPO, the August 2012 Knight Securities “glitch,” and the August 2013 NASDAQ SIP outage. In addition, during the financial crisis in 2008, the lack of direct access to audit trail data resulted in the Commission being unable to quickly and efficiently conduct analysis and reconstruction of market events. The state of OATS data in 2008 also limited FINRA's ability to analyze and reconstruct the market during the financial crisis because FINRA could not yet augment its OATS data with exchange data and OATS did not include market maker quotations. As a result, regulators had little information about the role of short sellers in market events and the identity of short sellers during the financial crisis, for example. Some of these shortcomings in regulatory data still apply today.

For background information on these events, see SEC Press Release, SEC Charges NASDAQ for Failures During Facebook IPO (May 29, 2013), available at http://www.sec.gov/News/PressRelease/Detail/PressRelease/1365171575032 ; In the Matter of Knight Capital Americas LLC, Securities Exchange Release Nos. 70694 (October 16, 2013); 73639 (November 19, 2014), 79 FR 72252, 72255, n.32 (December 5, 2014) (discussing NASDAQ SIP outage); see also Adopting Release, supra note 9, at 45732-33 (discussing difficulty of analyzing and reconstructing market events in absence of a consolidated audit trail).

See Short Sale Reporting Study, supra note 413. To resolve this lack of information, the Commission issued an emergency order creating a new filing requirement for 13f filers to report their short positions and short sales to the Commission weekly on Form SH. See former Rule 10a-3T; available at http://www.sec.gov/rules/other/2008/34-58591.pdf ; http://www.sec.gov/rules/other/2008/34-58591a.pdf ; http://www.sec.gov/rules/other/2008/34-58724.pdf ; http://www.sec.gov/rules/final/2008/34-58785.pdf ; http://www.sec.gov/news/press/2008/2008-209.htm ; http://www.sec.gov/divisions/marketreg/shortsaledisclosurefaq.htm . This data was kept confidential. After evaluating whether the benefits from the data justified the costs, the Commission let this requirement expire, replacing it with additional public data. See SEC Press Release, SEC Takes Steps to Curtail Abusive Short Sales and Increase Market Transparency (July 27, 2009), available at http://www.sec.gov/news/press/2009/2009-172.htm . This public data did not identify the short sellers as the Form SH data did. In addition, using data requested from SROs, the Commission conducted two studies on short selling during September 2008. These studies required data requests to select exchanges, took two months to complete and did not have information identifying short sellers. See “Analysis of a Short Sale Price Test Using Intraday Quote and Trade Data” available at http://www.sec.gov/comments/s7-08-09/s70809-368.pdf and “Analysis of Short Selling Activity during the First Weeks of September 2008” available at http://www.sec.gov/comments/s7-08-09/s70809-369.pdf .

For example, OATS still does not include all principal orders or option data. See Section IV.D.2.b(1)A, supra. Because FINRA collects some exchange data, FINRA is able to merge exchange quotes with OATS.

And although there is a proposed FINRA rule that will require FINRA members to report to OATS identification for their non-FINRA member customers who are broker-dealers, even after approval of this rule OATS will lack identification for customers who are not broker-dealers. See Section IV.D.2.b(1)B, supra.

More generally, regulators face significant difficulties in using some current data sources for a thorough market reconstruction. Some of the most detailed data sources, including sources like EBS and trade blotters that identify customers, are impractical for broad-based reconstructions of market events. In particular, including EBS data for a reconstruction of trading in the market for even one security on one day could involve many, perhaps hundreds, of requests, and would require linking that to SRO audit trail data or public data. Further, because EBS data lacks time stamps for certain trades, use of EBS data in market reconstructions requires supplementation with data from other sources, such as trade blotters.

See Section IV.D.2.b(3), supra (noting that in 2014, the SEC made 3,722 EBS requests which generated 194,696 letters to broker-dealers requesting EBS data). The Commission understands that FINRA makes about half this number of requests.

Large traders who file Form 13H with the Commission are assigned a “large trader identification number” by the Commission and must provide that number to their brokers for inclusion in the EBS records that are maintained by the clearing brokers. Rule 13h-1, subject to relief granted by the Commission, requires that execution time be captured (to the second) for certain categories of large traders. See Sections IV.D.2.a.(3) and IV.D.2.b, supra (discussing the EBS system and large trader reports and the limitations of these data sources in performing market reconstructions).

The Commission therefore expects that improvements in data completeness and accuracy from the Plan would enhance regulators' ability to perform analyses and to reach conclusions faster in the wake of a market event by reducing the time needed to collect, consolidate and link the data. The inclusion of Customer-IDs and consistent CAT-Reporter-IDs in CAT would allow regulators to more effectively and efficiently identify market participants that submit orders through several broker-dealers and execute on multiple exchanges and whose activity may warrant further analysis. This would be useful if regulators were interested in determining if a particular trader or category of traders had some role in causing the market event, or how they might have adjusted their behavior in response to the event, which could amplify the effects of the root cause or causes. Furthermore, the clock synchronization requirements of the Plan would improve the ability of regulators to sequence some events that happened in different market centers to better identify the causes of market events. Overall, the Commission preliminarily believes that, if the Plan is approved, regulators would have dramatically improved ability to identify the market participants involved in market events.

The Commission further believes that better data accessibility would significantly improve the ability of regulators to analyze and reconstruct market events. As noted above, CAT Data would improve data accessibility relative to every other data source because all SROs and the Commission would have direct access to CAT Data. If the Plan is approved, much of this information would be housed in the Central Repository with query capabilities that would allow regulators to access raw data beginning the day after an event. Further, as mentioned below in the SRO Surveillance Section, the CAT Data would link Reportable Events, which could allow regulators to respond to market events more rapidly because they would not need to process corrected and linked data before starting their analyses.

While the Commission recognizes that some data sources are currently available earlier, those data sources are so fragmented as to make collecting them for a broad-based market reconstruction infeasible.

Such benefits could be limited for market events that require linked data within five days of an event or if the linking algorithm in the Central Repository introduces data errors.

b. Market Analysis and Research

The Commission preliminarily believes that the CAT NMS Plan would benefit the quality of market analysis and research that is produced to increase regulatory knowledge and support policy decisions and would lead to a more thorough understanding of current markets and emerging issues. These expected benefits would stem from improvements in accessibility, accuracy, and completeness of regulatory data. Improvements in regulatory market analysis and research aimed at informing regulatory decisions would benefit investors and market participants by improving regulators' understanding of the intricacies of dynamic modern markets and how different market participants behave in response to policies and information. These more nuanced and more thorough insights would help regulators to identify the need for regulation that specifically tailors policies and interventions to the diverse landscape of market participants and conditions that characterize current financial markets, as well as assist them in conducting retrospective analysis of their rules and pilots.

A lack of direct access to necessary data, along with inaccuracies in the data that are available, currently limits the types of analyses that regulators can conduct. These data limitations constrain the information available to regulators when they are considering the potential effects of regulatory decisions. For example, in January 2010 the Commission published a concept release on equity market structure that discusses how the markets have rapidly evolved from trading by floor-based specialists to trading by high-speed computers. The concept release poses a number of questions about the role and impact of high-frequency trading strategies and the movement of trading volume from the public national securities exchanges to over-the-counter trading venues such as dark pools. Over the past five years there has been considerable discussion about these topics by regulators, market participants, the media, and the general public. Nevertheless, limitations in the completeness and accessibility of the available data have limited the research that followed the concept release.

See Concept Release on Equity Market Structure, supra note 733; see also Adopting Release, supra note 9, at 45733 (discussing the Concept Release on Equity Market Structure).

The Commission preliminarily believes that the CAT NMS Plan improves this situation, benefiting market analysis and research in support of SRO and Commission rulemaking. It would provide direct access to data that currently requires an often lengthy and labor-intensive effort to request, compile, and process. Additionally, the expected improvements in accuracy and completeness could benefit efforts to analyze the activities of particular categories of market participants, understand order routing behavior, identify short selling and short covering trades, issuer repurchases, and related topics. The requirement to store the data in a uniform format in the Central Repository is particularly important, as linking and normalizing data from disparate sources in different formats is a major component of completing many types of analyses and currently requires a significant amount of time. The Plan would provide direct access to data that regulators could use to more directly study issues such as high frequency trading, maker-taker pricing structures, short selling, issuer repurchases, and ETF trading.

The CAT NMS Plan could improve market analysis and research concerning HFT by providing regulators with direct access to more uniform and comprehensive data that identifies HFT activity more precisely compared to existing academic research that regulators currently utilize. Existing academic research on high frequency trading cannot precisely identify high frequency traders or their trading activity and more comprehensive regulatory analysis on high frequency trading currently relies on fragmented data that is cumbersome to collect and process. For both academics and regulators, studying high frequency traders is currently difficult because these traders typically trade across many exchanges, and often off-exchange as well. NASDAQ distributes a trade and quote dataset to researchers for the purposes of performing academic studies on high frequency trading. This dataset identifies the trading and quoting activity of a group of high frequency traders identified by NASDAQ, but only includes activity from the NASDAQ exchange. Other exchanges and market centers currently do not provide such data to academics or the public. As a result, studies of high frequency trading have been limited in their ability to examine thoroughly such strategies and their impact on the market. Because data on high-frequency trading tends to be fragmented across many data sources, it is difficult even for regulators to thoroughly analyze their aggregate activity level, study how their activity on one exchange affects their activity on another, and study the effect of particular high frequency strategies on market quality.

See High Frequency Trading, literature review, available at https://www.sec.gov/marketstructure/research/hft_lit_review_march_2014.pdf .

Even if other exchanges did provide such data, the NASDAQ data fields do not include the identities of the high frequency traders. As a result researchers would not be able to study the activity of the same high frequency trader across exchanges.

See infra note 724.

The Plan also would provide information on how various broker-dealers route their customer orders and would allow regulators to study whether access fees and rebates drive routing decisions as much as execution quality considerations. This could inform debates about effects of conflicts of interest created by such maker-taker pricing. Studies of maker-taker pricing require information on routing decisions and how routing affects execution quality. Current academic studies of maker-taker pricing rely on data that provide imprecise information that cannot directly link routing and execution quality, and current similar research carried out by some regulators is often hindered by the significant amount of time it takes to obtain the relevant data from all market centers. However, the Plan would provide regulators with direct access to a data source that would link order lifecycle events together in a way that would allow regulators to more thoroughly analyze how and where broker-dealers route various order types. This could assist regulators in analyzing the importance of fees to the routing decisions and the ultimate impact on investors of any conflicts of interest in broker-dealer routing decisions. Such analysis could inform debates regarding whether maker/taker pricing structures are harmful to market structure.

Similarly, the Plan would provide regulators with data to better understand the nature of short selling. Existing studies of the effects of short selling lack the ability to associate short selling activity with customer-level data, and also lack the ability to distinguish buying activity that covers short positions from buying activity that establishes new long positions. The Plan would allow regulators to examine, for example, how long particular types of traders hold a short position and what types of traders short around corporate events.

The Plan, in requiring information about a Customer, would also facilitate studies of how certain entities other than natural persons trade and the market impact of their trading. For example, existing information on repurchases is aggregated at the monthly and quarterly level while the CAT Data on issuer repurchases would be much more granular. CAT Data would provide information that could determine the size and timing of issuer repurchases, for example. In addition, CAT Data would provide information that could help identify open market repurchases whereas existing data does not distinguish the type of repurchase. As such, the Plan would facilitate research that addresses the timing of issuer repurchases around corporate events or stock option grants and exercises, the extent to which issuers use the safe harbor in Rule 10b-18, and how aggressively issuers trade in the market. In addition, CAT Data on the trading of leveraged ETFs, particularly the end of day rebalancing, could shed light on how the leveraged ETFs relate to market volatility. In addition, Customer information should facilitate analyses of the secondary market trading of ETF Authorized Participants in their ETFs. This could help regulators better understand the arbitrage process between an ETF and its underlying securities and the limitations of that arbitrage.

The CAT NMS Plan does not include requirements to record or report information on the creation or redemption of ETF shares.

The Commission preliminarily believes that CAT Data would also better inform SROs and the Commission in rulemakings and assist them in conducting retrospective analysis of their rules and pilots. In particular, SROs would be able to use order data that is currently not available to examine whether rule changes are in the interest of investors. For example, direct access to consolidated audit trail data that identifies trader types could help an SRO examine whether a new rule improved market quality across the entire market and whether it benefitted retail and institutional investors specifically. Further, CAT Data would allow SROs to examine whether a rule change on another exchange was in the interest of investors and whether to propose a similar rule on their own exchange.

c. Surveillance and Investigations

The Commission preliminarily believes that the enhanced surveillance and investigations made possible by the implementation of the CAT NMS Plan could allow regulators to more efficiently identify and investigate violative behavior in the markets and could also lead to market participants that currently engage in violative behavior reducing or ceasing such behavior, to the extent that such behavior is not already deterred by current systems. The current markets are characterized by surveillance systems that identify violators so that regulators may address these violations. Given that violative behavior is identifiable in current markets, and potential violators know that there is a positive probability that they would be caught by surveillance should they commit a violation, fewer potential violators commit violations than would do so in markets that had no surveillance. Potential violators' expected probability of being caught influences their likelihood of committing a violation. It then follows that any system change that increases the likelihood of violative behavior detection would increase potential violators' expected probability of being caught and thus reduce the likelihood that potential violators would commit a violation.

It is well established in the economics and political science literature that common knowledge among market actors can lead to the deterrence of behaviors; see, e.g., Schelling, Thomas, “The Strategy of Conflict: Prospectus for a Reorientation of Game Theory,” Journal of Conflict Resolution, Vol. 2 No. 3 (1958) and Ellsberg, Daniel, “The Crude Analysis of Strategic Choices,” American Economic Review, Vol. 51, No. 2 (1961). Therefore, market participants with knowledge of improvements in the efficiency of market surveillance, investigations, and enforcements, and consequently the increased probability of incurring a costly penalty, could be deterred from participating in violative behavior.

Specifically, if market participants believe that the existence of CAT, and the improved regulatory activities that result from improvements in data and data processes, increase the likelihood of regulators detecting violative behavior, they could reduce or eliminate the violative activity in which they engage to avoid incurring the costs associated with detection, such as fines, legal expenses, and loss of reputation. Such a reduction in violative behavior would benefit investor protection and the market as investors would no longer bear the costs of the violative behavior that would otherwise exist in the current system. Many of the improvements that would result from CAT could also allow regulators to identify violative activity, such as market manipulation, more quickly and reliably, which could improve market efficiency by deterring market manipulation and identifying and addressing it more quickly and more often when it occurs.

For example, as discussed in Section IV.E.2.c(1), the Plan would allow regulators to more efficiently conduct cross-market and cross-product surveillance relative to surveillance using current data sources, and the requirement that data be consolidated in a single database would assist regulators in detecting violative (but not obvious) activity. To the extent that market participants are aware of the current challenges to regulators in performing cross-market surveillance and aggregating data across venues, and to the extent that they believe that their violative behavior is more likely to be detected if regulators' ability to perform those activities improves, they may reduce or eliminate violative behavior if the CAT Plan is approved.

(1) SRO Surveillance

The Commission preliminarily believes that the CAT NMS Plan would result in improvements in SROs' surveillance capabilities and that many of the benefits to SRO surveillance stem from improvements to data completeness. These benefits encompass a number of improvements to surveillance, including: detection of insider trading; surveillance of principal orders; cross-market and cross-product surveillance, and other market surveillance activities.

Rule 613(f) requires SROs to implement surveillances reasonably designed to make use of the CAT Data. Further, data improvements resulting from the Plan would improve regulators' ability to perform comprehensive and efficient surveillance. As a result, the market surveillances required by Rule 613(f) could identify a broader and more nuanced set of market participant behaviors. As such, the CAT would also provide the opportunity for development of more effective and efficient surveillance system. It is also possible that the CAT Data and tools would enable further innovations in market surveillance beyond those currently contemplated. These innovations could be in response to new developments in the market over the next few years or to the new capabilities for regulators.

CAT Data would include additional fields not currently available in data used for surveillance. The inclusion of Customer-IDs in the CAT would significantly improve surveillance capabilities, including surveillance designed to detect market manipulation and insider trading. Because currently available data do not include customer identifiers, SROs performing insider trading and manipulation surveillance could be unable to identify some suspicious trading and must undertake multiple steps to request additional information after identifying suspect trades. The ability to link uniquely identified customers with suspicious trading behavior would provide regulators with better opportunity to identify the distribution of suspicious trading instances by a customer as well as improving regulators' ability to utilize customer-based risk assessment. This enhanced ability to link customers with behaviors would enable detection of market abuses that are perpetrated by customers trading or quoting through multiple accounts or on multiple trading venues.

As noted in Section IV.D.1.c, this economic analysis considers surveillance to be SROs running processing on routinely collected or in-house data to identify potential violations of rules or regulations.

The Commission understands that SRO surveillances on topics such as insider trading and market manipulation do not incorporate data that identifies customers. Based on alerts from their surveillances, SROs may open a review that runs through several stages of data requests before identifying a customer. As discussed above, the Commission notes that SRO audit trails typically do not provide customer information but a recent FINRA rule change would require its members to report to OATS non-FINRA member customers who are broker-dealers. See supra note 407.

Furthermore, having direct access to data could assist an SRO in its surveillance activities by potentially facilitating quicker responses to suspicious trading activity. Additionally, the inclusion of the principal orders of members would enable regulators to better identify rule violations by broker-dealers that have not previously had to provide audit trail data on their unexecuted principal orders. The evolution of the market has increased the importance of surveillance on principal orders. Many of these principal orders originate from algorithmic or high frequency trading firms who have been the recent subject of regulatory interest. Further, some rules and regulations provide for differential treatment of the principal orders of broker-dealer market makers. Yet, some current data sources used for SRO surveillance exclude unexecuted principal orders, limiting the surveillance for issues such as wash sales. As a result, many surveillance patterns are unable to detect certain rule violations involving principal orders.

See Securities Exchange Act Release No. 61358 (January 14, 2010), 75 FR 3594 (January 21, 2010) (“Concept Release on Equity Market Structure”); Exemption for Certain Exchange Members, supra note 394.

See Section IV.D.2.b(1), supra.

The Plan would also improve regulators' efficiency in conducting cross-market and cross-product surveillance. The Plan would particularly enhance regulators' ability to perform cross-market surveillance, across equity and options markets, by enabling any regulator to surveil the trading activity of market participants in both equity and options markets and across multiple trading venues without data requests. Regulators would also have access to substantially more information about market participants' activity, and the requirement that the data be consolidated in a single database would assist regulators in detecting activity that may appear permissible without evaluating data from multiple venues. Likewise, it would assist regulators in detecting activity that may not appear violative without evaluating data from multiple venues.

For example CAT Data would include Customer information, subaccount allocation information, exchange quotes, trade and order activity that occurs on exchanges, trade and order activity that occurs at broker-dealers that are not FINRA members, and trade and order activity that occurs at FINRA members who are not currently required to report to OATS. In addition CAT Data would require reporters to report data in milliseconds and would be directly available to non-FINRA regulators much faster than OATS is currently available to them. See Section IV.E.1.a, supra.

See Section IV.E.1.c(2), infra. The Commission notes that while this is a benefit allowed by consolidation of data in the Central Repository, linked data would not be available in the Central Repository until T+5, which may delay the completion of surveillance activities.

Increasing market complexity and fragmentation has increased the importance of cross-market surveillance. The Commission noted in its Regulation of NMS Stock Alternative Trading Systems proposing release that, “[i]n the seventeen years since the Commission adopted Regulation ATS, the equity markets have evolved significantly, resulting in an increased number of trading centers and a reduced concentration of trading activity in NMS stocks.” However, because market data are fragmented across many data sources and because audit trail data lacks consistent customer identifiers, regulators cannot run cross-market surveillance tracking particular customers. Furthermore, routine cross-product surveillance is generally not possible with current data. The potential enhancements in market surveillance enabled by the CAT NMS Plan are likely to result in more capable and efficient surveillance which could reduce violative behavior and protect investors from harm.

See Securities Exchange Act Release No. 76474 (November 18, 2015), 80 FR 80998 (December 28, 2015), at 81000.

As noted in the above, SROs currently do not conduct routine surveillance that tracks particular customers because data currently used for surveillance does not include customer information.

(2) Examinations

The Commission preliminarily believes that availability of the CAT would also improve examinations and that these improvements would benefit investor protection, and the market in general, by resulting in more effective supervision of market participants. The Commission conducted 493 broker-dealer examinations in 2014 and 484 in 2015, 70 exams of the national securities exchanges and FINRA in 2014 and 21 in 2015. In addition, the Commission conducted 1,237 investment adviser and investment company examinations in 2014 and 1,358 in 2015. Virtually all investment adviser examinations and a significant proportion of the Commission's other examinations involve analysis of trading and order data. Currently some data that would be useful to conduct risk-based selection for examinations, such as trade blotters, are not available in data sources available for pre-exam analysis. Further, data available during exams often require regulatory Staff to link multiple data sources to analyze customer trading. For example, some customer identities are present in EBS data, but time stamps are not. To evaluate the execution price a customer received, it is necessary to know the time of the trade to compare the price of the customer's execution with the prevailing market prices at that time. This requires linking the EBS data with another data source that contains trades with time stamps (such as the trade blotter). These linking processes can be labor-intensive and require the use of algorithms that may not link with 100% accuracy. Finally, for investment adviser examinations, examiners sometimes use non-trading data such as Form PF, Form 13-F, Form ADV, and clearing broker reports as a proxy for trading data when selecting investment advisers for examinations. The CAT would improve examinations in the following specific ways.

Regulators can obtain detailed equity transaction data by requesting a trade blotter from a particular firm; however, the data would only show the activity of that firm.

First, the Commission preliminarily believes that the expected improvements in the data qualities discussed above would enhance the ability of regulators to select market participants for focused examinations on the basis of risk. The direct access to consolidated data in a single location would dramatically improve regulators' ability to efficiently conduct analyses in an attempt to select broker-dealers and investment advisers for more intensive examinations based on identified risk. Having CAT Data stored in the Central Repository in a linked format would allow examiners to access much more data directly through a query and without performing the linking process on an ad-hoc basis than is currently available before an exam. The ability to use Customer Account Information in the process for selecting investment advisers for exams, for example, could allow those selection models to incorporate trading data directly instead of imperfect proxies for trading data. This could lead to improved outcomes for risk-based examinations, such as more regulatory resources invested in examining market participants who are at an elevated risk of violating federal securities laws, rules, and regulations, and SRO rules, and a reduction in the proportion of examinations that might not have been necessary if a more complete view of the market participant's activity had been available. Compliant market participants could benefit from a reduction in the relative frequency of burdensome examinations. Improvements in the breadth and effectiveness of risk-based examination would help protect investors by increasing the likelihood of identifying market participants who are violating laws, rules and regulations.

Second, the Commission preliminarily believes that with the CAT, regulators would be able to examine market participants more effectively. In particular, regulators would be able to conduct certain types of exams more efficiently because of the inclusion of Customer-IDs in CAT. In addition, direct access to CAT Data would provide examination Staff with the ability to conduct more analysis prior to opening an examination because data would be available without the need to make a formal data request. In addition, the clock synchronization provisions of the Plan could aid regulators in sequencing some events more accurately, thereby facilitating more informed exams. In sum, the Plan would allow the data collection portion of examinations to be completed more quickly with fewer formal data requests. More efficient examinations would help regulators better protect investors from the violative behavior of some market participants and could reduce examination costs for market participants who would have otherwise faced examinations that are less focused and more lengthy.

See Sections IV.D.2.b(2), supra and IV.H.2.a(1), infra.

(3) Enforcement Investigations

Many Commission enforcement actions involve trade and order data. The Commission preliminarily believes that the improvements in data qualities that would result from the CAT NMS Plan would significantly improve the efficiency and efficacy of enforcement investigations, including insider trading and manipulation investigations. The Commission believes that more efficient and effective enforcement activity is beneficial to both investors and market participants because it deters violative behavior that degrades market quality and that imposes costs on investors and market participants.

See supra note 711 and accompanying text.

See Section IV.E.1, supra.

Dramatic expected benefits come from improvements to the accuracy, accessibility, timeliness, and completeness of the data. As noted above, compiling the data to support an investigation often requires a tremendous amount of time and resources and requires multiple requests to multiple data sources and significant data processing efforts, for both SROs and the Commission. While individual SROs have direct access to the data from their own markets, their investigations often require access to the data of other SROs because firms trade across multiple venues. Some enforcement investigations, including those on insider trading and manipulation, require narrow market reconstructions that allow investigators to view actions and reactions across the market. Currently, the data fragmentation and the time it takes to receive requested data, makes these market reconstructions cumbersome and time-consuming. Further, new data fields related to Customer information and the Allocation Reports should improve the completeness of the data available to investigators.

See Sections IV.D.2.b(3) and IV.D.2.b(4), supra.

Under the CAT NMS Plan, the data for an enforcement investigation initiated at least five days after an event would be processed, linked, and available for analysis within 24 hours of a query, instead of the current timeline of weeks or longer. Further, some of the data processing steps that are now performed on an ad-hoc basis during an investigation would be systematically performed by the Plan Processor in advance. The availability of uncorrected data by noon on T+1 could improve the Commission's chances of preventing asset transfers from manipulation schemes because regulators could use the uncorrected data to detect the manipulation and identify the suspected manipulators. These improvements could shorten the times required to collect the data for investigations.

See Section IV.E.1.d(4), supra.

See Section IV.D.2.b(4), supra.

Other expected benefits stem from improvements in the accuracy and completeness of the data. The inclusion and expected improvement in the accuracy of customer identifying data could allow regulators to review the activity of specific market participants more efficiently; currently, identifying the activity of a single market participant across the market is cumbersome and prone to error. This information would be particularly helpful in identifying insider trading, manipulation and other potentially violative activity that depends on the identity of market participants. Customer information could also be helpful to regulators in more efficiently identifying investors who qualify for disgorgement proceeds and in estimating such disgorgement proceeds.

See Section IV.D.2.b(2)D, supra.

The Commission also believes that increasing the proportion of market events that could be sequenced under the CAT NMS Plan could yield some benefits in enforcement investigations, improving investigations of insider trading, manipulation, and compliance with Rule 201 of Regulation SHO and Rule 611 of Regulation NMS. The expected improvements in completeness could also benefit investigations by allowing regulators to observe in a consolidated data source relevant data that are not available in some or all current data sources, including time stamps, principal orders, non-member activity, allocations, and the identification of whether a trade increases or decreases an existing position. This data could be important, for example, when investigating allegations of market manipulation or cherry-picking in subaccount allocations. Having disaggregated information about allocations and issuer repurchases also could facilitate new ways to investigate allegations of unfair allocations and new ways to investigate and monitor manipulation through issuer repurchases.

Again, benefits associated with the ability to sequence events may be limited in some cases because many order events would not be able to be sequenced completely with the standards established in the CAT NMS Plan. See Section IV.D.2.b(2)B.i, supra.

(4) Tips and Complaints

The Commission preliminarily believes that the CAT NMS Plan would improve the process for evaluating tips and complaints by allowing regulators to more effectively triage tips and complaints, which could focus resources on behavior that is most likely to be violative. The SROs and Commission evaluate thousands of tips and complaints regarding trading behavior each year. In fiscal years 2014 and 2015, the Commission received around 15,000 entries in its TCR system, approximately one third of which related to manipulation, insider trading, market events, or other trading and pricing issues. As stated in the Baseline Section, the analysis of tips and complaints follows three general stages. The Commission expects that the Plan would improve the second and third stages, the third in ways described in the Examinations and Enforcement Investigations Sections. The second stage in the evaluations of tips, which help regulators determine the credibility of a tip or complaint, is limited by a lack of direct access to the most useful data; specifically, customer information and cross-market data. The availability of the CAT Data would drastically increase the detail of data available to regulators for the purposes of tip assessment. This access would assist the SROs and Commission in identifying which tips and complaints are credible, would help ensure that regulators open investigations or examinations on credible tips and complaints, and would limit regulatory resources spent on unreliable tips and complaints. Likewise, regulated market participants would likely benefit from a reduction in unnecessary burdens placed upon them by inquiries that are related to tips that the CAT Data could show are not credible.

See SEC Office of the Whistleblower, What Happens to Tips, https://www.sec.gov/about/offices/owb/owb-what-happens-to-tips.shtml .

See Sections IV.D.2.a(4), supra.

Cross-market data is especially key to market manipulation complaints, because regulators may need to examine a broad range data to see if a complaint is valid.

3. Other Provisions of the CAT NMS Plan

The Commission notes that there are a number of provisions of the CAT NMS Plan that provide for features that are uniquely applicable to a consolidated audit trail or otherwise lack a direct analog in existing data systems. Therefore, rather than analyze the benefits of these provisions as compared to existing NMS Plans or data systems, the Commission has analyzed these provisions in comparison to a CAT NMS Plan without these features. The Commission preliminarily believes that these provisions of the CAT NMS Plan increase the likelihood that the potential benefits of the CAT NMS Plan described above would be realized.

a. Future Upgrades

Several provisions in the Plan seek to ensure that the CAT Data would continually be updated to keep pace with technological and regulatory developments. For example, the Plan would require that the Chief Compliance Officer review the completeness of CAT Data periodically, that the Central Repository be scalable to efficiently adjust for new requirements and changes in regulations, and that Participants provide the SEC with a document outlining how the Participants could incorporate information on select additional products and related Reportable Events. The Commission preliminarily believes that these provisions would allow the CAT to be updated if and when the applicable technologies and regulations change.

See CAT NMS Plan supra note 3, at Sections 4.12(b)(ii), 6.2(a)(v)(E). The Chief Compliance Officer would be required to perform reviews on matters including the completeness of information submitted to the Plan Processor or Central Repository and report findings periodically to the Operating Committee.

See id. at Appendix D, Section 1.1.

See id. at Section 6.11. This document is due within six months of the Effective Date of the CAT NMS Plan.

Specifically, Rule 613(b)(6)(ii) and (iii) require that the Plan include a provision requiring a report at least every two years that details potential improvements in the CAT, such as incorporating new technology to improve system performance. Such a report would also include the costs of any such improvements. The CAT NMS Plan delegates responsibility for the report to the Chief Compliance Officer.

Section 6.1(d)(iv) of the Plan, with respect to new functionality, requires the Plan Processor to “design and implement appropriate policies and procedures governing the determination to develop new functionality for the CAT including, among other requirements, a mechanism by which changes can be suggested by Advisory Committee members, Participants, or the SEC,” as well as providing for the escalation of reviews of proposed technological changes and upgrades to the Operating Committee, and for addressing the handling of surveillance.

With respect to upgrades to maintain existing functionality, the Plan Processor could evaluate and implement potential system changes and upgrades to maintain and improve the normal day-to-day operating function of the CAT System; material system changes and upgrades are to be performed by the Plan Processor in consultation with the Operating Committee. The Plan Processor may on its own discretion initiate changes or upgrades to ensure compliance with applicable legal requirements. Regular reports on the operations and maintenance of the CAT System are to be provided by the Plan Processor to the Operating Committee, including reports on system improvements contemplated in Appendix D, Upgrade Process and Development of New Functionality.

See id. at Section 6.1(j).

See id. at Section 6.1(k).

See id. at Section 6.1(o).

Section 11 of Appendix D sets out the obligations of the Plan Processor with respect to the requirements discussed above (e.g., to develop a process to add functionality to CAT, including reviewing suggestions submitted by the SEC). The Plan Processor must create a defined process for developing impact assessments, including implementation timelines for proposed changes, and a mechanism by which functional changes that the Plan Processor wishes to undertake could be reviewed and approved by the Operating Committee. The Plan Processor “shall not unreasonably withhold, condition, or delay implementation of any changes or modifications reasonably requested by the Operating Committee.” There must be a similar process to govern the changes to the Central Repository discussed above—i.e., business-as-usual changes that could be performed by the Plan Processor with only a summary report to the Operating Committee, versus infrastructure changes that would require approval by the Operating Committee. Finally, a process for user testing of new changes must be developed by the Plan Processor.

See id. at Appendix D, Section 11.1.

See id at Appendix D, Section 11.2.

See id. at Appendix D, Section 11.3.

Appendix C notes that the Plan Processor must ensure that the Central Repository's technical infrastructure is scalable (to increase capacity to handle increased reporting volumes); adaptable (to support future technology developments so that new requirements could be incorporated); and current (to ensure, through maintenance and upgrades, that technology is kept current, supported, and operational).

See id. at Appendix C, Section A.5(a).

These provisions are designed to ensure that the Participants consider enhancing and expanding CAT Data shortly after initial implementation of the CAT NMS Plan and that the Participants consider improvements regularly continuing forward. The Commission preliminarily expects that, in addition to these provisions, the CCO review would further facilitate proactive expansion of CAT to account for a regulatory change or change in how the market operates, or should there be a need for regulators to have access to new order events or new information about particular order events. To the extent that the Participants determine that an expansion is necessary and it is approved by the Commission, the Plan's scalability provision promotes the efficiency of the implementation of that expansion such that it could be completed at lower cost and/or in a timely manner.

Taken together, these provisions could also provide a means for the Commission to ensure that improvements to CAT functionality are considered so as to preserve its existing benefits, or that expansion of CAT functionality is undertaken in order to create new benefits. These methods are not certain, but the Commission does retain the ability to modify the Plan, if such a step becomes necessary to ensure that future upgrades are undertaken as necessary. Moreover, the focus on scalability, adaptability, and timely maintenance and upgrades promotes a system that could be readily adapted over time, versus one that is difficult or costly to expand or modify. The Commission preliminarily believes that the provisions outlined above would allow the CAT Data to be continually updated to keep pace with technological and regulatory developments.

b. Promotion of Accuracy

The Commission notes that the Plan contains specific provisions designed to generally promote the accuracy of information contained in the Central Repository. The CCO is required, among other responsibilities, to perform reviews related to the accuracy of information submitted to the Central Repository and report to the Operating Committee with regard thereto, and there is a special Compliance Subcommittee of the Operating Committee, which is established to aid the CCO with regard to, among other things, issues involving the accuracy of information. The Plan also contains certain other provisions intended to monitor and address Error Rates.

See CAT NMS Plan, supra note 3, at Section 6.2(a)(v)(E).

See id. at Section 4.12(b).

See id. at Appendix C, Section A.3(b).

The Operating Committee is responsible for adopting policies and procedures regarding the accuracy of CAT Data, which the Plan Processor shall be responsible to implement. The Plan Processor in turn must provide regular reports regarding accuracy issues to the Operating Committee, specifically Error Rates relating to the Central Repository, including (to the extent the Operating Committee deems necessary or advisable) Error Rates by day, changes in the Error Rates over time, and Compliance Thresholds by CAT Reporter, by Reportable Event, by age before resolution, by symbol, by symbol type, and by event time. The Plan documents an initial Error Rate tolerance of 5%, but requires that, at least annually, the Plan Processor review the Error Rates and make recommendations to the Operating Committee for proposed changes to the maximum Error Rate; and requires that the Operating Committee set and periodically review the maximum Error Rate.

See id. at Section 6.5(d).

See id. at Section 6.5(d)(i).

Under the Plan, the Plan Processor would also provide details to each CAT Reporter on the number of rejected records and the reasons for their rejection on a daily basis. And on a monthly basis, the Plan Processor would publish report cards that would allow CAT Reporters to compare their Error Rates with those of industry peers; this is similar to the process used by FINRA for OATS reporting. The Plan Processor would notify each CAT Reporter that exceeds the maximum Error Rate, and provide the specific reporting requirements that they did not fully meet. Participants and the SEC could request reports on Error Rates from the Plan Processor. The Plan Processor would also provide statistics on each CAT Reporter's Compliance Thresholds—the CAT Reporter's specific Error Rate, which could serve as the basis for a review or investigation into the CAT Reporter's performance by the Participants or the SEC for failure to comply with CAT reporting obligations—to the Participants or the SEC.

In addition to providing CAT Reporters data on their Error Rates, the Plan states that the Participants believe that in order to meet Error Rate targets, industry would require certain resources, including a stand-alone testing environment, and time to test their reporting systems and infrastructure. The Technical Specifications must also be well-written and effectively communicated to CAT Reporters with sufficient time to allow proper systems updates. Finally, the Plan notes that reporters may be subject to penalties or fines for excessive Error Rates, to be defined by the Operating Committee.

See id. at Appendix C, Section A.3(b).

See id. at Appendix C, Section A.3(b), n.101.

The Commission preliminarily believes that these provisions to document Error Rates and promote data accuracy are reasonably designed to improve the overall accuracy of CAT Data relative to the exclusion of such provisions; however, the Commission also preliminarily believes that certain procedures outlined in the Plan may not incentivize all firms to further improve the quality of the data they report. The Commission recognizes that providing feedback to individual CAT Reporters on their individual Error Rates and information that compares Error Rates to industry peers could motivate firms with high Error Rates to reduce those rates, to avoid accruing penalties and fines associated with being a high Error Rate CAT Reporter. However, it is not clear what incentive, if any, would be provided to firms with median Error Rates to improve their regulatory data reporting processes; this could collectively limit industry's incentives to reduce Error Rates. Furthermore, the Commission notes that, under the Plan, proposals to adjust the maximum allowable Error Rate are to originate from the Plan Processor. The Commission preliminarily believes that the Participants (as data users) have incentives to pursue lower Error Rates as data errors could complicate their efforts to perform their regulatory responsibilities. However, the Commission preliminarily believes that the Plan Processor would also have to allocate resources to error resolution, so could be incentivized to pursue Error Rate reduction.

The Commission understands that OATS has an analogous feedback system, but not all current data sources have such a system.

The Commission notes that the Plan includes provisions requiring the establishment of a symbology database that will also foster accuracy. The Plan requires the Central Repository to create and maintain a symbol history and mapping table, as well as provide a tool to regulators and CAT Reporters showing the security's complete symbol history, along with a start of day and end of day list of reportable securities for use by CAT Reporters, in .csv format, by 6:00 a.m. on each trading day. This resource will assist regulators in accurately identifying all trading activity of securities across venues, many of which do not natively follow listing exchange symbology.

See CAT NMS Plan, supra note 3, at Appendix D, Section 2.

Regarding the Plan's business clock synchronization requirements, the Plan also discusses the expectation that Participants and their Industry Members will each be required to maintain a five-year running log, or comparable procedure, documenting the time of each clock synchronization performed and the result of such synchronization. These practices would reveal the parameters of any discrepancies, between Business Clocks and NIST, that exceed 50 milliseconds. As mentioned above, there is currently uncertainty regarding clock offsets, clock drift, and synchronization practices of Participants and Industry Members and the required practice of systematically maintaining five-year logs regarding these details should improve regulatory and industry understanding of these dynamics, which should provide a clearer foundation for evaluating the standards set in the Plan upon which future improvements could be considered.

Id. at Appendix C, Section A.3(c).

c. Promotion of Timeliness

In addition to the specific timeliness benefits discussed in the foregoing Sections, the Plan contains some provisions that promote performance of the Central Repository, and that therefore could indirectly improve the timeliness of regulator access to or use of the CAT Data. These are found in capacity requirements for the Plan Processor, disaster recovery requirements to ensure the availability of the system, and in supervision and reporting of timeliness issues.

Specifically, first, the Plan Processor must measure and monitor Latency within the Central Repository's systems, must establish acceptable levels of Latency with the approval of the Operating Committee, and must establish policies and procedures to ensure that data feed delays are communicated to CAT Reporters, the Commission, and Participants' regulatory Staff. The Plan further provides that “[a]ny delays will be posted for public consumption, so that CAT Reporters may choose to adjust the submission of their data appropriately. . . .” The Plan Processor must also provide relevant parties, as well as to the public, with approximate timelines provided for system restoration. Moreover, the Central Repository is required to be designed to meet certain capacity standards, including handling above-peak submission volumes, storing data for a sliding 6 year window (more than 29 petabytes of raw, uncompressed data), and the ability to add capacity quickly and seamlessly if needed.

See CAT NMS Plan, supra note 3, at Appendix D, Section 8.3.

Id.

Id.

See id. at Appendix D, Section 1.3.

Second, the Plan Processor must develop disaster recovery and business continuity plans to support the continuation of CAT business operations. Business continuity planning must include a secondary site for critical staff, capable of recovery and restoration of services within 48 hours, with the goal of next day recovery. The secondary site must have the same level of availability, capacity, throughput and security (physical and logical) as the primary site—i.e., it must be fully redundant. Thus, in the event of a widespread disruption, delays to CAT processing and regulator access to CAT of greater than a day or two could likely be prevented.

See id. at Appendix D, Sections 5.3-5.4.

Id.

Id.

Third, the Chief Compliance Officer of the Plan Processor must conduct regular monitoring of the CAT System for compliance, including with respect to the reporting and linkage requirements in Appendix D. Moreover, the Plan Processor must provide the Operating Committee with regular reports on the CAT System's operations and maintenance, including its capacity and performance, as set out in Appendix D.

See id. at Section 6.2(a)(v)(J).

See id. at Section 6.1(o)(i).

Finally, one caveat on the foregoing discussion is that system performance would in part be dependent on a series of SLAs to be negotiated between the Plan Participants and the eventual Plan Processor, including with respect to linkage and order event processing performance, query performance and response times, and system availability. As these have not yet actually been negotiated, some of the key timeliness benefits anticipated to accrue from implementation of the Plan could be subject to the successful negotiation on an acceptable basis of the terms of the SLAs.

See id. at Appendix D, Section 8.5.

d. Operation and Administration of the CAT NMS Plan

There are certain elements of the CAT NMS Plan's governance that, like the other factors discussed in this subsection, are uniquely applicable to a consolidated audit trail, and that the Commission therefore analyzed in comparison to a CAT NMS Plan without these features (or that implements those features in a different way). The Commission preliminarily believes that these provisions of the CAT NMS Plan increase the likelihood that the potential benefits of the CAT NMS Plan described above would be realized.

(1) Introduction

In adopting Rule 613, the Commission established certain requirements for the governance of the CAT NMS Plan, stating that those “requirements are important to the efficient operation and practical evolution of the [CAT], and are responsive to many commenters' concerns about governance structure, cost allocations, and the inclusion of SRO members as part of the planning process.” The Commission did not, in Rule 613, establish detailed parameters for the governance of the CAT NMS Plan, but rather allowed the SROs to develop specific governance provisions, subject to a small number of requirements. Recognizing that Rule 613 left Plan Participants with wide latitude to determine how to structure the Plan's governance, the Commission in the Adopting Release also stated that “[a]fter the SROs submit the NMS plan, the Commission and the public will have more detailed information in evaluating the NMS plan.”

See Adopting Release, supra note 9, at 45787.

Id. at 45787-45788.

The Plan's governance is described in greater detail in Section III.A.3. above, but generally consists of a Delaware LLC, which is to “create, implement, and maintain the CAT and the Central Repository,” and which is to be managed by the Operating Committee, consisting of one voting representative of each SRO Participant. The Operating Committee acts by majority or Supermajority Vote, depending on the issue. An Advisory Committee that includes a mix of broker-dealers, as required by Rule 613, is to “advise the [Operating Committee] on the implementation, operation and administration of the central repository.” These features are analyzed in greater detail below.

See Rule 613(b)(7). Whereas Section 4.13(b) requires that the Operating Committee select representatives of different types of broker-dealers, it specifies that Advisory Committee representatives would “serve on the Advisory Committee on behalf of himself or herself individually and not on behalf of the entity for which the individual is then currently employed.” See CAT NMS Plan, supra note 3, at Section 4.13(b).

The Commission preliminarily believes that the governance provisions identified in the Adopting Release continue to be important to the efficient operation and practical evolution of the Plan, particularly given that there are a range of possible outcomes with respect to both the costs and benefits of the Plan that depend on future decisions. The way in which the identified governance provisions have been incorporated into the Plan, as discussed in greater detail below, could help facilitate better decision-making by the relevant parties. This, in turn, means that the Commission could have greater confidence that the benefits resulting from implementation of the Plan would be achieved in an efficient manner and that costs resulting from inefficiencies would be avoided.

The Commission notes that it can monitor whether the benefits of CAT are being achieved. For example, certain Operating Committee actions are subject to Commission approval. The Commission also retains the ability to modify the Plan as it may deem necessary or appropriate. To enable the Commission to exercise its oversight authority in an informed manner and to make its views known, representatives of the Commission are permitted to attend meetings of the Operating Committee, although the Commission representatives may be excluded from Operating Committee Executive Sessions. Moreover, the Commission is entitled to receive information regarding the performance of the Central Repository, including a Regular Written Assessment of the operation of the Central Repository at least every two years, or more frequently in connection with any review of the Plan Processor's performance. The assessment would cover the performance metrics specified in Rule 613(b)(6)(i). The Commission is also entitled to receive any reports prepared in connection with the Operating Committee's annual performance review of the Plan Processor.

See CAT NMS Plan, supra note 3, at Section 4.3 (stating that actions authorized by Majority and Supermajority Vote of the Operating Committee are subject to approval by the Commission whenever such approval is required under the Exchange Act and the rules thereunder).

See CAT NMS Plan, supra note 3, at Section 4.4(a).

See 17 CFR 242.613(b)(6)(i). Rule 613(b)(6) requires the Participants to provide the Commission with a written assessment of operation of the CAT at least every two years, along with a detailed plan, based on the assessment, that indicates any potential improvements to the performance of the CAT and includes an estimate of the costs and potential impacts of such improvements on competition, efficiency and capital formation, as well as an estimated implementation timeline for such potential improvements.

See CAT NMS Plan, supra note 3, at Section 6.1(n). The review may be more frequent than annually if at the request of two non-affiliated Participants. The Commission also has other means of accessing information (e.g., through books & records requirements).

(2) Key Factors Relating to Governance

Two factors identified by the Commission in the Adopting Release as “important to the efficient operation and practical evolution of the [CAT]” are voting within the Operating Committee and the role and composition of the Advisory Committee. Voting thresholds that result in Operating Committee decision-making that balances the ability of minority members to have alternative views considered with the need to move forward when appropriate to implement needed policies can promote achievement of the Plan's benefits in an efficient manner. Similarly, an Advisory Committee that is balanced in terms of membership size and composition, as well as in its ability to present views to the Operating Committee, can result in better performance of its informational role, and thus more efficient achievement of the benefits of the Plan.

A. Voting

In adopting Rule 613, the Commission found that one Commenter's concerns about unanimous voting in the context of the CAT NMS Plan “have merit.” Specifically, the Commission stated that “an alternate approach” to voting involving “the possibility of a governance requirement other than unanimity, or even super-majority approval, for all but the most important decisions” should be considered, as it “may be appropriate to avoid a situation where a significant majority of plan sponsors—or even all but one plan sponsor—supports an initiative but, due to a unanimous voting requirement, action cannot be undertaken.” The Commission “urge[d] the SROs to take into account the need for efficient and fair operation of the NMS Plan governing the consolidated audit trail” in setting voting thresholds.

See Adopting Release, supra note 9, at 45787.

Id.

The Plan sets forth two voting thresholds for most matters to be decided by the Operating Committee. Majority approval of the Operating Committee is sufficient to approve routine matters, arising in the ordinary course of business, while non-routine matters, outside the ordinary course of business, would require a supermajority (two-thirds) vote of the Operating Committee to be approved.

As noted in Section IV.G.4, infra, the Plan requires unanimous voting in only three circumstances: A decision to obligate Participants to make a loan or capital contribution, a decision to dissolve the Company, and a decision to take an action by written consent instead of a meeting.

See CAT NMS Plan, supra note 3, at Section 4.3; Appendix C, Section B.8(d). (specifying actions of the Operating Committee that require a Supermajority Vote); see also id. at Appendix C, Section D.11(b).

The Plan generally eschews a unanimous voting threshold, except for the three clearly-defined circumstances noted above. Unanimity as a voting threshold may confer greater influence on holders of minority views, but it may also give a small faction the ability to extract private benefits inconsistent with Plan objectives by acting as holdouts. In a hold-out dynamic, one member may be able to block action that all the other members agree should move forward. While this dynamic may occasionally be used productively, to produce better decision-making through fostering discussion and compromise, it also may give one member the power to stand in the way of needed change.

There are other governance-related trade-offs for majority voting versus supermajority voting; these are discussed in greater detail in the Plan. See CAT NMS Plan, supra note 3, at Appendix C, Sections B.8(d) and D.11(b).

The ability of a single member to prevent action with regard to the Plan could be particularly troublesome if that member were motivated by a conflict of interest. The Plan requires recusal of the member representing such a Participant from voting in the Operating Committee on matters that raise a conflict of interest, defined as any matter subject to a vote that interferes, or is reasonably likely to interfere, with the member's objective consideration of the matter, or that is, or would reasonably likely be, inconsistent with the regulatory purpose and objectives of CAT. Recusal of a member could also be compelled by a supermajority of the Operating Committee. If conflicts of interest were the cause of all unproductive holding-out (i.e., holding out that does not contribute to better decision-making), then a robust conflict of interest provision could mitigate some of the negative features of unanimous voting.

That there are potential conflicts of interest between Participants acting in their self-regulatory capacities and Participants acting in the other capacities in which they serve is well-documented; see, e.g., Peter M. DeMarzo, Michael J. Fishman, and Kathleen M. Hagerty, “Self-Regulation and Government Oversight,” 72 Review of Economic Studies 687 (2005); see also David Reiffen and Michel Robe, “Demutualization and Customer Protection at Self-Regulatory Financial Exchanges,” Journal of Futures Markets (2011) and Securities Exchange Act Release No. 50700 (November 18, 2004), 69 FR 71256 (December 8, 2004) (Concept Release Concerning Self-Regulation); John W. Carson, Conflicts of Interest in Self-Regulation: Can Demutualized Exchanges Successfully Manage Them? (World Bank Policy Research Working Paper 3183, December 2003). These conflicts could be further complicated if the individual employee of the Participant SRO who represents the Participant SRO on the Operating Committee sought to advance a private gain for the individual employee that is inconsistent with the Plan's regulatory objective or the objective of the Participant SRO. Indeed, the idea that an agency conflict between a natural person and the entity that the person represents has been discussed extensively in the academic literature on the governance of corporations; see, e.g., Jonathan Berk and Peter DeMarzo, 2011, Corporate Finance, Second Edition, Prentice Hall (Section 2.1: Corporate Governance and Agency Costs).

See CAT NMS Plan, supra note 3, at Section 4.3(d) (recusal requirement) and Section 1.1 (definition of Conflict of Interest). Section 4.3(d) also automatically recuses a member from voting with respect to matters relating to the selection or removal of the Plan Processor if they or their affiliates are, or are bidding to be, the Plan Processor. Id.

See CAT NMS Plan, supra note 3, at Section 4.3(d).

Majority voting as a voting threshold strikes a different balance between the rights of members than does unanimous voting. Majority voting avoids the hold-out problem of unanimity, but can result in decisions that bear less concern for the interests of the minority members. Whether it does so or not may depend at least in part on voting dynamics on the Operating Committee. Under the Plan, each member has only one vote within the Operating Committee, and so an individual member—and represented Participant—could not unilaterally advance a position that benefits only the Participant under the Plan. That said, however, some individual members could exercise more influence than others over the outcome of the voting process. Participant SROs that are affiliated with one another could vote as a bloc by designating a single individual to represent them on the Committee. Individuals who represent more than one SRO would then in principle exercise more influence than other individuals on the Operating Committee. The Chair of the Operating Committee also could exercise more influence than other members on the Committee, even though the Chair only has one vote, through influence over Committee processes. Ultimately, however, no individual would have unilateral control over vote outcomes, even at a majority voting threshold. Whether the threshold results in adequate attention to the rights of minority members could therefore depend on the ease with which a majority coalition can be formed, whether those coalitions are fluid or static, and whether in practice decision-making is collegial or contentious. While majority voting could pose a risk of disregard for minority positions, that risk here is mitigated in that majority voting only applies to the less important matters that could arise in the operations of the Plan.

See CAT NMS Plan supra note 3, at Section 4.2(a) (“One individual may serve as the voting member of the Operating Committee for multiple Affiliated Participants, and such individual shall have the right to vote on behalf of each such Affiliated Participant.”) Even if separate representatives were appointed for each voting member, such individuals could agree to vote in a bloc; see also Section IV.G.1, infra, (discussing how many affiliated groups would need to vote together to reach a majority or supermajority).

By enabling a single individual (i.e., natural person) to vote on behalf of groups of Affiliated Participant SROs, the Plan reduces the share and number of individuals needed to approve a committee action below the share and number of votes required for approval. For example, as few as two individuals (who would possess more than one-third of member votes) may be sufficient to block an action that requires a two-thirds (a supermajority) vote for approval of an action of the Operating Committee under the Plan. This casting of multiple votes by a single group is limited for some decisions under the Plan, however. See CAT NMS Plan, supra note 3, at Section 4.4(a) (Meetings of the Operating Committee: special and emergency meetings); see also Section IV.G.1, infra (discussing, in n.1077, the various affiliated exchanges among the 20 members of the Operating Committee, which could appoint a single individual to represent them).

Specifically, see CAT NMS Plan, supra note 3, Section 4.2(b) which establishes that there shall be elected a Chair from among the members of the Operating Committee, and states that the Chair's powers are those that the Operating Committee may from time to time prescribe. For example, the Chair may be granted the power to set the agenda of Operating Committee meetings, and thereby advance agenda items favorable to the Chair. Id. Section 4.2(b) also specifies that the Chair is not entitled to a tie-breaking vote and that the Chair may be removed by Supermajority Vote of the Operating Committee.

The Plan's supermajority voting requirement for more important matters represents an intermediate ground between majority and unanimous voting, requiring more than a bare majority of members to agree to support a position, which therefore enhances the ability of members of the minority to seek to have their views reflected in the ultimate decision, while limiting the ability of minority members to act as holdouts. That said, the supermajority voting requirement may also have some disadvantages: To the extent that rules and practices already in place require correction, a supermajority voting requirement may make it more difficult to assemble the votes necessary to make needed changes. For example, supermajority voting could have the indirect effect of locking in the preferred business practices of the inaugural members of the Operating Committee. For decisions later in the Plan implementation, this lock-in effect of supermajority voting could make it more difficult for the Operating Committee to take non-routine actions, such as replacing the Plan Processor after the initial selection decision.

See id. at Section 4.3(i). Supermajority voting as a governance mechanism in the CAT NMS plan is distinct from an analysis of supermajority voting rules in other settings.

B. Advisory Committee

Rule 613(b)(7) requires that the Plan designate an Advisory Committee. Specifically, Rule 613(b)(7) calls for the formation of an Advisory Committee to advise the plan sponsors on the implementation, operation, and administration of the Central Repository, as detailed above in Section III.A.3 of this Notice. Under Rule 613(b)(7)(i), the Advisory Committee must include representatives of member firms of the plan sponsors (broker-dealers), acting in their own capacities as individuals on the Committee. Under Rule 613(b)(7)(ii), plan sponsors must give members of the Committee access to information and permit them to express their views and attend meetings of the Operating Committee. Also under Rule 613(b)(7)(ii), the Operating Committee has the right to exclude members of the Advisory Committee from its deliberations by meeting in Executive Session by a Majority Vote of its members.

See Section III.A.3 (Requirements Pursuant to Rule 608(a)), supra; see also Section IV.G.4.a, infra, for a discussion of the effects of the Advisory Committee on the efficiency of the Plan.

The Adopting Release states that the “provision requiring the creation of an Advisory Committee, composed at least in part by representatives of the plan sponsors,” was “[i]n response to the comment requesting that the broker-dealer industry receive a `seat at the table' regarding governance of the NMS plan.” In addition, the Commission “encourage[d] the plan sponsors to, in the NMS plan, provide for an Advisory Committee whose composition includes SRO members from a cross-section of the industry, including representatives of small-, medium-and large-sized broker-dealers.” Rule 613 does not give broker-dealers a vote on the Operating Committee itself. In the Adopting Release, the Commission stated that the structure of Rule 613 as adopted “appropriately balances the need to provide a mechanism for industry input into the operation of the central repository, against the regulatory imperative that the operations and decisions regarding the [CAT] be made by SROs who have a statutory obligation to regulate the securities markets, rather than by members of the SROs, who have no corresponding statutory obligation to oversee the securities markets.”

See Adopting Release, supra note 9, at 45786.

In implementing these provisions of Rule 613, the Plan requires the Advisory Committee to have diverse membership. Section 4.13 of the Plan requires an Advisory Committee with a minimum of six broker-dealers of diverse types and six representatives of entities that are not broker-dealers. That is, five of twelve seats on the initial Advisory Committee would be filled by representatives, respectively, of the client of a registered broker or dealer, two types of institutional investors, and two others with academic and regulatory expertise. Terms of Advisory Committee members would not exceed three years, and memberships would be staggered so that a third of the Committee would be replaced each year.

See CAT NMS Plan, supra note 3, Section 4.13(b).

See id. at Section 4.13(b)(i) through (xii).

See id. at Section 4.13(c).

The Commission believes that the Plan's provisions regarding the Advisory Committee advance the goals of the Advisory Committee articulated in the Adopting Release: To allow the Operating Committee to receive the benefit of members' expertise with respect to “expected or unexpected operational or technical issues” and “help assure the Commission and market participants that any requirements imposed on SRO members will be accomplished in a manner that takes into account the burdens on SRO members.”

Given the primary purpose of the Advisory Committee as a forum to communicate important information to the Operating Committee, which the Operating Committee could then use to ensure its decisions are fully-informed, the Plan's choices in implementing Rule 613 do reflect some trade-offs. One factor in the ability of the Advisory Committee to collect relevant information for the Operating Committee is the quality and depth of the expertise, and the diversity of viewpoints, of the Advisory Committee's membership. A larger and more diverse Advisory Committee may have better access to expertise and diversity of viewpoints from among members for use in advising the Operating Committee. But, members of a larger and more diverse Advisory Committee would face potentially greater difficulties in working among themselves to identify and convey the information that is available to them. The Plan balances these considerations by providing the Advisory Committee with sufficient membership to be able to generate useful information and advice for the Operating Committee, while being at a sufficiently low size and diversity level to permit the members to be able to work together without undue obstacles that could otherwise limit the Advisory Committee's effectiveness in conveying their views.

In a role similar to that of the Advisory Committee, outsiders on corporate boards of directors can bring expertise and independence to board actions, thereby enhancing board effectiveness. Trade-offs in determining the optimum size and composition of boards is the subject of extensive academic research. For example, Lehn, Kenneth, Sukesh Patro, and Mengxin Zhao, 2009, “Determinants of the size and structure of corporate boards: 1935-2000,” Financial Management, 747-780, consider the size and composition of the board to be determined by trade-offs associated with the information the directors bring to boards, which facilitate their monitoring and advisory role, and the coordination costs and free-rider problems associated with their presence. Harris, Milton and Raviv, Artur, 2008, “A Theory of Corporate Control and Size,” 21 Review of Financial Studies, 1797-1832, model the trade-off between benefits of greater expertise that outside directors bring versus the costs of an aggravated free-rider problem to arrive at the optimum number of outside directors on the board. Collective-action and communication problems can limit the effectiveness of a board as it gains members as explored by Harris and Raviv (2008) and Lehn, Patro, and Zhao (2009), in addition to Raheja, Charu, 2005. “Determinants of Board Size and Composition: A Theory of Corporate Boards,” 40 Journal of Financial and Quantitative Analysis, 283-306, and Yermack, David, “Higher Market Valuation for Firms with a Small Board of Directors,” Journal of Financial Economics, XL (1996), 185-211; see also Jerayr Haleblian and Sydney Finkelstein, “Top Management Team Size, CEO Dominance, and Firm Performance: The Moderating Roles of Environmental Turbulence and Discretion,” The Academy of Management Journal, Vol. 36, No. 4 (August, 1993), 844-863.

For related literature that expressly examines trade-offs and consequences of “diverse” boards, see Baranchuk, Nina, and Phil Dybvig, 2009, “Consensus in diverse corporate boards,” Review of Financial Studies 22(2), 715-747; and Malenko, Nadya, 2014, “Communication and Decision-Making in Corporate Boards,” Review of Financial Studies 27(5), 1486-1532.

Another factor that may bear on the Advisory Committee's ability to assemble a diverse range of views is the Plan's provisions that Advisory Committee members sit in their individual capacity, rather than as a representative of their employer. This may give Advisory Committee members greater freedom to speak to issues common to similarly-situated entities (e.g., large broker-dealers), rather than potentially-idiosyncratic views of the individuals' employers, which broader views in turn could better inform the Operating Committee about issues or impacts associated with the operation of the CAT.

Another factor in the ability of the Advisory Committee to advise the Operating Committee is whether the Advisory Committee, having assembled a diverse set of views, could effectively communicate those views to the Operating Committee. Two Plan provisions, relating to the staggering of member terms and the limits on participation of the Advisory Committee under Rule 613(b)(7)(ii), bear on this communication.

See CAT NMS Plan, supra note 3, at Section 4.13(b) and (c).

First, the Plan provides for Advisory Committee members to serve for staggered three-year terms in order to provide “improved continuity given the complexity of CAT processing.” Staggering of terms would prevent the entire Advisory Committee or large numbers of its members from turning over in any given year, which could enhance the cohesion of the Advisory Committee, and thereby its effectiveness in communicating member viewpoints to the Operating Committee. Second, the Plan gives the Advisory Committee varying roles with respect to the different actions to be taken by the Operating Committee. While the Advisory Committee members may attend meetings and submit views to the Operating Committee on matters prior to a decision by the Operating Committee, the Operating Committee may exclude Advisory Committee members from Executive Sessions.

See id. at Section 4.13(c); Appendix C, Section D.11(b) (“Governance of the CAT . . . Industry Members also recommended a three-year term with one-third turnover per year . . . to provide improved continuity given the complexity of CAT processing.”).

See CAT NMS Plan, supra note 3, at Section 4.13(d).

An additional factor that bears on the ability of the Advisory Committee to advise the Operating Committee is a feedback loop: Whether the Advisory Committee could receive sufficiently detailed information on the operations of the Plan so that the Advisory Committee members can, in turn, provide decision-useful information to the Operating Committee. Here, the Plan specifies that the Advisory Committee has the right to receive from the Operating Committee information necessary and appropriate to the fulfillment of its functions, but that the scope and content of the information is to be determined by the Operating Committee. Thus, the Commission notes that the Operating Committee could act to limit the effectiveness of the Advisory Committee—for example, if the Operating Committee were to fail to provide Advisory Committee members with notice of the items to be deliberated and voted upon by the Operating Committee with sufficient time and particularity for the Advisory Committee to be able to adequately fulfill its function, or fail to provide other pathways for Advisory Committee members to become aware of topics of interest or concern to the Operating Committee.

Id. at Section 4.13(e).

One other determinant bears on the effectiveness of the Advisory Committee in ensuring that the Operating Committee makes decisions in light of diverse information—whether the Operating Committee actually takes into account the facts and views of the Advisory Committee before making a decision. Although the Plan expressly provides for Advisory Committee input, it does not contain a mechanism—such as requiring the Operating Committee to respond to the Advisory Committee's views, formally or informally, in advance of or following a decision by the Operating Committee—to ensure that the Operating Committee considers the views of the Advisory Committee as a part of the Operating Committee's decision-making process.

(3) Conclusion

The Commission preliminarily believes that the governance provisions discussed above, which the Commission identified as being “important to the efficient operation and practical evolution of the [CAT], and . . . responsive to many commenters' concerns about governance structure, cost allocations, and the inclusion of SRO members as part of the planning process,” could help promote better decision-making by the relevant parties. These provisions thus could mitigate concerns about potential uncertainty in the economic effects of the Plan by giving the Commission greater confidence that its expected benefits would be achieved in an efficient manner and that costs resulting from inefficiencies would be avoided.

4. Request for Comment on the Benefits

The Commission requests comment on all aspects of the discussion of the potential benefits of the CAT NMS Plan. In particular, the Commission seeks responses to the following questions:

257. Do Commenters agree with the Commission's assessment of the potential benefits of the CAT NMS Plan? Why or why not?

258. To what extent do the uncertainties related to future decisions about Plan implementation impact the assessment of potential benefits of the Plan? Please explain.

259. Do Commenters agree that the inclusion of the data fields in one centralized data source in the CAT NMS Plan described above would result in more complete data than what is currently available to regulators? Which elements of the Plan would deliver improvements to completeness? Are there any elements of the Plan that would degrade the completeness of regulatory data? Please explain.

260. The Commission reviewed gap analyses that examine whether the CAT Data would contain all important data elements in current data sources and concluded that certain information is not included (e.g., OATS data fields that allow off-exchange transactions to be matched to their corresponding trade reports at trade reporting facilities and certain EBS elements). Please identify any such data elements that are missing under the Plan.

See SEC Rule 613—Consolidated Audit Trail (CAT) OATS—CAT Gap Analysis and SEC Rule 613—Consolidated Audit Trail (CAT) Revised EBS—CAT GAP Analysis, available at http://www.catnmsplan.com/gapanalyses/index.html . The Commission acknowledges that the Participants are continuing to study gaps between current regulatory data sources and the Plan as filed. CAT NMS Plan , supra note 3, at Appendix C, Section C.9

261. The Commission also seeks comment on the significance of the gaps identified in the analyses. If there are particular fields that are identified in the gap analyses that should not be incorporated into CAT, please identify them and explain.

262. The Commission expects that, pursuant to the requirements of the Plan, any missing elements that are material to regulators would be incorporated into CAT Data prior to the retirement of the systems that currently provide those data elements to regulators. Do you agree? Why or why not? Do you agree that CAT Data would include the audit trail data elements that currently exist in audit trail data sources? Why or why not?

The Plan requires that, prior to the retirement of existing systems, the CAT Data must contain data elements sufficient to ensure the same regulatory coverage provided by existing systems that are anticipated to be retired. See CAT NMS Plan, supra note 3, at Appendix D, Section 3.

263. Do Commenters agree that the CAT NMS Plan would improve the accuracy of the data available to regulators? Which elements of the Plan would deliver these improvements? Are there any elements of the Plan that would degrade the accuracy of regulatory data relative to today? Are there any elements of the Plan that would prevent or limit improvements in the accuracy of regulatory data? Are the provisions of the Plan related to accuracy appropriate and reasonable in light of the goal of improving data quality? Please explain.

264. Do Commenters believe that procedural protections in the Plan, such as the requirement that the Technical Specifications be “consistent with [considerations and minimum standards discussed in] Appendices C and D,” the requirement to provide the initial Technical Specifications and any Material Amendments thereto to the Operating Committee for approval by Supermajority Vote, and the requirement that all non-Material Amendments and all published interpretations be provided to the Operating Committee in writing at least ten days before publication, can mitigate uncertainty regarding future decisions and help promote accuracy? Please explain.

Id. at Section 6.9(a). The Commission notes that the standards in Appendices C and D do not cover all decisions that would affect the accuracy of the data.

See CAT NMS Plan, supra note 3, at Section 6.9(c)(i).

265. Do Commenters believe that the Error Rate, validations, and error resolution processes described in the CAT NMS Plan would provide improvements in accuracy? Are these processes appropriate and reasonable in light of the goal of improving data quality? Please explain.

266. The Plan specifies an error correction process after initial reports are received and indicates that practically all errors identifiable by the validations used in the error correction process would be corrected by 8:00 a.m. Eastern Time on day T+5, stating that errors are expected to be “de minimis” after the error correction period. Do Commenters believe that this is a reasonable conclusion? Please explain.

See id. at Appendix C, Section A.3(b) n.102. “De minimis” is not defined and no numerical Error Rate is given. The Plan also includes a compliance program intended to help achieve this goal.

267. Do Commenters believe that the provisions in the CAT NMS Plan related to event sequencing would provide improvements in accuracy? To what degree does the 50 millisecond clock synchronization requirement enable or prevent regulators' ability to sequence events that occur in different Execution Venues? Are the provisions of the Plan related to event sequencing appropriate and reasonable in light of the goal of improving data quality? Please explain.

268. The Plan does not specify the approach that would be used to sequence events when time stamps are identical. Do Commenters believe that there is a way for the Plan Processor to sequence events with identical time stamps? How would this process, or the lack of a process, affect the quality of the CAT Data?

269. The Plan states that “the Participants plan to require that the Plan Processor develop a way to accurately track the sequence of order events [of a particular order] without relying entirely on time stamps.” Do Commenters believe it is feasible to properly sequence the events of a simple or complex order without relying entirely on time stamps? Please explain. If such a procedure could be developed, how accurate would it be?

See CAT NMS Plan supra note 3 at Appendix C, Section A.3(c).

270. The Plan further states, “For unrelated events, e.g., multiple unrelated orders from different broker-dealers, there would be no way to definitively sequence order events within the allowable clock drift as defined in Article 6.8.” Do Commenters believe it would be feasible for the Plan Processor to develop a way to accurately sequence such unrelated orders given the time stamp and clock synchronization requirements of the Plan? Please explain. If such a procedure could be developed, how accurate would it be?

See id. at Appendix C, Section A.3(c) n.110.

271. Do Commenters agree with the Commission's data analysis of the clock synchronization improvements from the Plan? If not, how could the Commission improve the data analysis? Do Commenters have their own data analysis that informs on the expected improvements from the Plan? If so, please provide. Do Commenters agree that the improvements to the percentage of sequenceable order events by Plan standards are modest and the requirements of the Plan may not be sufficient to completely sequence the majority of market events relative to all other events?

272. Do Commenters agree with the Plan's assessment of the industry standard for clock synchronization? Does this reflect the standards for all CAT Reporters, including exchanges, ATSs, and other broker-dealers? If not, what would be a more appropriate way to define the industry standard for clock synchronization?

273. Do Commenters believe that the provisions in the CAT NMS Plan related to linking data would result in improvements to the accuracy of the data available to regulators? Would the process for linking orders across market participants and SROs improve accuracy compared to existing data? Would the Plan Processor be able to develop expertise in linking data more efficiently than the regulatory staff members from each entity could on their own? Please explain.

274. Would the Error Rates associated with the linking process represent improvements to data accuracy? Would Approach 1 to data conversion result in a lower Error Rate than Approach 2? Would the Approach affect the Plan Processor's ability to build a complete and accurate database of linked data? Are the Error Rates associated with the linking process appropriate and reasonable in light of the goal of improving data quality? Please explain.

275. Do Commenters believe that the inclusion of unique Customer and Reporter Identifiers would increase the accuracy of information in data regulators use and provide benefits to a broad range of regulatory activities that involve audit trail data? Please explain.

276. Do Commenters agree that the CAT Data would provide less aggregated allocation information and less aggregated issuer repurchase information? Why or why not? Would these changes significantly affect regulatory activities?

277. Do Commenters agree that the CAT NMS Plan would improve the accessibility of the data available to regulators? Which elements of the Plan would deliver these improvements? Are there any elements of the Plan that would degrade the accessibility of regulatory data relative to today? Are there any elements of the Plan that would prevent or limit improvements in the accessibility of regulatory data?

278. Do Commenters believe that the minimum requirements for direct access ensure that the Plan would improve access to current data, including the process of requesting data? Would the direct access facilitated by the Plan provide sufficient capacity and functionality? Would direct access reduce the number of ad hoc data requests?

279. Do Commenters agree that the CAT NMS Plan would improve the timeliness of the data available to regulators? Which elements of the Plan would deliver these improvements? Are there any elements of the Plan that would degrade the timeliness of regulatory data relative to today? Are there any elements of the Plan that would prevent or limit improvements in the timeliness of regulatory data?

280. Do Commenters believe that the CAT NMS Plan will facilitate the ability of each national securities exchange and national securities association to comply with the requirement in Rule 613(f) that they develop and implement a surveillance system, or enhance existing surveillance systems, reasonably designed to make use of the consolidated information contained in the consolidated audit trail? If not, why not?

281. Do Commenters agree that the CAT NMS Plan will facilitate the ability of regulators to conduct risk-based examinations? Why or why not? How significantly would the Plan improve risk-based examinations? Please explain.

282. Do Commenters agree that the CAT NMS Plan will improve the efficiency of regulators' enforcement activities? Why or why not? Which specific regulatory activities would be most improved by the CAT NMS Plan? Please explain.

283. Do Commenters agree that the CAT NMS Plan will improve the ability for regulators to determine the credibility of tips and complaints? Please explain.

284. Overall, do Commenters agree that the surveillance, examination, and enforcement activities of regulators would improve with the CAT NMS Plan? Please explain. Would these improvements be significant enough to deter violative behavior? Please explain. What would be the economic effect of this deterrence?

285. Would such improvements reduce the percentage of activities that generate false positives (i.e., detection of behaviors that are not violative) and/or reduce the percentage of activities that are false negatives (i.e., not detecting behaviors that are violative)? Please explain. What would be the economic effect of any changes in false positives or false negatives?

286. Do Commenters agree with the Commission's assessment of the economic effects of the improvements to surveillance, examinations, and enforcement from the CAT NMS Plan? Please explain.

287. Do Commenters agree that the CAT NMS Plan would improve the efficiency and effectiveness of regulators conducting analysis and reconstruction of market events? Please explain. Do Commenters agree with the Commission's assessment of the benefits to investors and the market of more efficient and effective analysis and reconstruction of market events? Please explain.

288. Do Commenters agree that the CAT NMS Plan would facilitate market analysis and research that would improve regulators' understanding of securities markets? Please explain. Do Commenters agree with the Commission's assessment of the benefits to investors and the markets from regulators having a better understanding of the markets? Please explain.

289. Do Commenters believe that there are other features of the CAT NMS Plan uniquely applicable to a consolidated audit trail that increase the likelihood that the potential benefits of the CAT NMS Plan would be realized? Please identify these features and explain.

290. Do Commenters agree that provisions of the Plan related to future upgrades, promoting accuracy, and promoting timeliness increase the likelihood that the potential benefits of the CAT NMS Plan would be realized? Do current regulatory data sources have provisions similar to ones the Commission analyzed? If so, please describe such provisions.

291. Do Commenters believe that provisions of the Plan provide incentives to reduce reporting errors for a CAT Reporter that has an Error Rate that does not exceed the thresholds that would trigger fines under the Plan or possible enforcement actions by regulators? If so, what are the incentives? Could the Plan provide different incentives to reduce reporting errors? Please explain.

292. Under the Plan, proposals to adjust the maximum allowable Error Rate are to originate from the Plan Processor. Do Commenters agree with this approach? Please explain. Should others, such as the Operating Committee, or Advisory Committee be able to originate changes to the Error Rate? Please explain.

293. Do Commenters agree that communication of data feed delays for public consumption is beneficial to the operation and effectiveness of the CAT? If so, in what ways? What are the benefits and costs of such public disclosure?

294. Do Commenters agree that the governance provisions identified in the Rule 613 Adopting Release continue to be important to the efficient operation and practical evolution of the Plan, and therefore to the achievement of the Plan's benefits? Are there other aspects of the Plan's governance that might enhance (or detract from) the Plan's ability to achieve its intended benefits? Are there other governance aspects that the Plan does not address that might enhance, if included (or detract from, if not remedied) the Plan's ability to achieve its intended benefits? Please identify these other features and explain how they enhance (or detract from) the Plan's ability to achieve its intended benefits.

295. The Commission's analysis of the provisions of the Plan relating to voting assumes that these provisions will promote the benefits sought to be achieved by the Plan because, by assigning different voting thresholds to different actions, the Plan seeks to address potential conflict of interest and holdout problems, balancing dissenters' rights with the need to move forward with needed changes. Is this a complete and accurate list of the factors that could bear on whether the voting provisions of the Plan will promote the benefits sought to be achieved by the Plan, and did the Commission correctly weigh these factors in preliminarily concluding that the Plan's voting provisions could help promote better Plan decision-making and, thus, improve achievement of the Plan's goals? If the Commission should have considered other factors or weighed the identified factors differently, please explain how, and what the costs and benefits of an alternative approach would be.

296. The Plan provides that “[a]ll votes by the Selection Committee shall be confidential and non-public.” What are the effects of confidential voting as a means of limiting conflicts of interest and promoting accountability? Would expanding confidentiality in voting to other situations help or hinder the effectiveness of the Operating Committee and its Subcommittees in achieving the regulatory objectives of the Plan? Please explain and provide supporting examples and evidence, if available.

See CAT NMS Plan, supra note 3, at Section 5.1(b)(v).

297. Do Commenters believe that the size, membership, and tenure of Advisory Committee members is appropriately tailored to encourage the effective accumulation and communication of Advisory Committee member views to the Operating Committee, thereby improving Plan decision-making? If not, why not? Are there other factors that could bear on whether the provisions of the Plan relating to the Advisory Committee will promote better decision-making? If so, what other factors?

298. Are there any alternatives for Advisory Committee involvement that could increase the effectiveness of its involvement? What benefits would these achieve in terms of improving the Operating Committee's efficiency? Would these alternatives increase or decrease costs?

299. What obstacles to information-sharing between individual members of the Operating Committee and the Commission, if any, are likely to limit the Plan's effectiveness in meeting its regulatory objectives? Is there any information, such as regarding individual SRO clock synchronization standards, that members would need to share within the Operating Committee to achieve plan regulatory objectives but may be uncomfortable sharing with one another (or more comfortable sharing with the Commission than with one another)? Please be specific and explain what, if any, changes to the plan could mitigate obstacles from inadequate information-sharing.

300. Are there any other factors relating to the operation and administration of the Plan that the Commission should consider as part of determining whether to approve the Plan? If so, what are those factors and how could they influence the costs and benefits of the Plan? Does the Plan currently address these factors? If not, how could the Plan address these factors and what would be the relative costs and benefits of any changes to the Plan?

F. Costs

As noted above, at the time of the Adopting Release the Commission deferred its economic analysis of the creation, implementation, and maintenance of CAT until after submission of the CAT NMS Plan. Accordingly, the Commission deferred its detailed analysis of costs associated with CAT. In light of the SROs having submitted the CAT NMS Plan, this Section sets forth the Commission's preliminary analysis of the expected costs for creating, implementing, and maintaining the CAT, as well as the associated reporting of data.

See Adopting Release, supra note 9, at 45789.

As discussed in detail below, the Commission has preliminarily estimated current costs related to regulatory data reporting, anticipated costs associated with building and maintaining the Central Repository, and the anticipated costs to report CAT Data to the Central Repository. These preliminary estimates are calculated from information provided in the CAT NMS Plan as well as supplemental information. Currently, the 20 Participants spend $154.1 million annually on reporting regulatory data and performing surveillance. The approximately 1,800 broker-dealers anticipated to have CAT reporting responsibilities currently spend $1.6 billion annually on regulatory data reporting. If the Plan is approved, the Commission preliminarily estimates that the cost of the Plan would be approximately $2.4 billion in initial aggregate implementation costs and $1.7 billion in ongoing annual costs. Furthermore, the Plan anticipates that market participants would have duplicative audit trail data reporting responsibilities for a period of up to a maximum of 2.5 years, preceding the retirement of potentially duplicative regulatory data reporting schemes. Duplicative audit trail data reporting could cost broker-dealers $1.6 billion per year or more and could cost the Participants up to $6.9 million per year. The Commission preliminarily believes that the primary component of costs for CAT's estimated annual costs would be the estimated aggregate broker-dealer data reporting costs of $1.5 billion per year, whereas the Central Repository build costs are preliminarily estimated by the Participants to be no more than $92 million, with annual operating costs of no more than $135 million.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(ii)(B)(1).

See Section IV.F.1.c(2), infra.

See Section IV.F.2, infra.

See id.

As explained in detail below, the Commission believes, however, that there is significant uncertainty surrounding the actual implementation costs of CAT and the actual ongoing broker-dealer data reporting costs if the Plan were approved. Methodology and data limitations used to develop these preliminary cost estimates could result in imprecise estimates that may significantly differ from actual costs. The Commission has used its best judgment, however, in obtaining and assessing available information and data to provide the analysis and estimates included in this Notice. The Commission is also requesting comment on the methodology and any additional data Commenters believe should be considered.

Furthermore, the Commission notes that because some CAT design decisions (such as setting forth detailed Technical Specifications) have been deferred until the selection of the Plan Processor, the associated cost uncertainties could cause the actual costs to vary significantly from the estimates set forth in this analysis.

The Commission notes that the cost estimates set forth in this analysis are updated from the cost estimates provided in the Proposing Release. In the Proposing Release, the Commission estimated $4.3 billion in initial implementation costs and $2.3 billion in ongoing annual costs. The Commission has now updated its analysis and estimates $2.4 billion in initial implementation costs and $1.7 billion in ongoing annual costs. The Commission believes that several factors drive differences in cost estimates from the Proposing Release to the current cost estimates in this analysis. First, the scope of CAT as contemplated in the Proposing Release is different than the scope of CAT Data as would be implemented by the CAT NMS Plan. For example, the Commission notes that, unlike CAT Data envisioned in the Proposing Release, the proposed Plan includes OTC Equity Securities, which if included in CAT would facilitate the possible retirement of OATS as an audit trail data reporting system at a relatively earlier date. While the Commission's cost estimates do not explicitly incorporate cost savings from systems retirement, cost estimates provided in the Plan and based on surveys of broker-dealers, participants and service providers may reflect some of these savings. For example, because respondents anticipate incorporating resources that would be devoted to OTC equity data reporting to CAT reporting, cost estimates may be lower than they would be if OTC equity data were excluded from CAT but were still reported to OATS on an ongoing basis. Thus, after all CAT Reporters start reporting to the Central Repository and the resolution of any data gaps between OATS and CAT, FINRA would not need to maintain OATS solely to fulfill its regulatory responsibilities relating to OTC Equity Securities. Additionally, the Commission's updated cost estimates are based on data submitted with the Plan, which was unavailable when the Commission first estimated the costs of CAT in the Proposing Release, as well as certain additional information obtained by Commission Staff. Furthermore, the Plan also integrates exemptive relief extended to the Participants regarding (1) Options Market Maker quotes; (2) Customer-IDs; (3) CAT-Reporter-IDs; (4) linking of executions to specific subaccount allocations on Allocation Reports; and (5) time stamp granularity for Manual Order Events. The Commission preliminarily believes that this exemptive relief contributes to reductions in cost of the Plan relative to those estimated in the Proposing Release. The Commission has incorporated this additional information into its current cost analysis.

See Proposing Release supra note 9, at 32596-602. The $4.3 billion and $2.3 billion cost estimates can be calculated using individual cost estimates from the Proposing Release. The Proposing Release expressed some cost estimates on a per-Participant basis. The Plan, however, breaks out costs to Participants by (i) single-exchange-operating Participants and (ii) Affiliated Participants that operate multiple exchanges. To validly compare the Commission's preliminary cost estimates to the cost estimates set forth in the Plan, the Commission's analysis aggregates costs to all Participants for these cost estimates. The Proposing Release anticipated 1,114 SRO members would report data to the Central Repository directly, and 3,006 broker-dealers would report data through a service provider. The Plan anticipates that approximately 1,800 broker-dealers would have CAT reporting obligations; the Commission preliminarily believes that the majority of these broker-dealers would rely on service bureaus to perform their regulatory data reporting. Again, to validly compare the different cost estimates, the Commission aggregates the cost estimates across all broker-dealer CAT Reporters.

Similarly, in the Adopting Release, the Commission explained that “the methodology that the Commission used in the Proposing Release to estimate the costs of creating, implementing, and maintaining a consolidated audit trail may no longer be suitable” and that certain “assumptions may no longer be valid since several of the specific technical requirements underlying the Proposing Release's approach have been substantially modified.” See Adopting Release, supra note 9, at 45781.

If FINRA were unable to retire OATS, the costs of duplicative reporting (discussed in Section IV.F.2, infra), would continue indefinitely. The Commission preliminarily believes this outcome is unlikely because the Plan discusses the Participants' plans to retire OATS if the Plan is approved. See CAT NMS Plan, supra note 3, at Appendix C, Section C.9.

See Proposing Release, supra note 9, at 32601-02.

As discussed further below, the Commission's analysis also incorporates data obtained from FINRA and information from discussions with broker-dealers and service bureaus arranged by FIF and staff. See infra notes 880 and 899.

The Commission's revised cost estimates are generally substantially lower than those presented in the Proposing Release. See Proposing Release, supra note 9, at 32601-02. The Proposing Release's estimate of total industry implementation costs is 40.45% higher than the current estimate, and the Proposing Release's estimate of ongoing total industry costs is 57.99% higher than the current estimate. Reductions in cost estimates are primarily driven by lower broker-dealer implementation and ongoing reporting costs that are largely attributable to a reduction in the number of broker-dealers anticipated to incur CAT reporting responsibilities, as the Proposing Release assumed that all 4,120 broker-dealers would be CAT Reporters but the Plan estimates that only 1,800 broker-dealers would incur CAT reporting responsibilities. The Proposing Release also presented higher estimates of the number of broker-dealers that are likely to be insourcers; these broker-dealers have significantly higher implementation and ongoing costs that outsourcing broker-dealers. The Proposing Release estimated Central Repository implementation costs that are 23.33% higher than current estimates; ongoing Central Repository costs were lower by 33.56%; SRO implementation costs were 82.21% higher in the proposing release; SRO ongoing costs were estimated to be 31.79% lower than current estimates. The Proposing Release did not recognize costs to Service Bureaus related to CAT.

1. Analysis of Expected Costs

The Plan provides estimates of the expected costs associated with the Plan, including costs to build and operate the Central Repository and costs to Participants and CAT Reporters to implement and maintain CAT reporting. As explained below, the Commission has thoroughly reviewed the cost estimates contained in the Plan and other relevant information to develop the Commission's preliminary estimate of expected costs of the Plan. The Commission preliminarily believes that in some cases the estimates provided in the Plan are reliable estimates of the potential costs of certain aspects of the Plan. The Commission preliminarily believes, however, that in other cases the data and methodology underlying certain Plan estimates are unreliable and, in such cases, the Commission has preliminarily evaluated and provided separate estimates based on alternative data or a different methodology.

Because the Plan does not provide data that permit partitioning costs associated with the Central Repository between Participants and broker-dealer CAT Reporters, this analysis discusses the Central Repository costs separately.

For example, the Commission preliminarily believes that cost estimates in the Plan relating to the costs that would be borne by broker-dealers are unreliable due to limitations of certain survey response data. These limitations and the Commission's alternative cost estimate are discussed in detail below. See Section IV.F.1.c, infra.

In this Section, the Commission provides preliminary estimates of the individual elements that constitute the estimated expected total cost associated with implementing and maintaining the CAT, including the costs of operating and building the Central Repository, the costs to Participants, the costs to broker-dealers, and other costs considered in the CAT NMS Plan.

a. Costs of Building and Operating the Central Repository

The Plan's estimates of the costs to build the Central Repository are based on Bids that vary in a range as high as $92 million. The Plan's estimates of annual operating costs are based on Bids that vary in a range up to $135 million. The eventual magnitude of Central Repository costs is dependent on the Participants' selection of the Plan Processor, and may ultimately differ from estimates discussed in the Plan if Bids are revised as the bidding process progresses. The Plan discusses these costs both as (i) one-time and ongoing costs as well as (ii) a five-year total cost, to help evaluate economic trade-offs between initial build costs and operating costs. The Plan anticipates that Participants and their members would bear the costs of building and operating the Central Repository. The Commission preliminarily believes that these estimates are reliable because they are the result of a competitive bidding process, although the Commission recognizes that the Bids are not legally binding on bidders. In particular, the Commission preliminarily believes that a Bidder would not likely decline a contract to be Plan Processor that was based on the Bid it submitted because that Bidder might lose future business due to reputational consequences of its actions. Furthermore, Bidders have invested considerable time and effort in evaluating the RFP and preparing their Bids and thus if a Bidder were unwilling to serve as Plan Processor according to the terms outlined in its Bid, the time and effort expended to prepare the Bid would be wasted resources. As explained further below, however, the Commission believes that these cost estimates associated with building and operating the Central Repository are subject to a number of uncertainties.

To estimate the one-time total cost to build the Central Repository, the Plan uses the Bids of the final six Shortlisted Bidders. The Bidders' implementation cost estimates range from $30 million to $91.6 million, with a mean of $53 million and a median of $46.1 million. The Plan also estimates the ongoing costs of the Central Repository. The Bids of the final six Shortlisted Bidders estimate annual costs to operate and maintain the Central Repository range from $27 million to $135 million, with a mean of $51.1 million and a median of $42.2 million. The Plan's summary statistics show that annual costs are not expected to be constant year-over-year for all Bidders, but the Plan does not provide further details on how the costs are expected to evolve over time or how many of the Bids have time-varying annual costs. Although the Commission preliminarily believes that costs provided by Bidders are reliable, the Commission recognizes that these ongoing costs could increase over time due to inflation or changes in market structure such as a significant increase in message traffic. It is also possible these costs could decrease due to improvements in technology, reductions in message traffic, and innovation by the Plan Processor.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(B). The Plan does not reflect any more specific cost ranges that result from narrowing the range of Bidders from six to three. See supra note 35.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(B).

Id.

Id.

The Plan also provides information based on the Bids on the total five-year operating costs for the Central Repository because the annual costs to operate and maintain the Central Repository are not independent of the build cost. In particular, it is plausible that the Bidders with the lowest build costs trade off lower build costs for higher recurring annual costs. To account for this possibility, the Plan presents the range of total five-year costs across Bidders using the Bids of the final six Shortlisted Bidders. The methodology takes the sum of the annual recurring costs over the first five years (discounted to the present with a discount rate of 2%) and adds the upfront investment. Across the six Shortlisted Bidders, the total five-year costs to build and maintain CAT range from $159.8 million to $538.7 million. This information is less granular than other Bidder cost information provided in the Plan, and no mean or median is provided or can be calculated with the information provided.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(B). The five-year presentation of Central Repository costs is converted into implementation and annual costs by using the maximum build cost and maximum annual operating cost over the five year period in the Bids. The Commission preliminarily believes that this presentation is conservative in the sense that it avoids underestimating the Central Repository costs that must be borne by industry. However, the Commission preliminarily believes that it is likely that this presentation overestimates the actual Central Repository costs because most individual Bids forecast variation in operating expenses year by year, with costs in some years lower than the maximum used in this presentation. Because the Central Repository costs are, in aggregate, significantly lower than the aggregate costs broker-dealers would incur in reporting CAT Data, the Commission preliminarily believes that this overestimation would not materially affect the magnitude of aggregate costs for the Plan to industry.

See supra note 836, and CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(B).

The Plan provides that costs associated with building and operating the Central Repository would be borne by both Participants and their members. In particular, the Plan provides for fixed-tiered fees based on ranges of activity levels to be levied on Execution Venues (i.e., the Participants (including a national securities association with trade reporting facilities, and ATSs)) based upon the Execution Venue's market share of share volumes, with options and equity venue fees determined by separate schedules set by CAT's Operating Committee. Furthermore, the Plan provides for fixed-tiered fees for Industry Members (broker-dealers) based on the message traffic generated by the member, including message traffic associated with an ATS operated by the member. The Plan also provides for the establishment of other fees for activities such as late, inaccurate, or corrected data submission by CAT Reporters. The Plan does not present information on the potential magnitude of these fees, but the Commission preliminarily believes they are likely to be a minor expense for CAT reporters, who should be able to avoid these fees by fulfilling their normal reporting responsibilities under the Plan. The Plan does not provide information on the relative allocation of these fees between transaction-based fees, message traffic-based fees, and other fees.

See CAT NMS Plan, supra note 3, at Section 11.

See id. at Section 11.3.

See id. at Section 11.3(b).

See id. at Section 11.3(c).

The economic analysis treats estimates of costs associated with building and operating the Central Repository separately from estimates of costs to Participants and other CAT Reporters to report CAT Data. While the costs of building and operating the Central Repository would be borne by the Participants and Industry Members, the allocation of the costs between and among those entities would be determined by the CAT Funding Model, which has not yet been finalized. See Section IV.C.2, supra. However, these costs are included in the Commission's estimate of the total costs to industry if the Plan is approved.

The Commission believes that a range of factors would drive the ultimate costs associated with building and operating the Central Repository and who would bear those costs. The Plan explains that the major cost drivers identified by Bidders are (1) transactional volume, (2) technical environments, (3) likely future growth in transactional volumes, (4) data archival requirements, and (5) user support/help desk resource requirements. The Plan does not present information on how sensitive the cost estimates are to each of these factors. Further, how Bidders propose to satisfy the RFP requirements could materially affect the ultimate cost to the industry to operate the Central Repository and who would bear those costs. For instance, some Bids may provide more extensive user support from the Plan Processor than others, effectively shifting user support costs from CAT Reporters to the Plan Processor, where such support might be more efficiently provided. However, the Plan does not provide information about how the Bidders propose to address each of the RFP requirements; thus, uncertainties exist around who would bear certain costs and how such costs could change if each Bidder's proposal related to these factors change.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(B).

The Commission is mindful that the cost estimates associated with building and operating the Central Repository are subject to a number of additional uncertainties. First, the Participants have not yet selected a Plan Processor, and the Shortlisted Bidders have submitted a wide range of cost estimates for building and operating the Central Repository. Second, the Bids submitted by the Shortlisted Bidders are not yet final. Participants could allow Bidders to revise their Bids before the final selection of the Plan Processor. Third, neither the Bidders nor the Commission can anticipate the evolution of technology and market activity with complete prescience. Available technologies could improve such that the Central Repository would be built and operated at a lower cost than is currently anticipated. On the other hand, if anticipated market activity levels are materially underestimated, the Central Repository's capacity could need to increase sooner, increasing the actual costs to operate the Central Repository than currently anticipated in the Bids. The Commission notes that costs to build and operate the Central Repository are relatively small compared to total industry costs if the CAT NMS Plan were approved; consequently, the Commission preliminarily believes that these uncertainties are unlikely to materially affect the final cost of the Plan to industry, if it is approved.

b. Costs to Participants

The Commission preliminarily believes that the Plan's estimates for Participants to report CAT Data are reliable because all of the SROs provided cost estimates, and most SROs have experience collecting audit trail data as well as expertise in the both the requirements of CAT as well as their current business practices. The Plan provides estimated costs for the Participants to report CAT Data. These estimates are based on Participant responses to the Costs to Participants Study (“Participants Study”) that the Participants collected to estimate SRO CAT-related costs for hardware and software, full-time employee staffing (“FTE costs”), and third-party providers. Respondents to the Participants Study also estimated the costs associated with retiring current regulatory data reporting systems that would be rendered redundant by CAT.

See id. at Appendix C, Section B.7(b)(iii)(B)(2). In addition to the costs the Participants would incur implementing and maintaining CAT, the Participants would also incur and would continue to incur costs associated with developing the CAT NMS Plan. The Participants estimate such costs to be $8,800,000. The Commission does not include these costs in its estimates of the costs associated with CAT if the CAT NMS Plan is approved because these costs have already been incurred and would not change regardless of whether the Commission approves or disapproves the CAT NMS Plan. Further, the Commission assumes that the CAT NMS Plan's implementation cost estimates include any additional CAT NMS Plan development costs that would be incurred by Participants if the CAT NMS Plan were approved.

The Participants Study delineates Participant responses into two groups. The first group consists of affiliated Participants, which includes single entities that hold self-regulatory licenses for multiple exchanges. The second group consists of Participants that hold a single self-regulatory license, including FINRA, the sole national securities association. Id. at Appendix C, Section B.7(b)(i)(A)(1).

Third-party provider costs are generally legal and consulting costs but may include other outsourcing. The template used by respondents is available at http://catnmsplan.com/PastEvents/ under the Section titled “6/23/14” at the “Cost Study Working Template” link.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(B)(2).

The Plan estimates costs for the Participants as an aggregate across all Participants (the six single-license Participants and the five Affiliated Participant Groups). The implementation cost estimate for Participants is $17.9 million, including $770,000 in legal and consulting costs and $10.3 million in full-time employee costs for operational, technical/development, and compliance-type functions. Annual ongoing costs are estimated to be $14.7 million, including $720,000 in legal and consulting costs and $7.3 million in full-time employee costs. Other than legal and consulting costs and full-time employee costs, the Plan does not specify the other categories of implementation and ongoing costs, but based on discussion with the Participants, the Commission preliminarily believes that much of the remaining costs would be attributed to IT infrastructure, including hardware and software costs.

Id. at Appendix C, Section B.7(b)(iii)(B)(2).

Id.

Id.

The Plan also provides estimates of the costs Participants currently face in reporting regulatory data. The Plan anticipates that some, but not all, of these reporting systems would be retired after implementation of the Plan. The Plan reports that aggregate annual costs for current regulatory data reporting systems are $6.9 million across all Participants.

Id.

Id. As required by Rule 613(a)(1)(ix), 17 CFR 242.613(a)(1)(ix), the CAT NMS Plan includes a plan to eliminate existing rules and systems that would be rendered duplicative under CAT. Id. at Appendix C, Section C.9. Among other things, this plan requires that within 18 months after Industry Members are required to begin reporting data to the Central Repository, each Participant will complete an analysis of whether its rules and systems related to monitoring quotes, orders, and executions collect information that is not rendered duplicative by CAT. Id. Each Participant must also analyze whether any such non-duplicative information should continue to be separately collected, incorporated into CAT, or terminated. Id. Therefore, depending on the results of these analyses, some existing regulatory reporting systems may continue to be in place after the implementation of CAT.

Id. at Appendix C, Section B.7(b)(ii)(B)(1).

In addition to data reporting costs, Participants face costs associated with developing and implementing a surveillance system reasonably designed to make use of the information contained in CAT Data as required by Rule 613(f). The Plan provides estimates of the costs to Participants to implement surveillance programs using data stored in the Central Repository. Participants would incur expenses, including full-time employee (“FTE”), legal, consulting and other costs to adapt their surveillance systems to utilize data in the Central Repository. The Plan provides an estimate of $23.2 million to implement surveillance systems for CAT, and ongoing annual costs of $87.7 million. The Plan does not provide information on why Participants' data reporting costs would substantially increase if the Plan were approved, nor does it provide information on why surveillance costs would decrease.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(B)(2). Rule 613 requires the SROs to file updated surveillance plans within 14 months of CAT implementation. 17 CFR 242.613(f). The Commission assumes that the CAT NMS Plan's estimate is limited to adapting current surveillance programs to the Central Repository. The Commission believes this is a conservative assumption because if other expenses were included in the estimate, the Commission would be overestimating the costs Participants would incur to implement and operate CAT if the CAT NMS Plan is approved.

The Commission preliminarily believes the data reporting cost estimates are reasonable because the Commission expects that Participants would be required to implement new technology infrastructure to report data to the Central Repository and support specialized personnel to maintain this infrastructure and respond to inquiries from the Plan Processor and users of CAT Data. The Commission likewise preliminarily believes that the surveillance cost estimates are reasonable, even though the annual estimate of $87.7 million is lower than the $147.2 million Participants, in aggregate, currently spend on surveillance programs annually because Participants could realize efficiencies from having data standardized and centrally hosted that could allow them to handle fewer ad hoc data requests. In addition, the Plan could allow Participants to automate some surveillance processes that may currently be labor intensive or processed on legacy systems, which could reduce costs because the primary driver of these costs is FTE costs.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(ii)(B)(1).

Table 6 summarizes the Participants' estimated costs, both current and CAT-related, that are set forth in the Plan. Currently, Participants spend approximately $154 million per year on data reporting and surveillance activities. The Participants estimate that they would incur $41 million in CAT implementation costs, and $102 million annually in ongoing costs to report CAT Data and perform surveillance as mandated under Rule 613.

Table 6—Participants' Cost Estimates

Current CAT implementation CAT ongoing
Data Reporting $6,900,000 $17,900,000 $14,700,000
Surveillance 147,200,000 23,200,000 87,700,000
Total 154,100,000 41,100,000 102,400,000

c. Costs to Broker-Dealers

(1) Estimates in the Plan

The Plan estimates total costs for those broker-dealers expected to report to CAT. In particular, the Plan relies on the Costs to CAT Reporters Study (“Reporters Study”), which gathered from broker-dealers the same categories of cost estimates used in the Participants Study—i.e., the hardware and software costs, full-time employee staffing costs, and third-party provider costs that CAT Reporters would incur if the Commission approves the Plan. The Reporters Study surveyed broker-dealers to respond to two distinct approaches for reporting CAT Data to the Central Repository. Approach 1 assumes CAT Reporters would submit CAT Data using their choice of industry protocols. Approach 2 assumes CAT Reporters would submit data using a pre-specified format. The Participants distributed the Reporters Study to 4,406 broker-dealers and received 422 responses, of which the Participants excluded 180 deemed materially incomplete and 75 determined to be erroneous. The Plan's cost estimate calculations are based on the remaining 167 responses. In aggregating the cost estimates across all broker-dealers expected to report CAT Data to the Central Repository, the Plan assumed that the characteristics of survey respondents (firm size and OATS reporting status) were representative of the approximately 1,800 broker-dealers expected to have CAT reporting obligations.

See id. at Appendix C, Section B.7(b)(i)(A)(2).

See id.

See id.

Not all broker-dealers are expected to have CAT reporting obligations; the Participants report that approximately 1,800 broker-dealers currently quote or execute transactions in NMS Securities, Listed Options or OTC Equity Securities and would likely have CAT reporting obligations. The Commission understands that the remaining 2,338 registered broker-dealers either trade in asset classes not currently included in the definition of Eligible Security or do not trade at all (e.g., broker-dealers for the purposes of underwriting, advising, private placements). The Plan describes the process of determining that 1,800 broker-dealers would report to the Central Repository in Appendix C. See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(ii)(B)(2).

Based on the Reporters Study survey data, the Plan estimates implementation costs of less than $740 million for small firms and approximately $2.6 billion for large firms, for a total of $3.34 billion in implementation costs for broker-dealers. For annual ongoing costs, the Plan estimates costs of $739 million for small firms and $2.3 billion for large firms, for a total of $3.04 billion in annual ongoing costs for broker-dealers. For both large and small broker-dealers, the Plan suggests that the primary cost driver for projected CAT reporting costs for broker-dealers is costs associated with full-time employees. For the reasons discussed below, the Commission preliminarily believes that the broker-dealer cost estimates in the Plan are in part unreliable, based on limitations with the Plan's underlying data in estimating costs. As discussed below, the Commission preliminarily believes that cost estimates in the Plan for large broker-dealers may be reliable, and the Commission has incorporated large firm data from the Plan into the Commission's estimates outlined below.

Survey respondents were instructed to classify themselves as “small” if their Total Capital (defined as net worth plus subordinated liabilities) was less than $500,000. See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(ii)(C) n.188. This is consistent with the definition of “small business” or “small organization” used with reference to a broker or dealer for purposes of Commission rulemaking in accordance with provisions of Chapter Six of the Administrative Procedure Act (5 U.S.C. 601 et seq.). See 17 CFR 240.0-10(c).

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iv)(A)(3).

Id.

See id. at Appendix C, Section B.7(b)(iii)(C)(2).

While the estimates presented in the Plan assume that the proportion of large versus small broker-dealers that responded to the Reporters Study is representative of the relative number of large versus small broker-dealers that are expected to incur CAT reporting obligations, the Commission's cost estimates do not embed this assumption. Instead, the Commission relies on data from FINRA to determine which firms are likely to outsource, and models those firms' costs based on information gleaned from FIF-organized discussions with industry. This is discussed further below, but this estimation results in relatively fewer firms' costs being estimated using “large” firm cost estimates presented in the Plan.

The Commission preliminarily believes, however, that the cost estimates for small broker-dealers provided in the Plan, which are based upon responses set forth in the Reporters Study, do not prove reliable estimates of smaller CAT Reporter costs for a number of reasons. First, some respondents classified as small in the Reporters Study appear to have responded numerically with incorrect units, with such responses resulting in annual estimated cost figures that would be 1,000 times too large. Second, maximum responses in certain categories of costs suggest that some large broker-dealers may have misclassified themselves as small broker-dealers. Third, methods used to remove outliers are likely to have introduced significant biases. Finally, the response rate to the Reporters Study survey was low and is likely to have oversampled small broker-dealers who currently have no OATS reporting obligations.

The Plan presents summary statistics such as average, median and maximum for each survey response. See CAT NMS Plan supra note 3, at Appendix C, Section B(7)(b)(ii)(C), Table 5. In the left most column, $14 million is the maximum response for “Hardware/Software Current Cost.”

In reaching these preliminary conclusions, the Commission reviewed the detailed discussions of the Reporters Study survey methodology in the Plan and the survey form and instructions provided to respondents. See 6/23/14 entry on CAT NMS Plan Web site, available at http://www.catnmsplan.com/pastevents/index.html . The Commission staff also discussed with the Participants potential methodology adjustments in aggregating the CAT Reporters Study data. After Commission staff discussions with the Participants, the Commission concluded that no methodology could address these fundamental issues with the survey data.

First, the Commission preliminarily believes that the respondents to the Reporters Study survey are likely to have used different units in their responses and that the survey precision is materially affected because inconsistent use of reporting units across respondents introduces an upward bias to the Reporters Study's findings. The survey collected cost estimates in $1,000 increments; however, there is evidence that some respondents did not provide estimates in $1,000 increments as requested. Rather, survey results in the Plan reveal, for example, that one small firm reported current annual hardware/software costs for current regulatory data reporting to be $14,000,000 per year. Because small firms responding to the survey by definition have no more than $500,000 in total capital, an annual $14,000,000 estimate for hardware/software costs for current data reporting seems unreasonable. Furthermore, a small survey respondent cited $3,500,000 in hardware/software retirement of systems costs, which seems unreasonable for a broker-dealer with less than $500,000 in total capital. These are only a few examples, but they raise the question of how many other respondents recorded incorrect units in their responses, particularly if screening methodologies have difficulty detecting such incorrect units. In light of these unreasonable results, the Commission preliminarily believes that the Plan's cost estimates for small broker-dealers reporting data to CAT has an upward bias because some firms did not correctly respond to the survey in $1,000 increments.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.(7)(b)(ii)(C), Table 5.

The Plan notes that it is possible that the firm intended to report that it had $14,000 in annual expenses for hardware/software. See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(ii)(C), n.193.

Because of errant responses of this type, the Plan recommends using medians instead of averages; however, for nearly all estimated cost categories in the Reporters Study, the median response was zero, which the Commission believes underestimates the costs that CAT Reporters are likely to face in most categories of costs. Consequently, the Commission is unable to adjust for these biases.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7.(b)(ii)(C), n.194.

In addition, the Commission preliminarily believes that the small firm cost survey information in the Reporters Study is unlikely to be representative of the small broker-dealers that would have CAT reporting responsibilities in part because the Commission also believes preliminarily that some survey respondents misclassified their firm's size, which renders the Plan's separate presentation of results for large and small broker-dealers imprecise. In particular, the Commission believes that at least one large firm misclassified itself as a small firm. The CAT NMS Plan Table 6 reveals that one firm designated as a small firm responded to the Reporters Study survey with it having 68 full-time employees dedicated to performing regulatory data reporting activities for a yearly cost of $27,300,000. The Commission believes, however, that a firm with 68 full-time employees reporting regulatory data could not be small (again, as defined by the survey to include firms with less than $500,000 in total capital) because such a firm would lack the working capital to support that level of employee expense. The presence of large-firms with significantly higher costs in the small-firm sample significantly biases the small-firm cost estimates upward.

See id. at Appendix C, Section B.(7)(b)(ii)(C), Table 6.

See id. at Appendix C, Section B.7.(b)(i)(C), n.188.

Moreover, the Commission preliminarily believes that the methodologies implemented to remove outliers in the Reporters Study introduce cost estimate biases. Based on discussions with the Participants, the Commission understands that to identify and remove outliers, the Participants first determined if each survey item's maximum response was a potential outlier because it was more than twice the value of the next highest response; the Participants then individually reviewed potential outliers and omitted those deemed errant. While the Commission recognizes that this methodology may mitigate the precision bias discussed above by removing a single response that is 1,000 times too high, it may not remove such outliers when two or more firms errantly report values 1,000 times too high, in which case an upward bias to the cost estimates would remain. Furthermore, if one firm genuinely incurs expenses that are more than twice those of the next highest respondent, such survey response might be removed under this methodology, even though such a response may accurately identify expenses expected by the respondent, which in turn introduces a downward bias to the cost estimates. For example, only 21 large OATS reporting firms are represented in the Reporters Study survey responses. If most of these 21 firms perform the majority of their regulatory data reporting functions in house, but one firm outsources all of its regulatory data reporting, that single firm could have outsourcing costs far higher than its peers. Under the Plan's cost estimate methodology, this outsourcing response in the Reporters Study might be removed as an outlier, unless another large, OATS reporting firm responded to the Reporters Study with at least half of the outsourcing costs. The Commission considered whether to request that the Participants provide updated cost estimates under a methodology that did not remove Reporters Study outlier responses, but the Commission preliminarily believes that this approach would exacerbate the precision problem discussed above and possibly increase the number of errant responses that are 1,000 times too high to the cost estimate data set.

See id. at Appendix C, Section B.7.(b)(i)(B)(ii)(C).

Finally, the Commission believes that the Reporters Study response rate is not adequate to be representative of the population of broker-dealers that would report to CAT. The survey was delivered to 4,025 broker-dealers. After removing erroneous and materially incomplete responses, only 167 responses remained of the 4,025 broker-dealers who were sent the survey. To be representative of the broker-dealers that would report to CAT, a final response rate of 4.15% seems low considering the diversity of these broker-dealers. The majority of broker-dealers are small and smaller broker-dealers are diverse along many dimensions relevant to the likely magnitude of their expected CAT costs, including business practices; tendency to centralize technology; specialization in market segments, such as options versus equities; and the range of products and markets in which individual broker-dealers participate. Because broker-dealer diversity is great, a survey of expected broker-dealer costs would ideally have a higher response rate to ensure a representative sample. Furthermore, of the 167 responses incorporated into the Plan's cost estimates, 118 respondent firms were classified as small in the Reporters Study, and 88 of these 118 small firms were identified as having no current OATS reporting responsibilities. The Commission preliminarily believes that small firms that anticipate limited CAT reporting responsibilities may have been oversampled by the Reporters Study survey because for nearly all categories of cost estimates, the median small firm response was zero, suggesting that they do not expect to have CAT reporting responsibilities. Consequently, the Commission preliminarily believes that the small firms that responded to the study cannot be statistically representative of the small firms that would incur CAT reporting obligations, because the Commission believes that most small broker-dealers would incur significant costs in reporting to CAT. These costs are estimated below.

Small firms may have no OATS reporting responsibilities because they do not engage in activities that would incur OATS reporting obligations, or they may be excluded or exempted under FINRA's OATS reporting rules. See Section IV.D.2.b(1)A, supra.

The Commission notes that small firms currently excluded from OATS reporting due to their size would have CAT reporting responsibilities under the Plan because the Plan makes no provision to exempt or exclude them, as FINRA does with OATS reporting. The Commission preliminarily believes that these firms are likely to experience higher implementation costs than other small firms because CAT reporting would likely necessitate establishing business relationships with service providers if they do not already have such relationships. The Commission preliminarily believes that most small firms that would have CAT reporting obligations but do not currently have OATS reporting obligations would not have the IT and regulatory personnel infrastructure to accomplish this reporting in-house. The Commission's estimation of these firms' costs to implement CAT includes higher estimates of employee costs to implement CAT to account for this increased burden.

Although the Commission has preliminarily concluded that the small broker-dealer cost estimates presented in the Plan are unreliable, the Commission also preliminarily believes that the cost estimates in the Plan for large broker-dealers may be reliable. The Commission preliminarily believes that problems with the Reporters Study data are less likely to affect the Plan's large broker-dealer cost estimates for several reasons. First, if a large broker-dealer were to respond to the Reporter Study survey with the incorrect level of units (resulting in estimates that were 1,000 times too large as was the case for some small broker-dealer responses), then these errant cost survey responses would result in estimates that likely would be denominated in billions of dollars. The maximums presented in the Plan's tables describing the Reporters Study data do not include responses denominated in billions; notably, under the Plan's cost estimate methodology, if such responses were generated, these responses likely would have been removed as outliers. Second, although it is possible that small broker-dealers misclassified themselves as large broker-dealers in the Reporters Study data, such misclassification does not seem to have biased the cost estimate results for large broker-dealers to the degree that the Commission preliminarily believes has occurred for the small broker-dealer Reporters Study data. Cost estimates for large broker-dealers, particularly those that do not have current OATS reporting obligations, are not inconsistent with information gathered by the Commission in discussions with broker dealers and service providers, although the Commission preliminarily believes that averages presented in the Plan generally fall between the expenses that a very large and complex broker-dealer would experience and those of a more typical broker-dealer in the same category. For example, the Plan estimates that the average large OATS-reporting broker-dealer currently spends $8.7 million annually to comply with current data reporting requirements. The Commission preliminarily believes that this estimate is likely to be substantially lower than the actual data reporting costs incurred by the largest and most complex broker-dealers that currently report to OATS; these very large and complex firms are assumed to spend far more than this estimate. There are, however, only a limited number of exceptionally large OATS-reporting broker-dealers. Similarly, the Plan's estimate is likely to significantly overestimate the costs incurred by the majority of firms classified as large by the Plan because most large firms are not as large or as complex as these limited number of exceptionally large broker-dealers. Summary statistics on activity levels of OATS reporting firms are discussed in detail below.

FIF arranged a group discussion with a small number of broker-dealers whose identities were not provided to Commission staff and individual discussions with five service bureaus whose identities were not provided to Commission staff. Also, staff arranged individual discussions with five additional broker-dealers. When market participant identities were unknown, FIF provided demographic information that allowed Commission staff to gauge a firm's size, complexity, and general market activities. Broker-dealers outside of the group discussion and service bureaus were asked for specific cost information that related to their regulatory data reporting costs; most broker-dealers and some service bureaus shared general estimates, particularly of staffing levels, and provided information on cost drivers and obstacles that firms face in accomplishing their regulatory data reporting, particularly challenges that they face in implementing changes to these requirements. Most, but not all, firms participating in discussions with Commission staff discussed OATS as their most challenging data reporting requirement. Some firms named LOPR and EBS as additional sources of regular challenges and significant costs. It is our understanding from these discussions, that some data reporting requirements, such as Rule 605 and Rule 606 reporting, are nearly always outsourced.

See infra note 882.

The Plan presents cost estimates for large broker-dealers' current regulatory data reporting costs and costs they would incur to implement and maintain CAT Data reporting. The Plan estimates that an OATS-reporting large broker-dealer has current data reporting costs of $8.7 million per year. A non-OATS reporting large broker-dealer is estimated to spend approximately $1.4 million annually. The Plan estimates that OATS-reporting large broker-dealers would spend approximately $7.2 million to implement CAT Data reporting, and $4.8 million annually for ongoing costs. For non-OATS reporting large broker-dealers, the Plan estimates $3.9 million in implementation costs and $3.2 million in annual ongoing costs. According to the Plan, the magnitude of each of these cost estimates is primarily driven by FTE costs.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.(7)(b)(ii)(C), Table 3. The $8.7 million figure was calculated by summing the average hardware/software cost, third party/outsourcing cost, and full-time employee costs using the Commission's estimated cost per employee of $424,350.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.(7)(b)(ii)(C), Table 4. The $1.4 million figure was calculated by summing the average hardware/software cost, third party/outsourcing cost, and full-time employee costs using the Commission's estimated cost per employee of $424,350.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.(7)(b)(iii)(C)(2)a., Table 9; and at Appendix C, Section B.(7)(b)(iii)(C)(2)b., Table 15.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.(7)(b)(iii)(C)(2)a., Table 10; and at Appendix C, Section B.(7)(b)(iii)(C)(2)b., Table 16.

(2) Commission Cost Estimates

The Commission's broker-dealer cost estimates incorporate some broker-dealer data from the Plan, but to address issues in the Plan's Reporters Study data, the Commission's cost estimates also include other data sources. As previously discussed, the Commission preliminarily believes that the small firm cost estimates presented in the Reporters Study are unreliable. As a result, the Commission has re-estimated the costs that broker-dealers likely would incur for CAT implementation and ongoing reporting. As with the Plan's cost estimates, the Commission's re-estimation relies on classifying broker-dealers based on whether they currently report OATS data. However, the re-estimation further classifies broker-dealers, as in the Commission's cost estimates presented in the Proposing Release, based on whether the firm is likely to use a service bureau to report its regulatory data, or, alternatively, whether the firm might choose to self-report its regulatory data. In this updated analysis, the Commission preliminarily estimates that the 1,800 broker-dealers expected to incur CAT reporting obligations currently spend approximately $1.6 billion annually to report regulatory data. If the CAT NMS Plan is approved, the Commission preliminarily believes that these broker-dealers would incur approximately $2.2 billion in implementation costs and $1.5 billion in ongoing data reporting costs.

Discussions below present information on data obtained from FINRA and gleaned from discussions with broker-dealers and service bureaus arranged by FIF and staff. See supra notes 880 and 899.

To the extent that the CAT NMS Plan underestimates the number of broker-dealers that would incur CAT reporting obligations, the Commission's updated estimates understate the actual costs Reporters would face if the CAT NMS Plan is approved.

These figures cover only broker-dealer costs. Industry-wide costs are summarized below in Section IV.F.2.

The Commission preliminarily believes classifying broker-dealers based on their manner of reporting provides a more accurate estimate of the costs firms will incur because, as noted below, costs differ based on whether the firm insources or outsources reporting responsibilities and insourcing/outsourcing does not necessarily correlate with firm size. Accordingly, the Commission begins its estimation of costs using the number of OATS Reportable Order Events (“ROEs”) reported by firms that report to OATS. The Commission preliminarily believes that because OATS reportable events, such as order originations, routes, and executions are also CAT Reportable Events, these two measures are likely to be highly correlated, making the number of OATS records a proxy for the anticipated level of CAT reporting. Based on discussions with broker dealers and service providers, however, the Commission preliminarily believes that firms that report high numbers of OATS ROEs decide to either self-report their regulatory data or outsource their regulatory data reporting based on a number of criteria, including potential costs. Thus, simply using the number of OATS ROEs as a proxy for firm size may not provide an accurate picture of the reporting costs for such firms. As a result, the Commission goes a step further in its estimation of costs by segmenting firms into two groups—those that insource and those that outsource their regulatory data reporting—and estimates costs separately for each group. Empirical evidence supporting this approach is detailed further below.

In other words, the Commission preliminarily believes that the higher the number of OATS ROEs reported, the higher the anticipated number of CAT records to report. As noted below, however, the Commission anticipates that the number of CAT records would exceed the number of OATS ROEs.

As explained further below, the Commission believes that firms reporting relatively few OATS ROEs would be unlikely to have the infrastructure and specialized employees necessary to insource regulatory data reporting and would almost certainly outsource their regulatory data reporting functions.

The Commission in its cost calculation uses the number of OATS ROEs as a measure of firm size, rather than traditional measures of firm size based on a single metric, such as capital level, or OTC dollar volume. The Commission preliminarily believes that the use of OATS ROEs provides a more accurate predictor of firm reporting behavior. Data provided by FINRA, for example, reveals that some firms with extremely high levels of OATS reporting activity have relatively low capital levels; furthermore, many firms that report exceptionally high numbers of OATS ROEs have no OTC dollar-volume. See infra note 893.

The Plan also separates industry costs of current OATS reporting firms from those that currently have no OATS reporting obligations, recognizing that the group of non-OATS reporting firms are diverse in size and scope of activities. The Commission maintains this approach in its re-estimation, as firms that do not currently report to OATS would face a different range of costs to implement and maintain CAT reporting because firms that currently do not report to OATS may have little to no regulatory data infrastructure in place. Broker-dealers that do not currently report to OATS may have higher or lower costs than firms that do report to OATS, depending on whether they do not report because of SRO membership status or lack of equity market activity or because of size and scope of activity within equity markets. For example, an electronic liquidity provider (“ELP”) may trade extensively both on and off-exchange, yet not report to OATS because it is not a FINRA member; such a firm could incur high data reporting costs under CAT because it has a high volume of records to report. Conversely, a small equity trading firm might be excluded or exempted from OATS reporting due to its size and scope of activities; such a firm could have relatively low CAT reporting costs, although still higher than its existing regulatory reporting costs, because it has few Reportable Events and is assumed to outsource its reporting responsibilities. Recognizing this diversity in non-OATS firms, the Commission's re-estimation anticipates a large range of firm activity levels in non-OATS CAT reporters and treats them differently when estimating their costs. This is discussed further below.

The Commission's re-estimation of costs assumes that firms that are currently excluded or exempted from OATS reporting are Outsourcers. By definition, OATS-reporting Outsourcers report fewer than 350,000 OATS ROEs per month. However, firms that are not FINRA members are not assumed to be Outsourcers; many of these firms are in the business of proprietary trading as ELPs or are Options Market Makers, which are assumed to be typical of large non-OATS reporters discussed in the Plan. The identification of these firms and their estimated costs of CAT reporting are discussed further in Section IV.F.1.c(2)B.i, infra.

In sum, the framework for the Commission's re-estimation is as follows. First, the Commission identifies those OATS-reporting firms that insource (“Insourcers”) and those that outsource based on an analysis of the number of OATS reporting ROEs combined with specific data provided by FINRA on how firms report. Furthermore, the Commission identifies firms that do not currently report to OATS but are likely to insource based on their expected activity level by identifying Options Market Makers and ELPs. Based on that analysis, the Commission preliminarily estimates that there are 126 OATS-reporting Insourcers and 45 non-OATS reporting Insourcers; these estimates are discussed further below. The Commission's re-estimation classifies the remaining 1,629 broker-dealers that the Plan anticipates would have CAT Data reporting obligations as “Outsourcers,” based on outsourcing practices observed in data obtained from FINRA and discussed further below. The Commission preliminarily believes that most of these firms would accomplish their CAT Data reporting through a service bureau. Next, to determine costs for Insourcers, the Commission relies upon cost estimates for firms classified as “large” in the Reporters Study. For Outsourcers, the Commission uses a model of ongoing outsourcing costs (“Outsourcing Cost Model”) to estimate both current regulatory data reporting costs and CAT-related data reporting costs Outsourcers would incur if the CAT NMS Plan were approved.

A. Broker-Dealer Reporting Practices

Although the Commission's analysis segregates broker-dealers into two groups (Insourcers and Outsourcers), within those groups, broker-dealer data reporting methods currently vary widely across firms, and these varied methods affect the data reporting costs that broker-dealers incur. As discussed previously, depending on the business in which broker-dealers participate, broker-dealers can have a wide range of reporting responsibilities.

There are two primary methods by which broker-dealers accomplish data reporting: Insourcing, where the firm reports data to regulators directly; and outsourcing, where a third-party service provider performs the data reporting, usually as part of a service agreement that includes other services. Firms that outsource retain responsibility for complying with rules related to outsourced activity. Based on data from FINRA and conversations with market participants, the Commission preliminarily believes that the vast majority of broker-dealers outsource most of their regulatory data reporting functions to third-party firms. Data provided by FINRA shows that 932 broker-dealers reported at least one OATS ROE between June 15 and July 10, 2015. Of these 932 firms, 799 reported at least 90% of their OATS ROEs through a service bureau. Broker-dealers generally used a single service bureau (497 firms) to report OATS, but some broker-dealers used multiple service bureaus (up to 9 service bureaus).

The Commission analyzed data on broker-dealer OATS reporting received from FINRA. This data source included the number of OATS ROEs reported by each individual broker-dealer, as well as counts of how many ROEs were reported by the firm directly and how many ROEs were reported through service bureaus, and the number of service bureaus that reported data for the firm. The dataset includes the firms' minimum net capital required and actual net capital as well as the number of registered persons associated with the firm. Factors that affect broker-dealers' insourcing/outsourcing decision are discussed below. Because market activity is highly correlated with volatility, this four-week period was chosen to have a typical level of volatility (as measured by VIX level) for the period September 16, 2010 through September 15, 2015.

Often, service bureaus bundle regulatory data reporting services with an order-handling system service that provides broker-dealers with market access and order routing capabilities. Sometimes regulatory data reporting services are bundled with trade clearing services. A broker-dealer's decision to insource/outsource these functions and services can be complex, and different broker-dealers reach different solutions based on their business characteristics. To illustrate, some broker-dealers self-clear trades but outsource regulatory data reporting functions; some broker-dealers have proprietary order handling systems, self-clear trades, and outsource regulatory data reporting functions. Other broker-dealers outsource order-handling, outsource clearing trades, and self-report regulatory data. The most common insource/outsource service configuration, however, for all but the most active-in-the-market broker-dealers is to use one or more service bureaus to handle all of these functions.

In most, but not all, cases, service bureaus host their client broker-dealer's order-handling system on the service bureau's servers while the broker-dealer has software serving as a “front end” for this system running on the broker-dealers' local IT infrastructure. For broker-dealers whose order-handling systems are thus hosted on their service bureau's servers, their service bureaus would handle many elements of CAT implementation, including clock synchronization. These broker-dealers would still incur some CAT implementation costs because some CAT Data, such as Customer information (including PII), is likely to reside outside of the broker-dealer's order handling system; consequently, such broker-dealers would need to develop technical and regulatory infrastructure to provide such CAT Data to its service bureaus. Further, broker-dealers that outsource could still need to adapt their in-house software systems to address order-management system changes. In addition to the resources needed to reprogram the system, any order-handling system change is likely to require significant staff training. Furthermore, broker-dealers that outsource would need to update their internal monitoring of their service bureau's reporting to ensure it meets the requirements of the Plan.

In discussions arranged by FIF, broker-dealers cited a number of factors that influence a broker-dealer's decision on whether to handle regulatory data reporting in-house. Generally, smaller broker-dealers (with relatively few registered persons and limited capital) do not have the business volume required to support the IT infrastructure and specialized staff that is necessary to perform in-house regulatory data reporting; these broker-dealers may have no business choice but to rely upon third-party service providers to provide order handling and market connectivity, as well as clearing services. For larger broker-dealers, outsourcing is more likely to be a discretionary business decision. In discussions with staff, larger broker-dealers cited a number of reasons to outsource. First, it may be a strategic choice; some broker-dealers view regulatory data reporting as a function that offers no competitive advantages and a costly distraction from other business activities, as long as an alternative solution satisfies reporting requirements. For these firms, compliance might be achieved at a lower-cost in-house, but the firms prefer to outsource the data reporting function to focus key resources on business functions. Second, some broker-dealers outsource these functions to reduce costs associated with demonstrating regulatory compliance. Multiple broker-dealers stated that using a regulatory reporting service that was familiar to regulators allowed more efficient regulatory examinations, because an in-house regulatory reporting system might require more staff time invested in facilitating examinations and demonstrating compliance. Third, some broker-dealers cited that keeping current with regulatory requirements drove their decision to outsource. These broker-dealers may have insourced initially, but they relayed that over time they experienced accelerating regulatory rule changes, which led to an escalation in their compliance costs. For these firms, the pace of regulatory rule changes drove the decision to outsource where they had at one time insourced, because the firm could fulfill its regulatory responsibilities at a lower cost by outsourcing and monitoring the service bureau's compliance.

In conversations with market participants, several broker-dealers suggested that for very small firms, establishing these service bureau relationships could be difficult. These firms might “piggy back” on another broker-dealer's infrastructure, essentially relying on them to act as an introducing broker. This would generally add another cost layer for these very small firms but could be more cost effective than establishing stand-alone service bureau relationships.

The Commission notes that an Industry Member CAT Reporter remains responsible for compliance with the requirements of the CAT NMS Plan and Rule 613, as reflected in the Compliance Rule of the SRO(s) of which it is a member, regardless of whether it has outsourced some or all of its regulatory data reporting functions to a third party.

On the other hand, some broker-dealers choose to insource their regulatory data reporting functions. In discussions arranged by FIF, broker-dealers cited a number of reasons supporting their decision to self-report. First, some broker-dealers cited ancillary benefits to constructing the IT infrastructure necessary to accomplish their regulatory data reporting. Data collected in a central location for regulatory data reporting and the software necessary to manipulate the regulatory data facilitates self-monitoring and business reporting, providing other benefits to the firm. Second, some broker-dealers cited protecting their proprietary strategies as a motivator to self-report regulatory data. These broker-dealers felt that sharing their trading data with a service bureau was potentially too revealing of their proprietary trading strategies. Third, some broker-dealers cited operational complexity as a driver of their insourcing decision. For these very large broker-dealers that traded in a wide range of assets, outsourcing would involve multiple service provider contracts. At least one broker-dealer stated that it did not believe service bureaus could meet all of its requirements due to its complexity. Finally, while some broker-dealers preferred to outsource to reduce the costs of demonstrating compliance, others stated that outsourcing would increase compliance costs because they could not conduct their own compliance checks to ensure the reports comply with relevant regulations.

Current costs of outsourcing regulatory data reporting vary widely across broker-dealers. Whether data reporting is provided on behalf of a broker-dealer by the provider of an order-management system or another third-party firm, a broker-dealer generally enters into long-term agreements with its service provider to obtain a bundle of services that includes regulatory data reporting, and costs to change service bureaus are high. Furthermore, based on discussions with service providers, the Commission understands that switching service bureaus can be costly and involve complex onboarding processes and requirements, and that systems between service bureaus may be disparate; furthermore, changing service bureaus may require different or updated client documentation. The Commission preliminarily believes that annual costs for provision of an order-handling system (including market connectivity, routing and regulatory data reporting) range from $50,000 to $180,000 annually for very small broker-dealers. Costs for very large broker-dealers that outsource these functions begin at $1 million to 2.4 million annually.

See Section IV.G.1.d, infra, for a discussion of the potential effects of the Plan on the market to report regulatory data.

These estimates are based on Staff discussions with service bureaus that were arranged by FIF. See supra note 895 and accompanying text. The $1 million per year figure contemplated a very large broker-dealer that provided its own order management system and market connectivity, so it likely represents a rough estimate of the regulatory data reporting costs of a very large firm. Because service bureaus did not provide an OATS activity level corresponding to “very large,” the Commission relies on an analysis of FINRA data on OATS reporting to calibrate its definition of “very large” in terms of OATS activity level and seeks comment on what activity level should correspond to cost estimates for “very large” broker-dealers. The Commission notes that because there are relatively few broker-dealers that report at medium activity levels, the Commission's estimation of outsourcing costs is not particularly sensitive to this definition because most broker-dealers whose costs are estimated using the Outsourcing Cost Model have very low OATS reporting levels. Finally, estimates of total reporting costs include provision of an order-management system and market connectivity.

For broker-dealers that perform regulatory data reporting in-house, implementation costs are likely to vary widely. Some very large broker-dealers that self-report regulatory data have a centralized IT infrastructure and trade in relatively few asset classes. Some of these broker-dealers carry no customer accounts, simplifying their regulatory data reporting obligations. The Commission preliminarily believes that such broker-dealers could incur relatively low CAT implementation costs because they have a centralized IT infrastructure that captures all broker-dealer activity and specialized personnel who are dedicated to broker-dealer-wide data reporting. At the other end of the spectrum, large broker-dealers may be very complex, facilitating complex multi-leg transactions and operating within a non-centralized structure. These broker-dealers would be likely to experience CAT implementation costs far higher than broker-dealers with less complex structures for several reasons. First, some of these broker-dealers do not have a centralized IT infrastructure; instead, orders could originate from many locations in the broker-dealer and may be handled by diverse legacy systems, each of which the broker-dealer would need to adapt for CAT Data reporting. Second, broker-dealers that accommodate more complex transactions that involve multiple asset classes would likely need to invest more time in understanding new regulatory requirements. In discussions with market participants, several broker-dealers noted, among other concerns, that determining the correct regulatory treatment for unusual trades can be a significant cost-driver in implementing regulatory rule changes and can delay implementation of system changes or precipitate a second round of changes once regulatory treatment of these trades is clarified. Third, broker-dealers that lack a centralized IT infrastructure would likely incur higher costs to comply with clock synchronization requirements because more servers may be handling orders than in firms with a more centralized IT infrastructure.

In discussions with market participants, some broker-dealers indicated that they operate more than a dozen instances of a third-party's order handling system, suggesting they originate orders at more than a dozen places within the broker-dealer, yet they handle data reporting in-house. Firms such as these are likely to incur far higher costs to implement CAT compared to broker-dealers with a centralized IT infrastructure and fewer legacy systems because there are more systems that require changes to comply with new data reporting requirements.

B. Re-Estimation

i. Count of Firms Likely To Rely Upon Service Bureaus for Data Reporting

To separately examine the costs to broker-dealers that outsource and to aggregate those costs across all broker-dealers, Commission Staff first established a count of CAT Reporters likely to outsource their regulatory data reporting functions. For this, the Commission analyzed data provided by FINRA.

See supra note 893 and accompanying text.

The FINRA data allows the Commission to examine how broker-dealers' current outsourcing activities vary with the number of ROEs reported to OATS. Figure 1 shows the percentage of OATS ROEs that are self-reported for five size categories of broker-dealers with the following OATS reporting activity levels for a four-week period from June 15-July 10, 2015: More than 1 billion records; 1 million to 1 billion records; 350,000 to 1 million records; 100,000 to 350,000 records; and 100,000 records or fewer. The bars for each category represent the percentage of total OATS ROEs reported by broker-dealers in the category that were reported directly by the broker-dealers.

The group that reports one billion records or more comprises 77.90% of OATS records; the group that reports one million records to one billion comprises an additional 22.05% of OATS records. The remaining three groups comprise just 0.05% of all OATS records. Overall, firms self-report 65.44% of OATS ROEs.

Based on this analysis of FINRA data, the Commission preliminarily believes that the 126 broker-dealers that reported more than 350,000 OATS ROEs between June 15 and July 10, 2015 make the insourcing-outsourcing decision strategically based on the broker-dealer's characteristics and preferences, while the remaining OATS reporters are likely to utilize a service bureau to accomplish their regulatory data reporting. The categories of broker-dealers assumed to outsource their data reporting are marked with an asterisk (*) in Figure 1.

The Commission preliminarily believes this decision is strategic and discretionary because FINRA data reveals that while many broker-dealers at these activity levels self-report most or all of their regulatory data, other broker-dealers outsource most or all of their regulatory reporting at these activity levels. At lower activity levels, most, but not all, broker-dealers outsource most if not all of their regulatory data reporting. The Commission is cognizant that some broker-dealers reporting fewer than 350,000 OATS ROEs per month can and do opt to self-report their regulatory data. However, based on conversations with broker-dealers, the Commission preliminarily believes that most broker-dealers at these activity levels do not have the infrastructure and specialized staff that would be required to report directly to the Central Repository, and electing to self-report would be cost-prohibitive in most but not all cases. See Section IV.F.1.c(2)A, supra.

As seen in Figure 1, broker-dealers in the highest OATS-reporting category insourced reporting for more than 60% of the OATS ROEs reported. More specifically, the FINRA data shows that 16 broker-dealers reported more than a billion OATS ROEs each between June 15 and July 10, 2015; most of these broker-dealers (11) self-reported nearly all of their regulatory data, but 3 used service bureaus for 100% of their OATS reporting.

Figure 1 also shows that broker-dealers that report between 1 million and 1 billion OATS ROEs during the four-week period insourced reporting for more than 70% of the OATS ROEs they reported in aggregate. Thirty-six of these 89 broker-dealers used service bureaus to report at least 90% of their OATS data while 42 of these 89 broker-dealers self-reported over 99% of their regulatory data.

For the 21 broker-dealers that reported more than 350,000 but fewer than 1 million OATS ROEs during the sample period, Figure 1 shows that they insource approximately 27% of their aggregate OATS ROEs reporting. Thirteen of these broker-dealers use service bureaus for more than 99% of their OATS reporting while 7 of these 21 broker-dealers self-reported more than 98% of their OATS data.

For the 806 broker-dealers that reported fewer than 350,000 OATS ROEs during the sample period, approximately 88.9% of those OATS ROEs were reported through service bureaus, with 730 broker-dealers reporting more than 99% of their OATS ROEs through one or more service bureaus. These broker-dealers are represented in the two right-most bars in Figure 1 that are identified with asterisks (*) in their labels. Because of the extensive use of service bureaus in these categories of broker-dealers, the Commission assumes that these broker-dealers are likely to use service bureaus to accomplish their CAT Data reporting.

Although most of these broker-dealers report nearly all of their ROEs through a service bureau, there are broker-dealers, both large and small, that self-report nearly all of their OATS data at all activity levels, including a broker-dealer that self-reported two OATS ROEs during the sample. Despite this variation, the Commission believes that its assumptions regarding which firms are likely to outsource and which firms have discretion are appropriate because (1) small firms that insource likely do so because it is less costly so the assumption simplifies the analysis and overestimates costs and (2) the cost information for the other firms already accounts for both insourcing and outsourcing.

ii. Estimation of Outsourcing Costs

The Commission has estimated ongoing costs for outsourcing firms using a model based on data gleaned from discussions with service bureaus and broker-dealers and implementation costs using information learned in conversations with industry. Service bureaus that provide order-handling systems, market connectivity and regulatory data reporting services estimated that a very small broker-dealer was likely to currently spend $50,000-$180,000 per year for these services; they suggested that current annual costs for very large broker-dealers would likely be $1,000,000-$2,400,000 but could be greater in some cases. The Commission assumes that a very small broker-dealer would report a single OATS ROE per month and a very large broker-dealer would report 100 million OATS ROEs per month.

See supra note 880.

Estimates are based on FIF-arranged conversations with service bureaus. See supra note 880.

The Commission preliminarily believes that firms that report more than 350,000 OATS ROEs per month outsource on a discretionary basis. If the estimate of activity level for very large firms is too large (100 million ROEs is used in the model estimation), the Commission's model would underestimate the costs of all firms that report fewer than 350,000 OATS ROEs per month currently. The Commission preliminarily believes the 100 million ROEs per year size estimate to be reliable because although most firms at activity levels between 40 million and 300 million OATS ROEs (15 firms) self-report, several use service bureaus.

Based on discussions with market participants, the Commission assumes that the cost function for outsourcing is concave. This type of function is appropriate when costs increase as activity level increases, but the cost per unit of activity (e.g., cost per report) declines as activity increases. Volume discounts can create such cost functions. Alternatively, if the Commission estimates outsourcing costs as a linear function using the two point-estimates (very small firms and very large firms) obtained from service bureaus, that outsourcing cost model would underestimate the costs of broker-dealers that are neither very large nor very small due to the concavity of the function. As shown in Figure 2, a concave function is greater than the linear function that connects its endpoints. To illustrate the underestimation concern, if the estimated pricing function was a straight line but the actual pricing function was concave, the estimates would be too low. Lacking data on outsourcing costs faced by broker-dealers with activity levels that are neither very small nor very large, which would assist the Commission in estimating the degree of concavity of the pricing function, the Commission's estimation assumes that service bureau pricing functions are similar in concavity to equity exchange pricing functions.

The Commission preliminarily believes that service bureau pricing functions are concave based on discussions with service bureaus arranged by FIF. See supra note 897.

The Commission relies on exchange pricing functions because the data is publicly available and because a broker-dealer's activity level on exchanges is correlated with the quantity of regulatory data it generates. If the pricing function for service bureau services is more concave than exchange pricing functions, the Commission's preliminary model would underestimate costs for broker-dealers that are neither very small nor very large because an increase in concavity would increase the distance between the concave and linear functions in Figure 2.

The Commission relies on a schedule of average charges to access liquidity and rebates to provide liquidity from four non-inverted exchanges to estimate the concavity of the exchange pricing function, which the Commission uses to approximate the concavity of the outsourcing cost model. On such exchanges, the party receiving liquidity in the transaction generally pays a fixed fee to do so; the party providing liquidity receives a rebate from the exchange. This rebate often marginally increases with the market participant's aggregate volume on the exchange. For liquidity providing firms, this pricing scheme would imply a concave function of the cost differential between taking and providing liquidity, which informs the Commission's estimation of the degree of concavity of the outsourcing cost model. The Commission preliminarily believes that estimating the shape of the function using exchange pricing functions is a reasonable approach because the same activities that determine a broker-dealer's access fees on exchanges—such as executing orders and the activities such as order submission that are requisite to those executions—would affect the broker-dealer's impact on a service bureau's infrastructure and thus the fee that a service bureau is likely to charge to provide services to the broker-dealer.

On many exchanges, the party posting a resting order earns a rebate when his order is executed. His counterparty, whose order immediately executes, pays a fee to the exchange, which exceeds the rebate the liquidity-providing party earned. The difference between the rebate and the fee represents the cost a market participant would incur to fill a resting order on the exchange, then immediately trade out of the position—a so-called “round-trip” cost. The magnitude of this round-trip cost is often a function of the market participant's trading activity on the exchange, with more active traders paying lower round-trip costs. On “inverted” exchanges, the party with the resting order pays a fee while her counterparty that receives immediate execution earns a rebate. The Commission's estimate of concavity relies on data from exchanges that do not feature inverted pricing.

The Commission obtained public fee schedule data from Web sites for NASDAQ, PSX, NYSE, and ARCA during October, 2015. For NASDAQ, the differential between access fees and liquidity rebates was calculated using the universal “take fee,” and rebates were for shares trading at greater than $1.00 per share ( http://www.nasdaqtrader.com/Trader.aspx?id=PriceListTrading2. For PSX, calculations used the Tape C remove charge less rebate to add displayed liquidity ( http://www.nasdaqtrader.com/Trader.aspx?id=PSX_Pricing ). For NYSE, calculations used the “Providing Tier 3/2/1” rebates versus the universal “take fee” (NYSE Trading Fees). For ARCA, calculations used charges and rebates for midpoint passive liquidity orders available at https://www.nyse.com/publicdocs/nyse/markets/nyse-arca/NYSE_Arca_Marketplace_Fees.pdf.

See supra note 908 for examples of exchange pricing schedules.

This estimation affects the shape of the function, and thus the relative prices that are estimated for each broker-dealer; the absolute level of prices is determined through the function's calibration, which is described below.

The Commission's estimation of the outsourcing cost model begins with construction of a tiered function based on the exchange pricing function; the incorporation of the exchange pricing function is the source of the concavity in the model. The Commission's estimation of exchange pricing assumes four activity level categories. The Commission preliminarily mapped OATS reporting activity levels to exchange fee break points, with the assumptions that only a very small minority of firms would qualify for the lowest-fee tier of services and all of the firms that reported so few OATS ROEs to be assumed to be Outsourcers would be at the highest-cost tier of service. Consequently, the Commission assumed the first fee break-point to be 350,000 OATS messages per month. A firm with 1 million messages per month is assumed to qualify for the third pricing tier. To qualify for the most favorable pricing tier, a firm would need to report more than 100 million OATS messages per month. The model is fitted by adding a constant to the implied cost of message traffic to bring firms with a single OATS ROE to the minimum $50,000 annual fee discussed by service bureaus. The fee for very large firms (for purposes of this model, 100 million plus records per month) is calibrated by multiplying the estimated exchange fee tiered function by a constant scale factor of 30. With these adjustments, the tiered function implies a firm with 20,000 OATS ROEs per month would incur a service bureau fee of $50,705 annually; a firm with 100 million OATS ROEs per month would incur a service bureau fee of $1.175 million annually; and a firm with 1 billion OATS ROEs per month firm would incur a service bureau fee of $11.3 million annually.

A tiered function often looks like a set of steps with points of discontinuity where the function appears to suddenly move up or down. Often, a tiered function's behavior is determined by the range of its independent variable (input value). For example, a firm that charges $1 per unit for orders of 100 units or less, or $.80 per unit for orders of more than 100 units prices according to a step function, with the number of units ordered being the independent variable. On exchanges, the round trip cost (access fee less rebate) is often a step function based on the firm's activity level during a given calendar period.

The Commission chose four tiers to strike a balance between incorporating as much information from exchange pricing models and having to extrapolate information from them. NASDAQ and PSX have five activity level tiers, while NYSE and ARCA have three activity level tiers. Building a model with only three tiers would ignore potentially significant information from NASDAQ and PSX while building a model with five tiers would require extrapolating information on nonexistent tiers on NYSE and ARCA, which adds imprecision to the function. For NASDAQ and PSX, the Commission used prices for the four most active tiers in the analysis; for NYSE and ARCA, the Commission used all three, with the middle activity level assumed constant over the two middle activity tiers in the outsourcing cost model. The aggregate exchange price function averages prices on those four exchanges.

The Commission preliminarily believes that this is a conservative assumption because all of the firms assumed to be outsourcing are assumed to be at the highest priced service level on a per record reported basis. This causes the Commission's estimate of their costs to be higher than other possible assumptions.

Estimates are outputs of the calibrated step function based on exchange pricing. Calculations are as follows: Outsourcing Cost = Fixed Fee ($50,000) + Monthly OATS ROEs × Fee per ROE. $50,705 = $50,000 + 20,000 × $0.03525; $1.175 million = $50,000 + 100MM × $0.01125; $11.3MM = $50,000 + 1B × $0.01125.

The final step in estimating the Outsourcing Cost Model is to smooth the tiered function by fitting it to a polynomial. As discussed previously, tiered functions are not continuous; the behavior of the function can change dramatically at a discontinuity, such as happens when moving from one activity level category to another. In the earlier illustrative example, a vendor offered pricing that would be characterized by a tiered function, in which the firm charges $1 per unit for orders of 100 units or less, or $.80 per unit for orders up to 400 units. In this example, a purchase of 100 units is more expensive than a purchase of 120 units. On exchanges, the pricing discontinuities may be acceptable to broker-dealers because the broker-dealers can more easily estimate a range of volume rather than actual volume, and thus pricing discontinuities may allow the broker-dealers to better forecast their expected exchange fees based on those volume ranges. For the Outsourcing Cost Model, however, such discontinuities are undesirable because service bureaus negotiate the contract with each customer individually and contracts generally cover a period of several years. Consequently, service providers provide custom quotations in consideration of the firm's business activities and likely capacity impact upon the provider's infrastructure. The Commission preliminarily believes that there are unlikely to be instances in which a service bureau's costs to service a customer would decrease if the customer were to become more active, and because the contract has a fixed cost, there is unlikely to be incentives to price with a tiered function to ease billing. To smooth the Outsourcing Cost Model, the Commission estimates a second degree polynomial to points imputed across the tiered function. This step essentially involves finding a smooth curve that closely tracks the tiered function, but smoothes its discontinuities.

In this illustrative example, 100 units would cost $100 (100 units × $1 per unit), while 120 units would cost $96 (120 units × $.80).

A first degree polynomial is linear; a second-degree polynomial includes a term raised to the power of two and defines a quadratic function. The Commission did not consider higher degree polynomials because they include inflection points, which would be undesirable in this model because there is unlikely to be a range in which costs per unit would be expected to increase with volume. Quadratic functions are characterized by curves with a single minimum or maximum and include concave curves that would be typical of cost curves with volume discounts. The estimated functional form of the outsourcing cost model used in cost estimates is based on OATS ROE activity levels expressed in millions of ROEs per month. The estimated function is: Cost estimate = −1.3939 ROEs 2. + 12,473 ROEs + 124,005. Model fit statistics, used to measure how well a model fits its underlying data, are not meaningful for this model because points used for the estimation are imputed rather than observed. This function is not monotonic (always increasing or always decreasing); it has a maximum at 4.47 billion ROEs. The Commission believes this is not a serious concern because the model is not used to provide cost estimates for firms that report more than 350,000 OATS ROEs per month.

The model's output in Figure 3 is an estimate of a broker-dealer's current cost to outsource data reporting services as part of a bundle of services from a service bureau; for smaller broker-dealers, it is assumed to include provision of an order management system and market connectivity.

In conversations with Commission staff, service bureaus related that some very large clients provide their own order-handling system and market connectivity. See supra note 880.

To estimate costs of CAT Data reporting by the service bureaus, the Commission preliminarily assumes that the current pricing function would apply for CAT Data reporting, but the costs in relation to the number of ROEs would increase because some events that are excluded from OATS (like proprietary orders originated by a trading desk in the ordinary course of a member's market making activities), would be included in CAT. The Commission estimates the expected increase in broker-dealer data by estimating the ratio of all SRO audit trail data (OATS and exchange data) to OATS data; with this methodology, the Commission estimates CAT Data ROEs reported by broker-dealers would increase from those reported to OATS by a factor of 1.9431. The Commission preliminarily believes that the assumption of the same cost function is reasonable for several reasons. First, the service bureaus that provide market access for broker-dealers already process the exchange traffic for most of these broker-dealers. Although the number of ROEs reported would increase, service bureaus already host most of the data that broker-dealers would report to the Central Repository. Second, although some broker-dealers would have to establish a process of hosting or processing their customer information at their service bureau, many broker-dealers already do so to allow their service bureau to prepare information for clearing. Consequently, most service bureaus have already established the infrastructure to host or process customer information. Third, the Plan requires broker-dealers to update customer information files, one of the additional data sources that broker-dealers would need to report to the Central Repository. While the costs of ensuring the appropriate security could be significant, these updates occur at a much lower frequency than the rate of a service bureau customer's market activity, and thus such updating activity would be unlikely to provide a technological stress on a service bureau's infrastructure.

Although the pricing function is assumed constant, broker-dealer costs would increase because the number of ROEs they report through their service bureaus would increase under the Plan. It is possible that, if the Plan is approved, data under CAT might be reported in a form other than ROEs; however, if a ROE is equivalent to a Reportable Event, the number of Reportable Events—regardless of the form of the event report—would increase by approximately the same adjustment factor.

To approximate the increase in reporting activity that broker-dealers would likely experience if the Plan were approved, the Commission relied on equity data from the week of September 15-19, 2014, previously provided by FINRA. This FINRA data includes all OATS data reported to FINRA, as well as SRO audit trail data from all equity exchanges effecting trades that week except the Chicago Stock Exchange. The adjustment factor was estimated by dividing the number of ROEs in SRO audit trail data hosted by FINRA for all exchanges and OATS, by the number of ROEs in OATS; this methodology is equivalent to assuming that all exchange message traffic would become reportable by broker-dealers. Because some exchange message traffic is already reported through OATS, this is a conservative assumption in the sense that it increases the adjustment factor and consequently increases estimates of broker-dealer reporting costs. To adjust for the missing exchange, data for the NASDAQ OMX BX (the lowest volume exchange with trading volume exceeding that of the Chicago Stock Exchange, based on trades reported through NYSE TAQ) was double-counted in the exchange activity total. Although this adjustment factor does not capture options data, the Commission preliminarily believes that the underestimation is not material in this application because the Plan assumes that Options Market Maker quotes (the most frequent option event) would not be reported by broker-dealers. Furthermore, the Commission notes that the largest group of events excluded by OATS but reportable under CAT's reporting rules (proprietary orders originated by a trading desk in the ordinary course of a member's market making activities) predominantly originate from insourcing firms for which the service-bureau model does not provide estimates of reporting costs. Consequently, the adjustment factor is likely to overestimate the increased regulatory data volume of outsourcing firms under CAT to a degree that should encompass the limited option activity reported by outsourcing broker-dealers.

Broker-dealers that self-clear but rely on a service bureau to perform their regulatory data reporting may not have infrastructure in place to share customer information with their service providers. However, service bureaus that provide regulatory data reporting services would need customer information to perform CAT reporting. The Commission preliminarily believes that service bureaus that do not currently collect customer information but provide regulatory data reporting services would need to change their business processes to continue to offer regulatory data reporting services; the Commission further assumes that the cost estimates presented in the Vendors Study encompass the expenses these service bureaus would incur to continue providing their current service offerings. In discussions with service bureaus arranged by FIF, some service bureaus that do not offer clearing services discussed additional costs, some related to security, that accompany hosting customer information. If these service bureaus were to stop offering regulatory data reporting services due to unwillingness to host customer information, their customers would be forced to establish new service bureau relationships or undertake self-reporting. The Commission cannot rule out that one or more service bureaus may choose to exit the market to provide data reporting services rather than change their business practices to satisfy their clients' responsibilities under the Plan. Any such event would potentially be very costly to the broker-dealer clients of the exiting service bureaus due to the switching costs that broker-dealers incur to change service bureaus. Such an event could also contribute to crowded entrances problems. See infra note 934. The Commission preliminarily believes that such service bureau exit events are unlikely because service bureaus should be able to pass costs associated with handling customer information on to their clients as part of a more comprehensive bundle of services. Furthermore, based on information from broker-dealer discussions arranged by FIF, the Commission preliminarily believes that the market for regulatory data reporting services is generally expanding and the trend is for more, not less, outsourcing. Consequently, the Commission believes that market share in this market is valuable and existing competitors are unlikely to voluntarily exit the market abruptly. The Commission preliminarily believes that most firms that report fewer than 350,000 OATS ROEs per month do not self-clear; smaller firms that do not self-clear are likely to already have relationships with service bureaus that host their customer information. It is possible that some of these firms have clearing arrangements that do not include regulatory data reporting; these firms may be forced to seek new service bureau relationships to satisfy their CAT reporting obligations, but it is also possible these clearing firms may either add CAT reporting as a service or establish a relationship with a service bureau to perform the function of providing customer information for CAT on behalf of its clients.

The Commission preliminarily believes this activity is unlikely to result in a service bureau pricing structure that significantly differs from the Commission's current outsourcing cost model. The Commission recognizes, however, that these new data sources create implementation costs for both broker-dealers and service bureaus, and preliminarily believes that these costs are reflected in cost estimates provided by service bureaus because service providers that responded to the Service Providers Study were presumably familiar with the requirements of CAT when they estimated the costs they could likely incur if the CAT NMS Plan is approved. The number of ROEs broker-dealers would report would likely increase because, for example, proprietary orders originated by a trading desk in the ordinary course of a member's market-making activities, currently excluded from OATS, would be included in a broker-dealer's audit trail data under the Plan. The increase in ROEs would drive an increase in service bureau costs that the Commission's model anticipates for broker-dealers that would outsource CAT Data reporting obligations. For illustration, consider two firms: Firm A reports the median number of OATS ROEs per month in the Outsourcers sample (1,251) and Firm B reports the maximum number of OATS ROEs per month (348,636). After CAT implementation, the estimation would assume that Firm A would report 2,431 ROEs of audit trail data per month and Firm B would report 677,435 ROEs of audit trail data per month. Using the outsourcing cost model discussed above, Firm A's annual cost would increase from $124,021 to $124,035. Firm B's average annual cost would increase from $128,353 to $132,454.

The Commission recognizes that OATS does not include options market activity. Because option quotes are not reportable by broker-dealers under the Plan, the Commission preliminarily believes that option related events would not significantly increase the number of events that would be included in regulatory data reporting for broker-dealers whose costs are estimated by the Outsourcing Cost Model. The Outsourcing Cost Model predicts costs only for broker-dealers that the Commission expects to outsource CAT reporting responsibilities. Because exchanges would report Options Market Maker quotes, the Outsourcing Cost Model would not predict the costs of reporting Options Market Maker quotes. See Exemption Order, supra note 18, at 11857-58.

In addition, the Commission recognizes that larger and more complex broker-dealers are likely to have significant regulatory reporting responsibilities related to their options activities, but the Commission preliminarily believes that these broker-dealers are likely to be included in the broker-dealers reporting more than 350,000 OATS ROEs per month. The Commission estimates these broker-dealers' costs using information from the Reporters Study in the Plan as opposed to the Outsourcing Cost Model, and those cost estimates presumably include costs related to options activity.

The Outsourcing Cost Model assumes that other CAT reporting tasks like providing customer information to the Central Repository are handled by the firms' service bureaus. In practice, some Outsourcers may have a service bureau that provides an order handling system and market connectivity, but does not currently host broker-dealers' customer information, while another service provider provides clearing services and hosts customer information. For broker-dealers with multiple service provider relationships, the clearing broker-dealer is assumed to provide services that include providing the Central Repository with the customer information for its broker-dealer clients. The Commission recognizes that not all clearing firms may plan to provide this service to their customers, and this may result in additional costs for broker-dealers that do not have relationships with service providers that will provide all services they need to comply with CAT, if it is approved. This is discussed further below in Section IV.G.1.d, infra.

Firm A: 2,431 = 1,251 × 1.9431. Firm B: 677,435 = 348,636 × 1.9431.

Firm A: $124,021 = −1.3939 × (0.001251) 2. + 12,473 × 0.001251 + 124005; $124,035 = −1.3939 × (0.002431) 2. + 12,473 × 0.002431 + 124,005. Firm B: $128,353 = −1.3939 × (0.348636) 2. + 12,473 × 0.348636 + 124,005; $132,454 = −1.3939 × (0.677435) 2. + 12,473 × 0.677435 + 124,005. The Commission notes that, as set forth, the outsourcing cost model's output is dominated by the fixed cost of maintaining service at low reporting levels. But if the service bureau cost model estimated a very large firm's outsourcing cost, a very large firm's cost increase due to CAT would be far more significant. For example, a firm that reported 1.05 billion OATS ROEs per month would have estimated current costs of $11.7 million annually; after CAT implementation, its costs would be estimated to be $19.8 million. However, the Commission does not assume that firms that report more than 350,000 OATS ROEs per month are Outsourcers nor does the Commission assume that they are necessarily Insourcers; instead, their costs are estimated using data from the Reporters Study.

Application of the model to data provided by FINRA allows the Commission to estimate current outsourcing costs for broker-dealers, as well as projected costs under the CAT NMS Plan. The Commission estimates that the 806 broker-dealers that monthly each currently report fewer than 350,000 OATS ROEs currently spend an aggregate $100.1 million on annual outsourcing costs. Under the CAT NMS Plan, the Commission estimates these 806 broker-dealers would spend $100.2 million on annual outsourcing costs. The Commission recognizes that the magnitude of this increase is quite small, but this is driven by the fact that the vast majority of firms that are assumed to outsource have very low regulatory data reporting levels currently. As mentioned previously, the median firm in this group reports 1,251 OATS ROEs per month; only 39 of these 806 firms currently reports more than 100,000 OATS ROEs per month. The Outsourcing Cost Model also does not include additional staffing costs that the broker-dealer is likely to incur for implementation and maintenance of CAT reporting; these are discussed further below, and are the primary cost driver of costs that Outsourcers are expected to incur if the Plan is approved. Furthermore, the Commission is cognizant that data reporting is normally part of a bundle of services provided by a service bureau; many of those services, including the provision of market access and an order handling system, are likely to contribute substantially to the costs service bureaus bear to service their clients. The Commission is cognizant that while the volume of transactions reported by broker-dealers assumed to be Outsourcers are unlikely to dramatically increase under CAT, the service bureaus would incur significant costs to implement changes required by CAT reporting. Those costs are discussed below. Assuming service bureaus pass those implementation costs on to their broker-dealer clients eventually, the Outsourcing Cost Model would change.

This data is described above. See supra note 893.

The average broker-dealer in this category reported 15,185 OATS ROEs from June 15-July 10, 2015; the median broker-dealer reported 1,251 OATS ROEs. Of these broker-dealers, 39 reported more than 100,000 OATS ROEs during the sample period.

See Section IV.F.1.d, infra.

This would constitute a transfer of costs between market participants, but would not affect the Commission's estimate of the total costs to industry. In particular, the Commission preliminarily believes that if service bureaus pass their implementation costs on to their broker-dealer clients, it would appear as higher ongoing costs for those clients, but the overall costs would not change.

Firms that outsource their regulatory data reporting still incur internal staffing costs associated with this activity. These employees perform activities directly related to regulatory data reporting such as answering inquiries from their service bureaus, investigating reporting exceptions, maintaining any systems that transmit data to their service providers, and overseeing their service bureaus' data reporting to ensure compliance. Based on conversations with market participants, the Commission estimates that these firms currently have 0.5 full-time employees devoted to regulatory data reporting activities. The Commission further estimates these firms would need one full-time employee for one year to implement CAT reporting requirements, and 0.75 full-time employees on an ongoing basis to maintain CAT reporting.

Other employees perform other compliance duties such as supervising associated persons, and creating and enforcing internal regulatory policies (e.g., personal trading, churning reviews, sales practice reviews, SEC filings and net capital compliance). Because these regulatory activities are not part of regulatory data reporting directly affected by the Plan, they are not included in activities that contribute to current regulatory data reporting costs in the Commission's analysis.

As previously discussed, the Commission preliminarily believes that small broker-dealer cost data in the Reporters Study is unreliable. Based on discussions with broker-dealers, the Commission preliminarily believes that very small broker-dealers are unlikely to have employees entirely dedicated to regulatory data reporting. Instead, other employees have duties that include dealing with service bureau matters and answering regulatory inquiries. The Commission assumes a full-time employee costs $424,350 per year. See Section V.D.2(2)A.i, infra.

In addition to broker-dealers that currently report to OATS, the Commission estimates there are 799 broker-dealers that are currently excluded from OATS reporting rules due to firm size, or exempt because all of their order flow is routed to a single OATS reporter, such as a clearing broker, that would have CAT reporting responsibilities. The Commission assumes these broker-dealers would have low levels of CAT reporting, similar to those of the typical Outsourcers that currently report to OATS. For these firms, the Commission assumes that under CAT they would incur the average estimated outsourcing cost of firms that currently report fewer than 350,000 OATS ROEs per month, which is $124,373 annually. Furthermore, because these firms have more limited data reporting requirements than other firms, the Commission assumes these firms currently have only 0.1 full-time employees currently dedicated to regulatory data reporting activities. The Commission assumes that these firms would require 2 full-time employees for one year to implement the CAT NMS Plan and 0.75 full-time employees annually to maintain CAT Data reporting.

In discussions with Commission Staff, FINRA has stated that there are currently 54 OATS-exempt broker-dealers and 691 OATS-excluded firms. The Commission's estimate of 799 new CAT-reporting broker-dealers is based on the counts of other broker-dealer types (current OATS reporters, ELPs, Options Market Makers, and floor brokers) and the 1,800 broker-dealer estimate provided in the Plan. Based on the FINRA information on OATS-excluded or OATS-exempt broker-dealers, there are 54 remaining broker-dealers in the 1,800 with an unknown type. The Commission preliminarily assumes that these broker-dealers are small and new reporters, although it is possible that they are floor brokers on exchanges other than the CBOE (CBOE floor brokers are accounted for directly as discussed below.) Floor brokers are assumed to have the same costs as new reporting small firms, so there would be no impact on the Commission's cost estimate if these firms were reclassified as options floor brokers.

Exemption or exclusion from OATS may be based on firm size or type of activity. Broker-dealers with exemptions or exclusions that relate to firm size are presumably relatively inactive. However, some firms may be exempted or excluded because they route only to a single OATS-reporting broker-dealer; this could encompass large firms that would be more similar to Insourcers.

The Commission assumes that these very small firms already have established service bureau relationships to provide an order handling system, market access, and clearing services. If any of these firms would have to establish these relationships to comply with CAT, they would likely face greater costs associated with implementing these relationships. Furthermore, the Commission notes that conversations with market participants revealed that establishing these relationships can be difficult for very small firms because their relatively low activity levels results in service bureau fees that may not make the relationship economically feasible for service providers. Faced with this constraint, some very small firms currently resort to establishing “piggy back” relationships with larger broker-dealers, essentially using another firm as its introducing broker. Such a relationship may add an additional layer of costs to those discussed here, but such an agreement may actually prove less costly for these small firms than establishing the service bureau relationships assumed in the cost estimation because the process of onboarding with a service bureau is costly.

The Commission recognizes that some broker-dealers that are categorized in its estimation as Outsourcers in fact currently self-report their regulatory data; there are 36 firms that the Commission categorized as Outsourcers that self-report more than 95% of their OATS ROEs. Some of these broker-dealers could find that the costs associated with adapting their systems to the CAT NMS Plan reporting would render self-reporting (insourcing) CAT Data reporting infeasible or undesirable; others could continue to self-report regulatory data. The Commission preliminarily believes that the estimated cost of outsourcing for these broker-dealers is reliable, but recognizes that some of these broker-dealers could choose to self-report for other reasons at costs that could exceed these estimates. If some of these broker-dealers choose to outsource under CAT, these broker-dealers would likely incur additional costs associated with establishing or re-negotiating service bureau relationships. The Commission does not have information on existing service bureau relationships for firms that currently self-report OATS data, so cannot estimate the costs these firms might face in aggregate. It would be, however, unlikely that many firms of this size do not have relationships with service bureaus that would provide this service because firms with limited OATS reporting are unlikely to be large enough to self-clear and support the IT infrastructure necessary to provide a proprietary order handling system and market access.

In addition to the 36 broker-dealers discussed above, it is possible that many of the 799 broker-dealers that are currently exempt or excluded from OATS reporting may seek to establish service bureau relationships to accomplish their regulatory reporting required under the Plan if it were approved. It is possible that this could precipitate a “crowded entrances” problem in the market for regulatory data reporting services, in which more broker-dealers wished to establish relationships than the market could accommodate. As discussed previously, the onboarding process for service bureaus is onerous and time-consuming, both for the broker-dealer and the service bureau. If a large number of broker-dealers seek relationships simultaneously, service bureaus might not accommodate them in time to meet CAT reporting requirements. In such a situation, smaller broker-dealers are more likely to fail to establish service bureau relationships because they are presumably less profitable for service bureaus to serve and so are likely to be seen as lower-priority when onboarding resources are constrained. Some small broker-dealers could be forced to establish relationships with larger broker-dealers and rely on their infrastructure, essentially using the larger partner as an introducing broker. This could add an additional layer of costs for the smaller broker-dealer. The Commission preliminarily believes that significant crowded entrances problems with service bureaus are unlikely for two reasons. First, in discussions with service bureaus arranged by FIF, several service bureaus stated that onboarding resources were not difficult to scale up. Consequently, it seems likely that service bureaus could deploy additional onboarding resources to accommodate new demand for their services. Second, the Commission preliminarily believes that most of the OATS exempt or excluded broker-dealers already have service bureau relationships which provide them with order handling systems and market access; it is likely that these service bureaus could add regulatory data reporting packages to their current bundle of services. Finally, the implementation timelines may help alleviate strained capacity because it would allow some time for expanding onboarding capacity and new entrants and would spread out onboarding somewhat. See Section IV.G.1.d, infra.

C. Aggregate Broker-Dealer Cost Estimate

The Commission's methodology to estimate costs to broker-dealers of implementing and maintaining CAT reporting varies by the type of broker-dealer. As discussed previously, the Commission preliminarily believes that the survey of small broker-dealers used in the Reporters Study is unreliable. The Commission does, however, rely on the Reporters Study's large broker-dealer cost estimates in estimating costs for Insourcers. Consequently, for broker-dealers that are FINRA members, the Commission relies on the Reporters Study data to estimate costs for broker-dealers that report more than 350,000 OATS ROEs per month (using estimates from the Reporters Study for large, OATS-reporting broker-dealers). For lower activity FINRA-member broker-dealers (including those that do not currently report to OATS due to exclusions and exemptions to OATS reporting requirements), the Commission relies on the Outsourcing Cost Model to estimate costs for CAT Data reporting.

See Section IV.F.1.c(1), supra.

The Commission's cost estimates assume that broker-dealers that currently reporter fewer than 350,000 OATS ROEs per month are likely to use one or more service bureaus to report their regulatory data. This is discussed further in Section IV.F.1.c(2)B.i, supra.

The Commission, however, preliminarily believes that there are three other categories of broker-dealers not reflected in the above detailed cost estimates that do not currently report OATS data but could be CAT Reporters. First, there are at least 14 ELPs that do not carry customer accounts; these firms are not FINRA members and thus have no regular OATS reporting obligations. The Commission preliminarily believes that it is likely that these broker-dealers already have self-reporting capabilities in place because each is a member of an SRO that requires the ability to report OATS on request. The second group of broker-dealers that are not encompassed by the cost estimates of FINRA member broker-dealers discussed above are those that make markets in options and not equities. Although not required by the CAT NMS Plan to report their option quoting activity to the Central Repository, these broker-dealers may have customer orders and other activity that would cause them to incur a CAT Data reporting obligation. Based on CBOE membership data, the Commission believes there are 31 options market-making firms that are members of multiple SROs but not FINRA. The third group comprises 24 broker-dealers that have SRO memberships only with CBOE; the Commission believes this group is comprised primarily of CBOE floor brokers and, further, preliminarily believes these firms would incur CAT implementation and ongoing reporting costs similar in magnitude to small equity broker-dealers that currently have no OATS reporting responsibilities because they would face similar tasks to implement and maintain CAT reporting. The Commission assumes the 31 options market-making firms and 14 ELPs would be typical of the Reporters Study's large, non-OATS reporting firms because this group encompasses large broker-dealers that are not FINRA members, a category that would exclude any broker-dealer that carries customer accounts and trades in equities. For these 45 firms, the Commission relies on cost estimates from the Reporters Study.

The category of Insourcers that do not currently report OATS data includes firms that have multiple SRO memberships that exclude FINRA. This category includes Options Market Makers and at least 14 ELPs; these are firms that carry no customer accounts and directly route proprietary orders to Alternative Trading Systems; further information on these firms including the methodology by which they are identified can be found in the 15b9-1 Proposing Release. See Proposed Amendments to Rule 15b9-1, supra note 498, at 18052. Because the Commission has identified at least 14 ELPs, it can consider these firms separately from Options Market Makers for analysis. However, the Commission recognizes that some firms that are classified as Options Market Makers may actually be ELPs, if they were not identified as ELPs previously and are members of CBOE; because the same cost estimates are used for these groups, this misclassification does not affect the Commission's aggregate cost estimates for broker-dealers. The Commission recognizes that some FINRA member firms also make markets in options; if these firms report more than 350,000 OATS ROEs per month, the Commission's estimate of these firms' costs would be based on the estimates for OATS-reporting large firms based on data in the Reporters Study, which are higher than estimates for non-OATS reporting large firms (which include Options Market Makers that do not currently report OATS). If FINRA member Options Market Makers report fewer than 350,000 OATS ROEs per month or are exempt or excluded from reporting, they would be incorrectly classified as Outsourcers. Furthermore, ELPs that were not included in the analysis for the 15b9-1 Proposing Release and are not CBOE members would be incorrectly classified as new Outsourcers.

Most if not all ELPs have SRO memberships that require them to report OATS data upon request. Consequently, these firms are likely to have infrastructure in place that would reduce their implementation costs for CAT. The Commission preliminarily believes that this is reflected in the lower CAT implementation costs that the Plan estimates for large firms that do not currently report OATS; these estimates form the basis of the Commission's estimates of costs that ELPs would face if CAT were approved.

See Section III.B.9, supra; see also Exemption Order, supra note 18, at 11857-58.

The Commission identified 39 CBOE-member broker-dealers that are not FINRA members, but are members of multiple SROs; 8 of these broker-dealers were previously identified as ELPs, leaving 31 firms with multiple SRO memberships that are unlikely to be CBOE floor brokers. These 31 firms are likely to include some ELPs. This methodology implicitly assumes that there are no Options Market Makers that are not members of the CBOE. Because the Commission uses the same cost estimates for ELPs and options market making firms, uncertainty in the classification of the 31 Non-FINRA member CBOE member firms does not impact the Commission's cost estimates. The Commission recognizes that Options Market Makers may be FINRA members, but preliminarily believes these broker-dealers would be identified as Insourcers using FINRA data discussed in Section IV.F.1.c(2)B.i and thus would not fall under cost estimates produced by the Outsourcing Cost Model.

The Commission recognizes that additional broker-dealers may be members of neither FINRA nor CBOE, yet may incur CAT reporting obligations if the Plan is approved. Indeed, the Plan estimates that 100 CAT Reporters are not currently FINRA members (B.7.(b)(ii)(B)(2)), while the Commission estimates 69 (24 floor brokers, 31 Options Market Makers, and 14 ELPs). The Commission has determined that categorizing additional broker-dealers that are currently classified as exempt or excluded FINRA members as non-FINRA members would not change the cost estimates because these groups have identical estimated per-firm costs.

The estimated costs in the Reporters Study for non-OATS reporting firms are lower than the Reporters Study's estimated costs for large OATS-reporting firms; in reviewing the Reporters Study data, the Commission considered the possibility that firms that do not currently report OATS may systematically underestimate the costs they would incur to initiate and maintain the type of comprehensive regulatory data reporting that OATS entails or the CAT NMS Plan would entail. After discussions with multiple broker-dealers, the Commission, however, preliminarily believes that large non-OATS reporting firms would likely have lower CAT Data reporting costs than current OATS reporting large firms because large non-OATS reporting firms tend to be cutting-edge technology firms that already have a centralized IT infrastructure; they are unlikely to have a fragmented structure with multiple legacy systems. A centralized IT infrastructure with cutting-edge technology would likely simplify their implementation of the CAT NMS Plan, as fewer of their systems would need altering and fewer servers would be subject to clock synchronization requirements.

The Commission presents cost estimates for individual broker-dealers in Table 7 that include estimates of current costs, CAT implementation costs, and ongoing CAT reporting costs. In addition, Table 7 presents cost estimates for three categories of costs: Hardware/software; staffing; and outsourcing. Table 7 also presents a total across these three categories. Current data reporting cost estimates range from $167,000 annually for floor broker and firms that are currently exempt from OATS reporting requirements to $8.7 million annually for firms that currently report more than 350,000 OATS ROEs per month (“Insourcers”). One-time implementation costs range from $424,000 for current OATS reporters that are assumed to outsource (“OATS Outsourcers”) to $7.2 million for Insourcers. Ongoing annual costs range from $443,000 annually for firms that are assumed to outsource (OATS Outsourcers, New Outsourcers and Floor Brokers) to $4.8 million for Insourcers.

The Commission preliminarily believes that “Hardware/Software” costs include technology such as servers and telecommunications infrastructure necessary to report data to the Central Repository, as well as software that must be acquired or costs to alter existing software. “Staffing” includes the costs of employees assigned to regulatory data reporting, and includes existing staff as well as staff that would need to be hired if the CAT NMS Plan is approved. “Outsourcing” includes costs of service bureau relationships, legal and technical consulting, as well as other services that firms would need to acquire from service vendors to accomplish CAT reporting.

Rounding may cause totals to vary from the sum of individual elements in Table 7.

Table 7—Cost Estimates for Individual Broker-Dealers by Type

Broker-dealer type Costs
Hardware/ software Staffing Outsourcing Total
Current Costs:
Insourcers $720,000 $7,587,000 $400,000 $8,707,000
ELPs 3,000 1,409,000 22,000 1,433,000
Options Market Makers 3,000 1,409,000 22,000 1,433,000
OATS Outsourcers 0 212,000 124,000 336,000
New Outsourcers 0 42,000 124,000 167,000
Floor Brokers 0 42,000 124,000 167,000
CAT Implementation:
Insourcers 750,000 6,331,000 150,000 7,231,000
ELPs 450,000 3,416,000 10,000 3,876,000
Options Market Makers 450,000 3,416,000 10,000 3,876,000
OATS Outsourcers 0 424,000 0 424,000
New Outsourcers 0 849,000 0 849,000
Floor Brokers 0 849,000 0 849,000
CAT Ongoing:
Insourcers 380,000 4,256,000 120,000 4,756,000
ELPs 80,000 3,144,000 1,000 3,226,000
Options Market Makers 80,000 3,144,000 1,000 3,226,000
OATS Outsourcers 0 318,000 124,000 443,000
New Outsourcers 0 318,000 124,000 443,000
Floor Brokers 0 318,000 124,000 443,000
Outsourcing costs are modelled on an individual broker-dealer basis. Category averages are presented here.

Table 8 presents aggregate total costs to broker-dealers by broker-dealer type. The Commission estimates that broker-dealers spend approximately $1.6 billion annually on current regulatory data reporting activities. The Commission estimates approximate one-time implementation costs of $2.1 billion, and annual ongoing costs of CAT reporting of $1.5 billion. The Commission notes that estimates of ongoing CAT reporting costs of $1.5 billion are slightly lower than current data reporting costs of $1.6 billion. This differential is driven by reductions in data reporting costs reported by large OATS-reporting broker-dealers in the Reporters Study survey. The Commission estimates that all other categories of broker-dealers would face significant increases in annual data reporting costs.

In the Reporters Study, Large OATS Reporters cite average current data reporting costs of $8.32 million and Approach 1 maintenance costs of $4.5 million annually.

Table 8—Aggregate Broker-Dealer Cost Estimates

Broker-dealer type Costs Count Individual total Aggregate total
Hardware/ software Staffing Outsourcing
Current Data Reporting Costs:
Insourcers $720,000 $7,587,000 $400,000 126 $8,707,000 $1,097,130,000
ELPs 3,000 1,409,000 22,000 14 1,433,000 20,068,000
Options Market Makers 3,000 1,409,000 22,000 31 1,433,000 44,437,000
OATS Outsourcers 0 212,000 124,000 806 336,000 271,113,000
New Outsourcers 0 42,000 124,000 799 167,000 133,137,000
Floor Brokers 0 42,000 124,000 24 167,000 3,999,000
Total 1,800 1,569,884,000
CAT Implementation Costs:
Insourcers 750,000 6,331,000 150,000 126 7,231,000 911,144,000
ELPs 450,000 3,416,000 10,000 14 3,876,000 54,257,000
Options Market Makers 450,000 3,416,000 10,000 31 3,876,000 120,141,000
OATS Outsourcers 0 424,000 0 806 424,000 342,026,000
New Outsourcers 0 849,000 0 799 849,000 678,111,000
Floor Brokers 0 849,000 0 24 849,000 20,369,000
Total 2,126,048,000
CAT Ongoing Costs:
Insourcers 380,000 4,256,000 120,000 126 4,756,000 599,285,000
ELPs 80,000 3,144,000 1,000 14 3,226,000 45,160,000
Options Market Makers 80,000 3,144,000 1,000 31 3,226,000 99,998,000
OATS Outsourcers 0 318,000 124,000 806 443,000 356,764,000
New Outsourcers 0 318,000 124,000 799 443,000 353,666,000
Floor Brokers 0 318,000 124,000 24 443,000 10,623,000
Total 1,465,496,000
Outsourcing costs are modeled on an individual broker-dealer basis. Category averages are presented here.

d. Costs to Service Bureaus

The Plan discusses costs that service bureaus would face to implement the CAT NMS Plan and maintain ongoing CAT reporting. The CAT NMS Plan's cost estimates for service bureaus are based on the Participant's Costs to Vendors Study (“Vendors Study”), which gathered data from third-party vendors. The Vendors Study requested information from thirteen (13) service providers about their potential costs for reporting CAT Data—five (5) service providers responded. The CAT NMS Plan cites aggregate implementation costs of $51.6 million to $118.2 million for service bureaus, depending on whether Approach 1 or Approach 2 is selected, where Approach 1 would be more costly to vendors. Aggregate ongoing annual cost estimates ranged from $38.6 million to $48.7 million.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(D), Appendix C, Section B.7(b)(iv)(A)(4).

See id. at Appendix C, Section B.7(b)(i)(A)(3); Appendix C, Section B.7(b)(iii)(D). The Commission preliminarily believes that most if not all market participants that responded to the Vendors Survey are service bureaus, but it is possible that some respondents are firms providing technology rather than service bureau services.

Approach 1 allows broker-dealers to submit data to the Central Repository using their choice of existing industry messaging protocols while Approach 2 would specify a pre-defined format. See Section IV.E.1.b(3), supra.

The Commission preliminarily believes that costs that service bureaus would face to implement CAT should be included as part of the aggregate costs of CAT. While the CAT NMS Plan does not require the use of service bureaus to report CAT Data, the Commission recognizes that the most cost effective manner to implement the CAT NMS Plan likely would be for most market participants to continue their current practice of outsourcing their regulatory data reporting to one or more service bureaus. By doing so, the roughly 1,600 broker-dealers predicted to outsource would avoid incurring a significant fraction of CAT implementation costs; instead, service bureaus would incur implementation costs on their behalf. Based on conversations with market participants, the Commission preliminarily believes that these implementation costs are likely to pass-through to broker-dealers that outsource data reporting, because service contracts between broker-dealers and service bureaus are renegotiated periodically, and approval of the CAT NMS Plan might trigger renegotiation as the bundle of services provided would materially change. Consequently, service bureaus likely would renegotiate their client agreements during the period of implementation of the CAT NMS Plan. The Commission preliminarily recognizes that service bureaus may, when re-negotiating these service contracts factor in the CAT implementation costs the service bureaus incurred; consequently, broker-dealers could see increases in costs that reflect a service bureau's efforts to recoup those costs. In its analysis of costs, the Commission includes these service bureau costs and separately identifies them as service bureau implementation costs, but the Commission recognizes that they are likely to ultimately be borne by broker-dealers.

Although the Commission preliminarily believes that service bureau implementation costs would ultimately be passed on to broker-dealers, the Commission believes these costs are not double-counted in this analysis because re-negotiation of service bureau's contracts with their clients is not explicitly factored in to the Outsourcing Cost Model. Instead, the Commission recognizes these costs as being borne by the service bureaus initially, and does not identify a specific mechanism by which they will ultimately be passed onto broker-dealers.

The Commission, however, preliminarily believes that the ongoing costs of CAT Data reporting by service bureaus would be duplicative of costs incurred by broker-dealers. The aggregate fees paid by outsourcing broker-dealers to service bureaus cover the service bureaus' costs of ongoing data reporting. To include ongoing service bureau costs as a cost of CAT would double-count the costs that broker-dealers incur for CAT Data reporting; thus, in aggregating the cost estimates for CAT, the Commission includes only the maximum implementation cost that vendors would likely face of $118.2 million.

2. Aggregate Costs to Industry

The Sections above provide four sets of cost estimates that together encompass the costs of the Plan. This Section discusses aggregation of these costs into the total costs of the Plan. The Plan provides estimates of the total costs to industry if the Commission approves the Plan. The Plan estimates initial aggregate costs to industry of $3.2 billion to $3.6 billion and annual ongoing costs of $2.8 billion to $3.4 billion, with system retirement costs of $2.6 billion. The Commission estimates that industry would spend $2.4 billion to implement CAT, and $1.7 billion per year in ongoing annual costs.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iv)(A)(5).

Using estimates discussed above, the Commission recalculated total implementation and ongoing annual costs, partitioned across market participant types as possible. Because the Plan does not discuss how Central Repository costs would be partitioned across Participants and CAT Reporters, the analysis here presents Central Repository costs separately from costs to Participants and costs to CAT Reporters. The Plan presents some costs related to constructing and operating the Central Repository as ranges; in these cases, the Commission uses range maximums in the total cost calculation. Where costs differ for Approach 1 and Approach 2, the Commission uses estimates for the approach that is more costly in aggregate.

Approach 1 aggregate costs are higher than those for Approach 2 for all market participants except in one case where service bureaus have lower ongoing costs for Approach 1. In its discussion of industry (broker-dealer) costs, the Plan states that the cost differences between these two approaches are not statistically significant and that there would likely be no incremental costs associated with either approach. See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(C)(2)e.

Table 9 presents estimates of aggregate current, implementation, and ongoing costs to the industry. The Commission notes that costs to broker-dealers are much greater than the costs of building and maintaining the Central Repository. In terms of magnitudes of aggregate costs, costs to the 126 largest broker-dealers that currently report OATS data is the largest driver of implementation costs, accounting for 38.3% of CAT implementation costs. Although these firms would face significant costs in implementing CAT, the Reporters Study survey results suggest that they anticipate lower ongoing reporting costs than they currently incur ($599 million annually in expected aggregate costs versus $1.1 billion annually in current aggregate regulatory data reporting costs). For all other categories of broker-dealers, the Commission estimates ongoing annual costs to be higher than currently reporting costs.

As discussed in Section IV.F.1.c(1), supra, the Commission preliminarily believes that cost estimates for Large Broker-Dealers presented in the Plan are reliable.

Table 9—Aggregate Data Reporting Costs to Industry

Number Current costs CAT
Implementation Ongoing
Central Repository 1 $0 $92,000,000 $134,900,000
Participants (all) 1 154,100,000 41,100,000 102,400,000
Service Bureaus (all, 13) 1 Unknown 118,200,000 Excluded
Broker Dealers:
Insourcers (126) 126 1,097,130,000 911,144,052 599,285,000
Outsourcers (806) 806 271,113,000 342,026,100 356,764,000
New Small Firms (799) 799 133,137,000 678,111,300 353,666,000
ELPs (14) 14 20,068,000 54,257,245 45,160,000
Options Market Makers (31) 31 44,437,000 120,141,043 99,998,000
Options Floor Brokers (24) 24 3,999,000 20,368,800 10,623,000
Total BD 1800 1,569,884,000 2,126,048,540 1,465,496,000
Total Industry 1,723,984,000 2,377,348,540 1,702,796,000

Although the Commission relied on an alternative to the Reporters Study data to estimate costs for most broker-dealers, the Commission's aggregate cost estimate is consistent with information presented in the Plan that suggests that ongoing costs under CAT would likely be lower than ongoing costs for current reporting systems. The Plan, however, also discusses significant costs ($2.6 billion) for retirement of current regulatory reporting systems.

See CAT NMS Plan, supra note 3, at Appendix C.

Id. at Appendix C, Section B.7(b)(iv)(A)(5).

The Commission has not included those costs in its estimate of the aggregate costs of the Plan for several reasons. First, for reasons discussed below, the Commission preliminarily believes that cost estimates provided in the Plan are unlikely to accurately represent the actual costs industry will face in retiring duplicative reporting systems. Second, the retirement of current regulatory reporting systems is not a requirement of the Plan and the timeline and process for their retirement is uncertain. While the Commission's cost estimates do not recognize explicit system retirement expenses, it also does not explicitly recognize savings from elimination of these systems, though they are recognized qualitatively as additional benefits of the Plan. The Commission preliminarily believes that this approach is conservative in the sense that (for reasons that are discussed below) system retirement costs are likely to be mitigated by incorporation of current reporting infrastructure into CAT reporting infrastructure, while cost savings associated with industry's need to maintain fewer regulatory data reporting systems are not explicitly recognized. Finally, while the Commission does not include explicit system retirement costs, the Commission does recognize that industry will experience a costly period of duplicative reporting if the CAT NMS Plan is approved, and the Commission believes it is possible that these costs may be conflated with actual retirement costs estimated in the Plan. These reasons are discussed further below. As discussed above, the Commission preliminarily believes that retirement costs are unlikely to reflect actual costs to industry in eliminating duplicative reporting systems for several reasons. First, for the majority of broker-dealers that outsource, system retirement would affect few in-house systems; these broker-dealers are likely to adapt the systems that interface with service bureaus for current regulatory data reporting to interface for CAT Data reporting. Consequently, the Commission believes that, for these broker-dealers, costs to implement CAT reporting are likely to implicitly accomplish the retirement of older regulatory data reporting systems because these older systems will be transformed—in whole or in part—into systems that accomplish CAT reporting. Second, for broker-dealers that self-report regulatory data, the Commission cannot determine the source of the costs of system retirement that are estimated in the Plan. At its simplest level, ceasing reporting activities would include scrapping IT hardware dedicated to the endeavor and terminating the employees responsible for such regulatory data reporting. The Commission recognizes that there are costs associated with those activities, but does not preliminarily believe their magnitude (estimated in the Plan as $2.6 billion) should approach or exceed the magnitude of costs of CAT implementation (estimated in this analysis as $2.4 billion). Although the Commission is uncertain what estimates were included in system retirement costs and the Commission recognizes that different survey respondents may have interpreted the question differently, the Commission preliminarily believes that the system retirement costs cited in the Plan might include industry estimates of an extended period of duplicative reporting costs, during which industry would report data to both CAT and to the systems that CAT would likely replace.

Id. at Appendix C, Section C.9.

Based on discussions with industry, the Commission believes that industry is likely to implement the CAT NMS Plan by repurposing systems and employees currently assigned to other regulatory data reporting. The cost of eliminating these resources, however, should provide an upper bound to what actual system retirement costs would be, because eliminating these resources is an available and effective means of retiring these systems; market participants could choose other methods if they are preferable in terms of reducing costs of system retirement or CAT implementation. See supra note 880.

The Commission preliminarily believes that the period of duplicative reporting would likely constitute a major cost to industry for several reasons. These reasons include the length of the duplicative reporting period; constraints on the capacity of industry to implement changes to regulatory reporting infrastructure that might cause market participants to implement changes using less cost-effective resources; and the inability of some market participants to implement duplicative reporting in house, necessitating that they seek service bureau relationships to accomplish their CAT reporting requirements.

Based on data provided in the Plan, the Commission believes that the period of duplicative reporting anticipated by the Participants is likely to last for 2 to 2.5 years. The Commission preliminarily believes that these estimates are reliable because they reflect the Participants' experience with their historical rulemaking activity, although the Commission preliminarily believes that some steps outlined by the Participants might happen concurrently with Commission rulemaking required to facilitate ending some duplicative reporting. The Plan outlines a timeline for eliminating duplicative reporting. The timeline begins when Industry Members (other than Small Industry Members) are required to begin reporting to the Central Repository. The elimination of duplicative reporting would require several steps: (1) The SROs would identify their respective duplicative SRO rules and systems; (2) the SROs would file with the Commission the relevant rule modifications or eliminations; (3) the Commission would review and consider such rule modification or elimination filings; and (4) subject to the requisite Commission approval, the SROs would then implement such SRO rule changes.

See CAT NMS Plan, supra note 3, at Appendix C, Section C.9. The elimination of duplicative reporting may or may not involve actually retiring IT systems. If current regulatory data reporting systems are adapted to report CAT Data, some of these systems may continue to also report duplicative data during the period of duplicative reporting. In such a case, system retirement would involve no longer using these systems to report the duplicative data and any savings may be associated with no longer requiring staff to maintain the software and systems that support the duplicative reporting.

According to the Plan, step (1)—SRO identification of duplicative SRO rules and systems—of the process could take 12 to 18 months from implementation. SROs have 12 months (in the case of duplicative rules and systems) or 18 months (in the case of partially duplicative rules and systems) to complete their analysis of existing rules and systems to identify which systems should continue collecting data, or whether data in the Central Repository could substitute for the information collected through rules and systems in place.

The Plan notes that if a Participant determines that sufficient data is not available to complete the analysis, a subsequent date could be identified for such a determination to be made.

Certain SRO rules or systems identified by the SROs in step (1) might first necessitate an SEC rule change before the SROs can properly modify or eliminate such SRO rule or system. If so, Commission rulemaking may be required. This step (1)—even for those SRO rule and system changes requiring Commission rulemaking—could still feasibly take less than 18 months total because the SRO's analysis of their rules and their corresponding SRO rule filings could be undertaken in parallel with any such related Commission rulemaking during this period.

See CAT NMS Plan, supra note 3, at Appendix C, Section C.9. For example, Commission rules that require broker-dealers to be able to report Large Trader or EBS data would prevent SROs from changing their rules to eliminate this capability. See id. Consequently, the timeframe for retirement of these systems may also be dependent on Commission rulemaking. The Commission recognizes that during the comment period of any SEC rulemaking, SROs might begin their analysis of their own rules and preparation of potential filings, possibly compressing this timeline further.

According to the Plan, step (2) of the process could take 6 months. After identifying the rules to eliminate or modify, the Plan provides the Participants with six months to file the proposed rule change with the Commission. It is possible for the Participants to file these sooner if their rule changes are not complex, but the Plan places an upper bound on this. Under this timeline, it could take 18 months to two years after the first broker-dealers start reporting to the Central Repository for Participants to file rules to eliminate duplicative reporting.

It could also take longer if the Participant determines that sufficient data is not available to complete such analysis by 12 or 18 months after Industry Member reporting to the Central Repository commences.

According to the Plan, step (3) of the process could take another 3 months to a year. The Commission recognizes that the approval process for Participant rule changes can take time. In particular, for the Commission to approve such rules could take another 3 to 12 months depending on how complex the rule change. However, the Commission preliminarily expects that as long as such rule changes would be fairly straight forward, approval would likely take 3 months or less. As such, the first three steps add up to 21 months to 27 months.

Step (4) involves implementing the Participant rule changes, which would eliminate duplicative reporting. The Plan states that Participants would, upon Commission approval of rule changes, implement the “. . . most appropriate and expeditious timeline . . . for eliminating such rules and systems.” The Commission preliminarily believes that the elimination of duplicative reporting will require significant planning and implementation, but believes that much of the required planning is likely to happen concurrently with the Commission approval process of the underlying SRO rules. Consequently, the Commission preliminarily believes that actual implementation could occur as soon as 90 days after approval, and is not likely to occur more than six months after approval. The Plan also states that Participants should consider in setting an implementation timeline, when the quality of CAT Data would be sufficient to meet surveillance needs. In addition, reducing some duplicative reporting could require changing Participant rules in response to the elimination or modification of Commission Rules.

See CAT NMS Plan, supra note 3, at Appendix C, Section C.9.

Based on the timelines for all four steps and the Commission's analysis of how this timeline would be affected by the need in some cases for Commission rulemaking, the Commission preliminarily believes that the period of duplicative reporting could last at least 2 years, and the period of system retirement could extend for up to 2.5 years after Industry Members begin reporting data, assuming SROs are not limited in their initial analysis by problems such as delays in Commission rulemaking or excessive Error Rates, and Commission approval of SRO rules is completed within 90 days of submission.

Second, industry-wide resources to update order-handling systems are limited. Based on conversations with market participants, the Commission preliminarily believes that while most Insourcers and service bureaus have permanent staff that specialize in these activities, some would rely on hiring additional staff or utilizing contractors to increase their capacity to implement changes to order handling and data reporting systems and support of duplicative reporting systems. Furthermore, multiple broker-dealers and service providers cited access to specialized staff as a constraint that limits their ability to implement regulatory rule changes, stating that while current and newly hired staff might be able to implement the CAT NMS Plan and continue supporting OATS, they would be unlikely to be able to continue to implement changes to both systems. Consequently, Insourcers and service bureaus would likely incur significant costs associated with hiring additional employees to implement the CAT NMS Plan and accomplish regulatory data reporting during any duplicative reporting period.

Third, the Commission preliminarily believes that some firms that are currently challenged to maintain their self-reporting of data may not have the resources to implement the CAT NMS Plan at the same time as current reporting absent a service bureau relationship. It is possible that a number of relatively large firms would seek to establish service bureau relationships to accomplish both CAT reporting and current reporting even as a number of very small firms that currently do not report OATS could seek to establish such relationships. This could precipitate a “crowded entrances” situation in the market to provide data reporting services. The establishment of these relationships would pose a significant cost to industry.

See supra note 934 and Section IV.G.1.d, infra.

The Commission expects that there would be some cost efficiencies with respect to current data reporting costs and CAT reporting costs during any period of duplicative reporting. For example, servers hosting software to produce records for CAT could possibly also host software to produce records for OATS during the duplicative reporting period because these regulatory reporting systems rely upon much of the same underlying data. However, the Commission does not currently have the necessary data to determine the extent of these efficiencies, which would vary across market participants. Therefore, the Commission cannot estimate duplicative reporting costs. The Commission preliminarily believes, however, that the current data reporting costs of $1.7 billion per year constitutes an estimate of the cost per year to industry of duplicative reporting requirements, as it represents the cost of duplicative reporting to industry if there are no efficiencies. The Commission notes, however, that staff required to implement changes to order handling systems are a limited resource. If market participants do not have adequate staffing to implement the changes required by CAT and maintain duplicative reporting, costs for duplicative reporting could exceed current reporting costs because market participants could have to rely on external staff (such as consultants) or contract through service bureaus to accomplish this reporting; this is likely to be more expensive than staff used for current reporting.

Further, the Commission does not believe that duplicative reporting costs should be added to the estimated aggregate costs of the CAT NMS Plan. The Commission believes that the aggregate costs above represent the total costs of the Plan and do not account for the differential between these costs and the costs the industry currently incurs for regulatory data reporting and maintenance. During the period of duplicative reporting, industry would incur the aggregate costs of accomplishing CAT reporting described above, plus the costs of current data reporting, which the Commission uses as an estimate of duplicative reporting costs. The Commission notes that market participants will incur costs equal to current data reporting costs if the Plan were not approved (because current regulatory data reporting would continue), or as duplicative reporting costs if the Plan were approved. Consequently, the Commission preliminarily believes these costs should not be considered as costs attributable to approval of the Plan, because market participants would bear these costs whether the Plan is approved or disapproved.

While broker-dealers are anticipated to bear the burden of the costs associated with CAT, including implementation costs, ongoing costs and duplicative reporting costs, the Commission does not know whether these costs would be passed on to investors, or whether these costs would be absorbed by the broker-dealers themselves. On one hand, it could be assumed that broker-dealers could pass on the costs associated with CAT to investors because broker-dealers currently already pass on certain regulatory fees to their customers. For instance, the SROs have adopted rules that require broker-dealer to pay Section 31 transaction fees, and some of these broker-dealers have in turn imposed fees on their customers in order to provide funds to pay for the fees owed to the SROs. However on the other hand, if the passing on of these costs is associated with higher fees, a given broker-dealer could decide to absorb these costs and not increase their fees, and by doing so, they may attract more customer order flow. The incremental order flow that the broker-dealer attracts from having lower fees relative to their competitors may indeed offset the costs associated with CAT that they incur by not passing these on to their customers. Other broker-dealers, cognizant that they could lose order flow to other broker-dealers that do not pass on the costs to their customers could strategically respond and thus, could also absorb these costs. Ultimately, the Commission does not know which situation is more likely to eventuate, primarily because the Commission generally does not know the cost structure of broker-dealers.

Under Section 31 of the Securities Exchange Act of 1934, SROs and all the national securities exchanges must pay transaction fees to the Commission based on the volume of securities that are sold on their markets. These fees are designed to recover the costs incurred by the government, including the Commission, for supervising and regulating the securities market and securities professionals. See “SEC Fee—Section 31 Transaction Fees,” available at https://www.sec.gov/answers/sec31.htm .

3. Further Analysis of Costs

a. Costs Included in the Estimates

In general, the CAT NMS Plan does not break down its cost estimates as a function of particular CAT NMS Plan requirements, although it does provide some cost information for certain requirements in the Plan. However, the Commission has considered which elements of the CAT NMS Plan are likely to be among the most significant contributors to CAT costs. The Commission preliminarily believes that significant sources of costs would include the requirement to report customer information, the requirement to report certain information as part of the material terms of the order, the requirement to use listing exchange symbology, and possibly, the inclusion of Allocation Reports. The Commission preliminarily believes that the clock synchronization requirements, the requirement that Options Market Makers send quote times to the exchanges, the requirement that the Central Repository maintain six years of CAT Data, and the inclusion of OTC Equity Securities in the initial phase of the implementation of the CAT NMS Plan are unlikely to be significant contributors to the overall costs of the Plan. Notably, the Commission believes that its estimates of the implementation costs and ongoing costs to industry above include each of the costs discussed in this Section because these provisions encapsulate major parts of the Plan.

The Commission preliminarily believes that the requirement in the CAT NMS Plan to report customer information for each transaction represents a significant source of costs. In particular, the adapting of systems to report customer information that is not included in current regulatory data on a routine basis could require significant and potentially difficult reprogramming because current audit trail data does not routinely provide this information. Consequently, this reprogramming could require gathering information from separate systems within a broker-dealer's infrastructure and consolidating it in one location, and redesigning an IT infrastructure to satisfy this requirement could interrupt other workflows within the broker-dealer, expanding the scope of systems that must be altered to accomplish CAT reporting. While the Commission preliminarily believes that the requirement to report customer information would be a significant source of costs, the Commission lacks the necessary information to estimate what proportion of the costs of the Plan are attributable to this requirement. The Plan does not provide information on the costs attributable to the reporting of customer information, and the Commission has no other data from which it can independently estimate these costs, because the Commission is not aware of any data currently available to it regarding the number of broker-dealers that would need to engage in significant reprogramming in order to report customer information as required in the Plan, or the costs of doing so. The Commission therefore seeks comment on the costs that would be attributable to the requirement to report customer information as set out in the CAT NMS Plan. The Commission also notes that the Plan reflects exemptive relief granted by the Commission in connection with this requirement. Specifically, as discussed further in the Alternatives Section, the Commission granted exemptive relief from certain requirements of Rule 613 to allow the alternative approach to customer information that leverages existing identifiers to be included in the Plan and subject to notice and comment. Based on cost survey data provided by the Participants, this approach would reduce quantifiable costs to the top three tiers of CAT Reporters by at least $195 million as compared to an approach that followed requirements of Rule 613 as adopted.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.1.a.iii.

See Exemption Order, supra note 18.

Id. at 17-18.

Similarly, the Commission preliminarily believes that the requirement to report material terms of the order that include an open/close indicator, order display information, and special handling instructions represents a significant source of costs. Not all broker-dealers are currently required to report these elements on every order and no market participants report an open/close indicator on orders to buy or sell equities. Thus, the adapting of some market participants' systems to report this information for each transaction could require significant and potentially difficult reprogramming that requires centralizing or copying information from multiple IT systems within the broker-dealer. As discussed above, redesigning a broker-dealer's IT infrastructure could disrupt multiple workflows and dramatically increase the costs associated with implementing the changes required by CAT. While the Commission preliminarily believes that this reprogramming would be a significant source of implementation costs, the Commission lacks the necessary information to estimate what proportion of the costs of the Plan are attributable to this requirement. The Plan does not provide information on the costs attributable to these elements of the Plan, and the Commission has no other data from which it can independently estimate the costs, because the Commission is not aware of any data currently available to it regarding the number of broker-dealers that would need to engage in significant reprogramming in order to report this information as required in the Plan, or the costs of doing so. The Commission therefore seeks comment on the costs that would be attributable to reporting the material terms of the order as set out in the CAT NMS Plan, including an open/close indicator, order display information, and special handling instructions.

See CAT NMS Plan, supra note 3, at Article I.

The Commission also preliminarily believes that the requirement to use listing exchange symbology in the CAT NMS Plan could represent a significant source of costs. The Plan requires CAT Reporters to report CAT Data using the listing exchange symbology format, which would also be used in the display of linked data; because broker-dealers do not necessarily use listing exchange symbology when placing orders on other exchanges or off-exchange, this requirement could require broker-dealers to perform a translation process on their data before they submit CAT Data to the Central Repository. The translation process could be costly to design and perform and result in errors that would be costly for the broker-dealers to correct. If other elements of the Plan were to necessitate a translation, then the listing exchange symbology could be fairly low cost because it would be just another step in the translation. However, if the Plan has no other requirement that would necessitate a translation, the costs of including listing exchange symbology on all CAT reports would include the costs of designing and performing the translation as well as the costs of correcting any errors caused by the translation. While the Commission preliminarily believes that the requirement to use listing exchange symbology could be a significant source of costs, the Commission lacks the necessary information to estimate what proportion of the costs of the Plan are attributable to this requirement. The Plan does not provide information on the costs attributable to this particular element of the Plan, and the Commission has no other data from which it can independently estimate these costs, because the Commission is not aware of any data currently available to it regarding the number of broker-dealers that would need to undertake the translation process, either as a result of this or other elements of Plan, or the costs of doing so. The Commission seeks comment on the costs that would be attributable to the requirement to report CAT Data using listing exchange symbology format as set out in the CAT NMS Plan.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.1.a.

For example, class A shares of ABC Company might be traded using ticker symbol “ABC A” on one exchange, “ABC_A” on another exchange, and “ABC.A” on a third. As written, the Plan would require all broker-dealers to use the listing exchange's symbol for its Central Repository reporting, regardless of the symbol in the order messages received or acted upon at the broker-dealer or exchange.

The Commission recognizes that industry would bear certain costs associated with Allocation Reports, particularly the requirement that the reports include allocation times. The Commission understands that some broker-dealers already record allocation times; broker-dealers that do not currently record these times will face implementation costs associated with changing their business processes to record these times. Implementation costs for allocation reporting may include significant costs associated with incorporating additional systems into their regulatory data reporting infrastructure to facilitate this reporting, if such systems would not already be involved in recording or reporting order events. Furthermore, Outsourcers could face significant implementation and ongoing costs associated with reporting Allocation Reports if their service bureaus do not extend their services to manage the servers that handle allocations. Because implementation costs for Allocation Reports would vary widely across broker-dealers and because the Plan does not break out costs associated with reporting allocation information, the Commission cannot separately estimate costs attributable to this reporting.

The Commission preliminarily believes that the clock synchronization requirements in the Plan represent a less significant source of costs. The CAT NMS Plan estimates industry costs associated with the 50 millisecond clock synchronization requirement, based on the FIF Clock Offset Survey. The FIF Clock Offset Survey states that broker-dealers currently spend $203,846 per year on clock synchronization activities, including documenting clock synchronization events. The FIF Clock Offset Survey states that firms expect the 50 millisecond requirement to increase those costs by $109,197 per firm.

See CAT NMS Plan, supra notes 3, at Section D.12, and note 127. The Commission notes that the survey has two limitations pertinent to specific cost estimates provided in the summary of survey results. First, cost estimates are likely to be significantly downward biased. Individual responses to cost data were gathered within a range; for example, a firm would quantify its expected costs as “Between $500K and less than $1M” or “$2.5M and over”. When aggregating these responses, FIF generally used the range midpoint as a point estimate; however, for the highest response, the range minimum was used (i.e., “$2.5M and over” was summarized as $2.5M.) This is likely to have produced a significant downward bias in aggregate survey responses. Second, the survey includes only broker-dealers and service bureaus, thus the data excludes exchanges. The Commission preliminarily believes this limitation would not significantly impact industry costs because all exchanges currently maintain clock synchronization standards finer than those discussed as alternatives.

See FIF Clock Offset Survey, supra note 127. This is based on the current practice of the broker-dealers who responded to the survey.

See id. at 16. The $109,197 figure is obtained by subtracting the cost of maintaining current clock offsets of $203,846 annually from the estimated per-firm annual cost of maintaining a 50 millisecond clock offset of $313,043; see also id. at 7 (“Even where firms were at the target clock offset, many firms cited additional costs associated with compliance including logging and achieving greater degrees of reliability”).

Based on discussions with industry, the Commission preliminarily believes that the majority of broker-dealers (Outsourcers) would not face significant direct costs for clock synchronization because time stamps for CAT Data reporting would be applied by service bureaus. However, the Commission preliminarily estimates there are 171 firms that make the insourcing-outsourcing decision on a discretionary basis; if these firms decide to insource their data reporting under CAT, each of these firms is likely to face costs associated with complying with new clock synchronization requirements. The Commission preliminarily estimates that industry-wide implementation costs for the 50 millisecond clock synchronization requirement would be $268 million, with $25 million annually in ongoing costs. The Commission preliminarily believes that approximately $19.7 million in broker-dealer implementation costs would be attributable to clock synchronization requirements. The Commission also preliminarily believes that service bureaus would face similar clock synchronization costs if the CAT NMS Plan is approved. Using 13 as an estimate of the number of service bureaus, approximately $1.4 million in service bureau implementation costs would be attributable to clock synchronization requirements in the Plan.

See Section IV.F.1.d for discussion of service bureau costs and the degree to which those costs might be passed on to broker-dealers.

These are the 126 current OATS reporters that report more than 350,000 OATS ROEs per month; the 31 options market-making firms; and the 14 ELPs.

See Section IV.H.2.a(1), infra, for a discussion of how these implementation costs might vary for different clock synchronization standards.

See id., for discussion of costs attributable to the 50 millisecond clock synchronization tolerance proposed in the Plan, including the $109,197 estimate of per-firm implementation costs of the 50 millisecond clock synchronization requirement; see also CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(A)(3). 171 broker-dealers × $109,197 = $18,672,687.

The CAT NMS Plan states that the Vendor Study was distributed to 13 service bureaus or technology-providing firms identified by the DAG. See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(A)(3). 13 service bureaus × $109,197 = $1,419,561. The Commission believes clock synchronization costs are already included in cost estimates provided in the Vendor Study. As discussed above (see Section IV.F.1.d), the Commission believes it is likely that these costs would ultimately be passed on to service bureaus' broker-dealer clients.

Other Plan requirements that the Commission preliminarily believes are unlikely to represent major contributions to the overall costs of the Plan include the requirement that Options Market Makers report the quote times sent to the exchanges, which the Plan estimates would cost between $36.9 million and $76.8 million over five years; the requirement to maintain six years of data at the Central Repository, which the Plan estimates would cost $5.59 million, and the inclusion of OTC Equity Securities in the initial phase of the implementation of the CAT NMS Plan.

See FIF, SIFMA, and Security Traders Association, Cost Survey Report on CAT Reporting of Options Quotes by Market Makers (November 5, 2013), available at http://catnmsplan.com/web/groups/catnms/@catnms/documents/appsupportdocs/p601771.pdf ; see also CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iv)(B).

See CAT NMS Plan, supra note 3, Section 12(m).

See id. at Section 12(q). The Commission does not have the information necessary to precisely estimate the costs that are incurred by including OTC Equity Securities in the initial phase of the implementation of the CAT NMS Plan, because the Plan does not separately present the costs associated with OTC Equity Securities. Because of low trading activity in the OTC equity markets, any significant costs associated with including OTC Equity Securities would be in implementation costs. Further, broker-dealers that implement CAT Data reporting for NMS securities may not incur significant additional costs to implement CAT Data reporting for OTC Equity Securities.

There are many other categories of costs that contribute to the aggregated estimates of the costs of the Plan in addition to the items discussed above. For example, in addition to providing CAT Reporters data on their Error Rates, the Plan states that the Participants believe that in order to meet Error Rate targets, industry would require certain resources, including a stand-alone testing environment, and time to test their reporting systems and infrastructure. There are also likely to be costs related to the Plan Processor's management of PII. As noted above, the Commission does not have sufficient information to analyze each individual category of costs, because the available cost estimates do not reflect a detailed breakdown of the expected cost of each element of the CAT NMS Plan. However, the Commission preliminarily believes that its estimates of implementation costs and the ongoing costs of the CAT NMS Plan reflect all relevant costs to industry.

The Commission also acknowledges that the costs associated with handling PII could create an incentive for service bureaus not to offer CAT Reporting services. The Commission does not believe that this incentive would significantly alter the services available to broker-dealers. For further discussion, see supra note 920 and Section IV.G.1.e, infra. The Commission also notes that, pursuant to the exemptive relief granted by the Commission, the approach to the reporting of Customer information in the CAT NMS Plan could allow for the bifurcation of PII reporting from the reporting of order data. See Exemption Order, supra note 18, at 11858-63.

b. Fees

The Plan states that the Operating Committee would have the authority to levy ancillary fees on both broker-dealers reporting to, and regulators accessing, the Central Repository. The Commission believes that ancillary fees levied on broker-dealers are unlikely to be levied broadly, because discussion in the Plan associates these fees with late and/or inaccurate reporting. The Plan also discusses ancillary fees possibly levied on regulators associated with the use of Central Repository data. The Commission recognizes that costs estimated in Bids for constructing and operating the Central Repository already anticipate use of the CAT Data by regulators, and that additional fees to access the data might give regulators incentives to make less use of the data than anticipated in the Benefits Section. However, any fee schedule proposed by the Participants would be filed with the Commission. Consequently, the Commission does not believe that the provisions for ancillary fees would likely significantly impact the costs or benefits of CAT.

See CAT NMS Plan, supra note 3, at Section 11.3(c).

4. Second-Order Effects and Other Security-Related Costs

a. Security

As noted in the Adopting Release, Commenters have expressed concerns regarding the risk of failing to maintain appropriate controls over the privacy and security of CAT Data. The Commission recognizes that investors and market participants could face significant costs if CAT Data security were breached.

See Adopting Release, supra note 9, at 45725, 45756-58.

The Commission believes that it is difficult to form reliable economic expectations for the costs of security breaches, because there are few examples of security breaches analogous to the type that could occur under the CAT NMS Plan. However, the Commission can break down the expected costs of security breaches into two components: The risk of a security breach and the cost resulting from a security breach. Therefore, the Commission separates its discussion of the expected costs of security breaches into these two components. The Commission recognizes that security risks could give rise to second order costs as well where the costs come not directly from the security breach but rather from the actions of market participants attempting to avoid security risks.

(1) Costs of a Security Breach

The form of the direct costs resulting from a security breach would vary across market participants and could be significant. For broker-dealers, investment advisers, and other similar institutions, a security breach could leak highly-confidential information about trading strategies or positions, which could be deleterious for market participants' trading profits and client relationships. A data breach could also expose the proprietary information about the existence of a significant business relationship with either a counterparty or client, which could reduce business profits.

Although the Plan does not require reporting positions, observation of a broker-dealer's recent executions can offer information about their change in position, or, potentially, information about their actual position if the audit trail information breached contains all trading activity since the creation of the position.

A data breach could also potentially reveal PII of Customers. Because some of the CAT Data that would be stored in the Central Repository would contain PII such as names, addresses and social security numbers, a security breach could raise the possibility of identity theft, which currently costs Americans billions of dollars per year. Because PII would be stored in a single, centralized location rather than stored across multiple locations, a breach in the Central Repository could leak all PII, rather than a subset of PII that could be leaked if the information was stored in multiple locations. As such, these costs associated with the risk of a security breach could be substantial in aggregate.

According to survey data, the Bureau of Justice Statistics reported $24.7 billion in identity theft costs in 2012, available at http://www.bjs.gov/content/pub/press/vit12pr.cfm .

At a June 23, 2015 congressional hearing titled, “Government Personnel Data Security Review”, Office of Personnel Management (OPM) Director Katherine Archuleta estimated the direct costs of the OPM data breach at $19 to $21 million. Available at http://www.c-span.org/video/?326710-1/opm-director-katherine-archuleta-testimony-spending-data-security&start=3304 . This breach of PII of current and former federal employees exposed PII for approximately 4 million individuals. Available at http://www.federaltimes.com/section/OPM-Cyber-Report/ . The Commission recognizes that the number of individuals whose PII would be stored in the Central Repository far exceeds the number of federal employees whose data was exposed in the OPM breach, and that these costs include only the direct costs (such as the provision of credit monitoring services to affected individuals) incurred by OPM and do not reflect the total costs that these individuals may face as a result of the data breach, which could be far larger than the direct costs faced by OPM. These indirect costs may include the consequences of the breach as well as costs of credit fraud and legal services to address consequences of the data breach. There may also be second-order effects to such a breach, if investors reduce their engagement with the securities industry to avoid these costs. See Section IV.F.4.a(3), infra.

A breach that reveals the activities of regulators within the Central Repository, such as data on the queries and processes run on query results, could compromise regulatory efforts or lead to speculation that could falsely harm the reputation of market participants and investors. For example, a breach could result in an article that reports on regulators querying trading information of certain individuals or broker-dealers, which could harm those individuals or broker-dealers even if no regulators open investigations. Further, perpetrators of a breach could attempt to trade on information on regulatory queries to try to profit ahead of public information of an action, to the disadvantage of other investors.

(2) Risk of a Security Breach

The Commission preliminarily believes that the risks of a security breach may not be significant because certain provisions of Rule 613 and the CAT NMS Plan appear reasonably designed to mitigate these risks. However, the Commission notes that the considerable diversity in the potential security approaches of the bidders creates some uncertainty about the effectiveness of the eventual security procedures and hence, the risk of a security breach.

The Commission notes that, at a minimum, the security of the CAT Data must be consistent with Regulation Systems Compliance and Integrity under the Exchange Act (“Reg SCI”) (17 CFR 242.1000 to 1007).

Provisions of Rule 613 provide safeguards designed to prevent security breaches. Rule 613(e)(4) requires policies and procedures that are designed to ensure the rigorous protection of confidential information collected by the Central Repository, and Rule 613(iv) requires that the Plan contain a discussion of the security and confidentiality of the information reported to the Central Repository. Rule 613 also restricts access to use only for regulatory purposes, and requires certain provisions that are designed to mitigate these security risks such as the appointment of a Chief Compliance Officer and annual audits of Plan Processor operating procedures.

The Plan also includes provisions designed to prevent security breaches. First, governance provisions of the CAT NMS Plan could mitigate the risk of a security breach. Section 4.12 of the CAT NMS Plan provides for a Compliance Subcommittee whose activities could reduce the risk that information is released to unauthorized entities. Among the Subcommittee's responsibilities is “the maintenance of the confidentiality of information submitted to the Plan Processor or Central Repository.” Furthermore, the Plan Processor is required to submit a comprehensive security plan to the Operating Committee and update this security plan annually. The security plan must cover all components of CAT, including physical assets and personnel; the plan “must document how the Plan Processor would protect, monitor and patch the environment; assess it for vulnerabilities as part of a managed process, as well as the process for response to security incidents and reporting of such incidents.” In addition, Section 6.2(b) of the Plan establishes a Chief Information Security Officer who is responsible for monitoring and addressing data security issues for the Plan Processor. Second, the Plan includes specific provisions designed to ensure the security of data in flight. For instance, the Plan requires that bulk extract data be encrypted, password protected and sent via secure methods of transmission. Third, Section 6.7(g) of the Plan requires that the Participants establish, maintain, and enforce written policies and procedures reasonably designed to (1) ensure the confidentiality of the CAT Data obtained from the Central Repository; and (2) limit the use of CAT Data obtained from the Central Repository solely for surveillance and regulatory purposes. Finally, the Plan makes further provisions designed to provide security for PII. For example, regulators authorized to access PII would be required to complete additional authentications, and PII would be masked unless users have permissions to view PII.

See CAT NMS Plan, supra note 3, at Section 4.12.

Id. at Section 6.12.

See id. at Appendix D, Section 4.

See id. at Appendix D, Section 8.2.2.

See id. at Appendix C, Section A.2(c).

As discussed in the Plan, the Participants collected information from the Bidders regarding security and confidentiality during the RFP process, however, there was considerable diversity in the approaches proposed by the Bidders and the Participants chose to give the Plan Processor flexibility on many implementation details and state the requirements as a set of minimum standards. These requirements include both general security and PII treatment requirements. General security requirements are designed to address physical security, data security during transmissions, transactions, and while at-rest, confidentiality, and a cyber-incident response plan. PII requirements include a separate PII-specific workflow, PII-specific authentication and access control, separate storage of PII data, and a full audit trail of PII access. Because many of the decisions that define security measures for the Central Repository are coincident with the selection of the Plan Processor, there is a degree of uncertainty with regards to security measures that would be implemented by the Plan Processor. Consequently, there is uncertainty about the significance of the risks, the expected costs of a breach when considering the likelihood of a data breach, and the second-order effects.

See id. at Appendix C, Section A.4; Appendix D, Section 4.

See CAT NMS Plan, supra note 3, at Appendix D, Section 4.1.2-4.1.6.

One study of 62 U.S. companies experiencing data breaches in 2015 puts the average cost per stolen record containing personal or sensitive information at $217; the average number of breached records per incident was 28,070. See Ponemon Institute, 2015 Cost of Data Breach Study: United States (May 2015) (noting, however, that the study specifically excluded breaches of over 100,000 records as not representative of “typical” data breaches). As one example of a large data breach, Target Corporation's 2013 data breach affecting 40 million credit card numbers and 70 million other records containing PII had, as of January 2015, resulted in $252 million of related expenses for Target. See Target Corporation, Form 10-K for the Fiscal Year ended January 31, 2015 (March 13, 2015). Because it is not clear what the risk of a breach would be for CAT, in terms of either likelihood or magnitude, these types of numbers are simply indicative; it is impossible to estimate with any precision what the cost of a breach might be. For example, a complete breach of the CAT System, including the PII storage, might expose records an order of magnitude larger than the Target breach; however the types of records stored in CAT could be more difficult to exploit than credit card information, but their exploitation might prove far more damaging to individuals and entities whose trading information, for example, were compromised.

The Commission preliminarily believes the Plan marginally increases the threat of breach of broker-dealer trading and business strategies because although SROs currently receive this data from their own members, SROs are expected to have access to other SROs data more readily within the Central Repository. There is some risk that SROs could use this data improperly to gain information on how broker-dealers interact with other SROs' trading platforms. The Plan includes certain measures that mitigate this risk, however, by restricting the use of CAT Data reported by other entities for business purposes.

Rule 613(e)(4)(1)(A) states that Participants and the Plan Processor “agree not to use such data for any purpose other than surveillance and regulatory purposes, provided that nothing in this paragraph (e)(4)(i)(A) shall be construed to prevent a plan sponsor from using the data that it reports to the central repository for regulatory, surveillance, commercial, or other purposes as otherwise permitted by applicable law, rule, or regulation.” Similar language appears in the CAT NMS Plan. The Commission preliminarily believes this provision does not increase security risks because the data reported to the Central Repository by a Participant is already available to that Participant. See CAT NMS Plan, note 3, supra, at Section 6.5(f)(i)(A).

(3) Second Order Effects

The desire to avoid direct costs of a security breach could motivate actions that would result in second order effects of security breaches. For example, if service bureaus perceive the costs and risks of a security breach to be great enough because of the addition of PII in the data, which is not included in current data, some could decide not to provide CAT Data reporting services. This could increase the potential for a short term strain on capacity and exacerbate the costs of this strain described above and below. Further, investors or other market participants could move their activity off-shore or cease market participation altogether to avoid having sensitive information stored in the Central Repository. Consequences of changes in investor behavior in response to the threat of a breach include: Investors holding suboptimal portfolios; lost profits to the securities industry; and higher costs of raising capital for U.S.-based securities issuers, if the public's willingness to participate in capital markets is sufficiently reduced.

See supra note 934 and Section IV.G.1.d, infra.

See Section IV.G.3, infra.

Nonetheless, the Commission preliminarily does not believe that the effect of the Plan on the risk or costs of a data breach would be great enough to result in significant second order effects. As discussed above, the Commission preliminarily believes the Plan marginally increases the threat of breach of broker-dealer trading and business strategies. However, the Plan includes certain measures that mitigate this risk. In light of these provisions, the Commission preliminarily believes that the Plan is unlikely to significantly deter broker-dealers from participating in markets. In addition, in deciding whether to trade in the U.S. markets or abroad, investors and other market participants would continue to assess a multitude of potential trade-offs. While the expected costs of a security breach may factor in, so would the level of investor protections, which the Commission preliminarily believes would increase if it approved the Plan.

See Section IV.E.2, supra.

Another possible second order effect of avoiding the risk and cost of a security breach event could be the risk that one or more service bureaus could choose to exit the market in providing data reporting services rather than change their business practices to report PII to the Central Repository, in order to assist their client(s) in meeting their reporting responsibilities under the Plan. Specifically, while some service bureaus currently handle PII for their broker-dealer clients, others do not or do so only on an occasional and limited basis. To the extent service bureaus that do not already handle such PII were to stop offering regulatory data reporting services due to an unwillingness to host such customer information, their customers would be forced to establish new service bureau relationships, or undertake self-reporting. This potentially would be very costly to the broker-dealer clients of the exiting service bureaus due to the switching costs that broker-dealers incur to change service bureaus. Such an event could also contribute to crowded entrances problems. As noted above, however, the approach in the Plan to the reporting of customer information could allow for the bifurcation of PII reporting from the reporting of order data, which could affect a service bureau's decision whether to exit the market for reporting services to a broker-dealer client. While the Commission cannot rule out that one or more service bureaus could choose to exit the data reporting services market to avoid the costs of a potential security breach, the Commission preliminarily believes that such exits are unlikely. In addition, the Commission preliminarily believes that security breach risks are unlikely to result in service bureau exit because the market for regulatory data reporting services is generally expanding and the trend is for more, not less, outsourcing. Consequently, the Commission preliminarily believes that market share in this market is valuable and existing competitors are unlikely to voluntarily exit the market abruptly.

See supra note 934.

See supra note 979.

See Section IV.G.1.d, infra.

b. Changes to CAT Reporter Behavior

The Commission acknowledges that increased surveillance could potentially impose some costs by altering the behavior of market participants. Benefits could accrue to the extent that improved surveillance, investigation, and enforcement capabilities allow for regulators to better identify and address violative behavior when it occurs; and to the extent that common knowledge of improved capabilities deters violative behavior. Costs could accrue to the extent that some forms of market activity, which are permissible and economically beneficial to the market and investors, could come under higher scrutiny, which could create a disincentive to engage in that activity.

See Section IV.E.2.c, supra.

In particular, the Commission acknowledges that some market participants could reduce economically beneficial behavior if those market participants believe that, because of enhanced surveillance, their activities would increase the level of regulatory scrutiny that they bear. In other words, if market participants engaging in non-violative activity believe that such activity could increase the likelihood of examinations, inspections, and other interactions with regulators, those market participants could reduce or cease such activity to reduce the frequency and costs of interactions with regulators, including staff time to accommodate inspections, facilitate examinations and answer regulatory inquiries. Because facilitating regulatory inquiries is costly to firms, such a firm might conclude that certain permissible activities generate insufficient profits to offset costs associated with the regulatory scrutiny generated by these activities, even if the firm's behavior is permissible and no fines or other penalties result from these inquiries. To the extent that market participants could reduce activity that benefits the market, this could impose costs on investors and the market in the form of a reduction in the economic value of such activity.

Additionally, in an environment of improved surveillance, regulators could increase the number of inspections, examinations and enforcement proceedings that they initiate. To the extent that these activities result in a reduction in violative behavior, the market benefits in not bearing the costs of this behavior. To the extent, however, the additional regulatory activity increases the number of inspections, examinations and enforcement on permissible activities, market participants would incur the increased costs of facilitating these regulatory inquiries. The Commission preliminarily believes, however, that these costs would be offset by other effects of CAT such as fewer ad hoc data requests, improvement in regulators' precision in selecting firms for risk-based exams, and other efficiency improvements, and that the related savings would likely be greater than such costs in aggregate.

See Section IV.E.2.c, supra.

For example, the Commission preliminarily believes that the Plan would improve the efficiency and effectiveness of risk-based exams. However, because the efficiency could increase the total number of risk-based exams, the total number of exams on permissible activity could go up even if the percentage of exams on permissible activity goes down.

c. Tiered Funding Model

The Commission preliminarily believes that establishing a small number of discrete fee tiers, as occurs under the Plan, could create incentives for CAT Reporters to alter their behavior to switch from one tier to another, thereby qualifying for lower fees. Specifically, in the discussion of Consideration 7, the Plan states that CAT Reporters would be classified into a number of groups based on reporter type and market share of share volume or message traffic and assessed a fixed fee that is determined by this classification. The higher-activity groups would be assessed higher fees. Equity Execution Venues would be classified into 2-5 fee tiers based on market share of share volume, option Execution Venues would be classified into a separate set of 2-5 fee tiers based on market share of share volume, and Industry Members would be classified into another set of 5-9 fee tiers based on message traffic. That is, the Plan describes a funding policy with a tiered funding model that places market participants who fall into the lower tiers at a fee advantage over the market participants that fall into the higher tiers. The Plan states that this funding model is designed to reward the characteristics—small market share of share volume in the case of Execution Venues, low message traffic in the case of broker-dealers—that would enable CAT Reporters to qualify for the lower tiers. The potential effect of rewarding these characteristics is to incent market participants at the margins to reconfigure their operations so as to qualify for smaller tiers than would otherwise apply. The potential for such an effect would be greater among those CAT Reporters that fall at the low end of a tier and could most easily alter their operations to qualify for a smaller tier. Similarly, the funding model could create incentives for a firm that has an activity level near the top of a tier to avoid additional market activity that might move it to a higher fee tier. For example, to control its tier level, a market participant could reduce its quoting activity or cease providing services in a set of securities. Such activity could affect liquidity and the availability of trading services to investors. The Commission notes, however, that because this incentive is contingent on being near a fee-tier cutoff point, it preliminarily believes relatively few market participants would likely be affected and thus market quality effects would likely not be significant. Furthermore, for those market participants near a cutoff point, managing activity to avoid a higher fee tier would necessarily incur costs of lost business and potential loss of market share, and would possibly be difficult to implement, which should mitigate any effects on market quality.

See CAT NMS Plan, supra note 3, at Section 11.3 and Appendix C, Section B.7(b)(4)(C).

The CAT NMS Plan defines “Execution Venue” as “. . . a Participant or an alternative trading system (“ATS”) (as defined in Rule 300 or Regulation ATS) that operates pursuant to Rule 301 of Regulation ATS (excluding any such ATS that does not execute orders).” The Plan also defines Industry Member as “. . . a member of a national securities exchange or a member of a national securities association”. See CAT NMS Plan, supra note 3, at Article I, Section 1.1 for definitions. Classification of Execution Venues into tiers is based on transacted volume market share of share volume (in the case of NMS stocks and OTC Equity Securities) or contract volume (in the case of listed options). For Industry Members, classification into tiers is based on message traffic. Based on conversations with Participants, the Commission preliminarily believes message traffic would be based on CAT Reportable Events reported to the Central Repository. See id. at Article XI, Section 11.3 for discussion of assignment to funding tiers.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(v).

This argument assumes that activity levels used to determine funding tiers do not naturally cluster near cutoffs, and that if such natural cutoff points exist, the Operating Committee would avoid setting such funding tier cutoff levels near those activity levels.

The Commission recognizes that the tiering of fees also could create calendar effects within markets. Although the Plan does not detail the horizon at which CAT would measure activity levels, the structure ultimately approved by the Operating Committee could affect market participant behavior near the end of a measuring period. For example, high levels of market activity during a measuring period might cause CAT Reporters to limit their activity near the end of a measurement period to avoid entering a higher fee tier. If this translates into a reduction in quoting activity, market liquidity conditions could deteriorate at the end of activity measurement periods, and improve when a new measurement period begins, for example.

The Commission notes that the Operating Committee has discretion under the Plan governance structure to make the tier adjustments discussed in Section 11.1.d for individual CAT Reporters. This provision might mitigate incentives for individual market participants to alter market activities to reduce their expected CAT fees.

d. Differential CAT Costs Across Execution Venues

The funding model proposed in the Plan is a bifurcated funding model, in which costs are first allocated between the group of all broker-dealers and the group of all Execution Venues, then within these groups by market activity level. The proposed funding model treats Execution Venues differently from broker-dealers; this differential treatment could introduce inefficiencies to the market for execution services. As discussed in a recent academic paper, differential funding models in execution venues could influence how broker-dealers route customer order flow, possibly to the detriment of execution quality realized by investors. The Commission preliminarily believes that the bifurcated funding model proposed in the Plan almost certainly results in differential CAT costs between Execution Venues because it would assess fees differently on exchanges and ATSs for two reasons. First, message traffic to and from an ATS would generate fee obligations on the broker-dealer that sponsors the ATS, while exchanges incur almost no message traffic fees. Second, broker-dealers that internalize off-exchange order flow, generating off-exchange transactions outside of ATSs, would face a differential funding model compared to ATSs and exchanges. The cost differentials that result might create incentives for broker-dealers to route order flow to minimize costs, creating a potential conflict of interest with broker-dealers' investor customers, who are likely to consider many facets of execution quality (such as price impact of a trade and probability of execution in a venue in which the order is exposed) in addition to any of these costs that are passed on to them.

See CAT NMS Plan, supra note 3, at Article XI.

See Robert H. Battalio, Shane A. Corwin and Robert H. Jennings, Can Brokers Have It All? On the Relation between Make-Take Fees and Limit Order Execution Quality (2015 working paper), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2367462 . (“Battalio, Corwin, and Jennings”).

See CAT NMS Plan, supra note 3, at Section 11.3.(b): “For the avoidance of doubt, the fixed fees payable by Industry Members pursuant to this paragraph shall, in addition to any other applicable message traffic, include message traffic generated by: (i) an ATS that does not execute orders that is sponsored by such Industry Member; and (ii) routing orders to and from any ATS sponsored by such Industry Member.” The Commission notes that exchange broker-dealers would be subject to message traffic fees as Industry Members under the Plan. However, the Commission notes that based on its analysis of OATS data from September 15-19, 2014, these broker-dealers are minor contributors to overall message traffic, accounting for less than 0.03% of OATS ROEs.

See CAT NMS Plan, supra note 3, at Article XI.

This assumes that CAT fees would ultimately be borne by the broker-dealers that make routing decisions. Currently, exchange access fees are often borne by broker-dealers that make routing decisions, as discussed in Battalio, Corwin, and Jennings. Id. If Execution Venues were to absorb these fees rather than pass them on to customers, broker-dealer routing decisions might not be affected. It is also possible that some Execution Venues could incorporate some sort of rebate for broker-dealer message fees into their fee schedules, effectively making some venues less expensive for broker-dealers to access.

In addition to friction created by the bifurcated structure of the funding model, the Commission preliminarily believes that the CAT NMS Plan funding model shifts broker-dealer costs associated with the Central Repository to all broker-dealers and away from Options Market Makers. The CAT NMS Plan provides that broker-dealers would not report their options quotations to the Central Repository, while equity market makers would report their equity quotations to the Central Repository. This differential treatment of market making quotes affects costs of funding the Central Repository in two ways. First, the elimination of Options Market Maker quotes from the message traffic of broker-dealers decreases the number of messages that must be reported and stored, which presumably reduces the overall cost of building and operating the Central Repository. This reduction in the overall cost of the Central Repository reduces costs to both broker-dealers and Execution Venues. Second, because Options Market Maker quotes would not be in the message traffic which determines the allocation of broker-dealer costs of the Central Repository, broker-dealers that do not quote listed options would pay a higher share of broker-dealer-assessed CAT fees than they would if Options Market Makers' quotes were included in the allocation of fees. Also, Options Market Makers would pay relatively lower fees than they would if their quotations were included in CAT message traffic from broker-dealers.

See Section IV.H.1.a, supra for a discussion of an alternative that would require Options Market Makers to report their quotes.

Although this differential treatment would marginally increase the cost of providing other broker-dealers services relative to options market making, the Commission preliminarily does not believe that this would materially affect a market participant's willingness to provide broker-dealer services other than options market making for several reasons. First, many market participants participate in both equities and options markets because activity in one market (equities or options) could be used to hedge positions acquired in the other market. Consequently, many firms already find it cost effective to participate in both markets. Second, broker-dealers participating in equity markets have significant infrastructure in place for serving that market and switching costs to participate in options market making are high due to the need to establish quantitative infrastructure to quote options, market connectivity, IT infrastructure, and clearing/settlement arrangements required to transact in options; consequently, reducing the cost to make markets in options is unlikely to attract broker-dealers to change their business models. Finally, the Commission believes that the market to provide liquidity in the options market is already a competitive one because many broker-dealers participate in that market and market share that is sufficient to cover substantial fixed costs of making markets in options is valuable; consequently, options market participants have incentives to compete to win market share. Without a market change that significantly affects profits to be made in options market making, it seems broker-dealers would need a competitive advantage relative to existing competitors to successfully win market share from the existing competitors. The Commission preliminarily believes that that broker-dealers that currently focus on equity market making and other broker-dealer services unrelated to options market making are likely to continue to focus on the markets in which they participate because their competitive advantages relate to these activities.

5. Request for Comment on the Costs

The Commission requests comment on all aspects of the discussion of the potential costs of the CAT NMS Plan. In particular, the Commission seeks responses to the following questions:

301. Do Commenters agree with the Commission's assessment of the potential costs of the CAT NMS Plan? Why or why not?

302. To what extent do the uncertainties related to future decisions about Plan implementation impact the assessment of potential costs of the Plan? Please explain.

303. Do Commenters agree that the Plan's level of detail regarding the drivers of the costs to build, operate, and maintain the Central Repository is sufficient to assess the economic effects of the Plan? If more detail is needed, how can this information be obtained?

304. Do Commenters agree that using the cost estimates provided in Bids from the Shortlisted Bidders provides reasonable estimates of costs to build and operate the Central Repository? Why or why not?

305. Estimates in the Plan suggest that the Participants' data reporting costs will significantly increase while surveillance costs will significantly decrease if the Plan is approved. Do Commenters agree that these changes are likely to occur? Please explain.

306. Do Commenters agree with the Commission's characterization of the limitations in the cost studies? Do Commenters agree with the Commission's assessment that the Vendors Study and Participants Study have reliable cost estimates? Do Commenters agree that cost estimates for large OATS Reporters and large non-OATS Reporters are reliable? Do Commenters agree that cost estimates for small reporters are unreliable? Why or why not? Do Commenters have more precise estimates of the costs than provided in the cost surveys?

307. The Commission re-estimated aggregated costs under a different set of assumptions than the Plan. Do Commenters agree that the re-estimated costs better represent the expected costs of the CAT NMS Plan? Why or why not? Do Commenters agree that most broker-dealers that report fewer than 350,000 OATS ROEs per month are likely to report this data through a service bureau?

308. Do Commenters agree with the estimates of annual service bureau costs for a very small OATS-reporting firm of $50,000 to $180,000 per year, which assumes that the service bureau provides order routing and an order-handling system? If not, please provide alternate estimates.

309. Do Commenters agree that the pricing function for service bureaus is concave (increasing at a decreasing rate)? Why or why not? The Commission assumes in its re-estimation that service bureau cost functions are approximately as concave as exchange pricing functions. Do Commenters agree? Why or why not?

310. Will the requirement to provide customer information to the Central Repository be a significant cost-driver for Outsourcers? Why or why not? Is the need for encryption of this data a significant cost-driver?

311. Will the anticipated retirement of duplicative reporting systems such as EBS affect Outsourcer costs? Why or why not? Will the reduction in ad hoc data requests significantly affect the costs incurred by service bureaus in assisting their clients in responding to these requests? Why or why not?

312. Are there ways in which the Commission could better estimate the aggregate costs of the CAT NMS Plan? If so, please explain.

313. Do Commenters agree with the Commission's assumption that most firms that report fewer than 350,000 OATS ROEs per month are self-clearing? If not, please explain. Do Commenters believe that these firms would have significantly higher implementation costs due to their need to provide this information to any service bureaus they use for regulatory data reporting?

314. Do Commenters agree that broker-dealers that are exempt or excluded from OATS reporting are likely to be small and should have their costs estimated as Outsourcers? If no, how many of these broker-dealers currently participate in more than 350,000 events that would be OATS-reportable, were they not exempt or excluded, per month?

315. Are Commenters aware of options market making firms that are FINRA members and report fewer than 350,000 OATS ROEs per month, or that are exempt or excluded from OATS reporting rules? If so, are there ways that the Commission can identify these firms to better estimate their costs under the Plan?

316. Are Commenters aware of ELPs that are not CBOE members that did not trade on ATSs in 2014? If so, are there ways that the Commission can identify these firms to better estimate their costs under the Plan?

317. Do Commenters agree that FINRA member broker-dealers that are Options Market Makers are unlikely to be exempt or excluded from OATS-reporting requirements, and are likely to report more than 350,000 OATS ROEs per month? If not, how many FINRA member Options Market Makers exist that are exempt or excluded from OATS reporting requirements, or that report fewer than 350,000 OATS ROEs per month? Are there methods by which the Commission could improve its estimates of costs these broker-dealers are likely to face if the Plan is approved?

318. According to survey results, Approach 1 aggregate implementation and ongoing costs are higher than those for Approach 2 for CAT Reporters, though not statistically so. The Commission notes that this cost estimate does not seem intuitive because Approach 2 could result in extra data processing by CAT Reporters to translate data into a fixed format whereas Approach 1 would require no translation. Why is the cost of Approach 1 anticipated to be higher than Approach 2? Can this be explained by the use of service bureaus whom CAT Reporters expect to charge the same for either approach? Can this be explained by the need to process data under either approach to replace ticker symbols with listing exchange symbology?

Approach 1 assumes CAT Reporters would submit CAT Data using their choice of industry protocols. Approach 2 assumes CAT Reporters would submit data using a pre-specified format.

319. Do Commenters believe that duplicative reporting systems will be retired and, if so, when? What systems do Commenters expect to be retired? Are there any systems that cannot be retired? What are the costs associated with retiring duplicative reporting systems? What are the benefits of retiring duplicative reporting systems? Would there be cost savings as a result of retiring any duplicative reporting systems? How does the timeline for retiring duplicative reporting systems affect the costs and benefits? Please explain.

See supra note 856.

320. Do service bureaus handle EBS reporting for their clients? To what extent would EBS reporting contribute to duplicative reporting costs or system retirement costs and savings?

321. The Commission's analysis discusses the Plan's timetable for retirement of duplicative reporting systems (i.e., a maximum of 2.5 years). Is the timetable for retirement of these systems in the Plan realistic and/or reasonable? Are there ways that the timetable for duplicative reporting system retirement could be accelerated? If so, how?

322. Do Commenters believe that the period of duplicative reporting that would precede the retirement of certain current, anticipated to be retired, regulatory reporting systems would impose significant cost burdens on industry? Are the Commission's estimates of those costs accurate? Are there dimensions of these costs that the Commission has not recognized? If so, what are they and what are their magnitudes?

323. What milestones should CAT be required to reach before duplicative reporting systems can be retired?

324. What costs would service bureaus face in accomplishing a period of duplicative reporting during which both CAT and the regulatory data reporting systems that the Plan anticipates would be retired are operational? How many FTEs would be involved?

325. What costs would broker-dealers face in accomplishing a period of duplicative reporting during which both CAT and the regulatory data reporting systems that the Plan anticipates would be retired are operational? How many FTEs would be involved?

326. The CAT NMS Plan estimates that market participants would face significant costs of approximately $2.6 billion in connection with retiring duplicative reporting systems. What expenses does this estimate cover, and which systems account for which costs? For some broker-dealers, would implementation of CAT reporting accomplish the retirement of other regulatory data reporting systems? How do system retirement costs differ between broker-dealers that outsource their data reporting versus those who perform this function in-house?

327. Do Commenters believe that the CAT NMS Plan would deliver additional cost savings from sources other than the retirement of duplicative reporting systems and a reduction in the amount of ad-hoc data requests to regulated entities? Are there any changes to the CAT NMS Plan that would increase the potential cost savings?

328. Are SROs adequately incentivized to retire current regulatory reporting and surveillance systems that might be replaced by CAT? Do they have incentives to resist the retirement of these systems that this analysis fails to identify?

329. Do Commenters agree that costs associated with the Plan incurred by broker-dealers could be passed down to their customers? Why or why not? If so, do Commenters have estimates regarding what fraction of broker-dealer costs would be passed down?

330. The Commission preliminarily believes that the Vendors Study measures ongoing costs that would also be captured by the third-party outsourcing costs in the other surveys. As a result, the Commission does not add these to the aggregated cost estimates. Do Commenters agree with this approach? Is there any double counting of costs across the surveys, or can the individual survey estimates be aggregated into an industry-wide estimate? Please explain.

331. According to survey results, Approach 1 aggregate implementation costs are higher than those for Approach 2 for vendors and ongoing costs are lower. The Commission notes that this implementation cost result does not seem intuitive because Approach 2 could result in creating a whole new data translation process to implement the Plan whereas Approach 1 would require no translation. Why is Approach 1 costlier for vendors to implement than Approach 2? Can this be explained by the need to process data under either approach to replace ticker symbols with listing exchange symbology?

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(iv)(A).

332. The Commission assumes that cost estimates from Participants include all costs the Participants would incur if the Plan is approved, and that other costs related to development of the Plan are not avoidable if the Plan is not approved. Is it reasonable for the Commission to treat all costs related to development of the Plan that are not included in implementation and ongoing costs as sunk costs? Why or why not?

333. To what degree would industry's costs to implement and maintain CAT reporting be passed on to investors? Would competition between broker-dealers affect the passing on of costs to investors? Why or why not?

334. How significant to the total industry costs of the CAT NMS Plan are clock synchronization requirements, the requirement that Options Market Makers send quote times to the exchanges, the requirement that the Central Repository maintain six years of CAT Data, and the inclusion of OTC Equity Securities in the initial phase of the implementation of the CAT NMS Plan? Why?

335. How significant to the total industry costs of the CAT NMS Plan is the requirement to report customer information to the Central Repository? What elements of this requirement contribute to its significance of the potential costs of the Plan? Are there ways in which this data can be made available to regulators that would prove less costly to industry and investors? If so, what are they?

336. How significant to the total industry costs of the CAT NMS Plan is the requirement to report certain information as part of the material terms of the order? What elements of this requirement contribute to its significance of the potential costs of the Plan? Are there ways in which this data can be made available to regulators that would prove less costly to industry and investors? If so, what are they?

337. How significant to the total industry costs of the CAT NMS Plan is the requirement to report information to the Central Repository using listing exchange symbology? What elements of this requirement contribute to its significance of the potential costs of the Plan? Are there ways in which this data can be made available to regulators that would prove less costly to industry and investors? If so, what are they?

338. How significant to the total industry costs of the CAT NMS Plan is the requirement to report allocation information to the Central Repository? What elements of this requirement contribute to its significance of the potential costs of the Plan? Are there ways in which this data can be made available to regulators that would prove less costly to industry and investors? If so, what are they?

339. Are there other requirements of the CAT NMS Plan that would be significant sources of costs? If so, what are they? Are there ways in which those requirements could be made less costly? If so, what are they?

340. Do Commenters agree that ancillary fees levied by the Plan Processor on broker-dealers in response to late or inaccurate reporting are unlikely to broadly levied on broker-dealers? Do Commenters believe they would comprise a significant source of CAT costs to industry? Why or why not?

341. Do Commenters agree with the Commission's analysis of potential cost savings from a reduction in the number (and ultimately the cost) of data requests as a result of regulators having direct access to CAT Data?

342. Do Commenters agree with the Commission's analysis of the risk of a security breach? Do Commenters agree with the Commission's analysis of the potential costs of a security breach? Are there factors not covered in the analysis? What are they? Are the security measures outlined in the Plan appropriate and reasonable? Why or why not?

343. Do Commenters agree with the Commission's analysis of potential changes to CAT reporter behavior? Why or why not? Are there additional factors that should be considered?

344. Do Commenters agree with the Commission's analysis of the Plan's funding model? Why or why not? Are there additional factors that should be considered?

345. Do Commenters agree with the Commission's analysis of potential costs resulting from differential CAT costs across Execution Venues? Why or why not? Are there additional factors that should be considered?

346. Should the Plan require the inclusion of a web-based manual data entry option for initial CAT reporting in addition to updates and corrections? Please explain. How would a web-based manual data entry option affect the costs incurred by CAT Reporters? Do any current regulatory data reporting systems have a web-based manual data entry option? If so, which ones and how often do broker-dealers utilize that option for data submission?

G. Efficiency, Competition, and Capital Formation

In determining whether to approve the CAT NMS Plan, and whether the Plan is in the public interest, Rule 613 requires the Commission to consider the impact of the Plan on efficiency, competition and capital formation.

The Commission preliminarily believes that the Plan generally promotes competition. However, as explained below, the Commission recognizes that the Plan could increase barriers to entry because of the costs to comply with the Plan. Further, the Commission's analysis identifies several limitations to competition, but the Plan contains provisions to address some limitations and Commission oversight can also address the limitations.

The Commission preliminarily believes that the Plan would improve the efficiency of regulatory activities and enhance market efficiency by deterring violative activity that harms market efficiency. The Commission preliminarily believes that the Plan would have modest positive effects on capital formation and that the threat of a security breach at the Central Repository is unlikely to significantly harm capital formation.

The Commission notes that the significant uncertainties discussed earlier in this economic analysis also affect the Commission's analysis of efficiency, competition, and capital formation. For example, the Commission recognizes that the uncertainties around the improvements to data qualities can affect the strength of the Commission's conclusions on efficiency, and the uncertainty regarding how the Operating Committee allocates the fees used to fund the Central Repository could affect the Commission's conclusions on competition. Additionally, the Commission recognizes that the Plan's likely effects on competition, efficiency and capital formation are dependent to some extent on the performance and decisions of the Plan Processor and the Operating Committee in implementing the Plan, and thus there is necessarily some further uncertainty in the Commission's analysis. Nonetheless, the Commission believes that the Plan contains certain governance provisions, as well as provisions relating to the selection and removal of the Plan Processor, that mitigate this uncertainty by promoting decision-making that could, on balance, have positive effects on competition, efficiency, and capital formation.

1. Competition

As required by Rule 613, the Plan contains an analysis of its expected impact on competition. The Plan's analysis considers potential impacts of the CAT NMS Plan on competition related to technology, cost allocation across CAT Reporters, and changes in regulatory reporting requirements. The Plan splits its analysis between “Participants and broker-dealers communities” and concludes that the Plan generally would avoid placing an inappropriate burden on competition in U.S. markets. The Plan's analysis states the criteria for evaluating impacts on competition by outlining the channel of potential impacts as policy changes caused by the Plan that “burden a group or class of CAT Reporters in a way that would harm the public's ability to access their services” and states that such impacts “should be measured relative to the economic baseline.”

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8 (noting that Rule 613(a)(1)(viii) requires the Plan to include a discussion of an analysis of the impact of the Plan on competition, efficiency and capital formation).

See id.

See id.

See id.

The Commission's evaluation of competition reorients the Plan's approach to analyzing competition, expands upon it, and notes some limitations in the scope and conclusions of the Plan's analysis. In particular, the Commission's analysis of competition is organized and segmented by the particular markets in which competition among service providers of types of services exists. The Commission's analysis focuses on four distinct markets: The market for trading services, the markets for broker-dealer services, the market for regulatory services, and the market for data reporting services. In the context of the Plan, this allows the competition analysis to consider a more complex interaction between all market participants in a defined market than would be feasible by focusing solely on market participant types. This approach allows the Commission to determine whether a differential impact across competitors affects overall competition in the market. Much like the Plan's criteria for evaluation, the Commission recognizes that any effects on competition, with respect to each market, should be compared to a Baseline that characterizes the competitive environment without the CAT NMS Plan. In addition, the Commission considered uncertainty in the effect of the Plan on competition in any of these markets.

After analyzing the discussion of competition and the other relevant provisions of the Plan in the context of four affected markets, the Commission preliminarily believes that, while there could be effects on individual competitors, these effects would not lead to changes to competition as a whole in affected markets in a way that would generate significant adverse effects. In sum, and as discussed in detail below, the Commission preliminarily believes that the Plan poses a risk for competition for trading services, but provisions in the Plan and Commission oversight could mitigate this risk. Additionally, the Plan could have a differential impact on the ability of smaller broker-dealers and broker-dealers subject to CAT reporting to compete in the various markets for broker-dealer services, but these differential impacts may not be significant enough to affect overall competition in the markets for broker-dealer services. Moreover, the Plan generally promotes competition to be the Plan Processor and competition for regulatory services, but friction in those markets could limit the competition. Finally, the Plan could have a harmful effect on competition in the market for data reporting services, at least in the short term, because of capacity constraints, but the prolonged implementation for small broker-dealers could limit these harmful effects.

a. Market for Trading Services

The Commission analyzed the CAT NMS Plan's economic effects on competition in the market for trading services, compared to the Baseline of the competitive environment without the Plan, and preliminarily believes that the Plan would not place a significant burden on competition for trading services. The Commission recognizes the risk for the Plan to have negative effects on competition and to increase the barriers to entry in this market, but preliminarily believes that Plan provisions and Commission oversight could mitigate these risks.

The market for trading services, which is served by exchanges, ATSs, and liquidity providers (internalizers and others), relies on competition to supply investors with execution services at efficient prices. These trading venues, which compete to match traders with counterparties, provide a framework for price negotiation and disseminate trading information. The market for trading services in options and equities consists of 19 national securities exchanges, which are all Plan Participants, and off-exchange trading venues including broker-dealer internalizers, which execute substantial volumes of transactions, and 44 ATSs, which are not Plan Participants. Since the adoption of Regulation NMS in 2005, the market for trading services has become more fragmented and competitive, and there has been a shift in the market share of trading volume among trading venues. For instance, from 2005 to 2013, there was a decline in the market share of trading volume for exchange-listed stocks on NYSE. At the same time, there was an increase in the market share of newer national securities exchanges such as NYSE Arca, BATS-Z, BATS-Y, EDGA and EDGX. During the same time period, the proportion of NMS Stocks trading off-exchange (which includes both internalization and ATS trading) increased; for example, during the second quarter of 2015, NMS Stock ATSs alone comprised approximately 15 percent of consolidated volume, and other off-exchange volume totaled 18 percent of consolidated volume over the same period. Aside from trading venues, exchange market makers provide trading services in the securities market. These firms stand ready to buy and sell a security “on a regular and continuous basis at publicly quoted prices.” Exchange market makers quote both buy and sell prices in a security held in inventory, for their own account, for the business purpose of generating a profit from trading with a spread between the sell and buy prices. Off-exchange market makers also stand ready to buy and sell out of their own inventory, but they do not quote buy and sell prices.

The Commission understands that ISE Mercury, LLC will become a Participant in the CAT NMS Plan and thus is accounted for as a Participant for purposes of this Notice. See supra note 3.

See Concept Release on Equity Market Structure, at 3598-3560, supra note 733 (for a discussion of the types of trading centers); see also Alternative Trading Systems with Form ATS on File with the SEC as of April 1, 2016, available at https://www.sec.gov/foia/ats/atslist0416.pdf .

See Securities Exchange Act Release No. 76474 at 81112, “Regulation of NMS Stock Alternative Trading Systems”, available at http://www.sec.gov/rules/proposed/2015/34-76474.pdf .

See id. at 81124.

See “Market Maker”, available at http://www.sec.gov/answers/mktmaker.htm (last visited April 18, 2016).

Laura Tuttle, OTC Trading: Description of Non-ATS OTC Trading in National Market System Stocks (March 2014), available at http://www.sec.gov/dera/staff-papers/white-papers/otc-trading-white-paper-03-2014.pdf .

The Plan examined the effect of the CAT NMS Plan on the market for trading services primarily from the perspective of the exchanges. The Plan asserts that distribution of regulatory costs incurred by the Plan would be distributed according to “the Plan's funding principles,” calibrated to avoid placing “undue burden on exchanges relative to their core characteristics,” and would thus not cause any exchange to be at a relative “competitive disadvantage in a way that would materially impact the respective Execution Venue marketplaces.” Likewise, the Plan asserts that its method of cost allocation would avoid discouraging entry into the Participant community because a potential entrant, like an ATS, would “be assessed exactly the same amount [of allocated CAT-related fees] for a given level of activity” both before and after becoming an exchange.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(a)(i); see also id. at Section 11.2 (for a discussion of the Plan's funding principles); Section, III.A.3.d, supra.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(a)(i).

The Commission also examined the effect of the funding model on competition in the market for trading services, including off-exchange liquidity suppliers and ATSs. In addition, the Commission considered the effect of implementation and ongoing costs of the Plan, whether particular elements of the Plan could hinder competition, and the effect of enhanced surveillance on competition in the market for trading services.

(1) The Funding Model

As noted above, the Operating Committee would fund the Central Repository by allocating its costs across exchanges, FINRA, ATSs and broker-dealers. The Operating Committee would decide which proportion of costs would be funded by exchanges, FINRA, and ATSs and which portion would be funded by broker-dealers. The Plan does not specify how the Operating Committee would select this allocation. However, the portion allocated to the exchanges, FINRA, and ATSs would be divided among them according to market share of share volume and the portion allocated to broker-dealers would be divided among them according to message traffic, including message traffic sent to and from an ATS. The Operating Committee would allocate fees for the equities market and options market separately based on market share in each market. The Operating Committee would file the fees resulting from its funding model with the Commission under the Exchange Act.

See id. at Article XI.

Id.

Any entity that becomes a new exchange would be required to join the CAT NMS Plan as a Participant. In addition, any new Participant to the Plan must pay a “Participation Fee,” to the Company “in an amount determined by a Majority Vote of the Operating Committee as fairly and reasonably compensating the Company and the Participants for costs incurred in creating, implementing, and maintaining the CAT.” This Participation Fee would be based on, among other potential factors, capital expenditures paid by the Company amortized over five years, costs incurred by the Company to accommodate the new Participant, and Participant Fees paid by other new Participants.

See id. at Section 3.3. The Commission notes that the Plan does not specify the Participation Fee. The Commission expects this fee to be filed as an amendment to the CAT NMS Plan under Rule 608 of Regulation NMS. See 17 CFR 242.608.

The Commission notes that Section 3.3(b)(v) of the CAT NMS Plan states, “In the event the Company (following the vote of the Operating Committee contemplated by Section 3.3(a)) and a prospective Participant do not agree on the amount of the Participation Fee, such amount shall be subject to review by the Commission pursuant to § 11A(b)(5) of the Exchange Act.” See CAT NMS Plan, supra note 3, at Section 3.3(b)(v); see also text accompanying notes 1038-1039, infra.

The Commission preliminarily believes that any impacts of such fees on competition in the market for trading services would manifest either through the model for the fees itself or through the later allocation of the fees across market participant types, across equity or options exchanges or, within market participant types and markets, through the levels of fees paid by each tier. Each of the different channels through which the Plan could have an adverse effect on competition is discussed separately below.

A. Funding Model

As discussed in Section IV.F.4.d, the Commission preliminarily believes that the structure of the funding model could provide a competitive advantage to exchanges over ATSs. The Plan states that an entity would be assessed exactly the same amount for a given level of activity whether it acted as an ATS or an exchange. However, FINRA would be charged fees based on the market share of off-exchange trading. ATSs, which are FINRA members, would presumably pay a portion of the FINRA fee through their broker-dealer membership fees. In addition, ATSs would pay a fee for their market share, which is a portion of the total off-exchange market share. Therefore, ATS volume would effectively be charged once to the broker-dealer operating the ATS and a second time to FINRA. This would result in ATSs paying more than exchanges for the same level of activity. Ultimately, if the funding model disadvantages ATSs relative to registered exchanges, trading volume could migrate to exchanges in response, and ATSs could have incentives to register as exchanges as well.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(iii)(C).

Id. at Section 11.3(b).

The Commission notes that ATSs currently incur a different set of regulatory fees than are incurred by exchanges, because ATSs are required to be members of a national securities association. FINRA charges its members fees to cover its regulatory costs. See FINRA Manual: Corporate Organization: By-Laws of the Corporation: Schedule A: Section 1—Member Regulatory Fees, available at http://finra.complinet.com/en/display/display_main.html?rbid=2403&element_id=4694 (“FINRA shall, in accordance with this Section, collect member regulatory fees that are designed to recover the costs to FINRA of the supervision and regulation of members, including performing examinations, financial monitoring, and policy, rulemaking, interpretive, and enforcement activities.”).

Additionally, the Commission preliminarily believes that the Participation Fee could discourage new entrants or the registration of an ATS as an exchange, increasing the barriers to entry to becoming an exchange. In particular, the factors listed in the Plan for determining the Participation Fee consider the previous costs incurred by the existing Participants but not the costs already incurred by the new Participant when it acted as an ATS. However, the Plan does not prescribe a set formula for determining the Participation Fee and the Plan does not preclude considering previous costs incurred by the ATS in the Participation Fee. In addition, although amendments designated by sponsors to an NMS plan as establishing or changing a fee may be effective upon filing with the Commission, the Commission may summarily abrogate the amendment that establishes (or in the future, changes) the Participation Fee within 60 days of its filing and require that the fee amendment be refiled in accordance with Rule 608(a)(1) and reviewed in accordance with Rule 608(b)(2) of Regulation NMS, if it appears to the Commission that such action is necessary or appropriate in the public interest, for the protection of investors, or the maintenance of fair and orderly markets, to remove impediments to, and perfect the mechanisms of, a national market system or otherwise in furtherance of the purposes of the Act.

See CAT NMS Plan, supra note 3, at Section 3.3(b).

See 17 CFR 242.608(a)(1); 608(b)(2); 608(b)(3)(i); and 608(b)(3)(iii). Pursuant to Rule 608(b)(2) of Regulation NMS, the Commission shall approve such amendment, with such changes or subject to such conditions as the Commission may deem necessary or appropriate, if it finds that such amendment is necessary or appropriate in the public interest, for the protection of investors and the maintenance of fair and orderly markets, to remove impediments to, and perfect the mechanisms of, a national market system, or otherwise in furtherance of the purposes of the Act. Approval of the amendment shall be by Commission order.

Further, because the funding model seems to charge ATSs more for their market share than exchanges, ATSs could pay relatively less for their market share as an exchange than as an ATS, countering this barrier to entry depending on the magnitudes of the two fee types.

B. Allocation of Fees

The Plan discusses the allocation of fees among market participants of different sizes within the same market participant type (Execution Venues versus broker-dealers), but does not discuss the allocation of fees across the different market participant types or markets. The Operating Committee would determine this allocation and would submit a filing to the Commission, which would be subject to Commission review and public comment. The Commission recognizes the potential for the Operating Committee to influence the market for trading services either by coordinating to favor one segment over another, or through an imbalance in the voting rights on the Operating Committee. The Commission also preliminarily believes that the Plan contains governance provisions that could mitigate such potential burdens on competition.

See supra notes 78 and 79 (describing how fee schedules for CAT could be filed and noting that they could take effect upon filing with the Commission).

The Commission recognizes that the potential for a burden on competition and effects on competitors in the market for trading services could arise from provisions relating to the allocation and exercise of voting rights. In particular, a concentration of influence over Committee decisions could directly and indirectly affect competition. The potential for concentration of influence over vote outcomes arises from proposed provisions to give one vote to each Plan Participant in an environment where some Participants are Affiliated SROs. Indeed, supermajority approval could be achieved through five of the 10 groups of Affiliated SROs and majority approval could be achieved with just four such groups. In light of this potential for concentration, voters could weigh some particular interests more than others. For example, the Participant groups with options exchanges could have the incentive to allocate a disproportionately low level of fees for options market share than for equity market share. Such an allocation could disadvantage competing Participants with only equities exchanges.

See CAT NMS Plan, supra note 3, at Section 4.3.

The CAT NMS Plan states that the Operating Committee shall consist of one voting member representing each Participant and that one individual may serve as the voting member of the Operating Committee for multiple Affiliated Participants and shall have the right to vote on behalf of each such Affiliated Participant. See id. at Section 4.2(a).

The twenty SROs that are Participants in the CAT NMS Plan include five sets of affiliated SROs (New York Stock Exchange LLC, NYSE Arca, Inc., and NYSE MKT LLC (the “NYSE Group”); The NASDAQ Stock Market LLC, NASDAQ OMX BX, Inc., and NASDAQ OMX PHLX LLC (the “NASDAQ Group”); BATS Exchange, Inc., BATS Y-Exchange, Inc., EDGX Exchange, Inc., and EDGA Exchange, Inc. (the “BATS Group”); Chicago Board Options Exchange, Incorporated and C2 Options Exchange, Incorporated (the “Chicago Options Group”); International Securities Exchange, LLC, ISE Gemini, LLC, and ISE Mercury, LLC (the “ISE Group”); and five independent SROs (National Stock Exchange, Inc.; Chicago Stock Exchange, Inc.; BOX Options Exchange LLC; Miami International Securities Exchange LLC; and Financial Industry Regulatory Authority, Inc.). The BATS Group has four votes, the NYSE Group, the NASDAQ Group and the ISE Group each have three votes, and the Chicago Options Group has two votes. See CAT NMS Plan, supra note 3, at Appendix C, Section D.11(b) (Affiliated Participant Groups and Participants without Affiliations). A majority approval requires eleven votes. This could include as few as four of the SROs and sets of affiliated SROs: the affiliated SROs that have four votes, two sets of affiliated SROs that have three votes, and one other SRO or set of affiliated SROs. Supermajority approval requires fourteen votes. This could include as few as five SROs and sets of affiliated SROs: the affiliated SROs that have four votes, three sets of affiliated SROs with three votes, and any additional SRO. Note also that as few as two sets of affiliated SROs could block a Supermajority approval by casting seven “no” votes: the affiliated SROs with four votes and any one of the affiliated SROs with three votes.

The inclusion of all exchanges on the Operating Committee could give the Plan Participants opportunities and incentives to share information and coordinate strategies in ways that could reduce the competition among exchanges or could create a competitive advantage of exchange trading over off-exchange trading. However, the Commission preliminarily believes that the Plan would limit these potential burdens on competition. In particular, the Plan includes provisions designed to limit the flow of information between the employees of the Plan Participants who serve as members of the Operating Committee and other employees of the Plan Participants. Additionally, the Plan includes provisions that guide the Operating Committee to set fees between exchanges and ATSs in a tiered fashion, based upon market share. Finally, Commission oversight could also mitigate any concerns that burdens on competition might arise as a result of this approach.

See infra note 1272. The Commission notes that FINRA could represent the perspectives of the off-exchange portion of the market, but FINRA would have only one vote and exchanges would have nineteen.

See CAT NMS Plan, supra note 3, at Section 9.6(a) (Participants may share Plan information with their employees and other Representatives on a need-to-know basis; their use of Plan information is restricted to what is needed to achieve plan regulatory objectives). Details on the implementation of these confidentiality provisions are not stated. However, see also id. at Section 9.6(c) (Participants may share information among themselves without Operating Committee approval in some instances).

See id. at Section 11.3; Appendix C, Section B.7(b)(iv)(C).

Additionally, the Commission agrees with the Plan's assessment that some governance features of the Plan would limit adverse effects on competition in the market for trading services. The governance structure of the Plan contains provisions to limit the incentive and ability of Operating Committee members to serve the private interests of their employers, such as rules regulating conflicts of interest. Such governance provisions could mitigate the potential for members of the Operating Committee to use their influence over the fee schedule to benefit their own enterprise in a way that unfairly harms the customers of competing exchanges and ATSs and places a burden on competition. Moreover, as discussed above, the Commission may summarily abrogate and require the filing of Plan amendments that establish or change a fee in accordance with Rule 608(a)(1) and review such amendments in accordance with Rule 608(b)(2) of Regulation NMS, if it appears to the Commission that such action is necessary or appropriate in the public interest, for the protection of investors, or the maintenance of fair and orderly markets, to remove impediments to, and perfect the mechanisms of, a national market system or otherwise in furtherance of the purposes of the Act. In such a case, if the Commission chooses to approve such amendment, it would be by order and, with such changes or subject to such conditions as the Commission may deem necessary or appropriate.

See supra note 796.

See supra note 1039.

(2) Costs of Compliance

Because all Participants but one compete in the market for trading services, the ability of affiliates to vote as a group could in principle allow a few large Participant groups to influence the outcome of competition in the market for trading services by making various decisions that can alter the costs of one set of competitors more than another set. Further, the Plan would allocate profits and losses from operating the Central Repository equally across Participants, which could advantage small exchanges in the event of a profit and disadvantage small exchanges in the event of a loss. This could negatively impact competition if the cost differentials are unnecessary in light of the cost-benefit trade-offs of alternatives and if the cost differentials are significant enough to alter the set of services that some Participants offer.

Generally, smaller competitors could have implementation and ongoing costs of compliance that are disproportionate relative to their size. Any choices that could exacerbate these differences could potentially result in the exit of smaller competitors. To lessen the impact of funding the Central Repository on smaller exchanges and ATSs, the Plan would apply a tiered funding model that charges the smallest exchanges and ATSs the lowest fees. Likewise, the Plan would apply a tiered funding model that would charge the smallest broker-dealers, including liquidity suppliers, the lowest fees. However, the Commission notes that the Plan does not indicate whether off-exchange liquidity providers would pay fees similar to similarly-sized ATSs and exchanges.

In addition, as noted above, the Plan provides that the Technical Specifications would not be finalized until after the selection of a Plan Processor, which would not occur until after any decision by the Commission to approve the Plan. The Commission recognizes that the costs of compliance associated with future technical choices or the selection of the Plan Processor could exacerbate the relative cost differential across competitors. For example, the Affiliated Participants on the Selection Committee could favor a Plan Processor that employs technology that would make implementation costs relatively higher for the exchanges that do not have affiliates. In addition, the Affiliated Participants, who have more votes on the Operating Committee, could be amenable to adding particular CAT Data items in the future that could expose violations on other exchanges, but not be amenable to CAT Data items that could expose violations on their own exchanges. While those groups could still use such data to surveil their own exchanges, if not in CAT Data, the data items would not be available for cross-market surveillance or efficient Commission examinations and enforcement. As such, the independent exchanges, which have only one vote on the Operating Committee, could face higher regulatory costs than exchanges of the Affiliated Participants. However, for the same reasons as stated above, the Commission preliminarily believes that the governance provisions of the Plan and Commission oversight could help to mitigate such effects on these competitors in the market for trading services.

See Section IV.C.2, supra.

(3) Enhanced Surveillance and Deterrence

The Commission also preliminarily believes that the CAT NMS Plan could promote competition in the market for trading services through enhanced surveillance and the deterrence of violative behavior that could inhibit competition. Should the Plan deter violative behavior, passive liquidity suppliers, such as on or off-exchange market makers could increase profits as a result of reduced losses from others' violative behavior. This increase in profits could encourage new entrants or could spark greater competition, which reduces transaction costs for investors. For example, spoofing, which involves building up the apparent depth of the market to trigger particular trading patterns and then trading against those patterns, could cause confusion about bona-fide supply and demand for a particular security. Liquidity providers could compete less than is optimal to provide liquidity in that security out of fear that they could suffer a decline in profitability if they trade at inopportune times as a result of others' spoofing behavior. If the Plan facilitates surveillance improvements that deter spoofing, it could increase incentives to provide liquidity and promote lower transaction costs for investors, particularly in stocks that may lack a critical mass of competing liquidity providers or that could be targets for violative trading behavior.

See Section IV.E.2.c, supra, for a discussion of how the CAT NMS Plan would enhance surveillance and deter violative behavior.

b. Market for Broker-Dealer Services

The Commission analyzed the effect of the CAT NMS Plan on the market for broker-dealer services. For simplification, the Commission presents its analysis as if the market for broker-dealer services encompasses one broad market with multiple segments even though, in terms of competition, it actually may be more realistic to think of it as numerous inter-related markets. The market for broker-dealer services covers many different markets for a variety of services, including, but not limited to, managing orders for customers and routing them to various trading venues, holding customer funds and securities, handling clearance and settlement of trades, intermediating between customers and carrying/clearing brokers, dealing in government bonds, private placements of securities, and effecting transactions in mutual funds that involve transferring funds directly to the issuer. Some broker-dealers may specialize in just one narrowly defined service, while others may provide a wide variety of services.

The market for broker-dealer services relies on competition among broker-dealers to provide the services listed above to their customers at efficient levels of quality and quantity. The broker-dealer industry is highly competitive, with most business concentrated among a small set of large broker-dealers and thousands of small broker-dealers competing for niche or regional segments of the market. To limit costs and make business more viable, small broker-dealers often contract with larger broker-dealers or service bureaus to handle certain functions, such as clearing and execution, or to update their technology. Large broker-dealers typically enjoy economies of scale over small broker-dealers and compete with each other to service the smaller broker-dealers, who are both their competitors and their customers.

See Securities Exchange Act Release No. 63241 (November 3, 2010), 75 FR 69791, 69822 (November 15, 2010) (Risk Management Controls for Brokers or Dealers with Market Access).

There are approximately 1,800 broker-dealers likely to be CAT Reporters, while approximately 2,338 broker-dealers would not be CAT Reporters because their businesses do not involve reportable events in securities covered by the Plan. Further, broker-dealers that are anticipated to have CAT reporting obligations could compete with the broker-dealers that would not have CAT reporting responsibilities in various broker-dealer market segments that are unrelated to CAT reporting. Some broker-dealers may offer specialized services in one line of business mentioned above, while other broker-dealers may offer diversified services across many different lines of businesses. As such, the competitive dynamics within each of these specific lines of business for broker-dealers is different, depending on the number of broker-dealers that operate in the given segment and the market share that the broker-dealers occupy.

Examples of these business activities include underwriting and advising. See supra note 864.

The Commission preliminarily believes costs of compliance incurred by broker-dealers to comply with the Plan, particularly to report order events to the Central Repository, will differ substantially between broker-dealers and may affect competition between smaller and larger broker-dealers. As discussed previously in the Commission's analysis of Costs, broker-dealers that outsource regulatory data reporting activities are expected to see their costs of regulatory data reporting increase, while broker-dealers that Insource may see a decrease in their regulatory data reporting costs. The Commission preliminarily believes this dynamic may affect competition between Outsourcers (that tend to be smaller) and Insourcers (that tend to be larger), and may increase barriers to entry in some segments of this market.

See Section IV.F.1.c(2)C, supra.

The Plan discusses certain aspects of competition pertaining to broker-dealers that relate to costs and the allocation of fees. The Plan states, “[b]roker-dealer competition could be impacted if the direct and indirect costs associated with meeting the CAT NMS Plan's requirements materially impact the provision of their services to the public. Further, competition may be harmed if a particular class or group of broker-dealers bears the costs disproportionately . . . .” The Plan asserts that it would have little to no adverse effect on competition between large broker-dealers, and would not materially disadvantage small broker-dealers relative to large broker-dealers. Regarding small broker-dealers, the Plan states, “. . . . [the allocation of costs on broker-dealers based on their contribution to market activity] may be significant for some small firms, and may even impact their business models materially . . . .” and that the Participants were sensitive to the burdens the Plan could impose on small broker-dealers, noting that such broker-dealers could incur minimal costs under their existing regulatory reporting requirements “because they are OATS-exempt or excluded broker-dealers or limited purpose broker-dealers.” The CAT NMS Plan attempts to mitigate its impact on these broker-dealers by proposing to follow a cost allocation formula that (in expectation) charges lower fees to smaller broker-dealers; furthermore, Rule 613 provides them additional time to commence their reporting requirements.

The Commission preliminarily agrees with the Plan's general assessment of competition among broker-dealers, and also with the Plan's assessment of differential effects on small versus large broker-dealers. The Commission agrees that the Plan's funding model is an explicit source of financial obligation for broker-dealers and therefore an important feature to evaluate when considering potential differential effects of the Plan on competition in the market for broker-dealers. The Commission understands that the tiered funding model should result in the smallest broker-dealers paying the smallest fees, but the Plan does not outline how the magnitudes of fees would differ across the tiers. The Commission also recognizes that the potentially greater level of service specialization that may characterize small broker dealers and the potentially non-linear economies of scale may result in the compliance costs associated with the Plan competitively disadvantaging small broker-dealers, on average, relative to large broker-dealers.

However, the Commission preliminarily believes that the segments of the market most likely to experience higher barriers to entry are those that currently have no data reporting requirements of the type the Plan requires and those that would involve more CAT Reporting obligations, such as the part of the broker-dealer market that involves connecting to exchanges, because of the technology infrastructure requirements and the potential to have to report several types of order events. The opportunity to rely on service bureaus or other solutions to reduce the costs of complying with the Plan could limit any increases in the barriers to entry in this market. Nonetheless, the Commission preliminarily believes that any increases in the barriers to entry are justified because they are necessary in order for the CAT Data to include data from small broker-dealers. In the Adopting Release, the Commission explained that excluding small broker-dealers from reporting requirements would “eliminate the collection of audit trail information from a segment of the broker-dealer community and would thus result in an audit trail that does not capture all orders by all participants in the securities markets.” The Commission further noted that “illegal activity, such as insider trading and market manipulation, can be conducted through accounts at small broker-dealers just as readily as it can be conducted through accounts at large broker-dealers” and that “granting an exemption to certain broker-dealers might create incentives for prospective wrongdoers to utilize such firms to evade effective regulatory oversight through the consolidated audit trail.”

The majority of broker-dealers do not directly engage in exchange trading, and most broker-dealers are not expected to have CAT reporting obligations. See supra note 864.

See Adopting Release supra note 9, at 45749.

See id.

The Commission also recognizes that the Plan could affect the current relative competitive positions of broker-dealers in the market for broker-dealer services. To varying degrees, the economic impacts resulting from the Plan could benefit some broker-dealers and adversely affect others. The magnitude of these effects on broker-dealers could vary across and within categories of broker-dealers and classes of securities. However, there is no clear reason to expect these impacts, should they occur, to decrease the current state of overall competition in the market for broker-dealer services so as to materially burden the price or quality of services received by investors on average.

Regardless of the differential effects of the CAT NMS Plan on small versus large broker-dealers, it is the Commission's preliminary view that the CAT NMS Plan, in aggregate, would likely not reduce competition and efficiency in the overall market for broker-dealer services. Even if small broker-dealers potentially face a burden, this may not necessarily have an adverse effect on competition as a whole in the overall market for broker-dealer services. Under the CAT NMS Plan, broker-dealers would have greater reporting responsibilities than they would otherwise have. Broker-dealers could face high upfront infrastructure costs to set up a processing environment to meet reporting responsibilities. Because these infrastructure costs are upfront, fixed costs, the burden to bear these costs could be potentially greater for small broker-dealers. Instead of bearing these costs in-house, small broker-dealers could contract with outside technology vendors for reporting services. This outcome could lead to lower costs relative to not using a vendor for reporting services. For these reasons, even firms that currently do not report to OATS, but will be CAT Reporters under the Plan, could face manageable upfront costs that permit them to continue in their line of business without a severe setback in their profitability.

The Commission notes that a difficulty in assessing the likely impacts of the CAT NMS Plan on competition among broker-dealers is that competition in the markets for different broker-dealer services could be affected in different ways. As mentioned above, there is great diversity in the business activities of broker-dealers. Broker-dealer services that are likely to incur CAT reporting responsibilities include: executing orders, whether it be as an ATS or acting as a carrying broker-dealer; intermediating between customers and carrying/clearing brokers; effecting transactions in mutual funds that involve transferring funds directly to the issuer; writing options; and acting as an exchange floor broker. As noted above, these broker-dealers may also compete with the approximately 2,338 other broker-dealers in market segments that are not related to CAT reporting, such as dealing in municipal bonds or arranging private placements of securities. If CAT costs represent a significant increase in overall costs, the Plan could disadvantage broker-dealers who are CAT Reporters in the market segments that do not require CAT reporting. For example, broker-dealers that, in addition to providing services related to market transactions that are reportable to CAT, also compete to provide fixed-income order entry as a line of business may be at a relative disadvantage to competitors in the fixed-income market who do not provide broker-dealer services that are related to market activity that is reportable to CAT. Whether this disadvantage amounts to a substantial reduction in competition in various markets depends on the magnitude of the disadvantage and whether it affects the price and level of services available to investors.

See Section IV.F.1.c, supra.

The Commission recognizes that the CAT NMS Plan could result in fewer broker-dealers providing specialized services that trigger CAT reporting obligations. The Commission preliminarily believes that this potential effect on broker-dealer specialization depends on whether three key conditions are met. First, the effect requires that, compared to large broker-dealers, small broker-dealers disproportionately specialize in providing regional or niche services to a particular market segment of clients. Second, the effect requires that this specialization is correlated with business risk associated with changes in marginal cost. Finally, the effect requires that the compliance costs of the CAT NMS Plan could affect the ability for some small broker-dealers to provide these specialized services. This effect, in which fewer broker-dealers compete in specialized market segments, could thereby negatively affect the competitive dynamics in these market segments, especially if these segments currently contain relatively few broker-dealers. The Commission preliminarily believes that these conditions could hold, particularly for smaller broker-dealers, and result in fewer broker-dealers operating in specialized or niche markets if the Plan is approved.

The Commission recognizes, however, that fewer broker-dealers in a specialized segment of the market may not necessarily harm competition in that segment. In particular, the costs of compliance with the Plan may be less of a relative burden for large broker-dealers who may, compared to small broker-dealers, provide a larger portfolio of specialized services to clients. This portfolio may buffer large broker-dealers from business risk associated with specialization. Because of the lower relative burden, large broker-dealers are more likely to maintain their presence in specialized market segments. If a sufficient number of large broker-dealers, or all broker-dealers more generally, maintain their presence in specialized market segments, a net decrease in broker-dealers may not affect the competition in such market segments to a level in which the market segment offers fewer or lower quality services or higher prices. However, the Commission recognizes that negative effects on competition in specialized market segments could result if broker-dealers achieve a level of market concentration necessary to adversely affect prices for investors.

c. Market for Regulatory Services

SROs compete in the market for regulatory services. Regulatory functions include market surveillance, cross-market surveillance, oversight, compliance, investigation, and enforcement, as well as the registration, testing, and examination of broker-dealers. Although the Commission oversees exchange SROs' supervision of trading on their respective venues, the responsibility for direct supervision of trading on an exchange resides in the SRO that operates the exchange. Currently, SROs compete to provide regulatory services in at least two ways. First, because SROs are responsible for regulating trading within venues they operate, their regulatory services are bundled with their operation of the venue. Consequently, for a broker-dealer, selecting a trading venue also entails the selection of a provider of regulatory services surrounding the trading activity. Second, SROs could provide this supervision not only for their own venues, but for other SROs' venues as well through the use of Regulatory Service Agreements or a plan approved pursuant to Rule 17d-2 under the Exchange Act. Consequently, SROs compete to provide regulatory services to venues they do not operate. Because providing trading supervision is characterized by high fixed costs (such as significant IT infrastructure and specialized personnel), some SROs could find that another SRO could provide some regulatory services at a lower cost than it would incur to provide this service in-house. Until recently, nearly all the SROs that operate equity and option exchanges contracted with FINRA for some or much of their trading surveillance and routine inspections of members' activity.

FINRA is the SRO responsible for supervision of trading off-exchange, which includes trading occurring on ATSs.

17 CFR 240.17d-2.

Every equity exchange except CHX and NSX has an RSA with FINRA which allows FINRA to provide cross-market surveillance for nearly 100% of the equity markets. These RSAs differ in scope, but in every case these contracts represent a partnership between FINRA and the other SROs to provide a full set of effective regulatory services. Recently NYSE Group and NASDAQ OMX decided to significantly scale back their RSA with FINRA and directly resume most of their market surveillance and investigation regulatory obligations.

As a result, the market for regulatory services in the equity and options markets currently has one dominant competitor, FINRA. This may provide relatively uniform levels of surveillance across trading venues. One SRO having a competitive advantage in providing such services could also limit the incentives to innovate in surveillance. Hypothetically, increases in the competition to provide regulatory services could promote regulatory oversight of exchanges and investor protection for investors. To the extent that a regulator could improve on current regulatory oversight, this could result in a better functioning, more liquid, financial market. However, it is possible that increased competition between SROs to provide regulatory services could have negative effects on the market if SROs compete on the basis of providing light-touch regulation, which might be less likely to detect violative activity.

The Commission preliminarily believes that the Plan could provide opportunities for increased competition in the market to provide regulatory services. In particular, designated regulatory Staff from all of the SROs would have access to CAT Data, which would reduce the differences in data access across SROs. This could reduce barriers to entry in providing regulatory services because data would be centralized and standardized, possibly reducing economies of scale in performing surveillance activities. Furthermore, because some types of previously infeasible surveillance would become possible with the availability of additional data, SROs would have greater opportunities to innovate in the type of surveillance that is performed, and the efficiency with which it is performed. In addition, when as Rule 613(a)(3)(iv) requires, SROs implement new or updated surveillance within 14 months after effectiveness of the CAT NMS Plan, any SRO could reconsider its approach to outsourcing its own regulation and whether it wants to compete for regulatory service agreements.

Without a Central Repository, an SRO wishing to compete as a regulatory services provider would need to invest in the IT infrastructure and enter into the data access agreements necessary to surveil broadly beyond its exchanges' data resources. By providing access to consolidated trade and order data to all SROs, CAT may reduce barriers to entry for this market. See Exemption for Certain Exchange Members, supra note 394, at 18057-58 (describing the barriers to entry of potential new national securities associations).

The Commission recognizes that efficient access to data is not the only prerequisite for entering the market to provide regulatory services and that high barriers to entry may still characterize this market.

See Section IV.G.2.a, infra, for a discussion of the efficiency improvements for surveillance.

d. Market for Regulatory Data Reporting Services

The Commission analyzed the effect of the CAT NMS Plan on competition in the market for data reporting services with a focus on its impact on the costs incurred by broker-dealers to comply with the Plan. As discussed in the Costs Section above, the Commission preliminarily believes that many broker-dealers, particularly smaller broker-dealers, would fulfill their CAT Reporting obligations by outsourcing to service bureaus and that the fees charged by the service bureaus would be a major cost driver for these broker-dealers. Further, these fees would factor into the increase in barriers to entry in the market for broker-dealer services. Therefore, the Commission preliminarily believes that any effects on competition in the market for regulatory data reporting services could have a significant effect on the costs incurred by broker-dealers in complying with the CAT NMS Plan.

See Section IV.G.1.b, supra.

The Plan provides information on broker-dealers' use of third-party service providers to accomplish current regulatory data reporting. The Plan notes that while some broker-dealers perform their regulatory data reporting in-house, others outsource this activity. The Plan does not state what proportion of broker-dealers currently outsources their regulatory data reporting work. However, the Commission interviewed a variety of broker-dealers and service bureaus in order to gain insight into the scope of broker-dealers' use of data reporting services. As noted in the Costs Section, the Commission understands that most firms outsource the bulk of their regulatory data reporting to third-party firms. The Commission preliminarily believes that the competition in the market to provide data reporting services is a product of firms choosing to perform this activity in-house or to outsource it based on a number of considerations including cost, with some firms choosing to outsource this activity across multiple service providers.

See Section IV.F.1.c(2)A, supra.

The market for regulatory data reporting services is characterized by bundling, high switching costs, and barriers to entry. The high IT infrastructure costs of regulatory data reporting creates economies of scale that give rise to the data reporting services provided by service bureaus. Broker-dealers, instead of investing in the IT infrastructure necessary for regulatory data reporting, could share the costs of the IT infrastructure with other broker-dealers by paying for a service bureau to report for them. Often, service bureaus bundle regulatory data reporting services with an order-handling system service that provides broker-dealers with market access and order routing capabilities. Sometimes service bureaus bundle regulatory data reporting services with trade clearing services.

See Section IV.F.1.c(2)A, supra, for more information on broker-dealer use of service bureaus.

In discussions with Staff, service bureaus stated that switching service bureaus can be costly and involve complex onboarding processes and requirements, that systems between service bureaus may be disparate, and switching service providers may require different or updated client documentation. However, service bureaus stated that on-boarding operations were infrequent and that it was rare for broker-dealers to switch between service providers. Difficulty switching between service providers could limit the competition among service bureaus to provide data reporting services, and impact the costs that Outsourcers incur to secure regulatory data reporting services. Furthermore, the high IT infrastructure costs also give rise to barriers to entry, which could slow the entry of new market participants into the market. Despite this, the trend in the market is toward expansion.

See supra note 920.

The Commission preliminarily believes that the Plan could alter the competitive landscape in the market for data reporting services in several ways. It is not clear whether demand for regulatory data reporting services would increase or decrease; although more broker-dealers would be required to report regulatory data, it is possible that flexible reporting options allowed by the Plan could make preparing data for reporting less onerous, leading to fewer firms choosing to outsource this activity.

It is possible that the Plan would increase the demand for data reporting services by requiring regulatory data reporting by broker-dealers that may have previously been exempt due to size under individual SRO rules. Because more broker-dealers would be required to report regulatory data under the Plan, the Commission preliminarily believes there could be an opportunity for increased competition in this market which might benefit all broker-dealers that outsource their regulatory data reporting activity. However, it is also possible that the increase in demand for data reporting services could serve to entrench existing providers if they capture a large share of newly created demand; this could lead to relatively higher costs for broker-dealers than they would face in a more competitive market. The potential increase in demand for data reporting services could impact the capacity of already existing data reporting services to meet this increase in demand, and this in turn could have implications for competition and pricing in the market for data reporting services. Considering the barriers to entry that characterize the market for data reporting services and this potential increase in demand, service bureaus could have less incentive to compete for broker-dealer clients because these clients are no longer scarce, and as such, the CAT NMS Plan could result in a decline in the competition for data reporting services. It is possible that broker-dealers seeking to establish relationships with service bureaus could have trouble securing them because of the limited on-boarding capacity and need to on-board many broker-dealers at once. In the short-run these capacity constraints and the high demand could increase the costs of reporting through a service bureau. However, the two year implementation period for large broker-dealers and three year period for small broker-dealers could alleviate the reduction in competition due to the onboarding capacity strain because current service bureaus have time to increase their on-boarding capacity and new entrants have time to build the necessary IT infrastructure and a client base.

See, e.g., FINRA Rule 7470.

The CAT NMS Plan could also dramatically change the pool of firms demanding data reporting services, which would be skewed toward firms that are smaller and on average costlier to service, which could result in higher prices, which could eventually be passed onto investors. In addition to small and medium sized broker-dealers that previously self-reported, the CAT NMS plan would result in more broker-dealers having data reporting responsibilities and the Commission preliminarily believes that these broker-dealers would predominantly be small. For example, very small broker-dealers that are currently exempt from OATS reporting requirements could seek to establish service bureau relationships. In addition, because the Plan would require additional elements in regulatory data, particularly customer data, some broker-dealers that currently self-report could no longer find it economically feasible to continue to do so.

In addition to possibly increasing demand for data reporting services, the CAT NMS Plan may have a mixed effect on the number of firms offering data reporting services. This can impact the competitiveness of this market, and affect the costs broker-dealers bear in securing these services. On one hand, the number of firms offering data reporting services could decrease, because the need to secure PII might increase the likelihood of liability and litigation risks in the event of a security breach. On the other hand, it is possible that the number of service bureaus offering data reporting services would increase. New reporting requirements for numerous broker-dealers could create opportunities for new entrants to meet this demand. This could increase capacity and result in innovation in providing these services, which could benefit broker-dealers needing data reporting services by potentially reducing reporting costs, or at least reducing the potential for cost increases. Lower reporting costs for broker-dealers could in turn benefit the investors who are serviced by these broker-dealers, through reduced costs.

See Section IV.F.4.a(3), supra for a discussion of the potential exit of service bureau resulting from the risk of a security breach.

It is also possible that the Plan would decrease the demand for data reporting services. Many broker-dealers currently pay another firm (such as a service bureau) to fulfill their regulatory data reporting; this may be because these broker-dealers find it would be more expensive to handle the translation of their order management system data into fixed formats, such as is required for OATS. If the Plan Processor allows broker-dealers to send data to the Central Repository in the formats that they use for normal operations, in drop copies for example, these broker-dealers may no longer see a cost advantage in engaging the services of a regulatory data reporting service provider because one of the costs associated with regulatory data reporting—having to translate data into a fixed format—will have been eliminated. Without the cost of having to translate data, some broker-dealers that currently outsource OATS reporting could choose, at the margin, to insource their regulatory data reporting.

The Plan does not mandate the data ingestion format. See CAT NMS Plan, supra note 3, at Appendix C, at Section A.1(b). The Commission recognizes that the CAT Reporters Study found no difference in expected costs for a fixed format, but requests comment on why the costs may be similar when it would seem logical that allowing flexible data reporting formats would reduce costs for broker-dealers. See Request for Comment Nos. 318 and 331 in Section IV.F.5, supra.

The Commission preliminarily believes that this reduction in demand would not likely be realized and, if realized, would be unlikely to offset the increase in demand that would come from CAT reporters not subject to OATS reporting. As noted in the Costs Section, of the 1,800 expected CAT Reporters, 868 do not currently report to OATS. This means that the Commission expects a large proportion of CAT Reporters may be broker-dealers that currently do not have a service bureau for regulatory data reporting but would choose to engage one to manage their CAT reporting responsibilities. This is more than the Commission's estimate of 806 current outsourcing broker-dealers. Therefore, it is unlikely that the number of current Outsourcers that choose to become Insourcers would be larger than the number of non-OATS reporters that would elect to outsource. As a result, demand is more likely to increase. Further, the requirement for CAT reports to use listing exchange symbology could require pre-report data processing even if the Plan Processor allows for the receipt of reports in the formats that broker-dealers use for normal operations. As a result, the CAT NMS Plan is unlikely to eliminate the costs of processing data prior to reporting that data to the Central Repository.

The Plan estimates that 1,800 broker-dealers are expected to have CAT reporting obligations. Based on data from FINRA, 932 broker-dealers currently report OATS data. 1,800-932=868. See Section IV.F.1.c(2)A, supra.

Id.

See supra note 949.

2. Efficiency

The Commission has analyzed the potential impact of the Plan on efficiency. The Plan includes a discussion of certain efficiency effects anticipated if the Plan is approved; as part of its economic analysis, the Commission discusses these effects, as well as additional effects on efficiency anticipated by the Commission. The Commission preliminarily believes that the Plan as proposed is likely to result in significant improvements in efficiency related to how regulatory data is collected and used. The Plan also has the potential to result in improvements in market efficiency by deterring violative activity that could reduce market efficiency. The Commission notes, however, that efficiency gains from the retirement of duplicative and outdated reporting systems would be delayed for up to two and a half years and the interim period of increased duplicative reporting would impose significant financial burden on Industry Members.

The Commission has also analyzed the likely effect of the Plan on allocative efficiency of existing capital within the industry. These potential effects are discussed in Section IV.G.3, infra.

See Section IV.F.2, supra.

a. Effect of the Plan on Efficiency

The Commission has analyzed the possible effects of the CAT NMS Plan on efficiency. Specifically, building off the discussion in the Plan, the Commission analyzed the effect of the Plan on the efficiency of detecting violative behavior through examinations and enforcement, on the efficiency of surveillance, on market efficiency through deterrence of violative behavior, on operational efficiency of CAT Reporters, and on efficiencies through reduced ad hoc data requests and quicker access to data.

The current state of regulatory data collection and use provides ample opportunity for efficiency improvements. First, regulators' ability to efficiently perform cross-market surveillance is hindered by data fragmentation. Second, regulators' ability to efficiently supervise and surveil market participants and carry out their enforcement responsibilities is hindered by limitations in current regulatory data. Finally, there are a number of other inefficiencies associated with the current system of regulatory data collection. These include: Delays in data availability to regulators; lack of direct access to data collected by other regulators results in numerous ad-hoc data requests; and the need for regulatory Staff to invest significant time and resources to reconciling disparate data sources.

See Section IV.E.2.c, supra.

See Section IV.E.2.c, supra.

See Section IV.D.2.b, supra. These other inefficiencies are discussed above in the Baseline and Benefits Sections.

The Plan discusses a number of expected efficiency effects associated with the Plan, including both positive and negative effects. The Commission preliminarily agrees with the Plan's assessment and has identified additional efficiency effects as well. The Plan outlines several positive effects relating to efficiency in: Monitoring for rule violations; performing surveillance; and supporting fewer reporting systems. Some of these efficiencies are also discussed in the Benefits Section of this analysis.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(b).

See Section IV.E, supra.

The Plan concludes SROs would experience improved efficiency in the detection of rule violations, particularly for violations that involve trading in multiple markets. The Plan states an expectation that SROs would need to expend fewer resources to detect violative cross-market activity, and such activity would be detected more quickly. The Commission agrees that the Plan would result in improvements in efficiency in the performance of examinations of market participants by SROs and the Commission. Improvements to data availability and access through the Central Repository could allow SROs and the Commission to more efficiently identify market participants for examination. The Commission also agrees that the Plan would improve the efficiency of enforcement investigations. If regulatory data access improves, the quality and quantity of enforcement investigations could increase through improvements to the comprehensiveness and timeliness of data used to support investigations. As mentioned previously, it can take months for regulators to assemble the data necessary to comprehensively investigate a regulatory inquiry. To the extent that the Plan allows regulators to access more comprehensive data directly from the Central Repository, regulators would be able to collect data faster and start processing it sooner, resulting in a more efficient data analysis portion of an investigation. As a result, follow-up enforcement inquiries could be avoided entirely in situations where data from the Central Repository allows regulators to conclude an initial inquiry without initiating an enforcement investigation. This benefit would be observable to both regulators and subjects of investigations, for whom ongoing enforcement investigations can be costly and the source of uncertainty.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(b); see also Section IV.E.2, supra.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(b).

See Section IV.E.2.c, supra.

See Section IV.E.2.c, supra.

The Commission notes that this does not preclude an increase in total enforcement investigations, but rather that some enforcement investigations may determine earlier in the investigation that no violation occurred.

The Plan states that the Participants believe that the CAT NMS Plan could improve the efficiency of surveillance. According to the Plan, this improvement is due to a number of factors including: Increased surveillance capacity; improved system speed, which would result in more efficient data analysis; and a reduction in surveillance system downtime. The Plan also cites reduced monitoring costs, but the Commission notes that estimates in the Costs Section of the Plan predict increased surveillance costs if the Plan is approved. The increased surveillance costs predicted in the Plan could reflect more effective surveillance under the Plan. Although the Plan does not discuss the cost-benefit trade-off of increased surveillance directly, the Commission notes that achieving the level of surveillance that would be possible if the Plan is approved would likely be more expensive using currently available data sources, if it is achievable at all, due to the inefficiencies that currently exist in delivering regulatory supervision, discussed previously.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(b) (stating that the CAT NMS Plan could reduce monitoring costs, enable regulators to detect cross-market violative activity more quickly, provide regulators more fulsome access to unprocessed data and timely and accurate information on market activity, and provide CAT Reporters with long term efficiencies resulting from the increase in surveillance capabilities); see also IV.E.2.c, supra.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(b). The Participants surveyed the 10 exchange-operating SRO groups on surveillance downtime. In conversations with Staff, the Participants informed Staff that average surveillance downtime was 0.03% from August 1, 2014 to August 31, 2015, and ranges from 0 to 0.21% across SROs.

See id.

See Section IV.E.2, supra.

The Commission preliminarily believes that CAT may reduce violative behavior. The Plan states that CAT may serve a deterrent effect, thereby reducing investor losses attributable to such behavior. Improvements in the efficiency of market surveillance, investigations, and enforcement could directly reduce the amount of violative behavior by identifying and penalizing market participants who violate rules and who would more easily go undetected in the current regime. Furthermore, market participants' awareness regarding improvements in the efficiency of market surveillance, investigations, and enforcement (or perceptions thereof), and the resultant increase in the probability of incurring a costly penalty for violative behavior, could deter violative behavior. Reductions in violative behavior through both of these economic channels could improve market efficiency, assuming violative behavior receives diminishing marginal gains and generates increasing marginal harm.

See Section IV.E.2.c, supra.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(b).

See, e.g., Schelling, Thomas, “The Strategy of Conflict: Prospectus for a Reorientation of Game Theory,” Journal of Conflict Resolution, Vol. 2 No.3 (1958); Ellsberg, Daniel, “The Crude Analysis of Strategic Choices,” American Economic Review, Vol. 51, No. 2 (1961).

See, e.g., Becker, Gary and William Landes, “Essays in the Economics of Crime and Punishment,” Columbia University Press, (1974).

The Plan discusses increased efficiency due to reductions in redundant reporting systems. The Plan also discusses increases in system standardization, which would allow consolidation of resources, including the sunsetting of legacy reporting systems and processes, as well as consolidated data processing envisioned from the Plan. However, the Commission is aware that the Plan, as proposed, calls for a period of years during which Industry Members would face duplicative reporting systems before older regulatory data reporting systems are retired. This period of duplicative reporting would impose a considerable financial burden on Industry Members.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(C) (discussing benefits of CAT to broker-dealers).

See id. at Appendix C, Section B.8(b).

See id. at Appendix C, Section B.9.

See Section IV.F.2, supra for a discussion of duplicative reporting and whether broker-dealers would pass costs on to investors.

The Plan discusses two other efficiency improvements: a reduction in ad-hoc data requests and more fulsome access to raw data. The Plan predicts a reduction in ad-hoc data requests, which would free up resources previously used to service such requests. However, while the Plan anticipates a decrease in ad-hoc data requests as a result of Plan-related data improvements, the Commission notes that it is possible that some types of ad-hoc data requests might increase. For instance, even if enforcement investigations initially use CAT Data, later-stage investigations may involve requests for data not included in CAT Data, such as commissions paid or a locate identifier for a short sale. An increase in the efficiency of enforcement investigations could increase the total number of later-stage investigations. Such investigations could produce additional ad-hoc data requests and require other interactions with market participants. The Commission recognizes that these data request increases would partially offset the efficiency improvements from the reduction in data requests noted above, but the Commission preliminarily believes that the Plan would improve efficiency by reducing the total number of data requests. The Commission, however, acknowledges that this decrease in data requests may be partially offset in an increase in the number of investigations in general, because enhanced surveillance is likely to detect more potentially violative activity that would need to be investigated.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(b).

This does not preclude regulators determining sooner if the actions they are investigating are not violative. Rather, an increase in the total number of enforcement investigations due to efficiency improvements can result in more later-stage investigations even if regulators are better able to conclude some investigations earlier.

See Section IV.D.1.c, supra.

Furthermore, the Plan anticipates more robust access to unprocessed regulatory data, which could improve the efficiency with which SROs could respond to market events where they previously had to submit data requests and wait for data validation procedures to be completed before accessing data collected by other regulators. The Commission recognizes that unprocessed data may contain errors that would later be fixed. The Commission preliminarily believes the benefits of the greater timeliness of the unprocessed data may justify the lack of validations and corrections in such unprocessed data.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(b).

See Section III.B.10, supra.

See Section IV.E.2.c, supra, for an example of benefits from regulators accessing uncorrected data on T+1.

b. Effects of Certain Costs of the Plan on Efficiency

The Plan discusses several sources of inefficiency due to costs of the Plan that are difficult to quantify, and are transient in nature. First, the Plan anticipates that implementation would introduce new costs related to data mapping and data dictionary creation. Second, the Plan discusses needs for expenditures, such as staff time for compliance with encryption requirements associated with the transmission of PII. While the Commission recognizes that these are additional activities and costs that the Plan would require, it views these as additional costs rather than inefficiencies and, though the Commission cannot quantify the magnitude, these costs are likely to have relatively minor contributions to overall costs of the Plan because they impose technical requirements on systems that industry will need to significantly alter to comply with other provisions in the Plan. Furthermore, the Commission notes that the costs of data mapping and encryption requirements are likely to be included in costs covered by surveys conducted by the Participants while preparing the Plan because these requirements were known publicly at the time the surveys were conducted, and are anticipated to be small relative to other costs entailed in potentially complying with the Plan if it is approved.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(b).

See id.

See Section IV.G.2.a, supra.

See Section IV.F.1, supra.

The Plan notes that there could be a market inefficiency effect related to the funding proposal for the Plan. For example, the cost allocation methodology for the Plan could create disincentives for the provision of liquidity, which could impair market quality and increase the costs to investors to transact. The Plan notes that the funding principles set forth in the Plan seek to mitigate the risk of reduction in market quality resulting from allocation of costs from building and operating the Central Repository. The Commission preliminarily recognizes that negative effects on efficiency could result from the CAT Funding Model. First, data reporters could respond to the Funding Model by taking actions to limit their fee payments, such as exiting the market or reducing their activity levels. Second, the funding policy of the CAT NMS Plan of aligning fees closely with the amounts that are required to cover costs could create incentives for the Plan Processor or Operating Committee to propose a cost schedule for the CAT that matches a given fee schedule, but is not the most efficient cost schedule for meeting the CAT regulatory objectives.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(b).

See id. at Section 11.2, Appendix C, Section B.7(b)(iv)(C).

See id. at Appendix C, Section B.7.(b)(iv)(C).

See id. at Appendix C, Section B.7(b)(v)(B).

Economics research that dates back to Averch, Harvey, and Johnson, Leland L. (1962) (“Behavior of the Firm Under Regulatory Constraint,” American Economic Review 52 (5): 1052-1069) characterizes an incentive of regulated utilities to inflate their costs in order to establish larger rate bases and justify higher rates. An opposite effect would arise if the regulated utility were unable to justify sufficient fee revenue to pay the fixed cost of expanding the base.

3. Capital Formation

a. Enhanced Investor Protection

The Commission has examined the potential effects on capital formation discussed in the Plan in addition to other potential effects on capital formation that the Commission believes could result if the Plan is approved. The Commission preliminarily believes that the Plan would have a modest positive effect on capital formation.

The Plan's analysis regarding capital formation concludes that the Plan would generally not have a deleterious effect on capital formation and could bolster capital formation that could lead to increased investor participation in capital markets. The Plan's analysis provides several reasons why the Plan would not adversely affect capital formation. Specifically, it asserts that the Plan would not place any undue burden on primary issuances; would not pass along CAT related costs to “investors in a way that would limit their access to or participation in capital markets”; and would not discourage market participation as a result of data security concerns given the data security safeguards outlined in the Plan. The Commission preliminarily agrees with the rationale of the Plan's analysis, but addresses some additional considerations regarding the scope of the Plan's effects on capital formation, as well as the channels through which these effects could accrue.

See CAT NMS Plan, supra note 3, at Appendix C, Section C.8(c).

See id.

The Plan's analysis states that the Plan may improve capital formation by improving investor confidence in the market due to improvements in surveillance. As discussed previously, the Commission believes that the Plan would provide substantial enhancements to investor protection through improvements to surveillance, particularly for cross-market trading. As discussed throughout, improved surveillance, as well as other regulatory activities, could decrease the rate of violative activity in the market, reducing investor losses due to violative activity, to the extent that such behavior is not already deterred by current systems. If improved surveillance leads to expectations of fewer losses due to violative activity, this may increase capital formation by facilitating a market where investors could be more likely to mobilize capital into securities markets.

See Section IV.E.2.c, supra and CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(ii)(B)(1) and (2), B.7(b)(iii)(C).

FINRA currently provides cross-market surveillance, but limitations in the data (e.g. reliable cross-market linkages, customer identification, parent order identification) limit the scope and reliability of this surveillance.

For example, as discussed in Section IV.E.2.c, the Plan would allow regulators to more efficiently conduct cross-market and cross-product surveillance relative to surveillance using current data sources, and the requirement that data be consolidated in a single database would assist regulators in detecting activity that does not appear clearly violative until data is linked and evaluated from multiple venues. To the extent that market participants are aware of the current challenges to regulators in performing cross-market surveillance and aggregating data across venues, and to the extent that they believe that their violative behavior is more likely to be detected if regulators' ability to perform those activities improves, they may reduce or eliminate violative behavior if the CAT Plan is approved.

There is evidence in the academic finance literature that countries with weaker investor protections, considering both the character of rules as well as the quality of enforcement, have smaller and narrower capital markets in terms of investor participation. See La Porta, R. et al, “Legal Determinants of External Finance,” Journal of Finance, Vol. 52 No. 3 (1997).

The Commission preliminarily believes there could be additional increases in capital formation in the form of improvements in allocative efficiency of existing capital within the industry. If investors perceive an environment of improved surveillance, they could be willing to allocate additional capital to liquidity provision or other activities that increase market efficiency. Furthermore, an environment of improved surveillance efficiency could result in the reduction of capital allocated to violative activities that impose costs on other market participants, because these market participants may no longer find it possible to engage in such behavior that exposes them to regulatory action. In this scenario, this reallocation of capital could improve market quality and efficiency even if net capital formation changes little. In addition to the potential reallocation of capital currently mobilized toward violative activities, investor capital that may currently be diverted because of the risk of loss to violative activities could also be reallocated should the violative activities decrease. If the CAT NMS Plan reduces manipulative quoting activities, either through improved detection/enforcement or through deterrence of such activities, then investors are less likely to make capital allocation decisions in response to manipulative quoting activities. In this scenario, because manipulative quoting activities have been reduced, the contribution of manipulation to prices has been reduced and prices should therefore better reflect fundamentals. It would follow that, to the extent that displayed prices better reflect fundamentals rather than manipulation, investors could allocate capital more efficiently for their purposes. The Commission notes, however, that market participants engaging in allowable activity that might risk additional regulatory scrutiny under the Plan regime could allocate capital to other activities to avoid this scrutiny, because even when activity is not violative, interacting with regulators can be costly for market participants. This reallocation away from allowable activity to avoid regulatory interactions could result in capital allocations that are less efficient.

See Section IV.F.4.b, supra, for a discussion of the potential for the efficiencies in surveillance, examinations, and investigations to increase the number of regulatory activities, including the number of regulatory activities on conduct that turns out not to violate regulations.

The Commission is unable to estimate the magnitude of allowable economic activity that does not occur when market participants anticipate relatively high costs of demonstrating regulatory compliance in the course of normal regulatory interactions such as exams and inquiries because this activity is not observable. However, Section IV.F.1.c(2) discusses how some broker-dealers avoid self-reporting regulatory data because of expectations of higher costs to demonstrate compliance, providing an example of an allowable activity that is perceived as costly due to the risk of compliance costs. See Section IV.F.1.c(2), supra.

The Plan states that the costs from CAT are unlikely to deter investor participation in the capital markets. The Commission notes, however, that the final costs of the Plan and the funding mechanism for CAT are not wholly certain at this time; thus, it is the Commission's view that there is uncertainty concerning the extent to which investors would bear Plan costs and consequently to what extent Plan costs could affect investors' allocation of capital. As mentioned above in the Costs Section, the Commission preliminarily does not know whether Plan costs incurred by the industry are likely to be passed on to investors. Competition in the market for broker-dealer services could mitigate some of these costs, but it may not minimize costs passed on to retail investors. Despite these potential costs to investors, investors could believe that the additional benefits they receive from the potential of a market that is more effectively regulated justify any additional costs they pay to access capital markets.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(c).

See Section IV.F.2, supra.

b. Data Security

The Commission preliminarily agrees with the Plan's assessment that data security concerns are unlikely to materially affect capital formation. In its discussion of capital formation, the Plan recognizes that data security concerns could potentially impact capital formation through market participants' perception that sensitive proprietary data might be vulnerable in case of a data breach at the Central Repository. The Plan's analysis discusses the security measures that are required by Rule 613 and the manner in which they have been implemented in the Plan. It concludes that these security measures are sufficient and that it is unlikely market participants would reduce their participation in markets in a manner that would affect capital formation.

As noted above, the Commission agrees that concerns regarding data security are unlikely to substantially affect capital formation, but that some uncertainty about the risks exist because of the variations in the potential security solutions and their resulting effectiveness. The Commission notes that the consequences of a data breach, nonetheless, could be quite severe. It is inherently difficult to form reliable economic expectations given that security breaches of the form that could occur under the CAT NMS Plan occur infrequently. Therefore, as described in Section IV.F above, even if a CAT Data security breach is unlikely with the safeguards required by the Plan, the scope of the potential consequences of such a breach in the event that one should occur is important to evaluating the risk to capital formation.

See Section IV.F.4.a, supra.

See id. for a more thorough discussion of the costs and risks of security breaches of the Central Repository.

A data breach could also substantially harm market participants by exposing proprietary information, such as a proprietary trading strategy or the existence of a significant business relationship with either a counterparty or client. The Commission notes, however, that broker-dealers already bear such risks in transmitting regulatory data to SROs. The Commission preliminarily believes that the marginal increase in the risks to broker-dealers associated with a data breach would be unlikely to deter broker-dealers from participating in markets.

A data breach could potentially reveal PII of investors. To address the potential for harm to the investing public and the health of capital markets through such a breach, the Plan has enhanced requirements for security around PII. Those requirements include a separate PII-specific workflow, PII-specific authentication and access control, separate storage of PII data, and a full audit trail of PII access. The Commission preliminarily believes that these risks will not materially affect investors' willingness to participate in markets because they already face these risks with PII shared with broker-dealers, though not in one centralized location. However, the risk and costs of a security breach would be only one factor that market participants would consider in deciding whether to participate in the market. Another consideration would be investor protection, which the Commission preliminarily believes would increase under the CAT NMS Plan.

See CAT NMS Plan, supra note 3, at Appendix D, Sections 4.1.1-4.1.6. The Commission notes that there is considerable diversity in the approaches proposed by the Bidders. Further, the Participants chose to give the Plan Processor flexibility on many implementation details and the Plan states the requirements as a set of minimum standards. Consequently, the final PII security solution cannot be evaluated—only the minimum standards specified in the Plan.

See Section IV.F.2, supra.

See Section IV.E.2, supra.

4. Related Considerations Affecting Competition, Efficiency and Capital Formation

The Commission recognizes that the Plan's likely effects on competition, efficiency and capital formation are dependent to some extent on the performance and decisions of the Plan Processor and the Operating Committee in implementing the Plan, and thus there is necessarily some uncertainty in the Commission's analysis. Nonetheless, the Commission believes that the Plan contains certain governance provisions, as well as provisions relating to the selection and removal of the Plan Processor, that mitigate this uncertainty by promoting decision-making that could, on balance, have positive effects on competition, efficiency, and capital formation.

a. The Efficiency of Plan Decision-Making

As noted in several places above, future decisions of the Operating Committee could significantly alter the economic effects of the Plan. As a result, this economic analysis also considered whether the process by which the Operating Committee would make such decisions promotes efficiency. According to the Plan, the inability of the Operating Committee to act in a timely manner could create consequences for efficiency, competition, and capital formation. On the other hand, the Commission notes that consequences also could arise if the Operating Committee makes decisions so quickly that it does not consider all relevant information. This Section analyzes whether the decision-making processes in the Plan promote timely decisions that consider all relevant information of value. While the Plan considers the potential for inefficiencies in the decision-making process, the Commission preliminarily believes that certain governance provisions in the Plan could create some inefficiencies in the decision-making process, but that these inefficiencies are limited or exist to promote better decision-making. The Plan discusses two areas where the proposed governance structure impacts the efficiency of the decision-making process: (1) Voting protocols and (2) the role of industry advisers. The Commission also considered the efficiency implications of the level of detail included in the Plan and the scalability of the Plan.

See, e.g., Section IV.C.2, supra.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(d).

See id.

The Plan specified three types of voting protocols and determines when each protocol applies. The Plan requires unanimous voting in only three circumstances: A decision to obligate Participants to make a loan or capital contribution, a decision to dissolve the Company, and a decision to take an action by written consent instead of a meeting. Further, the Plan requires supermajority voting in instances considered by the Participants to have a direct and significant impact on the functioning, management, and financing of the CAT System, such as selection and removal of the Plan Processor and key officers, approving the initial Technical Specifications, approving Material Amendments to the Technical Specifications proposed by the Plan Processor, and approving direct amendments to the Technical Specifications proposed by the Operating Committee. The Plan considers other matters as routine matters that arise in the ordinary course of business and would be subject to majority voting. As a practical matter, Majority Vote is the default standard for decisions other than those requiring supermajority or unanimous voting.

See Section III.A.3.a(3), supra, for a discussion of the management of the Company, including the definitions of the voting protocols and details on their application.

See CAT NMS Plan, supra note 3, at Appendix C, Section D.11(b), Voting Criteria of the Operating Committee.

See id. at Appendix C, Section D.11(b).

See id. at Appendix C, Section D.11(b). The Plan also requires supermajority voting on matters outside the ordinary course of business, such as modifications to a Material Contract, incurring debt, making distributions or tax elections, or changing the fee schedules.

The Plan balanced the efficiency of the decision-making process against the value of considering minority and dissenting opinions in proposing these voting protocols. In particular, the Plan recognizes that some voting protocols might impede the effective administration of the CAT System. From a mechanical perspective, voting protocols determine a threshold for a passing vote. Unanimity requires a threshold of 100% yes votes while majority voting requires a threshold of more than 50% yes votes and Supermajority requires two-thirds or more. The Plan explains that too-high a threshold for decision-making, such as may be the case in applying unanimity to all voting matters, could limit the ability of the Operating Committee to adopt broadly agreed upon provisions. For example, in the extreme, requiring unanimity in voting could result in one dissenting opinion holding up the entire decision-making process. Conversely, the Plan explains that a threshold that is set too low might limit the opportunities for the consideration of dissenting or minority opinions and alternative approaches. For example, if voting thresholds were too low, a set of Participants could potentially adopt provisions that might provide them a competitive advantage over other Participants.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(d).

See id.

See id.

See id.

The Commission preliminarily agrees with the discussion on the need to balance efficiency in the voting protocols in the Plan. The Commission notes that the speed and ability to make a decision are key components of whether the Plan promotes efficiency in its operations. High-vote thresholds may result in an increase in the effort needed to obtain enough votes to make a decision. Further, in addition to the drawn out discussions necessary to obtain a unanimous vote, a unanimous vote might also require compromises that reduce the efficiency of the decision-making process. This could be particularly costly in situations in which the Operating Committee must make a decision by a particular date. It could also result in inaction for decisions related to making discretionary changes that could improve data qualities, such as updates, if the Participants disagree among the various alternatives.

Furthermore, while the decision-making processes with a very low voting threshold would be faster, the resulting decisions might not consider all relevant information. As a result, the Commission preliminarily agrees that the inefficiencies in the voting protocols in the Plan are limited enough to strike a balance between the inefficiencies of the decision-making process and the quality of the decisions.

See Section IV.E.3.d, supra, for a discussion of how certain governance provisions could help promote better decision-making by the relevant parties.

The Plan also discusses the role of industry representation as part of the governance structure. Section 4.13 of the Plan requires an Advisory Committee that contains twelve members, including representatives from 7 types of broker-dealers, 2 institutional investors, and 3 individuals. In addition, the Plan says that the Advisory Committee is “intended to support the Operating Committee and to promote continuing efficiency in meeting the objective of the CAT.” The Plan also indicates that it is important to include industry representation to assure that all affected parties have representation.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.8(d).

See id. at Section 4.13 (Advisory Committee).

See id. at Appendix C, Section B.8(d).

The Commission preliminarily agrees with the discussion in the Plan that including industry representation might result in a more efficiently designed CAT, but adds that an Advisory Committee also adds operational inefficiencies. As discussed above, the Commission preliminarily believes that an Advisory Committee could add more diverse viewpoints to the debates surrounding Operating Committee decisions and thus reduce the risk that members of the Operating Committee could make decisions without first obtaining a full understanding of the underlying facts or the likely impact of its decisions. The Commission also recognizes, however, that including an Advisory Committee in the decision-making process might add complexity to the process and decisions might require more time relative to allowing the Operating Committee to make decisions without the input of an Advisory Committee. The inclusion of an Advisory Committee could thereby potentially adversely affect the efficiency of the Plan's operation. In general, the Commission preliminarily believes that as long as the Advisory Committee adds sufficiently useful information, the benefits from the Advisory Committee would justify any operational inefficiencies from the inclusion of the Advisory Committee.

See Section IV.E.3.d(2)B, supra.

The Commission considered an additional source of potential efficiencies in the decision-making process. The Plan specifies minimum standards for particular provisions or solutions in Appendix D of the Plan instead of specifying the solutions themselves in the Plan. While this creates uncertainty in the costs and benefits of the Plan and reduces the transparency for the bidders, the Commission recognizes that decisions to not specify certain solutions in the Plan could promote efficiency in the decision-making process of the Operating Committee. The Operating Committee and/or Selection Committee would effectively decide upon the unspecified details when selecting the Plan Processor and when approving the Technical Specifications. As such, certain technical details may not appear in the Plan and may not be subject to Commission approval or, potentially, to public comment. Instead, the Operating Committee could implement such decisions much more quickly and at a potentially lower cost. The Commission believes that the Commission and public review process could add value to the decision-making process, particularly in assuring that the decisions consider costs and benefits. However, a notice and comment process for certain technical changes could be cumbersome and time-consuming, and may not therefore be justified in the context of certain technical issues. The Plan therefore may be more agile and efficient in its ability to upgrade and improve systems quickly. On the other hand, the cost of this efficiency comes in the form of the significant uncertainties surrounding the economic effects of the Plan during the approval process.

For example, the Plan provides minimum standards for regulator access to CAT Data but does not propose any particular method for regulatory access. Nor does the Plan specify whether the regulators would have work space on servers at the Central Repository or whether regulators would have to download the results of every query before being able to process such results.

For example, the Selection Committee would decide on the details of regulator access in conjunction with selecting the Plan Processor or in subsequent negotiations with the selected Plan Processor.

Provisions of the Plan should also promote efficiently implementing expansions to the CAT Data. Appendix C of the Plan notes that the Plan Processor must ensure that the Central Repository's technical infrastructure is scalable and adaptable. These provisions should reduce the costs and time needed for expansions to the Central Repository.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.5(a).

b. Selection and Removal of the Plan Processor

The CAT NMS Plan uses a request for proposal (“RFP”) to select the Plan Processor that would design, build, and operate the Central Repository. The winning bidder becomes the sole supplier of the operation of the Central Repository. The Commission preliminarily believes this is necessary to achieve the benefits of a single consolidated source of regulatory data.

The competitiveness of the selection process influences the ultimate economic effect of the Plan because those effects depend in large part on the efficiency and effectiveness of the Plan Processor. In particular, many of the details of the Plan would be determined either by the winning bid or in negotiations with the Plan Processor after selection. The Plan Processor exercises control over the future costs of operating and maintaining the Central Repository in this context and the Plan Processor chooses its performance level, subject to the minimum standards in the Plan and with oversight from the Operating Committee.

Given the effects associated with the selection process for the Plan Processor, the Commission considered whether the Plan promotes a competitive process and whether the Plan contains provisions that would create incentives for the chosen Plan Processor to set costs and performance competitively. As explained below, the Commission preliminarily believes that the selection process generally promotes competition but that there are also a few potential limitations on competition. Moreover, the Commission recognizes that a competitive bidding process does not necessarily mean that the selected bidder would behave competitively after being selected as the Plan Processor. But the Commission preliminarily believes that the Plan could control the costs of the Central Repository and the performance of the Plan Processor if the Plan included sufficient competitive incentives for the selected Plan Processor. While the Commission preliminarily believes that threat of replacement of the Plan Processor could incentivize them to set costs and performance competitively, the high cost of replacement could limit these incentives.

See Goldfine and Vorrasi, “The Fall of the Kodak Aftermarket Doctrine: Dying A Slow Death in the Lower Courts,” 74 Antitrust Law Journal No. 1 (2004), p. 209 (stating that “competition in the primary market, as a matter of law, does not necessarily preclude the possibility of market power (and anticompetitive conduct) in the aftermarkets for parts and services,” and citing Eastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451 (1992)). Economic theories of the relation between primary markets and aftermarket are the focus of other literature as well; see infra note 1149. (In the context of the Plan, the “primary market” would be the initial selection of the Plan Processor while in the “aftermarket,” the selected Plan Processor would supply a performance level for the given revenues received from the Company.)

Under the theory of contestable markets, it is possible for the sole supplier of a service to behave as if there multiple suppliers, and thus not exercise monopoly power. Necessary conditions include the absence of entry and exit costs. William J. Baumol, John C. Panzar, Robert D. Willig (1982), Contestable Markets and the Theory of Industry Structure. When the conditions needed to support contestable markets are not met, the presence of alternative suppliers may not be sufficient to prevent the costly exercise of monopoly power, post-selection. For example, if the supplier cannot make complete and binding commitments to the price and quality of its post-selection services, and the buyer becomes locked into the sole supplier (e.g., due to switching costs or other sources of friction), a competitive selection process may lead to monopoly outcomes, post-selection; see, e.g., Carl Shapiro, 1995, “Aftermarkets and Consumer Welfare: Making Sense of Kodak,” Antitrust Law Journal, and Borenstein, Severin, Jeffrey K. Mackie-Mason, and Janet S. Netz, 1995, Antitrust Policy in Aftermarkets, Antitrust Law Journal 63: 455-82. For a recent survey of alternative theories, see section 3.1, Dennis W. Carlton and Michael Waldman, 2014. “Robert Bork's Contributions to Antitrust Perspectives on Tying Behavior,” Journal of Law & Economics.

(1) Competitiveness of the Plan Processor Selection Process

The Commission believes that two elements determine the competitiveness of the bidding process. The first relates to the voting process and the second relates to the degree of transparency in the bidding process. The Commission preliminarily believes that the Plan provisions relevant to these two factors could promote competition in the bidding process and limit the risk that selection of the Plan Processor would be affected by a conflict of interest, thereby promoting better decision-making.

The CAT NMS Plan outlines a bidding process whereby a Selection Committee votes on bidders during several rounds of voting that each narrow the potential bidders until one bidder is selected. Pursuant to the Plan, the bidders compete to be selected by proposing solutions to comply with Rule 613 and documenting the anticipated costs of doing so. The Plan also contains provisions for revising Bids if the Commission approves the Plan.

See CAT NMS Plan, supra note 3, at Section 5.2 (Bid Evaluation and Initial Plan Processor Selection).

Id. at Section 5.2(e).

The Participants received 31 Intent to Bid forms during the RFP process; 13 of the potential bidders withdrew before January 30, 2014; the Participants reported receiving 10 Bids by April 2, 2014. Six of these Bidders were shortlisted through the selection process in July 2014, including one SRO that is also a Bidder. In November 2015, the shortlist was further narrowed to three Bidders.

For details on the progression of the CAT RFP process, see RFP Process, SEC Rule 613: Consolidated Audit Trail (CAT), available at http://catnmsplan.com/process/ (last visited November 19, 2015).

See supra note 35.

In considering how competitive the voting process is, the Commission has considered whether conflicts of interest could limit competition in the bidding process through the proposed participation of a bidder representative on the Selection Committee. The Plan includes provisions that mitigate this conflict but that have not eliminated it completely. In particular, the Plan requires recusal of an SRO from any selection round if that SRO or its affiliate has submitted a bid—or is included as a material subcontractor as part of a bid—that is still under consideration in such round. Similarly, the Plan creates information barriers between the Staff at the SRO selecting the bidder and the Staff undertaking the bidding. These provisions promote a level playing field for all bidders because the SRO bidder does not know any more than a non-SRO bidder and so has no informational advantage in submitting a bid that the Selection Committee may find favorable. Further, the information barriers prevent those working on the bid from attempting to persuade members of the Selection Committee toward their bid in a way that other bidders cannot. The Commission recognizes, however, that there is a residual risk in having an SRO among the bidders; it is possible that voting Participants would be biased for or against that SRO either because they compete with that SRO in another market (and could gain a competitive advantage in that market by acting as Plan Processor) or because of repeated interactions with that SRO.

See CAT NMS Plan, supra note 3, at Section 4.3(d), at Section 5.1(b).

See id. at Section 5.1(d).

The Commission also recognizes that, to the extent the Operating Committee has specific preferred solutions as to how the Plan should be implemented, the degree to which the Committee is transparent about those preferences in the bidding process would affect the competitiveness of that process. For example, if the Commission were to approve the Plan and bidders were thereafter given the opportunity to revise their bids, the Operating Committee could promote competitiveness in the bidding process by outlining its preferences. Transparency into the Operating Committee's views regarding potential optimal solutions could assist a bidder in revising its bid to inform how that bidder could supply those optimal solutions, and the Selection Committee could then compare all bidders on those particular solutions. To the extent that the Operating Committee has strong preferences toward particular solutions but did not specify those preferences directly in the Plan, the bidder may not know that it could improve its chances of winning the bid by proposing a different solution and the Selection Committee would not know whether the bidder is capable of delivering the preferred solution more efficiently than the other bidders. On the other hand, the Commission notes that specifying a preferred solution also has the potential to discourage bidders from competing on innovation by proposing novel approaches that may deliver superior outcomes.

The Commission has no reason to believe that the Operating Committee has preferred solutions beyond what is in the Plan that would significantly impact the competitiveness of the Plan Processor selection process. Indeed, Appendix D of the Plan details numerous minimum standards not included in the RFP. In addition, the Plan also provides details on the range of solutions proposed by bidders and why the Operating Committee may not have a preference and therefore did not select a particular solution. This provides transparency to the bidders on the criteria the Selection Committee may use to compare bidders.

(2) Competitive Incentives of the Selected Plan Processor

The Plan could create competitive incentives for the selected Plan Processor by detailing strong requirements for the Plan Processor and providing an efficient mechanism to remove the selected Plan Processor and introducing an alternative Plan Processor in the event of underperformance. As described below, the Commission preliminarily believes that the Plan provides the selected Plan Processor with competitive incentives because the Plan contains defined procedures for monitoring and removing the Plan Processor for failure to perform functions adequately or otherwise. However, the ease with which the Operating Committee could remove the Plan Processor and the costs of switching to another Plan Processor could limit these competitive incentives.

The Plan contains several provisions that would allow the Operating Committee to remove the Plan Processor. By Supermajority Vote, the Operating Committee could remove the Plan Processor for any reason. The Operating Committee may, by Majority Vote, remove the Plan Processor if it determines that the Plan Processor has failed to perform its functions “in a reasonably acceptable manner” or if the Plan Processor's expenses “have become excessive or are not justified.” The consideration of such poor performance or excessive expenses would include (1) responsiveness to requests for technological changes or enhancements, (2) results of assessments performed pursuant to Section 6.6 of the Plan, (3) staying up-to-date on reliability and security of operations, (4) compliance with the requirements of Appendix D, and (5) other factors the Operating Committee may determine to be appropriate.

See CAT NMS Plan, supra note 3, at Section 6.1(q), (r), (s).

The Commission preliminarily believes that the ability of the Operating Committee to remove the Plan Processor for poor performance with only a Majority Vote incentivizes the Plan Processor to perform well enough to avoid being removed. The Commission further preliminarily believes that the performance of the Plan Processor would depend significantly on strong oversight by the Operating Committee.

See Section IV.E.3.d, supra, for a discussion of the incentives of the Operating Committee in overseeing the Plan Processor.

The Commission recognizes that the effort required to remove a Plan Processor could be significant, which would limit the incentives of the Plan Processor to perform well. To subject a removal to a Majority Vote, the Operating Committee would presumably need to demonstrate the Plan Processor's performance and determine that it was not “reasonably acceptable.” If not, the removal would be subject to Supermajority Vote, which could also take significant effort and a removal would be less likely to pass.

In addition, significant switching costs could influence whether removing a Plan Processor despite poor performance makes economic sense. In other words, the Operating Committee could wait for significant performance issues before initiating a vote to remove the Plan Processor. Additionally, before removing a Plan Processor, the Operating Committee would need to select a new Plan Processor. This would likely be a lengthy process taking significant time and effort by the Operating Committee. Moreover, switching Plan Processors could entail a complete rebuild of the Central Repository and significant implementation costs for CAT Reporters and Participants, potentially amounting to the initial implementation costs of the Plan. These costs would be higher if the Plan Processor's solutions include proprietary technologies that no other potential replacement (competitor) could supply. The costs would be lower if the new Plan Processor could implement the existing Technical Specifications. The benefits of switching could also depend on the benefits from technological advancements that these competitors could supply. In light of these costs, the competitive incentives of the Plan Processor to maintain top performance could be limited. Specifically, the Plan Processor may only need to perform well enough to keep the inefficiencies associated with their performance from exceeding the cost to switch to another Plan Processor. Despite the limitations on competitive incentives due to switching costs, however, the Commission preliminarily believes that the threat of replacement still provides an incentive to stay relatively current on technology advancements to avoid falling significantly behind potential competitors.

5. Request for Comment on Efficiency, Competition, and Capital Formation

The Commission requests comment on all aspects of the discussion of the effects of the CAT NMS Plan on efficiency, competition, and capital formation. In particular, the Commission seeks responses to the following questions:

347. The Participants state in the Plan that they believe the Plan would avoid disincentives such as placing an inappropriate burden on competition in the U.S. securities markets. In its analysis, the Commission concludes that competition is unlikely to be harmed to a degree that would affect investors. Do Commenters agree with the conclusions discussed in the Plan? Why or why not? Do Commenters agree with the Commission's conclusion regarding the Plan's impact on competition? Why or why not?

348. Do Commenters agree with the Commission's characterization of the relevant markets that the CAT NMS Plan affect? Why or why not? Do Commenters agree with the identified level of competition in each of the relevant markets in the Commission's analysis? Why or why not?

349. Do Commenters agree with the Commission's discussion of the Baseline for the market for trading services? Why or why not?

350. Do Commenters agree with the Commission's analysis of competition in the market for trading services under the Plan? Why or why not?

351. Do Commenters agree with the Commission's analysis of effects of the Plan's funding model on competition? Why or why not? Would the funding model as outlined in the Plan affect competition in the market for trading services between exchanges and ATSs? If so, how? Do Commenters agree with the Commission's analysis of the effects on competition of the Plan's allocation of CAT fees across market participants? Why or why not? Would the Participation Fee outlined in the Plan serve as a barrier to entry for ATSs that might otherwise register as exchanges? Why or why not?

352. Do Commenters believe that the allocation of voting rights among the Participants may serve to affect competition between Participants that operate options exchanges and those that do not? Why? Do governance provisions outlined in the Plan provide controls that could prevent burdens on competition due to the allocation of voting rights among Participants? If not, are there controls that could achieve this?

353. Do Commenters believe that the allocation of voting rights among the Participants may serve to affect competition between exchanges and ATSs in the market for trading services? Why or why not?

354. Do Commenters agree with the Commission's analysis of the effects on competition of costs of compliance with the Plan? Why or why not?

355. Do Commenters agree with the Commission's analysis of the effects on competition of the Plan's enhanced surveillance and deterrence? Why or why not?

356. Do Commenters agree with the Commission's analysis of the Baseline for competition in the market for broker-dealer services? Why or why not?

357. Do Commenters agree with the Commission's analysis of the effects on competition in the market for broker-dealer services of the Plan? Why or why not? Are these effects different for smaller broker-dealers? How? How significant are these impacts?

358. Do Commenters agree with the Commission's analysis of the competition to be Plan Processor? Why or why not?

359. Do Commenters believe that any elements of the CAT NMS Plan may affect competition among the bidders? Do Commenters believe that any decisions by the Operating Committee that are allowable or likely under the proposed Plan may affect competition among the bidders in the market to be Plan Processor? If so, how would these competitive dynamics affect CAT as outlined in the Plan?

360. Do Commenters agree with the Commission's analysis of competition in the market to be Plan Processor post-selection? Why or why not?

361. Do Commenters agree with the Commission's analysis of the Baseline for competition in the market for regulatory services? Why or why not?

362. Do Commenters agree with the Commission's analysis of competition in the market for regulatory services of the Plan? Why or why not?

363. Do Commenters agree with the Commission's analysis of the Baseline for competition in the market for data reporting services? Why or why not? Do Commenters believe that capacity constraints in this market may affect broker-dealers' ability to comply with data reporting requirements under the Plan?

364. Do Commenters agree with the Commission's analysis of competition in the market for data reporting services under the Plan? Why or why not?

365. If some or all of the Participants decide to share the Raw Data they collect pursuant to the CAT NMS Plan and use the combined data for commercial purposes, how do Commenters believe that might affect competition in the markets described above?

366. In the Plan, the Participants state that they believe the Plan would have a net positive effect on efficiency. The Commission's analysis states that the Commission preliminarily believes the Plan would have a significant positive effect on efficiency. Do Commenters agree with the conclusions stated in the Plan? Why or why not? Do Commenters agree with the Commission's analysis? Why or why not?

367. Do Commenters agree that costs related to the Plan's requirements for data mapping, data dictionary creation, and encryption associated with the transmission of PII would not significantly affect efficiency? Why or why not?

368. Do Commenters agree with the Commission's analysis of the Plan's effects on the efficiency of market regulation and oversight? Why or why not?

369. Do Commenters agree with the Commission's analysis of the Plan's effects on market efficiency due to reductions in violative behavior? Why or why not?

370. Do Commenters agree with the Commission's analysis of the Plan's effect on efficiency related to reductions in ad hoc data requests from regulators? Why or why not?

371. Do Commenters agree with the Commission's analysis of the Plan's effect on efficiency due to reductions in duplicative reporting systems? Why or why not?

372. Do Commenters believe that the period of duplicative reporting that would precede the retirement of certain current, anticipated to be retired, regulatory reporting systems would significantly affect efficiency? Why or why not?

373. Do Commenters agree with the Commission's analysis of inefficiencies related to the funding model? Why or why not?

374. Do Commenters agree with the Commission's analysis of the likelihood of CAT fees being passed on to investors under the Plan? Why or why not?

375. Do Commenters agree with the Commission's analysis of the efficiency of Plan operations? Why or why not?

376. Do Commenters agree with the Commission's analysis of the effects of voting thresholds for Operating Committee decisions on efficiency? Why or why not?

377. Do Commenters agree with the Commission's analysis of the Advisory Committee's effect on efficiency under the Plan? Why or why not?

378. Do Commenters agree with the Commission's analysis of the effects on efficiency of the Participants' decision to specify or not specify certain aspects of CAT in the RFP? Why or why not?

379. Do Commenters believe that the CAT NMS Plan would impact investor confidence? If so, how? Do investors currently lack confidence because of the current state of regulatory data? Would the expected improvements to investor protection result in increased investor confidence? Please explain. What would be the expected effects of changes in investor confidence on allocative efficiency and capital formation? What would be the magnitude of the economic effects from expected changes to investor confidence? Please provide analysis.

380. The Plan states that the Participants believe that the Plan would have no deleterious effect on capital formation. Do Commenters agree with the Participants' conclusions stated in the Plan? Do Commenters agree with the Commission's preliminary belief that the Plan would not have a deleterious effect on capital formation? Why or why not?

381. Do Commenters agree with the Commission's analysis of the Plan's effects on capital formation due to enhanced market surveillance and regulatory activities? Why or why not?

382. Do Commenters agree with the Commission's analysis of effects on capital formation due to data security provisions of the Plan? Why or why not?

H. Alternatives

As a part of its economic analysis, the Commission is considering and soliciting comment on alternatives to certain approaches or elements of the CAT NMS Plan. The Commission analyzes alternatives that could have a direct and significant impact on costs or benefits deriving from at least one of the four data qualities discussed above: accuracy, completeness, accessibility, and timeliness. While the discussed alternatives are not the only alternatives that could significantly impact costs, benefits, or data quality, they are an attempt to identify reasonable options. Each has the potential to alter the Commission's preliminary conclusions regarding the economic effects of the CAT NMS Plan.

The analysis of alternatives is divided into three categories. First, the Commission analyzes alternatives to the approaches the Exemption Order permitted the Participants to include in the Plan. As noted in the Exemption Order, the Commission was persuaded to grant exemptive relief to provide flexibility such that the proposed approaches described in the Exemption Request can be included in the CAT NMS Plan and subject to notice and comment. Second, the Commission analyzes alternatives to certain specific approaches in the CAT NMS Plan, including alternative approaches to clock synchronization, time stamps, Error Rates, error correction timelines, the funding model, listing exchange symbology, data accessibility standards, and the intake capacity levels. Third, the Commission analyzes alternatives to the scope of certain specific elements of the Plan. Specifically, the Commission analyzes the impact of changing the scope of the CAT to exclude certain data fields. The Commission also analyzes alternatives to exclude OTC Equity Securities and the requirement to periodically refresh all customer information. Finally, the Commission solicits comment on the broad alternative of modifying OATS and/or another existing system to meet the requirements of Rule 613 instead of approving the Plan.

See Exemption Order, supra note 18.

Id.

1. Alternatives to the Approaches the Exemption Order Permitted To Be Included in the Plan

The Commission is soliciting additional comment on alternatives to the approaches the Exemption Order permitted the SROs to include in the CAT NMS Plan. Specifically, the Commission is soliciting comment on how the following alternatives (the “Rule 613 approach”), described in further detail below, would affect the costs and benefits of the CAT: (a) Requiring both Options Market Makers and Options Exchanges to report Options Market Maker quotations to the Central Repository, (b) requiring CAT Reporters to report a Customer-ID for each Customer upon the original receipt or origination of an order, (c) requiring CAT Reporters to report a universal CAT-Reporter-ID to the Central Repository for orders and certain Reportable Events, (d) requiring the reporting of the account number for any subaccount to which an execution is allocated, and (e) requiring that Manual Order Events be reported with a time stamp granularity of one millisecond.

Id.

a. Options Market Maker Quotes

The Commission is soliciting comment on how an alternative approach—the Rule 613 approach—to the reporting of Options Market Maker quotations might impact the costs and benefits of the Plan. Rule 613(c)(7) provides that the CAT NMS Plan must require each national securities exchange, national securities association, and any member of such exchange or association to record and electronically report to the Central Repository details for each order and each Reportable Event, including the routing and modification or cancellation of an order. Rule 613(j)(8) defines “order” to include “any bid or offer” so that the details for each Options Market Maker quotation must be reported to the Central Repository by both the Options Market Maker and the exchange to which it routes its quote. The SROs requested an exemption from Rules 613(c)(7)(ii) and (iv) and proposed an approach whereby only Options Exchanges—but not Options Market Makers—would be required to report information to the Central Repository regarding Options Market Maker quotations. The Commission granted exemptive relief to the SROs to allow the approach to collecting Options Market Maker quotations described in the Exemption Request to be included in the CAT NMS Plan and subject to notice and comment.

See Exemptive Request Letter, supra note 16, at 2-5.

See Exemption Order, supra note 18.

Pursuant to the exemptive relief granted by the Commission, the CAT NMS Plan provides that only Options Exchanges—but not Options Market Makers—would be required to report information to the Central Repository regarding Options Market Maker quotations. On the other hand, the Rule 613 approach would require that each Options Market Maker quotation be reported to the Central Repository by both the Options Market Maker and the exchange to which it routes its quote. The Commission preliminarily believes that the Rule 613 approach would increase certain costs associated with the implementation and operation of CAT as compared to the Plan as filed without providing any additional material information.

See CAT NMS Plan, supra note 3, at Appendix C, Background Section.

Under the Rule 613 approach, the reports from the Options Exchanges would be virtually identical to the reports coming from the Options Market Makers, with the exception that reports from the Options Market Makers would indicate the time that the Options Market Maker routes its quote, or any modification or cancellation thereof, to the exchange (“Quote Sent Time”). However, to ensure that regulators would receive all of the information contemplated by Rule 613(c)(7), the CAT NMS Plan requires that (1) Options Market Makers report to the relevant Options Exchange the Quote Sent Time along with any quotation, or any modification or cancellation thereof; and (2) Options Exchanges submit the quotation data received from Options Market Makers, including the Quote Sent Time, to the Central Repository without change. Under the CAT NMS Plan, therefore, regulators would have access to all the material information in CAT that would be provided under the Rule 613 approach. As such, the Commission preliminarily does not believe that there would be any additional benefits to using the Rule 613 approach.

Id. at Section 6.4(d)(iii).

Furthermore, the CAT NMS Plan estimates that the Rule 613 approach would increase the amount of records that must be handled by the Central Repository by 18 billion records per day, at an additional cost of between $2 million and $16 million for data storage and technical infrastructure over a five year period. A cost survey estimates the Rule 613 approach would cost all Options Market Makers between $307.6 million and $382 million over five years. Under the approach taken in the CAT NMS Plan, these costs would be avoided but the Options Market Makers surveyed would spend approximately $8.5 million to send Quote Sent Times to the exchanges and all Options Market Makers would spend $36.9M to $76.8M. In aggregate, the estimates provided suggest that the Rule 613 approach would add between $230.80 million and $345.10 million to industry costs over five years. The Exemption Request also notes that the additional costs would be disproportionately borne by smaller broker-dealers relative to their market share.

Id. at Appendix C, Section B.7(b)(iv)(B).

See FIF, SIFMA, and Security Traders Association, Cost Survey Report on CAT Reporting of Options Quotes by Market Makers (November 5, 2013), available at http://www.catnmsplan.com/industryfeedback/p601771.pdf ; see also CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iv)(B).

See FIF, SIFMA, and Security Traders Association, Cost Survey Report on CAT Reporting of Options Quotes by Market Makers 3-4 (November 5, 2013), available at http://www.catnmsplan.com/industryfeedback/p601771.pdf .

To be conservative, the Commission estimates the lower end of the range to be the lower cost to comply with a CAT NMS Plan without the exemption minus the higher cost to comply with a CAT NMS Plan with the exemption ($230.8M = $307.6 − $76.8M). Likewise, the higher end of the range is the higher cost to comply with a CAT NMS Plan without the exemption minus the lower cost to comply with a CAT NMS Plan with the exemption ($345.1M = $382M − $36.9M).

See Exemptive Request Letter, supra note 16, at 7.

The Commission notes that there are limitations to the cost estimation methodology presented in the Exemption Request. These limitations include the lack of quantified cost estimates for additional indirect cost savings associated with the exemption. However, the Commission preliminarily believes that the Rule 613 approach would increase certain costs associated with the implementation and operation of CAT as compared to the Plan as filed without providing any additional material information.

b. Customer-ID

The Commission is soliciting comment on how an alternative approach—the Rule 613 approach—to the reporting of customer information might impact the costs and benefits of the Plan. Rule 613(c)(7)(i)(A) requires that for the original receipt or origination of an order, a CAT Reporter report the “Customer-ID(s) for each Customer.” “Customer-ID” is defined in Rule 613(j)(5) to mean “with respect to a customer, a code that uniquely and consistently identifies such customer for purposes of providing data to the central repository.” Rule 613(c)(8) further requires that “[a]ll plan sponsors and their members shall use the same Customer-ID and CAT-Reporter-ID for each customer and broker-dealer.” The SROs requested an exemption from the requirements in Rule 613(c)(7)(i)(A) and Rule 613(c)(8), and proposed an approach whereby each broker-dealer would assign a unique Firm Designated ID to each trading account, which would be linked to a set of identifying information (the “Customer Information Approach”). Using the Firm Designated ID and the other information identifying the Customer that would be reported to the Central Repository, the Plan Processor would then assign a unique Customer-ID to each Customer. Upon original receipt or origination of an order, broker-dealers would only be required to report the Firm Designated ID on each new order, rather than using the Customer-ID. The Commission granted exemptive relief to the SROs to allow the alternative approach to Customer-IDs described in the Exemption Request to be included in the CAT NMS Plan and subject to notice and comment.

See Exemptive Request Letter, supra note 16, at 9. Because the Plan Processor would still assign a Customer-ID to each Customer under the Customer Information Approach, the SROs did not request an exemption from Rule 613(j)(5).

See Exemption Order, supra note 18, at 11863.

Pursuant to the exemptive relief granted by the Commission, the CAT NMS Plan provides for the use of the Customer Information Approach. The Commission is soliciting comment on the Rule 613 approach, which would require that broker-dealers report Customer information using a consistent, unique Customer-ID, as set out in in Rule 613(c)(7)(i)(A) and Rule 613(c)(8). The Commission preliminarily believes that the Rule 613 approach would increase certain costs associated with the implementation and operation of CAT as compared to the Customer Information Approach while providing substantially identical data.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.1(a)(iii).

The Commission also preliminarily believes that the Rule 613 approach would have no significant impact on the benefits of the CAT NMS Plan. The Participants maintain that, under the Rule 613 approach, there would be no gains in terms of accuracy or reliability, no effect on the ability to link records, and no effect on the time the data would be made available to regulators, as compared to the Customer Information Approach. The Participants also believe that there may be accuracy gains under the Customer Information Approach if it reduces errors that may otherwise occur if broker-dealers must adapt their systems and business processes to manage Customer-IDs.

See Exemptive Request Letter, supra note 16, at 15-18.

Id.

The Commission also preliminarily believes that the Rule 613 approach would increase the costs of the CAT NMS Plan. In their Exemption Request, the Participants discussed a number of reasons why the Customer Information Approach is less burdensome than the Rule 613 approach. First, it reduces the CAT implementation burden on market participants by eliminating the need for changes to their current customer identification systems. Currently, market participants have individual formats for their customer identifiers; under the Customer Information Approach, no standardization of form would be required. Second, the Customer Information Approach eliminates the need for centrally-assigned Customer-IDs to be assigned at the Central Repository and communicated back to market participants. Third, it allows the Plan Processor to implement modifications and technical upgrades to the Customer-ID generation process and infrastructure without the involvement of CAT Reporters. Fourth, the Customer Information Approach eliminates the need to train CAT Reporters on the Customer-ID management process and provide related technical support. Fifth, it potentially reduces delays faced by investors opening new accounts, who might not be able to transact until the Central Repository has assigned a Customer-ID and communicated it to the broker-dealer representing the Customer.

See id. at 17.

See id.

See id.

See id. at 16-17.

Based on cost survey data provided by the Participants, the Rule 613 approach would increase quantifiable costs to the top three tiers of CAT Reporters by at least $195 million. The Commission notes that this likely underestimates the increased costs to all CAT Reporters because the Rule 613 approach would likely increase costs to CAT Reporters outside the top three tiers also. Furthermore, the Bidders have indicated that the costs of building and operating the Central Repository under the Rule 613 approach would not be lower than the costs of the Customer Information Approach. The Commission therefore preliminarily believes that the Rule 613 approach would increase the costs of the CAT NMS Plan relative to the Plan's Customer Information Approach, while providing substantially identical data.

Id. at 17-18.

Id. at 17.

c. CAT-Reporter-ID

The Commission is soliciting comment on how an alternative approach—the Rule 613 approach—to the reporting of CAT Reporter information might impact the costs and benefits of the Plan. A CAT-Reporter-ID is “a code that uniquely and consistently identifies [a CAT Reporter] for purposes of providing data to the central repository.” Subparagraphs (c)(7)(i)(C), (ii)(D), (ii)(E), (iii)(D), (iii)(E), (iv)(F), (v)(F), (vi)(B), and (c)(8) of Rule 613 provide that the CAT NMS Plan must require CAT Reporters to report CAT-Reporter-IDs to the Central Repository for orders and certain Reportable Events. Additionally, Rule 613(c)(8) requires that CAT Reporters use the same CAT-Reporter-ID for each broker-dealer. To leverage existing infrastructure and business processes, the Participants requested an exemption from Rule 613(c)(7) and (c)(8) to allow a different approach to be included in the Plan; CAT Reporters would report existing SRO-assigned market participant identifiers when submitting data to the Central Repository (“SRO-Assigned Market Participant Identifiers”). The Central Repository would then generate a corresponding CAT-Reporter-ID for internal use to identify CAT Reporters. This approach—called the “Existing Identifier Approach”—allows the CAT-Reporter-IDs to be managed at the Central Repository by the Plan Processor without the involvement of the Reporters. The Commission granted exemptive relief to the SROs to allow the Existing Identifier Approach to be included in the CAT NMS Plan and subject to notice and comment.

17 CFR 242.613(c)(7)(i)(C), (ii)(D), (ii)(E), (iii)(D), (iii)(E), (iv)(F), (v)(F), (vi)(B), and (c)(8).

See Exemptive Request Letter, supra note 16, at 19.

Id.

See Exemption Order, supra note 18, at 11866.

Pursuant to the exemptive relief granted by the Commission, the CAT NMS Plan provides for the use of the Existing Identifier Approach. The Commission is soliciting additional comment on the Rule 613 approach, which would require that CAT Reporters use a consistent, unique CAT-Reporter-ID, as set out in in Rule 613(c)(7) and Rule 613(c)(8). The Commission preliminarily believes that the Rule 613 approach would increase certain costs associated with the implementation and operation of CAT as compared to the Existing Identifier Approach while providing substantially identical data.

See, e.g., CAT NMS Plan, supra note 3, at Sections 6.3(d) and (e), 6.4(d).

The Commission preliminarily believes that the Rule 613 approach would not result in more reliable or accurate data as compared to the Existing Identifier Approach. The Exemption Request states that “the proposed approach would not compromise the goal of Rule 613 to record and link Reportable Events to the CAT Reporter associated with the event.” The processed CAT Data would contain the CAT-Reporter-ID fields, and the Participants maintain that there would be no loss of accuracy or reliability, no effect on the ability to link records, and no effect on the time the data would be made available to regulators.

See Exemptive Request Letter, supra note 16, at 21.

Id. at 22-23.

In fact, the Commission preliminarily believes that the Rule 613 approach would reduce the quality of data obtained as compared to the Existing Identifier Approach. Specifically, the Rule 613 approach would reduce the granularity of information on departments, trading desks, and other business units within CAT Reporters, which would be captured under the Existing Identifier Approach. This additional granularity would be possible under the Existing Identifier Approach because identifiers currently in use are often assigned to entities that are defined more granularly than the CAT-Reporter-ID level. The Commission also preliminarily believes that the ability to leverage existing infrastructure and business processes may reduce the potential for delays and errors that could be associated with requiring CAT Reporters to modify their systems and workflows to handle the CAT-Reporter-IDs.

The Commission preliminarily believes that the Rule 613 approach would increase the costs of the CAT NMS Plan relative to the Existing Identifier Approach. The Participants estimate implementation costs for the top three tiers of CAT Reporters for the Rule 613 approach of $78 to $244 million, depending on how report types have to use the CAT-Reporter-IDs. The Exemption Request does not compare these costs to the Existing Identifier Approach allowed by the exemption and included in the Plan. The Participants note that these estimates are conservative because they are based on only 11% of broker-dealers. The Participants indicated that they have consulted with the bidders and the industry in compiling this analysis.

Id. at 24.

Id. at 24.

Id. at 25.

Id. at 22.

While the Commission preliminarily believes that the Rule 613 approach would increase certain costs associated with the implementation and operation of CAT as compared to the Existing Identifier Approach, the Commission notes that there are limitations associated with the cost estimation methodology presented in the Exemption Request. These limitations include the exclusion of SROs and smaller CAT Reporters from the survey, no apparent differentiation between initial, deferred, and recurring costs, and lack of support for the method used to extrapolate the estimates for large broker-dealers to the industry. Nor do the cost estimates address the broker-dealers who would be CAT Reporters but are currently not OATS reporters, including those that are currently not registered with FINRA, which may have a very different cost structure. However, it is likely that the dominant effect would be the exclusion of many CAT Reporters from the cost estimates, which would tend to underestimate the cost increases. The Commission currently has no data from which it can independently estimate the cost differential because it depends on information internal to each of a heterogeneous group of CAT Reporters, which is not compiled or stored anywhere and to which the Commission therefore does not have ready access. The Commission believes that these effects are not likely to alter its preliminary conclusion that the Rule 613 approach would significantly increase the costs of the CAT NMS Plan as compared to the Plan's Existing Identifier Approach. The Commission is requesting comment on this preliminary conclusion and any additional data Commenters believe should be considered.

d. Linking Order Executions to Allocations

The Commission is soliciting comment on how an alternative approach to the reporting of allocation information—the Rule 613 approach—might impact the costs and benefits of the Plan. Rule 613(c)(7)(vi)(A) requires each CAT Reporter to record and report to the Central Repository “the account number for any subaccounts to which the execution is allocated (in whole or part).” This information would allow regulators to link the subaccount to which an allocation was made to the original order placed and its execution. In the Exemptive Request Letter and April 2015 Supplement, the SROs requested an exemption from Rule 613(c)(7)(vi)(A) to include in the Plan an approach whereby CAT Reporters would instead submit information to the Central Repository that would allow regulators to link subaccount information to the Customer that submitted the original order. The Commission granted exemptive relief to the SROs to allow this approach to be included in the CAT NMS Plan and subject to notice and comment.

See Exemptive Request Letter, supra note 16, at 28-29; April 2015 Supplement, supra note 16, at 2.

See Exemption Order, supra note 18, at 11868.

Pursuant to the exemptive relief granted by the Commission, the CAT NMS Plan provides that, rather than providing the account number for any subaccounts to which the execution is allocated, CAT Reporters would submit information to the Central Repository in the form of an Allocation Report, in order to allow regulators to link subaccount information to the Customer that submitted the original order. The Allocation Report would include the Firm Designated ID for any account(s), including subaccount(s), to which executed shares are allocated, and provide the security that has been allocated, the identifier of the firm reporting the allocation, the price per share of shares allocated, the side of shares allocated, the number of shares allocated to each account, and the time of the allocation, which is information that is not currently required to be reported and/or retained by broker-dealers. There would not be a direct link in the Central Repository between the subaccounts to which an execution is allocated and the execution itself. However, CAT Reporters would be required to report each allocation to the Central Repository on an Allocation Report, and the Firm Designated ID of the relevant subaccount provided to the Central Repository as part of the Allocation Report could be used by the Central Repository to link the subaccount holder to those with authority to trade on behalf of the account. Further, the Allocation Reports used in conjunction with order lifecycle information in CAT would assist regulators in identifying, through additional investigation, the probable group of orders that led to allocations.

See CAT NMS Plan, supra note 3, at Section 6.4(d)(ii)(A)(1).

See id. at Section 1.1; see also Exemption Order, supra note 18, at 44-45.

See Exemption Order, supra note 18, at 45.

Id.

The Commission is soliciting comment on the Rule 613 approach, which would require CAT Reporters to record and report the account number for any subaccounts to which the execution is allocated, as described above. The Commission preliminarily believes that that the Rule 613 approach could provide the Central Repository with a way to link allocations to order lifecycles. This linkage would not be available under the current approach. However, based on estimates provided by the Participants, the Commission preliminarily believes that the Rule 613 approach would increase certain costs associated with the implementation and operation of CAT as compared to the Plan as filed by roughly $525 million.

In the Exemption Request, the SROs explained that under the Rule 613 approach allocations made from an average price account would not reflect a true one-to-one relationship between an execution and an allocation, and therefore the information provided would not directly link a single order execution and the subaccount to which an allocation was made. See Exemptive Request Letter, supra note 16, at 28. However, the Commission believes that under the Rule 613 approach, regulators would receive information that would identify each execution resulting from the original order placed, as well as the identity of all the subaccounts to which those executions were allocated. This information would provide regulators a finite list of executions from which the subaccount allocations could have been made.

The Participants estimate that the Plan's approach to allocation information would result in a reduction in implementation cost for the top three tiers of CAT Reporters of $525 million as compared to the Rule 613 approach. See Exemptive Request Letter, supra note 16, at 31.

The Commission preliminarily believes that either approach would allow regulators to link specific allocations, and the prices received on those allocations, with the aggregated executions that resulted in the allocations and their execution prices. Industry feedback received by the Participants indicates that existing business practices typically involve aggregating executions in an average price account before making allocations, and forcing a precise matching between orders and executions ex-post would be misleading. The Exemption Request maintains that, under the approach in the Plan, there would be no loss of accuracy or reliability, no effect on the ability to link order records, and no effect on the time the data would be made available to regulators as compared to the Rule 613 approach. The Exemption Request also states that there may be accuracy and reliability gains if the exemption reduces errors that may otherwise occur if broker-dealers were required to re-engineer their allocation handling systems and business processes to meet the requirements of Rule 613.

See Exemptive Request Letter, supra note 16, at 28 (“[T]his approach . . . introduces an artificial relationship between any one execution and one allocation. . . . Although, . . . the ultimate allocation of the shares executed that result from [an] aggregated order may be useful for regulatory surveillance purposes, tying these allocations to multiple different executions is of little regulatory benefit.”).

Id. at 30.

Id.

However, the Rule 613 approach would provide regulators access to allocations linked to specific disaggregated orders, which is not possible under the approach in the Plan. The Exemption Request notes that linking particular allocations to particular order lifecycles would be inaccurate in some circumstances, such as when many orders are allocated to many customers. The Commission is soliciting comment on whether such information would necessarily be inaccurate, and whether requiring the linking of allocations to order lifecycles would reduce accuracy for several reasons. First, in cases in which one order is allocated to one customer, the Rule 613 approach would provide an improvement in accuracy over the approach proposed in the CAT NMS Plan because the Rule 613 approach would allow the Central Repository to accurately link such allocations to order lifecycles whereas the approach proposed in the CAT NMS Plan might not. Under the CAT NMS Plan, for regulators to link the allocations to the order lifecycles, they would need to construct an algorithm that would rely on less information than the Central Repository would have under the Rule 613 approach. As a result, these regulator linkages would likely be less accurate than a Central Repository linkage. The Commission preliminarily believes that this is true for cases in which one order is allocated to many customers and when many orders are linked to one customer. For the many-to-many allocations, in which many customer orders are grouped and worked by the market participant using many orders to acquire the aggregate position ultimately used to fill the customer orders, the Commission notes that broker-dealers likely already maintain records that allow them to ensure that the allocations receive fair prices based on market executions. The Commission is soliciting comment on whether such information might be sufficient to link the many allocations to the many orders executed in an accurate manner. Such information would greatly aid investigations of fair allocations because it would allow regulators to reconstruct the manner in which allocations occur.

See id. at 28-30.

The Commission preliminarily believes that the Rule 613 approach would increase the costs of compliance with the CAT NMS Plan. According to industry feedback collected by the Participants, the Rule 613 approach would require broker-dealers to undertake a major re-engineering of their middle and back office systems and processes. The Participants estimate a reduction in implementation cost over the Rule 613(c)(7)(vi) Baseline for the top three tiers of CAT Reporters of $525 million; consequently, the Commission preliminarily believes that this alternative would cost at least $525 million more than the estimated costs of the CAT NMS Plan to implement. The Participants indicated that they have consulted with the bidders and the industry in compiling this analysis.

Id. at 27.

Id. at 31.

See id. at 30-31.

e. Time Stamp Granularity

The Commission is soliciting comment on how an alternative approach—the Rule 613 approach—to time stamps on “Manual Order Events” might impact the costs and benefits of the Plan. Rule 613(c)(7) and Rule 613(d)(3) require time stamps with a minimum granularity of one millisecond on all order events. The Participants requested an exemption from the requirement in Rule 613(d)(3) that for Manual Order Events each CAT Reporter record and report details for Reportable Events with time stamps that “reflect current industry standards and [are] at least to the millisecond.” The Commission granted exemptive relief to the SROs to allow the approach to recording and reporting time stamps for Manual Order Events described in the Exemption Request to be included in the CAT NMS Plan and subject to notice and comment.

“Manual Order Events” are defined to mean “non-electronic communication[s] of order-related information for which CAT Reporters must record and report the time of the event.” See CAT NMS Plan, supra note 3, at Section 1.1.

See 17 CFR 242.613(c)(7) (requiring use of time stamps pursuant to 17 CFR 242.613(d)(3)); 17 CFR 242.613(d)(3) (requiring time stamp granularity be “at least to the millisecond”).

See 17 CFR 242.613(d)(3); Exemptive Request Letter, supra note 16, at 32.

See Exemption Order, supra note 18, at 11869.

Pursuant to the exemptive relief granted by the Commission, the CAT NMS Plan provides that: (1) Each CAT Reporter would record and report Manual Order Event time stamps to the second; (2) Manual Order Events would be identified as such when reported to the CAT; and (3) CAT Reporters would report in millisecond time stamp increments when a Manual Order Event is captured electronically in the relevant order handling and execution system of the CAT Reporter (“Electronic Capture Time”). On the other hand, the Rule 613 approach would require that CAT Reporters record and report details for Manual Order Events with time stamps that are at least to the millisecond, as required by Rule 613(c)(7) and Rule 613(d)(3). The Commission preliminarily believes that the Rule 613 approach would increase the costs of implementing the CAT NMS Plan while providing little regulatory benefit relative to the current approach.

See CAT NMS Plan, supra note 3, at Section 6.8.

The Participants maintain in the Exemption Request that there would be little benefit, and possibly some adverse consequences, of capturing Manual Order Event time stamps in milliseconds. They note that determining the time of a manual event is inherently imprecise, due to the limits of human reaction time in completing a transaction and the time required to manually record the event. They claim human reaction time to visual stimulus is on the order of 400-500 milliseconds, making millisecond time stamps imprecise. The Commission preliminarily agrees that attempting to record the precise millisecond in which a manual event occurred would necessarily be imprecise. The Commission also preliminarily agrees that potential adverse consequences could arise from relying on time stamps with a misleading level of precision.

See Exemptive Request Letter, supra note 16, at 33.

Id. at 37.

Id.

The Commission notes that Manual Order Events are not clearly and exhaustively defined, and the definitions may not be available until the Technical Specifications are published. It may be possible for the Plan Processor to classify some types of order events as Manual Order Events that were not considered to be a Manual Order Event for the purposes of this analysis. This creates a degree of uncertainty as to whether the Rule 613 approach might yield some regulatory benefit.

The Participants discussed the costs and benefits of the proposed exemption in their Exemption Request. They estimated a minimum total cost to the industry of $10.5 million based on the cost of advanced OATS-compliant clocks with granularity of one second, and noted that clocks with millisecond granularity would likely be more expensive if available. The Participants also noted that the industry was consulted through the DAG and an unsuccessful attempt was made to find a commercially available time stamping device with millisecond granularity. Based on this information, the Commission preliminarily believes the Rule 613 approach to Manual Order Events would increase certain costs associated with the implementation and operation of CAT as compared to the Plan as filed without providing any significant additional benefit.

See Exemptive Request Letter, supra note 16, at 36-37.

Id. at 35.

2. Alternatives to Certain Specific Approaches in the CAT NMS Plan

The Commission has analyzed alternatives to specific approaches in the CAT NMS Plan with respect to clock synchronization, time stamps, error rates, the time within which errors must be corrected, the funding model, requirements regarding listing exchange symbology, data accessibility standards, and intake capacity levels.

a. Clock Synchronization

The Commission is soliciting comments on alternate approaches to clock synchronization as compared to those proposed in the CAT NMS Plan. First, the Commission is soliciting comment on alternatives to the Plan's one-size-fits-all definition of “industry standard.” Under these alternatives, “industry standard” would be defined in terms of the standard practices of different segments of the CAT Reporters, or by looking at information other than current industry practices. These alternative approaches could result in clock offset tolerances shorter than the CAT NMS Plan's proposed 50 millisecond standard for some or all CAT Reporters. The Commission preliminarily believes that these alternatives could substantially increase the benefits of CAT in regulatory activities that require event sequencing, such as analysis and reconstruction of market events, as well as market analysis and research in support of policy decisions, and cross-market surveillance, examinations, investigations, and other enforcement functions.

See Section IV.E.1.b(2), supra.

Second, the Commission is soliciting comment on two additional alternatives that could allow for more cost-effective clock synchronization standards. In particular, the Commission is soliciting comment on modifying the requirement to document clock synchronization activities such that only events that require clock adjustment would be required to be documented, and modifying the clock synchronization requirement such that clocks would not have to be synchronized at times when systems are not recording time-sensitive CAT Reportable Events, such as orders originated outside of market hours when they are not immediately actionable. The Commission preliminarily believes that reduced clock synchronization logging requirements might significantly reduce ongoing costs associated with clock synchronization compliance as compared to the Plan as filed, without losing any additional material information. In addition, the Commission preliminarily believes that more flexible clock synchronization standards outside of regular and extended trading hours may also reduce costs without a material loss to the ability of regulators to sequence order events as compared to the Plan as filed, without losing any additional material information. Each of these alternatives is outlined below.

(1) Alternative Clock Synchronization Standards

Rule 613(d)(1) requires synchronization of business clocks for the purposes of recording the date and time of Reportable Events consistent with industry standards. The CAT NMS Plan describes the “industry standard” in terms of the technology adopted by the majority in the industry. The Plan therefore bases its clock synchronization standard on current practices of the broker-dealer industry generally, and provides that one standard would apply to all CAT Reporters. The Commission is soliciting comment on an alternative interpretation of “industry standard” that would consider the standard practices of different segments of the CAT Reporters for the purposes of setting the clock synchronization requirements. The Commission is also soliciting comment on an alternative that would define industry standard by looking at information other than current industry practice; for example, the most accurate technology currently available in the industry, or the standard recommended by a particular authority or industry group.

The Commission did not define the term “industry standard” in Rule 613. In the Adopting Release, the Commission noted that it expected the Plan to “specify the time increment within which clock synchronization must be maintained, and the reasons the plan sponsors believe this represents the industry standard.” See Adopting Release, supra note 9, at 45774.

The benefits of alternative clock offset tolerances discussed in this Section may be dependent on time stamp granularity requirements. Related alternatives are discussed in Section IV.H.2.b, infra.

See CAT NMS Plan, supra note 3, at Appendix C, Section 12(p).

First, the Commission is soliciting comment on an alternative definition of industry standard that would consider the standard practices of different segments of CAT Reporters. Under this alternative, all systems within market participants that process CAT-Reportable Events would be required to comply with a clock synchronization requirement reflecting an industry standard particular to that market participant's segment of the industry. Currently, the Commission lacks the information necessary to reach a preliminary conclusion regarding the appropriate industry standards for all subsets of the industry. Specifically, neither the FIF Clock Offset Survey nor the Plan provides comprehensive data on the clock synchronization practices of firms within each of the relevant subsets of the industry, and the Commission has no data from which it can independently estimate the cost differential because the Commission is not aware of any such data available to it at this time. However, the Commission is soliciting comment on this approach, which it believes would result in a clock offset tolerance of less than 50 milliseconds for some market participants. The Commission seeks comment on the current practices for clock synchronization in various segments of the industry, including but not limited to broker-dealers that are introducing firms, institutional firms, retail firms that accept customer orders electronically, registered market makers and principal trading firms, as well as service bureaus hosting order management systems, exchanges and ATSs, and branches of broker-dealers that predominantly handle manual orders. The Commission also seeks comment on the costs and benefits of requiring varying clock offset tolerances within the industry.

The Commission notes that the current practices for exchanges and Execution Venues may differ from the industry standard for broker-dealers as defined by the Plan, and current practices for certain systems within broker-dealers may vary by the system within the broker-dealers. For example, a small clock offset tolerance may be nearly universally adopted for systems like ATSs that operate a matching engine, while systems involved in manual entry of orders may typically have larger clock offset tolerances. By defining industry standard based on practices of the broker-dealer industry generally, the Plan does not account for these differences.

Other information now available for the Commission and the public to study, particularly information from the FIF Clock Offset Survey, shows that several of the survey respondents that have a current clock offset tolerance of one second are clearing firms or service bureaus. According to the same survey, current clock offset tolerances vary from one second to five microseconds among the broker-dealers surveyed with 22% of respondents having multiple clock offset tolerances across their systems. Further, the FIF Clock Offset Survey shows that the firms with multiple clock offset tolerances typically engage in multiple lines of business. The fact that some broker-dealers maintain clock offset tolerances at different levels within the firm suggests that these broker-dealers believe that clock precision is more important for some systems; furthermore, based on conversations with market participants, the Commission preliminarily believes that market participants strategically upgrade certain systems and reallocate older technology within the firm to applications where up-to-date technology is less critical.

See FIF Clock Offset Survey, supra note 127.

See Section IV.D.2.b(2)B.i, supra for more information regarding the distribution of broker-dealer clock offset tolerances.

Based on FIF-organized conversations with broker-dealers and service bureaus. See supra note 880.

Systems that have greater clock offset tolerances may have technology that is too old to support smaller clock offset tolerances. The Commission preliminarily believes that if a shorter clock offset tolerance is important to these broker-dealers, they would update their systems to support newer technology capable of smaller clock offset tolerances.

Finally, exchanges and ATSs, as well as the SIPs, may have current clock offset tolerances that are significantly different from the clock offset tolerances at broker-dealers and could therefore achieve finer clock offset tolerances at lower cost than broker-dealers. According to FIF, all exchange matching engines meet a clock offset tolerance of 50 milliseconds or less while NASDAQ states that all exchanges that trade NASDAQ securities have clock offset tolerances of 100 microseconds or less. In conversations with Commission Staff, the Participants stated that absolute clock offset on exchanges averages 36 microseconds, further suggesting that certain business activities warrant smaller clock synchronization tolerances.

See supra notes 441 and 442. Specifically, the NASDAQ SIP Web site implies that exchanges reporting to the NASDAQ SIP synchronize their systems to 100 microseconds.

See Section IV.D.2.b(2)B.i, supra for more information on clock offset tolerances of exchanges and the SIPs.

See supra note 436.

Given this information, the Commission recognizes the possibility that some business systems and some CAT Reporter types would rarely be responsible for recording the date and time of reportable events and also recognizes that the time stamp precision of such rare events might not be as critical as for other events. For example, a system that routes customer orders to market centers may be considered critical for sequencing market events, while a system that facilitates manual input of orders received by telephone may not. Conversely, the clock synchronization practices of some CAT Reporters may be more critical to the overall benefits of CAT or could be less costly to implement. For example, a service bureau that provides an order-handling system hosted on its own servers is likely to route orders for many market participants and its clock synchronization practices would, thus, be critical to event sequencing. On the other hand, the precision of time stamps from systems of an isolated broker-dealer that routes customer orders to its service bureau or another broker-dealer for market access and conducts no proprietary trading may be less critical to event sequencing, especially if the receiving system at the service bureau would record a high-precision time stamp when the order is received. Furthermore, instituting higher clock precision at a single service bureau would be less costly than instituting that same level of clock precision at the service bureau and all of its broker-dealer customers as is required by the Plan as filed.

Relative to the proposed clock synchronization standard, the Commission preliminarily believes that an alternative approach that would consider the standard practices of different segments of the industry for the purposes of setting the clock synchronization requirements, and would require a smaller clock offset tolerance than in the Plan for certain business systems that are more critical to being able to accurately sequence order events, could have significant benefits. In other words, the Commission preliminarily believes that some business systems may be responsible for time stamping more time-sensitive order events than others, where more time-sensitive orders are those for which precise time stamps are more critical for event sequencing.

The Commission does not currently have the information necessary to specify which particular types of business system handle more time-sensitive orders because neither the FIF Clock Offset Survey nor the Plan provides this data. The Commission has no data from which it can independently estimate this because the Commission is not aware of any such data available to it. However, the Commission recognizes the potential for such an approach. For example, it is possible that almost all of the order origination events, routing events, modification events, and execution events, which are likely to be more time-sensitive than other CAT Reportable Events, occur on systems at broker-dealers that conduct certain types of businesses. The businesses that seem most likely to record these time-sensitive events include: Introducing broker-dealers; institutional broker-dealers; retail broker-dealers that accept customer orders electronically; registered market makers; principal trading firms; service bureaus that host order management systems; exchanges; and ATSs.

Further, some systems collect order events that either do not require a granular time stamp; other systems would not be required to record order events in real time. An example would be regional branches of broker-dealers that only handle manual orders which require a time stamp to the second until the broker enters the order into an electronic system. If the order entry hits a centralized system quickly, then perhaps the clock precision of the centralized system may be sufficient for sequencing.

The Commission is also soliciting comment on an alternative approach that would define industry standard by looking at information other than current industry practices; for example, by considering the most accurate technology currently available in the industry, or the standard recommended by a particular industry group or authority. Defining industry standards by majority practices may have the unintended effect of setting a standard that delays adopting advances in technology. The Commission preliminarily believes that this alternative approach could result in defining an industry standard for clock synchronization that would require a clock offset tolerance for all CAT Reporters that is lower than the 50 millisecond standard required by the Plan. The Commission seeks comment on any appropriate definitions of “industry standard” with respect to clock synchronization, including the costs and benefits of using any alternative definitions of “industry standard” for the purposes of setting clock synchronization requirements. The Commission also seeks comment on whether a definition of “industry standard” could set a maximum clock offset tolerance with an expectation that each CAT Reporter would be responsible for smaller clock offsets if the CAT Reporter is technically capable of such clock offsets.

The Commission conducted an analysis to assess the benefits of alternative approaches to defining industry standard that would result in smaller clock offset tolerances for some or all segments of CAT Reporters. The Commission evaluated the percentage of unrelated events that can potentially be sequenced under various clock offset tolerances, including the 50 millisecond tolerance outlined in the CAT NMS Plan. The Commission estimates that approximately 7.84% of unrelated orders for listed equities and 18.83% of unrelated orders for listed options can be accurately sequenced using a clock offset tolerance of 50 milliseconds. The Commission augmented this analysis by conducting a clock synchronization analysis to examine certain alternative clock offset tolerances from those examined in the FIF Clock Offset Survey. Table 10 shows the results of the Commission's analysis as a percentage of unrelated order events for equities that could be sequenced under various alternative clock offset tolerance.

See Section IV.E.1.b(2)A, supra. In general, events occur with such frequency that a 50 millisecond clock synchronization standard would not be sufficient to sequence all orders; see also CAT NMS Plan, supra note 3, at Appendix C, Section A.3(c) n.110 (“Events occurring within a single system that uses the same clock to time stamp those events should be able to be accurately sequenced based on the time stamp. For unrelated events, e.g., multiple unrelated orders from different broker-dealers, there would be no way to definitively sequence order events within the allowable clock drift as defined in Article 6.8 [of the CAT NMS Plan].”).

See Section IV.D.2.b(2)B, supra, for information on the Commission's clock offset tolerance analysis. Specifically, the analysis says that an order event can be sequenced if its time stamp is at least twice the clock offset tolerance from any other event on another venue.

Table 10—Sequencing Accuracy of Unrelated Events by Clock Offset Tolerance

Clock offset tolerance Percentage of unrelated events that can be sequenced
Equities (%) Options (%)
50 milliseconds 7.84 18.83
5 milliseconds 16.51 35.54
1 millisecond 22.08 50.70
100 microseconds 42.47 78.42

The Commission's analysis suggests that approximately 16.51% of unrelated order events for equities and 35.54% of unrelated order events for options could be sequenced under a clock offset tolerance of 5 milliseconds, 22.08% of orders events for equities and 50.70% of order events for options could be sequenced under a clock offset tolerance of 1 millisecond, and 42.47% of order events for equities and 78.42% of orders events for options could be sequenced under a clock offset tolerance of 100 microseconds. Given these results, the Commission believes that requiring a smaller clock offset tolerance than the Plan's proposed 50 milliseconds for some segments of the industry could improve the accuracy of event sequencing.

Relative to the Plan's proposed universal 50 millisecond clock offset tolerance, the Commission preliminarily believes that requiring a smaller clock offset tolerance for some segments of the industry would likely increase the costs of the CAT NMS Plan. Table 11 is from page C-126 of the CAT NMS Plan, and it provides the costs of the Plan's proposed clock offset tolerance (50 milliseconds) and alternative tolerances (100 microseconds, 5 milliseconds, and 1 millisecond). These costs assume that each clock offset tolerance is applied to all business systems. However, as noted above, the alternative the Commission is soliciting comment on is to require smaller clock offset tolerance for certain segments of the industry. So, the estimates below provide an upper bound on the potential cost if the Commission requires smaller clock offset tolerances in some cases.

Table 11 is from the CAT NMS Plan, supra note 3, at Appendix C, Section D.12(p) and it draws its numbers from the FIF Clock Offset Survey. See supra note 127.

Table 11—Implementation and Annual Ongoing Cost Estimates per Firm by Clock Offset Tolerance

Clock offset tolerance Estimated implementation cost (per firm) Estimated annual ongoing cost (per firm)
50 milliseconds $554,348 $313,043
5 milliseconds 887,500 482,609
1 millisecond 1,141,667 534,783
100 microseconds 1,550,000 783,333

The Commission understands that the cost figures in Table 11 do not net out the current ongoing costs of clock synchronization, which are $203,846. Table 12 shows the preliminary estimated annual ongoing cost increase (ongoing costs minus current costs) to comply with various alternative clock offset tolerances as well as the clock offset tolerance specified in the Plan.

See FIF Clock Offset Survey, supra note 127, at 16. This is based on current practice of the broker-dealers who responded to the survey.

Table 12—Annual Ongoing Cost Increases per Firm by Clock Offset Tolerance

Clock offset tolerance Estimated annual ongoing cost increases (per firm)
50 milliseconds $109,197
5 milliseconds 278,763
1 millisecond 330,937
100 microseconds 579,487

Based on these estimates, the Commission estimated aggregate clock synchronization costs for broker-dealers consistent with the estimation of their total CAT compliance costs as detailed in the Costs Section above. The Commission assumed that 171 broker-dealers would incur the full ongoing costs and full implementation costs indicated in the FIF Clock Offset Survey. Conversely, the remaining 1,629 broker-dealers that are already assumed to use service bureaus would rely on the 13 service bureaus to facilitate their clock synchronization, and therefore would pay lower implementation and ongoing costs than those in the FIF Clock Offset Survey. The Commission understands that broker-dealers that rely on service bureaus for order management systems and regulatory reporting usually use servers operated by their service bureaus and most would therefore not directly bear the costs to implement and comply with clock synchronization standards. For the implementation costs for those relying on service bureaus for clock synchronization, the Commission assumes 1/4 FTE for 50 milliseconds, 1/2 FTE for 5 milliseconds, 3/4 FTE for 1 millisecond, and 1 FTE for 100 microseconds. Under these assumptions, broker-dealers that outsource their order management and regulatory reporting obligations would incur costs (shown in Table 13) that are significant relative to the estimated implementation costs for broker-dealers that handle order management and reporting obligations in-house.

See Section IV.F.3.a, supra.

The 171 broker-dealers comes from the total of Insourcers, ELPs, and Options Market Makers.

See Section IV.F.1.d, supra for a discussion of service bureaus passing costs on to clients.

As in the Costs Section above (see Section IV.F.1.c(2)C), monetizing the FTE costs involves multiplying the number of FTEs by $424,350. See infra note 1487.

Table 13—Implementation Cost Estimates per Firm for Outsourcing Firms by Clock Offset Tolerance

Clock offset tolerance Estimated implementation costs (per firm) for outsourcing firms
50 milliseconds $106,000
5 milliseconds 212,000
1 millisecond 318,000
100 microseconds 424,000

With these implementation costs, the Commission aggregated implementation and ongoing costs as indicated in Table 14.

Table 14—Aggregated Implementation and Annual Ongoing Cost Estimates by Clock Offset Tolerance

Clock offset tolerance Estimated aggregate implementation cost Estimated aggregate annual ongoing cost
50 milliseconds $268 million $25 million.
5 milliseconds 497 million 63 million.
1 millisecond 714 million 75 million.
100 microseconds 956 million 131 million.

$268 million ≉ 171*$554,348 + 1,629*0.25*$424,350. $497 million ≉ 171*$887,500 + 1,629*0.5*$424,350. $714 million ≉ 171*$1,141,667 + 1,629*0.75*$424,350. $956 million ≉ 171*$1,550,000 + 1,629*$424,350.

$25 million ≉ 171*$109,197 + 13*4.2*$109,197. $63 million ≉ 171*$278,763 + 13*4.2*$278,763. $75 million ≉ 171*$330,937 + 13*4.2*$330,937. $131 million ≉ 171*$579,487 + 13*4.2*$579,487. 13 is the number of service bureaus and 4.2 is the ratio between the total incremental ongoing charges to broker-dealers and the total incremental ongoing costs to service bureaus derived from the cost estimates above. See Section IV.F.2, supra.

Table 14 suggests that the Plan's clock synchronization costs for the approximately 1,800 expected CAT Reporters would be approximately $268 million in estimated implementation costs and about $25 million in ongoing costs. To estimate the relative costs of each alternative compared to the Plan, the Commission subtracted the costs of the Plan from the costs of each alternative.

Table 15 provides estimates for how the costs of alternative clock offset tolerances applied to all business systems would be greater than those of the CAT NMS Plan if a different clock offset tolerance applied to all CAT Reporters.

Table 15—Aggregated Implementation and Annual Ongoing Cost Increases by Clock Offset Tolerance

Clock offset tolerance Estimated increase in implementation cost (aggregate) Estimated increase in annual ongoing cost (aggregate)
5 milliseconds $229 million $38 million.
1 millisecond 446 million 50 million.
100 microseconds 688 million 106 million.

The Commission recognizes that the benefits of clock synchronization of less than one millisecond are limited unless the time stamps are also more granular. Requiring more granular time stamps than the 1 millisecond in the Plan would increase the costs relative to those in Table 15.

The Commission does not have information on the implementation and ongoing costs to exchanges or ATSs of various alternative clock offset tolerances because trading venues were not included in the FIF Clock Offset Survey. The Plan does not provide this data, and the Commission has no other data from which it can independently estimate this, because the Commission is not aware of any such data available to it. However, exchanges may currently synchronize their clocks to within 100 microseconds. Consequently, the Commission preliminarily believes that any of the alternative clock offset tolerances discussed above would not materially increase costs to Participants relative to the costs they would incur under the Plan because their current clock synchronization procedures seem to satisfy any of the proposed clock offset tolerances. In the case of ATSs, these systems tend to be operated by large and complex broker-dealers that are unlikely to rely upon service bureaus to perform their clock synchronization responsibilities. Consequently, the Commission preliminarily believes that cost estimates for the broker-dealers surveyed by FIF are likely to include broker-dealers that operate ATSs and already reflect any additional clock synchronization costs attributable to operating ATSs. However, if Execution Venues (including ATSs) were to have smaller clock offset tolerances than other broker-dealer systems, broker-dealers operating ATSs would be expected to incur higher clock synchronization costs than other broker-dealers.

See Section IV.D.2.b(2)B.i, supra; see also supra notes 435 and 436.

As noted above, the Commission is soliciting comment on both an alternative that would consider the standard practices of different segments of the CAT Reporters for the purposes of setting the clock synchronization requirements, and an alternative that would define industry standard by looking at information other than current industry practice. The Commission preliminarily believes that if the CAT NMS Plan used an alternative interpretation of “industry standard” that considered the standard practices of different segments of the CAT Reporters for the purposes of setting the clock synchronization requirements, the cost increases associated with smaller clock offset tolerances might be lower than estimates presented in the tables above. In particular, if the clock synchronization requirements were only applied to the most time-sensitive systems, the costs increases would be lower than those presented. In addition, if the only broker-dealers required to comply with clock synchronization requirements were the ones accepting, routing, and executing orders, the costs could be lower than those presented above. The Commission does not have the information necessary to quantify how much lower the costs would be under an alternative that applied different clock offset tolerances to different segments of the CAT Reporters, because neither the Plan nor the FIF Clock Offset Survey break the cost estimates for changes in clock synchronization requirements down by business system types, and the Commission has no data from which it can independently estimate this, because the Commission is not aware of any such data available to it.

This belief is also consistent with information in the FIF Clock Offset Survey. See supra note 127, at 20. Specifically, the survey found that respondents would save on costs if the alternative clock offset tolerance were applied only to “server-side trading systems.”

The Commission recognizes that a clock offset tolerance smaller than 50 milliseconds would have differential cost across market participants. An alternate approach to defining “industry standard” that took into account the standard practices of different segments of CAT Reporters could mitigate those costs. All FIF Clock Offset Survey respondents that provided technology information use technology capable of 50 millisecond clock offset tolerances, but 36% of those respondents do not employ a technology capable of clock offset tolerances smaller than 50 milliseconds. Some survey respondents indicated that they employ software that is not capable of clock offset tolerances of less than 50 milliseconds or that desktop PCs would be a challenge with such clock offset tolerances. An alternative definition of “industry standard” that considered the practices of various segments of the industry could apply smaller clock offset tolerances to a subset of business systems; the Commission expects that applying smaller clock offset tolerances to a subset of systems would cost less than applying such clock offset tolerances to all systems. However, the benefits could also be limited in terms of the percentage of unrelated events that could potentially be sequenced, as compared to a definition of “industry standard” that a set a lower clock offset tolerance for all CAT Reporters.

(2) Alternative Logging Procedures

Rule 613(d)(1) requires synchronizing business clocks that are used for the purposes of recording the date and time of any Reportable Event. The CAT NMS Plan further requires that Participants and other CAT Reporters maintain a log recording the time of each clock synchronization that is performed and the result of such synchronization, specifically identifying any synchronization initiated in response to an observed discrepancy between the CAT Reporter's business clock and the time maintained by the NIST exceeding 50 milliseconds. According to the FIF Clock Offset Survey, costs in logging the synchronization events is a significant driver of overall clock synchronization costs.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(c).

Other cost drivers include hardware and software costs and costs in ensuring reliability.

A few survey respondents indicated that the number of logged events would go up significantly with a shorter clock offset, which requires a costly logging system. Therefore, the Commission is soliciting comment on an alternative that would require logging only exceptions to the clock offset (i.e., events in which a market participant checks the clock offset and applies changes to the clock). While logging every event, including clock offset checks, may be cost effective with longer clock synchronization tolerances, the Commission questions whether logging each event is cost efficient with finer clock offset tolerances, given the large number of events expected for the proposed and alternative clock synchronization standards. For example, if an investigation is relying on properly sequenced events, the investigation only would need to examine exception files to ensure the precision of the time stamps. The FIF Clock Offset Survey suggests that relaxing the logging requirement could reduce the burdens associated with clock synchronization.

See FIF Clock Offset Survey, supra note 127, at 19. One survey respondent noted that a log file for a one second clock offset would require 1 gigabyte of compressed storage each day but clock offset log files for 50 millisecond clock offset would increase the daily data storage 10 fold. Another survey respondent noted that its current system logs 86,000 events per day and that the proposed clock offset would require logging 35 million events per day; see also CAT NMS Plan, supra note 3, at Appendix C, Section A.3(c).

This is one of the alternatives suggested in the FIF Clock Offset Survey. See supra note 127.

The Commission cannot quantify the reduction in costs from this alternative because it lacks data on the proportion of clock synchronization costs that are associated with event logging and the proportion of those costs that could be avoided by alternative event logging requirements. The Commission preliminarily believes that any reduction in benefits from this alternative, as compared to the CAT NMS Plan's approach for clock synchronization, would be minor because the inclusion of clock synchronization checks that required no clock adjustment would not improve regulators' ability to sequence events. The Commission notes, however, that enforcement of clock synchronization requirements may be more difficult without comprehensive logging requirements that document firms' actions to comply with requirements; consequently, relaxing the logging requirement may also reduce incentives to comply with the clock synchronization requirements.

(3) Alternative Clock Synchronization Hours

The Commission is soliciting comment on alternative requirements for the times during which clock synchronization is required that would provide more flexibility than the requirements of the Plan. The clock synchronization requirement presented in the CAT NMS Plan makes no provision for reduced clock synchronization requirements at times during which systems are not performing tasks that produce time-sensitive CAT Reportable Events; in the FIF Clock Offset Survey, respondents identified that there were certain times during which maintaining clock synchronization is more costly. Survey respondents noted they would incur additional costs in maintaining clock offset “99.9% of the time” or with “100% reliability” and costs associated with managing “clock synch instability . . . after server reboot.” The Commission notes that maintaining 99.9% or 100% reliability may be unnecessary during times when the system does not record Reportable Events. Further, the Commission understands that generally a system does not record Reportable Events during server reboots. Therefore, the Commission preliminarily believes that an alternative that does not require synchronizing clocks when servers are not recording Reportable Events or when precise time stamps are not as important to sequencing, such as outside of normal trading hours, would not materially reduce benefits. Given the responses to the FIF Clock Offset Survey, the Commission preliminarily believes that this alternative could reduce costs because synchronization activities and log entries related to those events would not be as beneficial outside of normal trading hours. The Commission does not have information necessary to quantify the cost reduction because cost information available to the Commission is not broken down by time of day or server status.

b. Time Stamp Granularity

The Commission is soliciting comment on the benefits and costs of an alternative time stamp granularity requirement of less than one millisecond. Rule 613(d)(3) requires time stamp granularity consistent with industry standards and, as discussed above, the Plan requires time stamps that reflect industry standards and are at least to the millisecond. Furthermore, the Plan requires Participants to adopt rules requiring that CAT Reporters that use time stamps in increments finer than milliseconds use those finer increments when reporting to the Central Repository. As discussed in the Commission's analysis of alternative clock offset tolerance requirements, millisecond time stamps may be inadequate to allow sequencing of the majority of unrelated Reportable Events across markets. In addition, as discussed below, the Commission recognizes that the benefits of more granular time stamps would be limited unless the Plan were to require a clock offset tolerance far lower than is proposed in the Plan.

See Section IV.H.1.e, supra.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(c).

See Section IV.E.1.b(2)B, supra.

The Commission recognizes that regulators' ability to sequence events is dependent on both clock offset tolerance and time stamp granularity. If the Plan requires any or all CAT Reporters to implement clock offset tolerances of less than a millisecond, time stamps reported at the millisecond level would not capture the additional precision of the smaller clock offset tolerance and much of the benefits of this smaller clock offset requirement would be lost if time stamps were rounded or truncated due to a millisecond time stamp granularity requirement. The Commission notes that provisions in the Plan require that any Participant that utilizes time stamps in increments finer than the minimum required to be reported under the Plan utilize such increments in reporting data to the Central Repository. Also, the Commission notes that a sub-millisecond clock offset tolerance would not in itself require the reporting of sub-millisecond time stamps to the Central Repository.

See CAT NMS Plan, supra note 3, at Section 6.8(b).

A requirement for time stamps at resolutions finer than 1 millisecond would entail certain costs. Because some market participants already use time stamps at the sub-millisecond level and will be required to report this information under the Plan, such a requirement is unlikely to create significant additional costs for CAT Reporters. Furthermore, while some exchanges and broker-dealers are already required to report time stamps at the sub-millisecond level, implementation costs are likely to vary across CAT Reporters. The Plan does not provide data on the cost of requiring sub-millisecond time stamps, and the Commission has no other data from which it can independently estimate this, because the Commission is not aware of any such data currently available to it.

Requiring sub-millisecond time stamp reporting would bring certain benefits. However, the Commission preliminarily believes these benefits may be limited without requiring clock offset tolerances of less than one millisecond as well. For example, with a 50 millisecond clock offset tolerance, a time stamp can only pinpoint the time of an event to a 100 millisecond range. In this case, sub-millisecond time stamps provide little benefit to regulators attempting to determine the order of events occurring in venues with separate clocks. However, even with a 1 millisecond clock offset tolerance, a sub-millisecond time stamp granularity requirement could provide some benefit for regulators attempting to sequence events. For example, two events recorded at times 12:00:00.0001 and 12:00:00.0021 on different venues can be sequenced with a 1 millisecond clock offset, while if these time stamps were rounded or truncated to 12:00:00.000 and 12:00:00.002, they could not be sequenced with certainty, because it would be possible that both events occurred at 12:00:00.001. If the Plan were to require sub-millisecond clock offset tolerances, the additional benefits of this sub-millisecond clock offset tolerance would be significantly limited without time stamps that were similarly granular.

See Section IV.H.2.a(1), supra.

c. Error Rate

The Commission is soliciting comment on the benefits and costs of alternative maximum Error Rates. The Commission does not possess sufficient data to quantitatively assess the costs and benefits of an alternative to the maximum Error Rates specified in the CAT NMS Plan. However, the Commission is using information provided in the CAT NMS Plan to perform a qualitative assessment of the proposed maximum Error Rates.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.3(b).

The potential benefits from a lower maximum Error Rate than proposed in the CAT NMS Plan could be improved accuracy in the data, and a quicker retirement of OATS and other regulatory data reporting systems. However, the CAT NMS Plan states that errors would be de minimis by the morning of day T+5, therefore the improvement in accuracy does not seem to affect the data available to regulators starting on day T+5. Accordingly, the benefit of improved accuracy as a result of a lower maximum Error Rate comes primarily from regulatory use of the data prior to day T+5. While the Commission believes that most regulatory uses would involve data after day T+5, regulators also have essential needs for uncorrected data prior to day T+5. For example, as discussed in the Benefits Section, the availability of unprocessed data within three days of an event could improve the Commission's chances of preventing asset transfers from manipulation schemes. Therefore, a lower Error Rate in data available before day T+5 could, in certain regulatory contexts, be meaningful.

The Commission recognizes that a lower Error Rate could also lead to the same accuracy level as the proposed Error Rate, but more violations and consequences from those violations. This is likely to occur if the Error Rates in the Plan are lower than what every broker-dealer could reasonably obtain on the timeline; as a consequence, because broker-dealers are reporting the most accurate data they are currently able to report, a lower Error Rate cannot improve data quality, but it can produce additional costs in the form of penalties levied by the Plan Processor. However, as long as at least one broker dealer can reasonably obtain lower Error Rates than those in the Plan, a lower Error Rate would improve accuracy because the lower Error Rate would incentivize that broker-dealer to reduce its initial errors.

See id. at Appendix C, Section A.3(b), n.102.

See Section IV.E.3.d(3), supra.

Second, because OATS currently has a lower observed error rate than the CAT NMS Plan, a reduction in CAT Error Rates may accelerate the retirement of OATS because the SROs may find it advantageous to retain OATS until CAT Data is at least as accurate as OATS data. However, the CAT NMS Plan does not require a particular target Error Rate before OATS can be retired and the Plan does not estimate any cost savings associated with the retirement of OATS or other systems, beyond those resulting from the end of a period of costly duplicative reporting. Therefore, any acceleration in the retirement of OATS would not provide a direct benefit resulting from a lower Error Rate. Further, the error rates in OATS may not be comparable to the Error Rates in CAT Data because the algorithm that identifies errors in CAT Data is unlikely to be identical to the algorithm that identifies errors in OATS. In particular, the Plan requires some types of validation checks on CAT Data that OATS data does not go through. These additional validation checks will help to ensure the accuracy of information types not currently collected by OATS such as Customer Account Information, Firm Designated ID, and options information, or to ensure the accuracy of information necessary for the order lifecycle linking process. Consequently, the Commission cannot be sure of the specific CAT Error Rate that would accelerate retirement of OATS. In addition, the Commission does not have cost estimates for different maximum Error Rates because such information was not provided in the CAT NMS Plan.

See CAT NMS Plan, supra note 3, at Appendix C, Sections A.1(a)(iii) and A.3(a) and Appendix D, Section 7.2 for a discussion of the types of required validations of CAT Data.

While reducing error rates may have these potential benefits, the Commission recognizes that it would also come at a cost. In particular, reducing Error Rates could increase the implementation and ongoing costs incurred by CAT Reporters and the Central Repository as compared to costs estimated in the Plan, as filed. To achieve lower Error Rates, some CAT Reporters might have to run additional validation checks on their data before sending their data to the Central Repository. Such CAT Reporters would incur additional costs to code and test any additional validation checks prior to implementation. CAT Reporters might also have to monitor and adjust their validation checks to respond to Error Rate reports from the Central Repository, incurring additional ongoing costs. However, the CAT Reporters already achieving lower Error Rates might not require additional checks, adjustments, or monitoring. Additionally, the Commission preliminarily believes that costs incurred by CAT Reporters to reduce error rates prior to sending data to the Central Repository may ultimately result in lower costs associated with correcting errors after the data is sent. The Commission also notes that the costs incurred would depend in part on the format in which data is reported to the Central Repository, which has yet to be determined. If a solution is chosen that requires the reformatting of data, and this reformatting results in errors, then the costs could be higher. Conversely, a solution that does not require data reformatting could result in a lower Error Rate with lower costs to CAT Reporters.

Additionally, the Plan contains provisions that require the Plan Processor to monitor and address Error Rates. For example, the Plan Processor is required to notify each CAT Reporter that exceeds the maximum Error Rate, and provide the specific reporting requirements that they did not fully meet. Requiring a lower Error Rate could increase the costs of these provisions, as compared to the costs estimated in the Plan as filed, because more CAT Reporters would exceed the Error Rate at which penalties are levied by the Plan Processor.

d. Error Correction Timeline

The Commission is soliciting comment on an alternative error correction timeline to that proposed in the CAT NMS Plan. The CAT NMS Plan proposes a deadline of T+3 for submission of corrected data to the Central Repository. The CAT NMS Plan also discusses recommendations from FIF and SIFMA to impose a day T+5 deadline, which is the current standard for OATS. The Participants state in the CAT NMS Plan that they believe it is important to retain the day T+3 deadline in order to make data available to regulators as soon as possible.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.1(a)(iv).

Id.

Id.

The Commission is soliciting comment on whether the CAT NMS Plan should impose a day T+5 deadline rather than the day T+3 deadline. In comment letters submitted to the Participants, FIF and SIFMA maintain that the day T+3 deadline may not be feasible and would prove costly to market participants. The alternative of a day T+5 deadline could reduce the costs relative to the CAT NMS Plan for CAT Reporters. The Commission preliminarily believes that the delays in regulatory access from a day T+5 deadline would significantly reduce regulators' ability to conduct surveillance and slow the response to market events relative to the CAT NMS Plan. However, the Commission also believes that day T+5 error correction may reduce costs to industry relative to the CAT NMS Plan, although the Commission is unaware of any cost estimates that have been provided to date.

See Letter from Manisha Kimmel, Managing Director, FIF, to the Participants, dated November 19, 2014, available at http://www.catnmsplan.com/industryfeedback/p601972.pdf;; Industry Recommendations for the Creation of a Consolidated Audit Trail (CAT), SIFMA, March 28, 2013, available at http://www.catnmsplan.com/industryfeedback/p242319.pdf.

e. Funding Model

The mechanism by which CAT fees are allocated is important because it can potentially disadvantage particular business models. Although the Plan does not discuss the final details of the CAT funding model, it does provide some details, including a set of funding principles that the Participants have discussed with the Development Advisory Group. The Commission is soliciting comment on alternative mechanisms for allocating fees across Execution Venues and across Industry Members.

The CAT NMS Plan presents details regarding an allocation of costs between the Execution Venues and the other Industry Members (i.e., broker-dealers), but does not detail the proportions of fees to be borne by each group. Under the CAT NMS Plan, fees would be tiered by activity levels, with market participants within a given tier incurring a fixed fee. In the case of Execution Venues (exchanges and ATSs), market share of share volume would determine the tier of the Execution Venue. In the case of broker-dealers, fees would be allocated by message traffic. The Commission is cognizant that ATSs are operated by broker-dealers, complicating this division of fees between broker-dealers and Execution Venues. This is discussed further below.

For a discussion of the economic effect of the tiered structure, see IV.F.4.c, supra.

(1) Unified Funding Models

The Commission is soliciting comment on several unified funding models as alternatives to the Plan's bifurcated funding model. One of the alternative funding models the Commission is soliciting comment on is a unified funding model in which Central Repository costs are allocated across all market participants (including Execution Venues) by message traffic. The Commission expects that message traffic will be a primary cost driver for the Central Repository, because transactional volume (which is cited by the Plan as a primary cost driver for the Central repository) is highly correlated with message traffic. Consequently, assessing CAT costs on market participants by message traffic may have the benefit of aligning market participants' incentives with the Participants' stated goal of minimizing costs. However, the Commission is also aware that while a broker-dealer's choice of business model is likely to determine its level of message activity, the majority of an exchange's message traffic is passive receipt of quote updates. Because quotes must be updated on all exchanges when prices change, exchanges with low market share are likely to have more message traffic (incurring CAT fees) per executed transaction (generating revenue). Consequently, a model that charges exchanges for the passive receipt of messages from broker-dealers is likely to disadvantage the smaller exchanges relative to a model that charges for market share of executions.

Using MIDAS data, Commission staff analyzed the number of equity exchange proprietary feed messages and trades during the week of October 12, 2015. The message per trade ratio varied across exchanges from 38.46 to 987.17, with a median of 57.21.

Commission staff data analysis confirms this for the smallest exchanges. Except for the smallest exchanges, the trade to message ratios range from about 0.016 trades for every quote update to about 0.026 trades for every quote update and appear constant across market share levels. However, the smallest exchanges by market share have only about 0.001 trades for every quote update to about 0.009 trades for every quote update.

The Commission is also soliciting comment on an alternative approach to reporting market maker quotations on exchanges that could address this concern. In this approach, market makers (both equity and options) would not need to report their quotation updates. Exchanges (both equity and options) would report quotation sent times (as detailed in the Plan with regard to Options Market Makers and the Exemption Request ). Exchanges would not be assessed message traffic fees for these quotation updates; the broker-dealers who sent the quotes would be assessed for this message traffic. All other message traffic, regardless of which market participant initiated it, would be assessed fees associated with CAT using a common rate formula.

See Exemption Order, supra note 18, at 7-8.

The Commission is soliciting comment on this alternative for a number of reasons. First, it ties CAT costs to a primary driver of the magnitude of Central Repository costs: message traffic. Second, it substantially reduces the number of messages stored in the Central Repository. Third, it avoids disadvantaging smaller exchanges whose message traffic may be relatively large compared to their execution volume. Finally, this alternative avoids bifurcated fee approaches that may cause one Execution Venue to be relatively cheaper than another due to the manner in which CAT fees are assessed and may cause conflicts of interest for broker-dealers routing customer orders. However, this alternative assesses CAT fees based on messages rather than the revenue-generating activity of trades. This may provide market participants with incentives to change their business models to reduce CAT fees, which could lead to reduced quotation activity that could be detrimental to market liquidity levels. Furthermore, because the vast majority of message activity originates with broker-dealers, this approach necessarily shifts most of the ultimate CAT funding burden to broker-dealers.

See Section IV.F.1.a, supra, stating that transactional volume is a primary driver of the costs of the Central Repository. The Commission preliminarily believes that transactional volume is highly correlated with message traffic.

For example, if the CAT funding model were set to make ATS trades significantly more costly relative to exchange trades, the exchanges might benefit from increased market share because ATSs might be compelled to increase their access fees to offset the proportionately higher CAT charges that they would incur. In the extreme, some ATSs might cease operations or seek to register as exchanges. Most ATSs do not disseminate quotation information; exchanges are required to do so. Reorganizing an ATS as an exchange therefore involves significant changes to its business model. Consequently, the Commission believes it unlikely that many ATSs would register as exchanges to avoid proportionately higher CAT charges. If certain types of trades have lower costs when their trades execute on an ATS, their trading costs would increase if they are forced onto exchanges. If some trades would not happen in the absence of an ATS, this would drive down overall trading volumes (as opposed to a shift from ATS to exchange). Lower overall trading volumes would be considered welfare-reducing, as they indicate foregone gains from trade.

The Commission also is soliciting comment on a second alternative approach to CAT funding, a unified funding approach where the tiers in the funding model are based on market share of share volume. Under this approach, all market participants (both exchanges and broker-dealers) would qualify for a tier based on reported share volumes. Share volume would count equally toward the tier regardless of the Execution Venue selected by the broker-dealer originating the order. However, this approach does not align the costs of operating and maintaining the Central Repository, which would largely depend on message traffic, with the fees charged to market participants. Furthermore, it is possible that some Execution Venues could compete for order flow by not passing this fee on to their customers, generating the same limitations as discussed above for the funding model in the Plan.

See Section IV.F.4.c, supra.

A third alternative would be for the funding model to impose fees on every individual trade instead of imposing a fixed fee by tier. This approach has several benefits. First, the Commission preliminarily believes that implementation costs for this approach are likely to be lower than other alternatives because infrastructure already exists to levy fees on each trade (this is the mechanism by which Section 31 fees are levied). Second, it ties fees to the revenue-generating activity of trading, rather than quoting activity, which results in those more likely to afford high fees paying the higher fees. Quoting activity provides liquidity to the market, but often does not necessarily result in an execution that can bring revenue to the market participant placing the quote; consequently, levying CAT fees on trades avoids making a generally desirable activity (posting liquidity) more costly. Third, it avoids the problems that may accompany a bifurcated approach to CAT cost allocation. Because the fee is levied regardless of where the trade occurs, it limits incentives of market participants to route to exchanges to avoid message traffic fees within broker-dealers or to avoid exposing an order in multiple venues to try to find non-displayed liquidity. These liquidity-seeking activities might reduce a client's trading costs, but they also potentially incur message traffic fees, creating a conflict of interest for broker-dealers.

Under Section 31 of the Act, 15 U.S.C. 78ee, and Rule 31 thereunder, 17 CFR 240.31, SROs such as FINRA and the national securities exchanges must pay transaction fees to the SEC based on the volume of securities that are sold on their markets. These fees are designed to recover the costs incurred by the government, including the SEC, for supervising and regulating the securities markets and securities professionals. The SROs have adopted rules that require their broker-dealer members to pay a share of these fees. Broker-dealers, in turn, may impose fees on their customers that provide the funds to pay the fees owed to their SROs. See SEC, Section 31 Transaction Fees (September 25, 2013), available at http://www.sec.gov/answers/sec31.htm.

Some quoting behavior may be costly to the market, for example spoofing or layering. This analysis assumes that message traffic fees associated with this undesirable behavior would not be sufficient to reduce that behavior. If that assumption is false, funding models that assign fees to quotes have the additional benefit of reducing disruptive activity. The Commission preliminarily believes that the benefits of reducing disruptive quoting activity via levying fees on quotes would not justify the costs of reducing beneficial quoting activity through the same fees.

Assessing fees directly on trades entails certain costs as well. First, it does not provide incentives for market participants to limit their message traffic, which is a primary cost-driver for the Central Repository. Second, it does not provide the benefits of a tiered approach, which the CAT NMS Plan lists as including transparency, predictability and ease of calculation.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(v)(B).

(2) Bifurcated Funding Models

The Commission is also soliciting comment on alternatives to the funding model proposed in the CAT NMS Plan that would also be bifurcated. One alternative would be to allocate CAT costs to broker-dealers by market share of share volume while retaining the Plan's funding model for Execution Venues. A benefit of this alternative would be to avoid disincentives to liquidity provision operations, particularly for infrequently traded securities and high volatility securities. A disadvantage of this approach would be that it does not align the fees charged to a CAT Reporter with the costs those CAT Reporters impose on the Central Repository in terms of message traffic, potentially resulting in disproportionate charges to CAT Reporters because high message traffic broker-dealers would pay no more than low message-traffic broker-dealers with the same level of trading activity.

SROs currently fund their regulatory data collection through a number of mechanisms. The Commission notes that FINRA does not charge its members for OATS directly. Rather, it is funded from FINRA's regulatory budget, which is collected from its members through various membership fees. The options exchanges charge an Options Regulatory Fee (“ORF”), which is a pass-through exchange fee collected by OCC clearing members on behalf of the U.S. option exchanges. The stated purpose of the fee is to assist in offsetting exchange costs relating to the supervision and regulation of the options market (e.g., routine surveillance, investigations, and policy, rule-making, interpretive and enforcement activities). The fee was first adopted by CBOE in 2008. See Securities Exchange Act Release No. 58817 (October 20, 2008), 73 FR 63744 (October 27, 2008). Subsequently, PHLX (Securities Exchange Act Release No. 61133 (December 9, 2009), 74 FR 66715 (December 16, 2009), ISE (Securities Exchange Act Release No. 61154 (December 11, 2010, 74 FR 67278 (December 18, 2009)), BOX (See Securities Exchange Act Release No. 61388 (January 20, 2010), 75 FR 4431(January 27, 2010)), NYSEAmex (Securities Exchange Act Release No. 64400 (May 4, 2011), 76 FR 27114 (May 10, 2011), NYSE Arca (Securities Exchange Act Release No. 64399 (May 4, 2011), 76 FR 27114 (May 10, 2011), NASDAQ (Securities Exchange Act Release No. 66158 (January 13, 2012), 77 FR 3024 (January 20, 2012, C2 (Securities Exchange Act Release No. 67596 (August 6, 2012), 77 FR 47902 (August 10, 2012)), MIAX (Securities Exchange Act Release No. 68711 (January 23, 2013), 78 FR 6155 (January 29, 2013)), ISE Gemini (Securities Exchange Act Release No. 70200 (August 14, 2013), 78 FR 51242 (August 20, 2013)), and BATS (Securities Exchange Act Release No. 74214 (February 5, 2105), 80 FR 7665 (February 11, 2015)) also adopted an ORF. The OFR is currently assessed to customer orders at a rate of $0.0417 per U.S. exchange listed option contract. The ORF is assessed on all trades, both buys and sells. Further, FINRA charges fees for reporting to TRACE. Certain fees are based on the number of users and type of connection a firm has to the system, and others are based on size of the transaction. See FINRA Rule 7730.

The Commission is further soliciting comment on the alternative of requiring the CAT NMS Plan to treat ATSs only as broker-dealers for funding purposes, instead of treating ATSs as Execution Venues. Under this alternative, firms that operate ATSs would not be charged for both their ATS's market share of share volume (like an exchange) and its message traffic (as a broker-dealer). Instead, the firm operating the ATS would pay fees based on the ATS's message traffic as part of its operations as a broker-dealer, rather than as an Execution Venue as well, for fee purposes. As described in Section IV.F.4.d, the Commission preliminarily believes that under the current funding model in the CAT NMS Plan, the cost differentials that result might create incentives for broker-dealers to route order flow to minimize costs, creating a potential conflict of interest with broker-dealers' investor customers, who are likely to consider many facets of execution quality (such as price impact of a trade and probability of execution in a venue in which the order is exposed) in addition to any of these costs that are passed on to them. The Commission is aware that this alternative would, in effect, shift part of the Central Repository funding costs from broker-dealers to Execution Venues because volume transacted on ATSs would not be assessed a portion of the Execution Venue funding burden and this portion would instead be allocated to exchanges. Furthermore, the Commission is aware that it is possible that under this alternative approach, ATSs might pay less in fees than similarly situated exchanges, which could disadvantage exchanges relative to ATSs.

As explained in Section IV.F.4.c, supra, the Commission preliminarily believes that the bifurcated funding model proposed in the Plan results in differential CAT costs between Execution Venues because it would assess fees differently on exchanges and ATSs for two reasons. First, message traffic to and from an ATS would generate fee obligations on the broker-dealer that sponsors the ATS, while exchanges incur no message traffic fees. Second, broker-dealers that internalize off-exchange order flow, generating off-exchange transactions outside of ATSs, would face a differential funding model compared to ATSs and exchanges.

See CAT NMS Plan, supra note 3, at Article VIII.

The Commission is also soliciting comment on the alternative approach of not charging broker-dealers for message traffic to and from their ATSs while still assessing fees to ATSs as Execution Venues or exchange broker-dealers for their message traffic. Under this alternative, broker-dealers that operate ATSs would pay trading volume based fees on their ATSs volume in the same manner as exchanges' fees are assessed. However, the message traffic to and from the ATS would not be included in the message traffic used to calculate fees assessed to the broker-dealer that sponsors the ATS. The Commission preliminarily believes this alternative would help mitigate the broker-dealer routing incentives discussed above. The Commission is aware that because the volume executed on ATSs would be included in the portion of Central Repository funding assigned to Execution Venues, this funding approach would not shift part of the funding burden assigned to Execution Venues away from ATSs (and the broker-dealers that operate them) to exchanges as the previous alternative would.

The Commission preliminarily believes that either of these ATS-related funding alternative approaches would avoid disadvantaging ATSs relative to similarly situated exchanges, and would be less likely to result in the conflicts of interest in routing described above. Currently, the Commission lacks sufficient details on the fee structure to make this determination, because the fee structure has not yet been finalized.

The Commission is also soliciting comment on the alternative of excluding ATS volume from TRF volume for purposes of allocating fees across Execution Venues. Under this alternative, SROs that operate TRFs (currently only FINRA) would not pay Execution Venue fees for volume that originated from an ATS execution. This alternative would avoid the problem of double-counting ATS volume as share volume, which originates because each ATS trade is counted for fee-levying purposes as share volume associated with an ATS, then counted again as share volume when the trade is printed to a TRF. However, the Commission notes that other over the counter volume, such as occurs when orders are executed off-exchange against a broker-dealer's inventory, would be assessed share volume fees while the message traffic that resulted in this execution would also be subject to fees through the broker-dealers that had order events related to these transactions. This contrasts to executions that occur on exchanges, where the venue that facilitates the execution does not pay fees for message traffic that led to the execution.

The Commission is also soliciting comment on the alternative of not treating the Trade Reporting Facilities (“TRFs”) as FINRA Execution Venues. TRFs capture ATS share volume, which is already subject to fees allocated to Execution Venues, and non-ATS off-exchange share volume, which is subject to CAT fees allocated to broker-dealer message traffic. Consequently, under the approach in the Plan, the activity that generates a TRF trade report is already assessed CAT fees through the broker-dealers that facilitate the trade, or the ATSs that served as the Execution Venue. Under this alternative approach, FINRA would not pay any fees directly into the Central Repository, and broker-dealers would only incur fees directly levied on them by the Operating Committee, rather than also indirectly paying the TRF fees passed on to them by FINRA. If FINRA does not pay fees directly to the Central Repository, this could alter its incentives with respect to matters of cost voted on by the Operating Committee. However, it is possible that, since FINRA represents the viewpoints of its broker-dealer members, its incentives would be similar under either approach.

The CAT NMS Plan would allocate net profit or net loss from the operation of the CAT equally among the Participants, regardless of size, which could advantage small exchanges in the event of a profit and disadvantage small exchanges in the event of a loss. This could negatively impact competition if the cost differentials are significant enough to alter the set of services that some competitors offer. As an alternative, the Commission is soliciting comment on whether the profit or loss from operating CAT should be allocated across Participants by market share of share volume, consistent with how the CAT costs would be allocated under the Plan. The Commission preliminarily believes that this alternative would limit the possibility of extraordinary profits or losses from CAT resulting in a disproportionate advantage or disadvantage to exchanges with low trading volume.

Id.

Finally, the Commission is soliciting comment on requiring a strictly variable funding model, rather than the fixed-tiered model in the CAT NMS Plan. Under a variable funding model, each trade or message is subject to a fee, rather than a broker-dealer incurring a fixed fee that depends on that broker-dealer's volume tier. The Commission preliminarily believes that this alternative might increase administrative costs of the CAT NMS Plan as compared to an approach that uses the fixed-tiered funding model. However, the Commission also preliminarily believes that the fixed-tiered funding model can create incentives for market participants to change their behavior to avoid fees when their activity is near the boundary between two tier levels. The Commission preliminarily believes that a strictly variable funding model could reduce inefficiencies resulting from market participants changing their behavior to move into a lower fee tier.

For example, under a fixed-tiered funding approach, any broker-dealer with no more than 10,000 CAT Reportable Events in a given month might pay $100 in fees, even a broker-dealer reporting a single event. Under a strictly variable funding approach, every broker-dealer CAT message might be assessed one cent in fees. For a broker-dealer reporting 10,000 CAT Reportable Events in a given month, the same fee burden would be incurred, but a broker-dealer reporting a single CAT reportable event would pay only one cent.

See Section IV.F.3.b, supra.

f. Requiring Listing Exchange Symbology

The Commission is soliciting comment on an alternative to the CAT NMS Plan that would allow CAT Reporters to report using their existing symbologies, rather than listing exchange symbology. The Plan requires the Plan Processor maintain a complete symbology database, including the historical symbology. The CAT NMS Plan also requires CAT Reporters to report data using the listing exchange symbology format, which would be used in the display of linked data. The CAT NMS Plan also requires Participants to provide the Plan Processor with the issue symbol information, and validation of symbology would be part of data validation performed by the Plan Processor.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.1(a).

The Commission preliminarily believes that, in light of the proposed requirement for the Plan Processor to maintain a complete symbology database, the requirement that CAT Reporters report using listing exchange symbology may result in unnecessary costs to CAT Reporters. Therefore, the Commission preliminarily believes that the alternative of allowing CAT Reporters to use their existing symbologies for reporting purposes could significantly reduce the costs for exchanges and broker-dealers to report order events to the Central Repository, as compared to the approach in the CAT NMS as filed, without a significant impact on the expected benefits of the Plan or the costs to operate the Central Repository.

Currently, Execution Venues handle complex symbology in different fashions. Some common stocks, for example, have multiple classes of shares. Exactly specifying the issue to be traded involves identifying the ticker symbol and sometimes a share class. On some venues, the convention is that these security types are reported without a delimiter in the symbol; other venues use a delimiter, and delimiters can vary across venues. For example, assume a firm has a listing symbol of ABC, and has two classes of shares, A and B. An issue might be “ABC A” on one venue, “ABC_A” on another, and “ABCA” on a third. This can cause numerous problems for analyses that extend beyond a single trading venue, particularly if “ABCA” is the complete listing symbol for an unrelated security. As mentioned in the Benefits Section, the inclusion of the complete symbol history of a security and the requirement for queries, reports, and searches to automatically collect the appropriate data despite symbol changes promotes accurate query responses by ensuring the inclusion of order events that might have been excluded because of symbology differences and by excluding order events in unrelated securities. The Commission preliminarily believes that the CAT NMS Plan can achieve these benefits without requiring CAT Reporters to report using listing exchange symbology.

As discussed in the Costs Section, one potential cost driver to CAT Reporters is the need to process reports before submitting them to the Central Repository. If reports can contain drop copies from an order management system, CAT Reporters can aggregate their drop copies and send them without further processing the reports. If, on the other hand, CAT Reporters need to transform or add any fields to the report, those CAT Reporters would need to develop, test, and maintain code to run the transformation, and they would need to actually transform the data at least once a day. If CAT Reporters do not need to run this transformation at all, they could save money. The Commission preliminarily believes that the requirement to report in listing exchange symbology could be the only requirement that necessitates that CAT Reporters transform data before reporting it to the Central Repository. Therefore, the Commission preliminarily believes that eliminating this requirement could reduce costs relative to the CAT NMS Plan as filed.

See Section IV.F.3.a, supra.

See id.

Some broker-dealers may already have adequate computational resources to run the transformation, whether at once, in batches, or in real-time; others could have to invest in such resources—an investment that would be saved by eliminating the requirement to use listing exchange symbology. The degree of cost savings would depend on any requirements to transform the data prior to reporting, which depends on the allowable formats for transmission. The CAT NMS Plan does not specify the allowable formats or whether the Central Repository would require a fixed format. If the Technical Specifications require a fixed format, broker-dealers would most likely have to transform their data prior to reporting it to the Central Repository regardless of the requirement to use listing exchange symbology, and the listing exchange symbol requirement could add very little to the reporting costs. Therefore, the Commission recognizes significant uncertainty in the cost savings associated with this alternative.

Further, the Commission cannot estimate the degree to which eliminating this requirement could reduce costs as compared to those in the CAT NMS Plan as filed, because it lacks the data to do so. The Plan assumes the need to transform the data to match exchange symbologies and therefore does not separately itemize the cost for transformation as a separate step in the reporting process. The Commission has no data from which it can independently estimate the cost differential because it depends on information internal to each of a heterogeneous group of CAT Reporters (e.g., the symbologies their current systems use and whether those are readily transformed to match listing exchange symbologies), which information is not compiled or stored anywhere and to which the Commission therefore does not have ready access.

g. Data Accessibility Standards

The Commission is soliciting comment on alternative approaches to the manner in which the CAT NMS Plan provides data access to regulators. Section IV.E.1.c of the CAT NMS Plan summarizes the Central Repository's requirements to provide access to regulators. This access would include both an online targeted query tool and a user-defined direct query or bulk extract. The CAT NMS Plan also specifies minimum standards the Central Repository must meet, such as capacity to support 3,000 minimum regulatory users and minimum acceptable response times for queries of varying complexity and size. The CAT NMS Plan also requires that the Plan Processor provide an open API that allows use of regulator-supplied common analytic tools. As discussed above, the CAT NMS Plan could result in many improvements to regulatory activities such as surveillance, examinations, and enforcement, but these benefits may not be fully realized if access to data is cumbersome or inefficient. The Commission does not have information on the incremental benefits and costs of each aspect of regulator access as would be necessary to analyze specific alternatives to the many data access standards in the CAT NMS Plan.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.2(c).

Id. at Appendix C, Section A.1(b).

See Section IV.E.2, supra.

The Commission is generally soliciting comment on alternatives to each minimum data accessibility standard required in the CAT NMS Plan. With multiple standards that could each be adjusted in countless ways, the set of possibilities is infinite, which precludes their enumeration and discussion within this analysis. Instead, this Section discusses several examples and requests comment on alternative standards that might be adopted. Because query response time standards provide exact limits, the Commission uses those to illustrate how changing the standards could affect benefits and costs. The CAT NMS Plan requires query responses for various types of queries of 5 minutes, 10 minutes, 3 hours, and 24 hours, where the simplest queries involving scanning narrow sets of data would be required to return in 5 minutes and complex queries scanning multiple days of data and returning large datasets would be required to return within 24 hours.

The Commission notes that particularly large and complex data queries can take extensive computing resources. While the benefits of direct access to CAT Data depend on reasonably fast query responses, the Commission recognizes that faster query response times come at a cost. The Commission does not have detailed information on significant breakpoints in those costs to judge whether slightly longer response times than those in the Plan could significantly reduce the costs of developing, maintaining, and operating the Central Repository. For example, the Commission does not know whether a 48-hour response time on a query of 5 years of data is significantly less expensive than a 24 hour response time, but either maximum response time would provide a significant improvement in timeliness over current data. Likewise, the Commission does not know whether the response times could be faster without a significant increase in costs. The Commission recognizes that the detailed information on numerous other minimum standards regarding access to regulators is similarly unclear. Therefore, the Commission requests comment regarding all standards for regulatory access and whether technology creates natural breakpoints in costs such that a particular alternative could reduce the costs of the Plan without significantly reducing benefits or could increase benefits without significantly increasing costs.

h. Intake Capacity Levels

The Commission is soliciting comment on alternatives to the intake capacity level required in the CAT NMS Plan. The CAT NMS Plan requires that the Central Repository have an intake capacity of twice historical peak daily volume measured over the most recent six years and the ability to handle peaks beyond this Baseline level for short periods. In setting this requirement, the Participants could have selected any number of alternative intake capacity standards.

See CAT NMS Plan, supra note 3, at Appendix D, Section 1.1.

The Commission performed an analysis using MIDAS data and determined that, for equities, the daily message traffic volume would exceed two times the maximum daily message volume from the previous six years (2010 through 2015) with a probability of 0.033%, which amounts to the intake exceeding capacity levels about once every 81/3 years. Message volume measures all equity messages, including orders, order updates, executions and cancellations, from MIDAS exchange direct feeds, consolidated SIP feeds, and a small portion of the FINRA ATS feed.

The Commission collected daily message volume from MIDAS for six years (January 1, 2010 through November 19, 2015) and found that August 10, 2011 generated the highest message traffic with 8.6 billion messages. A Box-Cox transformation was applied to the data to fit it into a normal distribution. Using a probability density function to fit the transformed data into a normal distribution, the Commission found the probability that the daily message volume would exceed 17.2 billion (twice the maximum) messages is 0.033%. The MIDAS data used are all equity messages between 4 a.m. and 7 p.m. on trading days—including orders, order updates, executions, and cancellations—from exchange direct feeds, consolidated SIP feeds, and a small portion of the FINRA ATS feed. MIDAS does not receive messages before 4 a.m. and after 7 p.m. from its feed sources. The data is missing AMEX feeds from January 1, 2010 through October 4, 2010; however, on average AMEX messages represent only 0.26% of daily message volume from all feeds.

The Commission preliminarily believes that intake capacity level is likely to be a primary cost driver for the Central Repository. In selecting a standard, there is a trade-off between additional cost for constructing and operating the Central Repository and the risk that increased volume could exceed the Central Repository's capacity. If the capacity were exceeded, the Commission preliminarily believes that regulators' access to CAT Data could be significantly delayed. The Commission is cognizant that periods of heavy market activity are more likely to be periods with market events that would require regulatory investigation, so the risk that the Central Repository might not be able to provide timely access to data when it is most needed is concerning.

Transactional volume and the growth in transactional volume is likely a primary driver of the costs of the Central Repository. See Section IV.F.1.a, supra. The Commission believes that higher transactional volumes require higher intake capacity levels, higher storage capacity, and higher processing capacity.

The Commission is soliciting comment on requiring a different intake capacity level. Alternative intake capacity levels would result in costs and benefits that depend on the specific alternative capacity level and whether it is higher or lower than the proposed level. For an alternative with a lower intake capacity level, such as 1.5 times the historic peak capacity level, the cost of creating and operating the Central Repository might be lower, but the risk that the Central Repository would be unable to meet regulator's data needs would be higher than under the CAT NMS Plan, particularly following events similar to the Flash Crash and August 24th, which created both a high volume of trading records and a high demand for timely regulatory analysis.

An alternative with a higher required intake capacity level, such as 3 times the historic peak capacity level, would likely entail higher costs than the CAT NMS Plan, but higher intake capacity levels would reduce the risk of the Central Repository being unable to meet regulators' data needs and thus increase the benefits of the Plan.

The CAT NMS Plan does not provide sufficient information for the Commission to quantify the cost difference between alternative intake capacities and the intake capacity in the CAT NMS Plan and there are no analogous projects of this scope with publicly-available data from bidding or otherwise from which the Commission could extrapolate.

3. Alternatives to the Scope of Certain Specific Elements in the CAT NMS Plan

The Commission notes that Rule 613 sets forth the minimum elements the Commission believes are necessary for an effective consolidated audit trail. The Commission also notes that it adopted these elements after notice and comment, including analyzing comment letters submitted in response to the Rule 613 Proposing Release. Moreover, the Participants, pursuant to Rule 613, analyzed and proposed for inclusion in the CAT NMS Plan certain elements after consultation with their members, the Bidders and the DAG.

See Adopting Release, supra note 9.

See id.

See CAT NMS Plan, supra note 3.

While the Commission and the SROs have previously analyzed Rule 613, including the elements to be included in the CAT NMS Plan, the Commission now has the Plan, together with the cost and alternatives analysis provided by the Participants. The Commission has reviewed the Plan, including the cost estimates, and has performed its own economic analysis of the Plan. With the benefit of having reviewed and analyzed the Plan, the Commission believes that it is reasonable to solicit comment on alternatives to the scope of certain elements of the CAT NMS Plan because these alternatives could impact the cost and benefits of CAT, and given the passage of time, there may be market developments that could affect those costs and benefits that should be evaluated. These alternatives include: (1) Not requiring certain data fields that are currently required by the Plan; (2) requiring the Operating Committee to consider including more primary market transactions than it would otherwise be required to consider under the Plan; (3) removing from the Plan the OTC Equity Securities recording and reporting requirements; and (4) excluding certain Customer information periodic update requirements.

a. Data Fields

Rule 613 provides that the Plan must require the reporting of certain data fields. It also gives discretion to the Participants to require the reporting of data fields beyond the minimum set of fields mandated by Rule 613. The Commission is soliciting comment on whether there should be changes to the data fields that would be subject to CAT reporting. Specifically, the Commission is soliciting comment on whether any data fields that would be subject to CAT reporting under the Plan should be excluded.

Id.

The Commission is soliciting comment on whether any data fields that would be subject to CAT reporting under the Plan should be excluded. For example, Rule 613 required the Plan to include a unique customer identifier. As discussed further in Section IV.H.1 above the Commission granted the Participants an exemption from certain requirements in Rule 613 so that the Plan could include an approach whereby each broker-dealer would assign a unique Firm Designated ID to each trading account, which would be linked to a set of identifying information. The Commission preliminarily believes that this approach would reduce the costs of requiring the customer identifier as compared to the Rule 613 approach.

Using the Firm Designated ID and the other information identifying the Customer that would be reported to the Central Repository, the Plan Processor would then assign a unique Customer-ID to each Customer. Upon original receipt or origination of an order, broker-dealers would only be required to report the Firm Designated ID on each new order, rather than using the Customer-ID. See Exemption Order, supra note 18, at 14-15. Because the Plan Processor would still assign a Customer-ID to each Customer under the Customer Information Approach, the SROs are not requesting an exemption from Rule 613(j)(5).

See Section IV.H.1.b, supra.

As an alternative, the Commission could eliminate the requirement to report customer identifiers. In the Adopting Release, the Commission recognized that the implementation of the unique customer identifier requirement might be complex and costly, and that the reporting of a unique customer identifier would require SROs and their members to modify their systems to comply with the Rule's requirements. While the Commission preliminarily believes that eliminating the customer identifier would reduce certain costs to industry associated with the implementation and operation of CAT as compared to the Plan as filed, without providing any additional material information, the Commission preliminarily believes that such a change would limit the benefits of the Plan significantly. As the Commission noted in the Adopting Release for Rule 613, unique customer identifiers are vital to the effectiveness of the consolidated audit trail, and the inclusion of unique customer identifiers would greatly facilitate the identification of the orders and actions attributable to particular customers and thus substantially enhance the efficiency and effectiveness of the regulatory oversight provided by the SROs and the Commission. Further, without the inclusion of unique customer identifiers, many of the potential benefits of a consolidated audit trail would not be achievable.

See Adopting Release, supra note 9, at 45756.

Id.

The Commission could also consider the alternative of excluding the allocation time field from reporting requirements in the Allocation Reports. Although this field is not currently required for recordkeeping, some broker-dealers do already retain allocation time information at the subaccount level in their trade blotters, though the Commission does not have precise information on the prevalence of this practice. The Commission preliminarily believes that removing allocation time would significantly reduce the benefits of the Plan because regulators currently undergo significant difficulties to obtain allocation times and the allocation times would be useful for enforcement investigations. At the same time, given the uncertainty in the current practices and the lack of information on the costs of this field in the Plan, the Commission is not sure how significant the cost savings of excluding the allocation time field would be. The Commission preliminarily believes that the substantial benefits of having allocation time at the subaccount level available and relatively accessible for regulatory activities warrants the costs associated with requiring CAT Reporters to include this field in CAT Data and that these costs would be significantly mitigated to the extent that CAT Reporters already retain this information.

See Section IV.E.1.a, supra.

The Plan requires both the CAT-Reporter-ID for the broker-dealer routing an order and the CAT-Reporter-ID for the broker-dealer receiving a routed order to be reported to the Central Repository, both when the order is routed and again when the routed order is received. The Commission could eliminate the requirement to report the CAT-Reporter-IDs when the routed order is received. However, while the Commission preliminarily believes this might reduce the CAT Reporting burden on some broker-dealers as compared to the Plan as filed, without providing any additional material information, the Commission noted in the Adopting Release that it does not believe the information reported when the order is received would be duplicative. Instead, the Commission noted that information regarding when a broker-dealer received a routed order could prove useful in an investigation of allegations of best execution violations to see if, for example, there were delays in executing an order that could have been executed earlier. In addition, the Commission notes that if a market participant is required to report when it receives an order, regulators could solely rely on information gathered directly from that market participant when examining or investigating the market participant. The Commission also noted that it relies on such data to improve its understanding of how markets operate and evolve, including with respect to the development of new trading practices, the analysis and reconstruction of atypical or novel market events, and the implications of new market dynamics.

See Adopting Release, supra note 9, at 45763.

Id.

Id.

The Commission preliminarily believes that, with respect to the reporting of data fields required by Rule 613, the analysis in the Adopting Release is still applicable and the elimination of these data fields from the Plan would result in a failure to achieve many of the significant potential benefits of the Plan. However, as noted above, the costs or benefits of including particular fields in the Plan as filed, may have changed due to technological advances and/or changes in the nature of markets since Rule 613 was adopted. The Commission is therefore soliciting comment on the benefits and drawbacks of eliminating these and any other required data fields from the Plan.

b. Primary Market Transactions

The CAT NMS Plan does not require the reporting of any primary market information to the Central Repository. However, as required by Rule 613(i), the CAT NMS Plan commits to incorporating a discussion of how and when to implement the inclusion of some primary market information into a document outlining how additional Eligible Securities could be reported to the Central Repository (the “Discussion Document”), which would be jointly provided to the Commission within six months after effectiveness of the Plan. Additionally, as required by Rule 613(a)(1)(vi), the Plan includes a discussion of the feasibility, benefits and costs of including primary market transactions in the CAT NMS Plan.

See CAT NMS Plan, supra note 3, at Appendix C, Section C.9. Section 6.11 of the Plan satisfies a requirement in 17 CFR 242.613(i) to plan for expansion.

17 CFR 242.613(a)(1)(vi); CAT NMS Plan, supra note 3, at Appendix C, Section A.6.

In its discussion of primary market transactions, the CAT NMS Plan states that including some primary market allocation information in the CAT NMS Plan would provide significant benefits without unreasonable costs, while other allocation information would provide marginal benefits at significantly higher cost. Specifically, the discussion in the CAT NMS Plan divides the primary market allocation information into two categories: Top-account allocations and subaccount allocations. Top-account allocations refer to allocations during the book-building process to institutional clients and retail broker-dealers. These allocations are conditional and can fluctuate until the offering syndicate terminates. Top-account institutions and broker-dealers make the subsequent subaccount allocations to the actual accounts receiving the shares. The Plan concludes that, with respect to primary market information, only the subaccount allocations would provide significant benefits without unreasonable costs if they were to be incorporated into the CAT.

See id. at Appendix C, Section A.6(b)-(c).

Based on that discussion, the Plan states that “the Participants are supportive of considering the reporting of Primary Market Transactions, but only at the subaccount level, and would incorporate analysis of this requirement, including how and when to implement such a requirement, into their document outlining how additional Eligible Securities could be reported to the Central Repository, in accordance with SEC Rule 613(i) and Section 6.11 of the Plan.” The Plan therefore would limit the discussion of reporting primary market transactions in the Discussion Document to the subaccount level. As an alternative to the approach in the Plan, the Commission is soliciting comment on whether to broaden the required scope of the discussion of primary market allocation information in the Discussion Document to include an analysis of incorporating both top-account and subaccount information for primary market transactions into the CAT. The Commission preliminarily believes that the potential benefits of including top-account information in the CAT could be significant and that the costs of including top-account information could be lower than what is described in the CAT NMS Plan and appropriate in light of significant potential benefits. For these reasons, the Commission preliminarily believes that top-account information should not be excluded from the Discussion Document.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.6(c).

Some primary market information is currently available to regulators. FINRA collects primary market allocation information on the initial and final list of distribution participants in their Distribution Manager. Based on discussions with Participants, the Commission understands that issuers of IPOs are required to report primary market allocations to broker-dealers within the Distribution Manager, but reported information does not contain broker-dealer customer information on those allocations. Primary market allocations to market participants other than broker-dealers can be voluntarily reported to the system. FINRA uses this system in the course of investigations in response to complaints and in normal examinations of broker-dealers. The Commission can request data from the Distribution Manager. When the Commission or an SRO needs additional primary market information, they request it from underwriters and other broker-dealers in the offering process. These ad hoc data requests can take weeks for underwriters to process and, if requesting data from multiple underwriters or other broker-dealers, each could submit the data in a different format or with different data definitions, adding time to the process of combining the data across underwriters.

Primary market information currently assists regulators in examining underwriting practices and surveilling for violations of regulations regarding allocations in primary offerings. The information also is useful for conducting market analysis and research on policy issues such as allocation decisions, flipping, and secondary market price support and the analysis and reconstruction of market events such as the Facebook IPO or the Vonage IPO.

See Reena Aggarwal, Allocation of Initial Public Offering and Flipping Activity, 68(1) Journal of Financial Economics 111-135 (2003); Reena Aggarwal, Manju Puri and N. Prabhala, Institutional Allocation in Initial Public Offerings: Empirical Evidence 57 (3) Journal of Finance 1421-1442 (2002); Raymond P. Fishe, How Stock Flippers Affect IPO Pricing and Stabilization, Journal of Financial and Quantitative Analysis 319-339 (2002); and Raymond P. Fishe, Ekkehart Boehmer, Underwriter Short Covering in the IPO Aftermarket: A Clinical Study, Journal of Corporate Finance, 575-594 (2004). For background information on the Facebook IPO, see SEC Press Release, SEC Charges NASDAQ for Failures During Facebook IPO (May 29, 2013), available at http://www.sec.gov/News/PressRelease/Detail/PressRelease/1365171575032 . For background information on the Vonage IPO, see FINRA, FINRA Fines Citigroup Global Markets, UBS and Deutsche Bank $425,000, Orders Customer Restitution for Supervisory Failures in Vonage IPO (September 22, 2009), available at http://www.finra.org/newsroom/2009/finra-fines-citigroup-global-markets-ubs-and-deutsche-bank-425000-orders-customer .

The Commission preliminarily believes that including both top-account and subaccount allocation information for primary market transactions in CAT would make primary market information that identifies customers directly accessible to regulators, which would be beneficial. In particular, top-account information in addition to subaccount information would be necessary to surveil, without requesting data from underwriters, for prohibited activities in the book-building process and would improve the efficiency of investigations into such prohibited activities. For example, including top-account information in CAT Data would provide regulators efficient access to data relevant for investigations into tie-in arrangements because regulators would be able to correlate treatment in the primary offering with other trading activity to see if, for example, those who trade more in the aftermarket receive more of the initial public offering shares they request than others. Including such information in CAT Data would also provide efficient access to data that could identify potential allocations that preference some customers over others in the IPO allocation process because the SROs and Commission could examine the relationship between IPO initial allocations, initial indications of interest, and fluctuations in allocations and indications of interest during the book-building process. In the Adopting Release, the Commission noted several additional benefits of collecting top-account information in addition to subaccount information for primary market transactions. For example, examinations of “spinning,” “laddering,” and other “quid pro quo” arrangements would benefit from efficient access to such CAT Data, which would facilitate a comparison of those customers allocated shares in an offering to those who are not allocated shares in an offering and how the conditional allocations change during the book-building process. Book-building information, which is currently very difficult for regulators to assemble, would provide very useful insights into IPO and follow-on allocations in market analysis. Such insights would better inform rulemaking and other policy decisions.

The CAT NMS Plan estimates that for broker-dealers to implement a system to record and report top-account and subaccount allocation information for primary market transactions would take 36 months of staff time per firm at a cost of $234.8 million whereas just subaccount information would take 12 months of staff time per firm at a cost of $58.7 million. The inclusion of top-account allocation information accounts for the difference of $176.1 million. The CAT NMS Plan explains that including top-account information in the CAT would result in higher implementation costs because the top-account information is maintained in book-building systems in investment banking divisions of broker-dealers that differ fundamentally from secondary market systems.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.6(c). The estimated costs reflect the implementation cost of systems development needed to support top-account and subaccount information for primary market transactions to CAT. The $234.8 million figure assumes 36 months of staff time, with 21.741 days per month at a $1200 daily FTE rate for 250 firms. The $58.7 million figure assumes 9 months of staff time, with 21.741 days per month at a $1200 daily FTE rate for 250 firms. The estimates do not include any ongoing annual costs to maintain the reporting; the Commission assumes that these systems would be supported by staff already engaged to support CAT reporting.

Id. at Appendix C, Section A.6(a).

However, the Commission preliminarily believes that the costs of adding top-account allocation information may be lower than those estimated in the CAT NMS Plan, for several reasons. First, in combination with an alternative that would require less granular time stamps or a larger allowable clock offset on less time-sensitive systems, the costs for top-account information would be lower than indicated in the Plan. The Commission recognizes that the benefits from time stamp granularity and clock synchronization in the systems for reporting top-account information may be lower than those for secondary market systems because activity occurs far less frequently than it does on exchanges and regulators may not need to sequence primary market transactions relative to secondary market transactions within a second. The Commission is unable to estimate cost savings from alternative clock synchronization requirements because estimates presented in the Plan do not cite these specific costs. Second, the Plan's estimate is sensitive to the number of underwriters. In particular, the estimates assume 250 underwriters would need to implement changes to provide for top-account allocation information for primary market transactions. This is also the same number of underwriters assumed to need to implement subaccount allocation information. However, the Commission suspects that the number of underwriters that would need to implement changes for top-account information may be lower than the number that implement subaccount information for primary market transactions because the lead underwriters could have all of the information necessary to report the top-account information. If so, then only those underwriters that expect to lead an offering would need to implement systems changes to report top-account allocation information. Estimating costs only for lead underwriters could result in a much smaller estimate.

See Cost Estimate for Adding Primary Market Transactions into CAT (February 17, 2015), available at http://www.catnmsplan.com/industryfeedback/p602480.pdf .

The Commission does not have an estimate of the ongoing costs of underwriters reporting top-account information. However, the Commission preliminarily estimates an average of approximately 120 IPOs each year and 340 follow-on offerings each year from 2001 to 2014. Assuming each offering contains approximately 260 initial allocations, including all indications of interest, with 10 amendments from initial allocation to final allocation, each offering would generate 2,600 CAT Reportable Events for a total of 1.2 million per year. This total is much smaller than the number of Reportable Events in the secondary market (trillions). Therefore, while the Commission cannot estimate the costs of ongoing primary market reporting, the Commission believes the ongoing costs of reporting primary market transactions would be a fraction of the ongoing costs of secondary market reporting and would likely be supported by staff already engaged to maintain CAT reporting.

The Commission estimated the number of allocations per offering by averaging the data for the 11 IPOs made public along with an academic paper. See Jay R. Ritter and Donghang Zhang, Affiliated Mutual Funds and the Allocation of Initial Public Offerings, 86(2) Journal of Financial Economics 337-368 (2007) and http://bear.warrington.ufl.edu/ritter/Allocation08282012.xls . If the Commission assumes that each offering would generate 10 amendments to allocations prior to the subaccount allocations, there would be 2,600 reports per offering and 1.2 million reports per year using the number of offerings in 2014. If each offering instead generates 5 or 20 amendments, the number of reports per year would be 0.6 million or 2.4 million.

The Commission also recognizes that including top-account information in the CAT NMS Plan could change the competitive landscape of the market for underwriting services. In particular, some underwriters may choose to exit the market instead of report top-account information. The Commission preliminarily believes that the compliance costs themselves would be low compared to underwriting fees. Nonetheless, the Commission recognizes that some underwriters may exit rather than comply with the CAT NMS Plan requirements. Likewise, the Commission recognizes that the costs to implement CAT reporting of top-account allocation information could increase barriers to entry.

The primary market issued about $450 billion in common stock in 2014 and underwriters earned $5.2 billion in underwriting fees in 2014. This is high relative to the $176 million cost estimate above. The value of issuances comes from the Securities Data Corporation and information regarding the aggregate underwriting fees comes from FOCUS reports.

Finally, the Commission recognizes that requiring top-account information in the CAT NMS Plan could alter the way underwriters conduct their book-building activities. The Commission is not sure if these changes would be beneficial or harmful to issuers and investors. For example, issuers and investors could benefit if including top-account information in CAT deters book-building activity that violates Regulation M or FINRA Rule 5110, 5130 or 5131, though some particular investors may lose any gains from preferential treatment. However, the Commission is uncertain whether investors and issuers would benefit if underwriters altered their book-building activity in an effort to reduce their reporting burden. For example, if reporting every change to a conditional allocation proved cumbersome, underwriters may choose to update preliminary allocations less often. This could change the way that underwriters and investors interact with each other in the book-building process with implications for the potential success of the offering or investors' satisfaction with the outcome.

c. OTC Equity Securities

The CAT NMS Plan requires the reporting of data regarding OTC Equity Securities upon implementation of the CAT NMS Plan. The Commission is soliciting comment on the alternative of eliminating the requirement to report activity in OTC Equity Securities from the CAT NMS Plan, and instead requiring only that the SROs include a discussion of how OTC Equity Securities could be incorporated into the CAT in the Discussion Document that they are required to provide within six months after the effective date of the Plan pursuant to Rule 613(i). This was the approach taken with respect to OTC Equity Securities in Rule 613, because the Commission believed that limiting the scope of the CAT to NMS securities was a reasonable first step in implementing the CAT. Under this approach, the CAT NMS Plan would require each national securities exchange and national securities association, within six months after effectiveness of the national market system plan, to jointly provide to the Commission a document outlining in detail how OTC Equity Securities (along with certain other categories of securities) could be incorporated into the CAT information, including an implementation timeline and a cost estimate. The Commission preliminarily believes that excluding OTC Equity Securities from the CAT upon implementation would reduce costs of the CAT NMS Plan. But, the Commission also preliminarily believes that removing the requirement to report activity in OTC Equity Securities from the CAT NMS Plan would limit the regulatory benefits of the CAT NMS Plan significantly.

Id.; see also Adopting Release, supra note 9 at 45744. The Plan states that “[e]ven though SEC Rule 613 does not require reporting of OTC Equity Securities, the Participants have agreed to expand the reporting requirements to include OTC Equity Securities to facilitate the elimination of OATS.” See CAT NMS Plan, supra note 3, at Appendix C, Section C.9.

Under the alternative approach, OTC Equity Securities would be excluded from the Plan upon implementation. While they could still be incorporated into the Plan following the submission of the Discussion Document, the alternative approach would create uncertainty as to whether or not OTC Equity Securities would ultimately be incorporated into CAT NMS Plan and the timeline for that process.

Excluding OTC Equity Securities from the CAT NMS Plan could limit oversight of the OTC equity market relative to the oversight obtainable under the Plan. FINRA currently collects reports on OTC equity markets in its OATS data. The primary difference between OATS and CAT Data for OTC Equity Securities would be in completeness, due to the additional data fields in CAT Data that are not in OATS, particularly Customer-ID; in any accuracy improvements relative to OATS; in direct access for the Commission; and in the timeliness relative to OATS, particularly in having linked data that requires less time to process. Relative to the Plan, therefore, excluding OTC Equity Securities could reduce the efficiency and effectiveness of regulators overseeing the OTC market, conducting investigations of manipulation, pump and dumps, and improper penny stock sales. It could also reduce the efficiency of estimating disgorgement payments to harmed investors relative to the Plan.

The Commission has discussed the potential for fraudulent activity in the OTC market. See SEC, Microcap Fraud, available at http://www.sec.gov/spotlight/microcap-fraud.shtml.

See supra note 351 and related text.

The CAT NMS Plan states that including OTC Equity Securities could facilitate the retirement of OATS. If OTC Equity Securities are not included in the CAT NMS Plan upon implementation, including OTC Equity Securities at a later time would require an amendment to the CAT NMS Plan, which could take significant time and potentially delay the retirement of OATS. The Commission is cognizant that the period of duplicative reporting, during which both CAT and OATS would be reported by market participants, is likely to impose a significant cost on industry. The CAT NMS Plan states that the inclusion of OTC Equity Securities at CAT implementation is generally supported by industry to facilitate the retirement of OATS.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.1(a) n.16.

The Commission notes, however, that the incorporation of OTC Equity Securities is not the only hurdle needed to retire OATS, and other hurdles may remain open even after any approval of the CAT NMS Plan. For example, the Plan anticipates a period of 12-18 months during which the SROs would analyze rules and systems to determine which require duplicative information. The process and timeline for elimination of duplicative reporting systems is discussed in Section IV.F.2, supra.

See Section IV.F.2, supra.

See CAT NMS Plan, supra note 3, at Appendix C, Section C.9.

The Commission preliminarily believes that excluding OTC Equity Securities from the CAT upon implementation would reduce certain costs associated with implementation and operation of CAT as compared to the Plan as filed, without providing any additional material information, because less data would be reported, therefore requiring fewer resources to implement and maintain the CAT. The Commission further preliminarily believes that CAT Reporters and the Central Repository would avoid certain compliance costs if OTC equities were excluded. To the extent that market participants rely on separate IT infrastructure to handle activity in OTC as opposed to listed securities, delaying the inclusion of OTC Equity Securities in CAT postpones costs associated with updating these systems. Postponing these system modifications may allow these modifications to be more efficiently integrated into other planned system upgrades, reducing costs to industry. The Commission notes that, even under this alternative approach, market participants still may incur these costs eventually, because the approach contemplates that the CAT NMS Plan could be expanded to require the reporting of order events in OTC Equities following the submission of the Discussion Document.

For example, in February, 2016, the average daily number of trades in OTC securities is approximately 98,300, on an average of approximately 18,500 issues over that same period. While that volume of trades is not large, the number of distinct issues is.

Furthermore, the Commission preliminarily believes that the cost savings from delaying incorporating OTC Equity Securities in the CAT NMS Plan are likely to be lower than the increase in costs of duplicative reporting that result from a delay to OATS retirement. Any broker-dealers that trade both OTC Equity Securities and listed equity or option securities would have to comply with the Plan regardless of the inclusion of OTC equities, so the cost savings to these broker-dealers from the exclusion of OTC Equity Securities may not be significant. The Commission preliminarily believes that the number of broker-dealers that trade only OTC Equity Securities is small. Finally, the Commission expects that the duplicative reporting costs would be fairly significant and that extending the time until the retirement of OATS would be a significant additional cost.

The Commission cannot estimate the amount of the cost reduction from excluding OTC Equity Securities because it lacks the data to do so. The CAT NMS Plan presents data only on the aggregate costs of on-exchange and OTC equity reporting; it does not present data on the costs specifically attributable to OTC equity reporting. The Commission has no data from which it can independently estimate the cost differential because it depends on information internal to each CAT Reporter (e.g., how their systems would change for the alternative compared to the Plan), which is not compiled or stored anywhere, and to which the Commission therefore does not have ready access, and it depends on when OTC Equity Securities would otherwise be included and the status of OATS and other systems in the interim.

d. Periodic Updates to Customer Information

As noted above in Section IV.E.1.b(4), the Plan Processor is required to create a Customer-ID and map Firm Designated IDs to this Customer-ID so that records stored in the CAT Data link to the Customers. To facilitate this, the Plan requires CAT Reporters to submit an initial set of Customer information to the Central Repository and subsequent daily updates and changes to that Customer information. In addition to daily updates to reflect changes in Customer information required in Rule 613, the CAT NMS Plan also requires members to submit periodic full refreshes of all Customer information to the CAT. The Commission is soliciting comment on an alternative that would eliminate the requirement for periodic full refreshes.

See CAT NMS Plan, supra note 3, at Appendix C, Section A.1(a)(iii); Appendix D, Section 9.1.

See id., at Appendix C, Section A.1(a)(iii) n.33.

The CAT NMS Plan states that the purpose of these refreshes is to “ensure the completeness and accuracy of Customer information and associations.” Although the Commission believes that the Participants should ensure that customer information in the Central Repository is complete and accurate, the requirement for periodic full refreshes seems redundant if the initial list and daily updates are complete and accurate and would, therefore, provide no additional benefit. Further, not requiring these periodic refreshes could reduce the risk of a security breach of personally identifiable information. However, the Commission recognizes that periodic full refreshes of customer information could address any errors that are introduced in the daily update process, although the Commission preliminarily believes that such problems are likely to be quite rare. In addition, the Commission recognizes that not requiring the periodic full refreshes could reduce certain costs associated with implementation and operation of CAT as compared to the Plan as filed for CAT Reporters, although the Commission preliminarily believes that these cost reductions would be minor for two reasons. First, the quantity of data required to refresh the customer information table is very small compared to the size of market data files submitted regularly by most market participants. Second, because market participants would need to develop software and procedures to initially populate the customer information table, that software and procedure should be available to refresh the table periodically. Therefore, the Commission preliminarily believes that removing the requirements for periodic full refreshes of customer information could minimally reduce the cost of the Plan without materially reducing the benefits.

Id.

4. Alternatives to the CAT NMS Plan

The Commission is soliciting comment on the broad set of alternatives of modifying existing systems to reduce the data limitations described above instead of approving the CAT NMS Plan.

When it adopted Rule 613, the Commission noted that “the costs and benefits of creating a consolidated audit trail, and the consideration of specific costs as related to specific benefits, are more appropriately analyzed once the SROs narrow the expanded array of choices they have under the adopted Rule and develop a detailed NMS plan.” The Commission also noted that a “robust economic analysis of . . . the actual creation and implementation of a consolidated audit trail itself . . . requires information on the plan's detailed features (and their associated cost estimates) that will not be known until the SROs submit their NMS plan to the Commission for its consideration.” Accordingly, the Commission deferred its economic analysis of the actual creation, implementation, and maintenance of the CAT until after submission of an NMS plan.

See Adopting Release, supra note 9, at 45725-26.

Id. at 45726.

The Commission recognizes that approving the CAT NMS Plan is not the only available means of improving the completeness, accuracy, accessibility, and timeliness of the data used in regulatory activities. Alternatively, the Commission could mandate improvements to one or more existing data sources to address the data limitations noted in the Baseline Section. The Commission previously considered this set of alternatives when considering whether to adopt Rule 613. The Commission has now reviewed the CAT NMS Plan, including the cost estimates, and has performed its own economic analysis of the Plan. With the benefit of having reviewed and analyzed the Plan, the Commission is now soliciting comment on this set of alternatives.

Id. at 45739-41.

As an alternative to the CAT NMS Plan, the Commission could require modifications to OATS. In the Adopting Release, the Commission noted that it had received comments suggesting various ways that the OATS system could be modified to serve as the central repository for the consolidated audit trail. However, the Commission also noted that OATS would require significant modifications in order to provide the attributes that the Commission deems crucial to an effective audit trail. In particular, OATS excludes some exchange-based and other types of non-member activity; it does not collect market-making quotes submitted by registered market makers (in those stocks for which they are registered); it is not a central repository and therefore does not presently provide other regulators with ready access to a central database containing processed, reconciled, and linked orders, routes, and executions ready for query, analysis, or download; it does not presently collect options data; it does not afford regulators an opportunity to perform cross-product surveillance and monitoring; and it does not collect information on the identities of the customers of broker-dealers from whom an order is received.

Id.

Id. at 45741.

The Commission preliminarily believes that, as stated in the Adopting Release, the missing attributes identified above are crucial to improving the completeness, accuracy, accessibility, and timeliness of the data used in regulatory activities. Thus, any alternative to CAT based on OATS that does not address those deficiencies would limit the potential benefits of the alternative significantly. Given the modifications necessary, the Commission cannot estimate the potential cost savings, if any, from mandating an OATS-based approach as an alternative to the CAT NMS Plan, because the Commission does not have sufficient information to estimate the cost of modifying OATS to address some or all of these deficiencies, either separately or in combination. The Plan does not provide data on the cost of making each relevant modification to OATS, and the Commission has no other data from which it can independently estimate this, because the Commission is not aware of any such data currently available to it. The Commission notes, however, that Rule 613 provided flexibility to the SROs to propose an approach based on OATS and/or other existing data sources. Given that Rule 613 provided this flexibility to the SROs, the Commission preliminarily believes that the SROs could have utilized an OATS-based approach if that approach would have represented significant cost savings relative to the Plan's approach, and the SROs that operate those reporting systems had presented such a solution as a Bid. Furthermore, the Commission notes that an approach that modifies and expands OATS to satisfy the requirements of the CAT NMS Plan remains feasible under the current bidding process. The Commission seeks comment on the costs and benefits of requiring modifications to OATS as an alternative to the CAT NMS Plan.

Id. The Commission also notes that the current Plan could allow the Plan Processor to leverage some elements of the existing OATS infrastructure and/or other existing data sources in the implementation of the CAT.

Another alternative would be for the Commission to modify other data sources instead of, or in combination with, OATS. However, like OATS, all of the current data sources have limitations that would need to be addressed in order to provide the attributes that the Commission deems crucial to an effective audit trail. Furthermore, the Commission preliminarily believes that modifying any other single data source would be more costly than modifying OATS, which is currently the most comprehensive audit trail. While the Commission could require the modification of multiple data sources in combination, the Commission preliminarily believes that an alternative to the CAT NMS Plan that relied on multiple data sources, such as a combination of OATS, COATS, other SRO audit trail data and/or publicly available data, would eliminate the benefits associated with having a single, complete consolidated source from which regulators can access trade and order data, which the Commission considers to be very significant.

The limitations of the various data sources are discussed in Section IV.D, supra.

These benefits are discussed in Section IV.E, supra.

In summary, the Commission cannot estimate the potential cost savings, if any, from modifying one or more other data sources instead of, or in combination with, OATS, because the Commission does not have sufficient information to estimate the cost of modifying each of the currently available data sources to address their current limitations, separately or in combination. The Plan does not provide data on the cost of making each relevant modification to each current data source, and the Commission has no other data from which it can independently estimate this, because the Commission is not aware of any such data currently available to it.

However, the Commission preliminarily believes that mandating improvements to the completeness, accuracy, accessibility, and timeliness of current data sources without an NMS Plan that requires the consolidation of data and increased coverage across markets and broker-dealers would likely significantly limit the potential benefits, possibly without providing significant cost savings. The Commission seeks comment on the costs and benefits of modifying one or more currently available data sources, separately or in combination, as an alternative to the CAT NMS Plan.

5. Request for Comment on the Alternatives

a. Generally

383. Are there any other alternatives that the Plan should require? If so, please describe the alternative and the costs and benefits of the alternative relative to the Plan.

b. Alternatives to the Approaches Permitted by the Exemption Order

See also Sections III.B.5-III.B.9, supra, for additional requests for comment on the alternative Rule 613 approaches to the approaches the Exemption Order allowed to be included in the CAT NMS Plan.

384. Should the CAT NMS Plan require Options Market Makers to report their quotes to the Central Repository? Please explain. Do Commenters believe that the costs of the Rule 613 approach would be disproportionately borne by smaller broker-dealers? Why or why not? Please provide data supporting your position.

385. Should the Plan treat equity market makers the same as Options Market Makers for purposes of quotation reporting—i.e., equity market makers report only Quote Sent Time and exchanges to which the quote is routed report the other information? Why or why not? What are the relative costs and benefits of this alternative? Please provide cost estimates.

386. Should the Plan require an alternative approach to reporting market maker quotes on exchanges where both equity and Options Market Makers would not need to report their quotation updates, and instead the exchanges would report Quote Sent Times in their reports of receiving these quotation updates? Why or why not? How would such an alternative affect the costs of building and operating the Central Repository? How would such an alternative affect market-maker costs of implementing and continuing CAT reporting?

387. Should the CAT NMS Plan require that Allocation Reports provide sufficient information for the Central Repository to be able to link those allocations to order lifecycles? What are the costs and benefits of providing this information? Please explain and provide cost estimates.

388. How do broker-dealers currently track which customers should receive allocations from which set of orders and how do broker-dealers ensure that those orders receive the correct average price? Can these same systems provide a key that could accurately link the allocations to lifecycles in many-to-many allocations? Please explain.

389. Should the CAT NMS Plan require an alternative to the Customer Information Approach? If so, what alternative should the Commission require and what are the relative costs and benefits of the alternative? Please explain.

390. Should the CAT NMS Plan require an alternative approach to assigning CAT-Reporter-IDs? If so, what alternative should the Commission require and what are the relative costs and benefits of the alternative? Please explain.

391. Should the CAT NMS Plan provide for the use of the LEI or another unique identification code as an alternative to the CAT-Reporter-ID? What are the advantages and disadvantages of this approach?

392. Should the CAT NMS Plan require an alternative to the requirement to time stamp manual orders to the second? If so, what alternative should the Commission require? For example, should the Plan require millisecond time stamps or one-minute time stamps? Please explain and provide information on the relative costs and benefits of the alternatives.

c. Alternatives to Certain Specific Approaches in the CAT NMS Plan

See also Sections III.B.2, III.B.4, III.B.10, III.B.11, supra, for additional requests for comment related to alternatives to certain specific approaches in the CAT NMS Plan.

393. Should the “industry standard” for the purposes of the clock synchronization and time stamping be “one-size-fits-all”? Please explain. If not, how should the CAT NMS Plan structure variations in clock synchronization and time stamp requirements that are based on industry practices?

394. Should the “industry standard” for the purposes of the clock synchronization and time stamping requirements be defined based on industry practice? Please explain. If not, how should “industry standard” be defined? Should the “industry standard” consider information other than current industry practice, such as the most accurate technology currently available in the industry, or the standard recommended by a particular industry group or authority? Could a definition of “industry standard” set a maximum clock offset threshold with an expectation that each CAT Reporter would be responsible for smaller clock offsets if the CAT Reporter is technically capable of such clock offsets? Please explain and include information on the relative costs and benefits of such alternative definitions.

395. What benefits, if any, would derive from applying the same uniform clock synchronization standards to all market participants versus applying different standards to different participant types? Which approach is preferable? If applying different standards to different participant types, which participant types should have smaller clock offset tolerances and which should have larger clock offset tolerances and what are the industry standards for those participant types? Please explain and provide any supporting data.

396. Do Commenters agree with the Commission's cost estimates for clock synchronization alternatives? Are there CAT Reporters other than broker-dealers that would incur significant costs from increasing clock synchronization standards to allowable clock drifts of less than 50 milliseconds, such as 1 millisecond or 100 microseconds? At what level of clock synchronization would these costs become material? Please explain. Do Commenters have estimates of these costs?

397. Does the FIF Clock Offset Survey reflect the operational capabilities of all potential CAT Reporters? Please explain.

398. Do Commenters agree that an alternative that would relax the logging requirements such that CAT Reporters would only need to log exceptions and resulting synchronization events (and not every synchronization event) would reduce costs of the CAT NMS Plan without materially reducing its benefits? Why or why not? Do Commenters have an estimate of how much such an alternative would reduce costs, either in isolation or in combination with the alternative to not require synchronization outside of event recording times? Please provide supporting documentation for these estimates.

399. Is there a need for clock synchronization standards outside of regular and extended trading hours? Is clock synchronization beneficial for retail orders that come in overnight? Are there examples of times or events outside of regular and extended trading hours when clock synchronization is more beneficial? Do Commenters agree that an alternative that would not require synchronizing clocks outside of times when servers record Reportable Events would reduce costs of the Plan without materially reducing its benefits? Do Commenters have an estimate of how much such an alternative would reduce costs? Please explain and provide supporting documentation if possible.

400. Are some CAT Reportable Events more time-sensitive than other events? If so, what events are more time-sensitive and why? What systems are more likely to process these events, and where are those systems located (i.e., within broker-dealers, service bureaus, Execution Venues)? Please explain.

401. What market participant systems, if any, should have smaller clock offset tolerances? Why? What clock synchronization standard should these systems have? Why? What market participant systems, if any, should have smaller clock offset tolerances? Why? What clock offset tolerances should these systems have? Why?

402. Should the Plan require time stamps to be reported more granularly than the one millisecond required in the Plan? If so, what standard should be required? Do Commenters agree with the Commission's analysis of the costs and benefits of requiring finer time stamp resolution than 1 millisecond? Please explain.

403. Should the CAT NMS Plan require different Error Rates in CAT? For example, should the Plan require a lower initial Error Rate? If so, what initial Error Rate should the Plan require and why? What would be the costs and benefits of requiring a lower initial Error Rate? Should the Plan require a lower Error Rate at some time period after implementation? If so, what Error Rate should the Plan require and why and when? What would be the costs and benefits of requiring a lower Error Rate?

404. Should the CAT NMS Plan require a day T+5 error correction deadline instead of a day T+3 error correction deadline? What are the relative costs and benefits of different error correction deadlines? Please explain and provide cost estimates.

405. Should the CAT NMS Plan require an alternative to the funding model in which broker-dealers and Execution Venues pay fees on the same fee schedule? If so, how would that funding model be structured and what metric would determine the fee level? How would that funding model affect the costs and benefits of the Plan, including the effect on competition? Please explain.

406. The Plan cites “transactional volume” as a cost driver for the Central Repository, but uses “message traffic” to allocate Central Repository costs across Industry Members. Do Commenters agree with the Commission's assumption that these two metrics are highly correlated? Is one of these metrics preferable for allocating costs across Industry Members? Please explain.

407. Should the CAT NMS Plan require alternative metrics to the message traffic and market share metrics required by the Plan for determining the tiers of the funding model but still place Execution Venues on a different fee schedule than broker-dealers? If so, which metrics? How would these alternative metrics affect the costs and benefits of the Plan, including effects on competition? Could these alternative metrics create conflicts of interest? Please explain.

408. Do Commenters agree with the Commission's analysis of unified versus bifurcated funding models? Why or why not?

409. Should the Plan require a unified funding model wherein Central Repository costs are allocated across all market participants by message traffic? Why or why not?

410. Should the Plan require a unified funding model wherein the tiers of the funding model for all CAT Reporters would be based on market share of share volume? Why or why not?

411. Should the Plan require a unified funding model wherein a fixed fee is levied on every trade? Why or why not? Could such a funding model reduce implementation costs by utilizing infrastructure already in place to assess Section 31 fees?

412. Should the Plan require a bifurcated funding model wherein Central Repository costs are allocated across broker-dealers by market share of share volume? Why or why not?

413. Should the Plan require a bifurcated funding model wherein Central Repository costs treat ATSs as part of broker-dealers only, instead of including them as Execution Venues? Why or why not?

414. Should the Plan require a bifurcated funding model wherein broker-dealer message traffic to and from an ATS are not included in message traffic measures used to assess fees on broker-dealers? Why or why not?

415. Should the Plan require a bifurcated funding model wherein ATS volume is excluded from TRF volume for the purposes of assessing Execution Venue fees to operators of TRFs? Why or why not?

416. Should the Plan require a bifurcated funding model wherein TRFs are not counted as Execution Venues for purposes of assessing fees on Execution Venues? Why or why not?

417. Should the Plan require that profits or losses from operating the Central Repository be allocated across Participants by market share of share volume? Why or why not?

418. Should the Plan require a strictly variable, rather than tiered, funding model? Why or why not?

419. Should the CAT NMS Plan require any funding model alternatives that could result in ATSs and exchanges paying equivalent fees? If so, how should that funding model be structured and what metrics should determine the funding tiers? How would that funding model affect the costs and benefits of this alternative, including effects on competition? Could these alternatives create conflicts of interest and, if so, to what extent? Please explain.

420. How should the CAT NMS Plan distribute the profits and losses of the Company among Participants? What are the relative costs and benefits of alternative ways to divide the profits and losses among the Participants? Please explain.

421. Should the CAT NMS Plan require a strictly variable funding model in which the fees paid are a set percentage of message traffic or share volume instead of a tiered funding model in which fees are fixed for a tier that is determined by message traffic or market share of share volume? If so, how would that funding model be structured? What are the relative costs and benefits of that funding model, including the effect on competition? Please explain.

422. Should the CAT NMS Plan exclude the requirement to report listing exchange symbology and instead allow CAT Reporters to use existing symbologies? Please explain. Would excluding this requirement allow broker-dealers to report data to CAT without processing the data ahead of the report? Please explain. What would be the relative costs and benefits of removing this requirement from the Plan? Please provide any cost estimates.

423. Should the CAT NMS Plan require alternative minimum standards for access to the CAT Data to those proposed in the CAT NMS Plan? If so, what alternative minimum standards should the Commission require? For example, should the response time on the largest queries be longer or shorter than 24 hours? How would changes to the alternative minimum standards affect the costs and benefits of the Plan? Please be specific and provide cost estimates.

424. Should the CAT NMS Plan require an intake capacity level different from twice historical peak daily volume measured over the most recent six years? If so, what intake capacity level should the Plan require? What are the relative costs and benefits of this alternative intake capacity level?

425. The Plan proposes using a “daisy chain” approach for linking order events within the Central Repository. This approach was chosen in favor of an approach that would require a unique order ID to be assigned by the first market participant that receives an order, and that order ID to be passed to and used by any market participant that handles the order afterward (the “unique order ID” approach). Do Commenters believe that a unique order ID approach or any other alternative approach would produce more accurate linkages than a daisy chain approach or any other benefits? Please explain. According to the Plan, the daisy chain approach would minimize impact on existing OATS reporters because OATS already uses this type of linkage. Do Commenters believe that a unique order ID approach or any other alternative approach would increase the costs for CAT Reporters who currently report to OATS or have any other effect on the costs of the Plan? Please explain and provide estimates. Given that the Bids from potential Plan Processors all utilize the “daisy chain” approach, would adopting a unique order ID approach at this stage cause a significant disruption in the progress toward the implementation of a consolidated audit trail? Please explain. What would the costs of such a disruption be?

See CAT NMS Plan, supra note 3, at Appendix C, Section A.1(b).

Id.

426. The CAT NMS Plan requires that the Plan Processor make use of a commercially available file management tool. What are the benefits to CAT Reporters from this requirement? Does this requirement have any effects on the competition between bidders? For example, are any bidders, such as those that could more efficiently use a proprietary file management tool, disadvantaged by this requirement? Please explain. Does this requirement affect the ability of the Operating Committee to replace an under-performing Plan Processor? Are there other costs or benefits of this requirement? Please explain.

d. Alternatives to the Scope of Certain Specific Approaches in the CAT NMS Plan

427. Should the CAT NMS Plan require excluding any data fields currently required to be included in the CAT Data (e.g., unique customer identification, allocation time, and CAT-Reporter-IDs at both order routing and receipt)? If so, which ones? Please explain and provide information on the relative costs and benefits of excluding those data fields, including any cost estimates.

428. Should the CAT NMS Plan exclude primary market information? Why or why not?

429. Do Commenters agree with the analysis in the Plan of the feasibility, benefits, and costs of the inclusion of primary market information (including primary market transactions) in the CAT NMS Plan? Please explain.

430. Do Commenters have additional analysis relevant to the decision to include primary market information (including primary market transactions) in the CAT NMS Plan? If so, please describe that analysis, including any data.

431. Do Commenters agree with the Plan's decision to include subaccount allocation information for primary market transactions in the Discussion Document, which commits the Operating Committee to consider the implementation of this subaccount allocation information in the CAT NMS Plan? Please explain.

432. Do Commenters agree with the Commission's assessment of the costs and benefits of requiring top-account allocation information for primary market transactions? Please explain. Should the Operating Committee consider requiring top-account information? Please explain.

433. What are the implications of the SROs decision not to include top-account information for primary market transactions in the Discussion Document? Please explain.

434. Should the CAT NMS Plan exclude OTC Equity Securities? Please explain. Would the exclusion of OTC Equity Securities in the CAT NMS Plan delay the retirement of OATS? If so, by how long and what would be the added cost be? Please provide an estimate. What are the other costs and benefits of excluding OTC Equity Securities from the CAT NMS Plan?

435. The CAT NMS Plan requires that CAT Reporters provide periodic refreshes of all customer information to the Central Repository to maintain an accurate database of customer information. What intervals for updates would be appropriate and reasonable, and what information should be required to be updated? Should the CAT NMS Plan remove the requirement for periodic full submission of customer information beyond the daily updates sent when customer information changes? Please explain. Would broker-dealers reduce their costs if they did not have to report all customer information periodically? Would the removal of this requirement significantly reduce the risk of a security breach of personally identifiable information? Please explain.

e. Alternatives to the CAT NMS Plan

436. Do Commenters agree with the Commission's analysis of the broad alternatives to approving the CAT NMS Plan, such as modifying OATS and/or other data sources to meet the objectives of Rule 613? Please explain. Are there other alternative approaches that the Commission has not identified that it should consider? Please explain.

f. Alternatives Discussed in the CAT NMS Plan

The Commission recognizes that the Plan discusses many alternatives that the Commission does not analyze above, including alternatives in Consideration 12 therein. This Consideration (Rule 613(a)(1)(xii)) requires the Participants to discuss in the Plan any reasonable alternative approaches that the plan sponsors considered in developing the Plan, including a description of any such alternative approach; the relative advantages and disadvantages of each such alternative, including an assessment of the alternative's costs and benefits; and the basis upon which the plan sponsors selected the approach reflected in the CAT NMS Plan. Such discussions appear in Section 12 of Appendix C. The Commission reviewed these alternatives and has not included above a discussion of all of the specific alternatives addressed in the Plan. In some cases, the Commission, at this time, has no analysis to add beyond the analysis in the Plan. In other cases, the Plan does not require any specific alternative, so the Commission cannot analyze the effect on the Plan of selecting a different alternative. The Commission is soliciting comment on the alternatives discussed by the Participants in the Plan but not discussed above. The Commission requests comment on each of these alternatives, both in isolation and in combination, as well as any data that would assist the Commission in evaluating the costs and trade-offs associated with these alternatives.

437. Organizational Structure. According to the CAT NMS Plan, the Participants considered various organizational structures of the Bidders. The CAT NMS Plan notes that the Bidders have three general organizational structures: (1) Consortiums or partnerships (i.e., the Plan Processor would consist of more than one unaffiliated entity that would operate the CAT), (2) single firms (i.e., one entity would be the Plan Processor and that entity would operate the CAT as part of its other ongoing business operations), and (3) dedicated legal entities (i.e., Plan operations would be conducted in a separate legal entity that would perform no other business activities). The CAT NMS Plan notes that each type of organizational structure has strengths and weaknesses but does not discuss those strengths and weaknesses. The CAT NMS Plan concludes that the organizational structure should not be a material factor in selecting a bidder and does not mandate any specific organizational structure for the Plan Processor. The Commission requests comment on whether the CAT NMS Plan should mandate a particular organizational structure. Why or why not? How can the organizational structure of the Plan Processor affect the costs and benefits of the CAT NMS Plan? What are the relative strengths and weaknesses of the different organizational structures?

See CAT NMS Plan, supra note 3, at Appendix C, Section D.12(b).

Id.

438. Primary Storage. The CAT NMS Plan states that bidders proposed two methods of primary data storage: traditionally-hosted storage architecture and infrastructure-as-a-service. The CAT NMS Plan does not mandate a specific method for primary storage, but does indicate that the storage solution would meet the security, reliability, and accessibility requirements for the CAT, including storage of PII data, separately. The CAT NMS Plan also indicates several considerations in the selection of a storage solution including maturity, cost, complexity, and reliability of the storage method. The Commission requests comment on whether the CAT NMS Plan should mandate a particular data storage method. Why or why not? How can the storage method affect the costs and benefits of the Plan? What are the relative strengths and weaknesses of the different primary storage methods?

See CAT NMS Plan, supra note 3, at Appendix C, Section D.12(c). Traditionally-hosted storage architecture is a model in which an organization would purchase and maintain proprietary servers and other hardware to store CAT Data. Infrastructure-as-a-service is a provisioning model in which an organization outsources the equipment used to support operations, including storage, hardware, servers, and networking components, to a third party who charges for the service on a usage basis.

439. Personally Identifiable Information. The CAT NMS Plan discusses several requirements to reduce the risk of misuse of PII, such as multi-factor authentication and Role Based Access Control for access to PII; separation of PII from other CAT Data; restricted access to PII; and an auditable record of all access to PII data contained in the Central Repository. The CAT NMS Plan notes that all bidders proposed some of these requirements, but only some bidders proposed others. The Commission requests comment on whether the Plan should mandate any/all of these requirements. The Commission further requests comment on the alternatives to these requirements. What are the potential alternative ways to protect PII? What are the costs and benefits of those alternatives compared to the Plan? Please provide estimates or other data to support answers.

See CAT NMS Plan, supra note 3, at Appendix C, Section D.12(e). Multifactor authentication is a mechanism that requires the user to provide more than one factor (e.g., biometrics/personal information in addition to a password) in order to be validated by the system. Id.

See CAT NMS Plan, supra note 3, at Appendix C, Section D.12(e). Role Based Access Control (“RBAC”) is a mechanism for authentication in which users are assigned to one or many roles, and each role is assigned a defined set of permissions. Id.

See CAT NMS Plan, supra note 3, at Appendix C, Section D.12(e). Appendix D provides additional discussion of these PII requirements. See id. at Appendix D, Section 4.1-4.2.

440. Data Ingestion Format. The Plan discusses the trade-offs between requiring that the CAT Reporters report data to CAT in a uniform defined format or in existing messaging protocols. The Plan does not require either method. A uniform defined format would include the current process for reporting data to OATS. This is Approach 2 in the CAT Reporters Study. Several bidders proposed to leverage the OATS format and enhance it to meet the requirements of Rule 613. The Plan states that this could reduce the burden on certain CAT Reporters (i.e., current OATS Reporters) and simplify the process for those CAT Reporters to implement the CAT. Accepting existing messaging protocols would allow CAT Reporters to submit copies of their order handling messages that are typically used across the order lifecycle and within order management processes, such as FIX. This is Approach 1 in the CAT Reporters' Survey. The Plan states that using existing messaging protocols could result in quicker implementation times and simplify data aggregation. The Plan further notes that the surveys revealed no cost difference between the two approaches, but that FIF members prefer using the FIX protocol. Should the Plan specify a particular approach? Please explain.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(A)(2); Section D.12(f). These are also called “Approach 1” and “Approach 2” in the Costs Section herein.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(A)(2).

Id. at Appendix C, Section D.12(f).

Id. at Appendix C, Section B.7(b)(i)(A)(2).

Id. at Appendix C, Section D.12(f).

Id.

Id.

441. The Commission requests further information on the relative costs and benefits and strengths and weaknesses of these two data ingestion format approaches. Would either of these approaches produce more accurate data? For example, would using existing messaging protocols such as FIX be more accurate because CAT Reporters would send their messages without the possibility of adding errors when translating them to a different format? Alternatively, would using existing messaging protocols such as FIX be less accurate because the Central Repository would have to translate too many different and possibly bespoke formats into a uniform format for the CAT data? Would a hybrid approach produce the most accurate data? How else would the benefits of the CAT NMS Plan differ between these approaches?

A hybrid approach would allow data to be submitted in either a uniform defined format or using existing messaging protocols.

442. The Commission requests comment on the implementation costs of these two data ingestion format approaches. The Commission expects that broker-dealers would need to modify existing messaging protocols to implement CAT regardless of which approach the Plan requires for reporting order events. What additional implementation costs would CAT Reporters incur to report using existing messaging protocols? What additional implementation costs would CAT Reporters, both OATS and non-OATS reporters, incur to report using a uniform defined format such as a modification of OATS format? In what ways would the implementation costs incurred at the Central Repository differ for the two approaches? What is the estimated cost of implementing each approach for CAT Reporters, Participants, and the Central Repository?

443. The Commission requests comment on the ongoing costs of these two data ingestion format approaches. How would ongoing costs be different for the two approaches? Would CAT Reporters need to process the order messages before reporting using existing messaging protocols to comply with requirements such as using the listing exchange symbology? If so, how costly is that processing? How costly is the processing required to translate order messages into a uniform defined format such as OATS format? What other ongoing costs associated with these approaches would CAT Reporters incur and how would they differ for the two approaches? How do the ongoing costs incurred by the Central Repository differ for the two approaches? Would the translation process from existing messaging protocols into a uniform format be more costly for the Central Repository relative to putting reports submitted in a uniform defined format in a single dataset? Would the validation process associated with existing messaging protocols be more costly for the Central Repository than uniform defined format because of the complexity of validating data from many different and possibly bespoke messaging protocols? What are the estimated ongoing costs of each approach for CAT Reporters, Participants, and the Central Repository?

444. Process to Develop the CAT. Bidders proposed, and the Plan describes, several processes for development of the CAT: The agile or iterative development model, the waterfall model, and hybrid models. The CAT NMS Plan does not mandate a particular development process because any of the options could be utilized to manage the development of CAT. The CAT NMS Plan notes that the agile model is more flexible and more susceptible to the early delivery of software for testing and feedback, but that the agile model makes it more difficult to accurately estimate the effort and time required for development. The waterfall model would also facilitate longer-term planning and coordination among multiple vendors or project streams. The Commission requests comment on the strengths and weaknesses of each development process. The Commission further requests comment on whether the CAT NMS Plan should mandate a particular process and the impact on the relative costs and benefits of the mandating a particular process.

See CAT NMS Plan, supra note 3, at Appendix C, Section D.12(g). An agile methodology is an iterative model in which development is staggered and provides for continuous evolution of requirements and solutions. A waterfall model is a sequential process of software development with dedicated phases for Conception, Initiation, Analysis, Design, Construction, Testing, Production/Implementation and Maintenance. Id.

Id.

Id.

445. Industry Testing. The CAT NMS Plan requires a dedicated test environment that is functionally equivalent to the production environment and available 24 hours a day, six days a week. The CAT NMS Plan discusses alternative approaches for industry testing. Using the production environment for scheduled testing events on weekends or on specific dates would allow for realistic testing because multiple users are likely to test at the same time. However, CAT Reporters would not be able to test when it might be more convenient or less costly for them to test. The Commission requests comment on whether the Plan should mandate particular industry testing processes and the benefits and costs of these alternatives compared to the requirements of the CAT NMS Plan. How would either of these alternatives lead to more accurate data than the Plan? Would the alternatives otherwise affect the benefits of the CAT NMS Plan? How would either of these alternatives affect the costs of the CAT NMS Plan for CAT Reporters, Participants, and the Central Repository? Please provide estimates, if available.

See CAT NMS Plan, supra note 3, at Appendix D, Section 1.2.

See id, at Appendix C, Section D.12(h).

446. Quality Assurance (QA). The CAT NMS Plan mentions several alternative approaches to quality assurance, but does not select a particular approach. In particular, the CAT NMS Plan states that the Participants considered many approaches, including continuous integration, test automation, and industry standards such as ISO 20000/ITIL. Although the Plan does not mandate a particular approach, certain requirements were detailed in the RFP. In addition, the CAT NMS Plan discusses the trade-offs associated with the QA staffing level. The Commission requests comment on whether the CAT NMS Plan should mandate a particular QA approach. Why or why not? If so, which approach should the Plan mandate? How can the QA approach affect the costs and benefits of the CAT NMS Plan? For example, how does the QA approach affect the accuracy and accessibility of the CAT Data? What are the relative strengths and weaknesses of the different quality assurance approaches?

See id., at Appendix C, Section D.12(i).

See RFP, supra note 29, at 31. Specifically, the RFP requires that Bidders' responses include both the functional and non-functional testing that includes the following: System testing, integration testing, regression testing, software performance testing, system performance testing, application programming interface (API) testing, user acceptance testing, industry testing, interoperability, security, load and performance testing, and CAT Reporter testing.

See CAT NMS Plan, supra note 3, at Appendix C, Section D.12(i). Bidder QA staffing levels range from 2 to 90. Id.

447. User Support and Help Desk. The CAT NMS Plan discusses several alternatives related to how the Plan Processor provides a CAT Help Desk that would be available 24 hours a day, 7 days a week and be able to manage 2,500 calls per month. Specifically, alternatives relate to the number of user support staff members, the degree to which the support team is dedicated to CAT, and whether the help desk is located in the US or offshore. The CAT NMS Plan discusses the benefit and cost trade-offs, but does not mandate any of the particular alternatives. Instead, the CAT NMS Plan commits to considering each bidder's user support proposals in the context of the overall bid. The Commission requests comment on whether the CAT NMS Plan should specify the standards for user support. How would the various alternatives affect the benefits of CAT? How would the various alternatives affect the implementation costs of CAT? How would the various alternatives affect the ongoing costs of CAT for CAT Reporters, Participants, and the Central Repository? Please explain and provide estimates, if available.

See CAT NMS Plan, supra note 3, at Appendix C, Section D.12(j). The RFP specified these standards. Id.

See id. The Plan states that a larger support staff could be more effective, but would be more costly. Further, a dedicated CAT support team would have a deeper knowledge of CAT but would be more costly. Finally, a U.S.-based help desk could facilitate greater security and higher quality service, but would be more costly. Id.

448. CAT User Management. The CAT NMS Plan discusses several alternatives to manage users, but does not require a specific approach or standards. Specifically, the CAT NMS Plan discusses help desk creation of accounts, user creation (by broker-dealers or regulators), and multi-role solutions. Generally, there are trade-offs in terms of convenience and security in the approaches. The Commission requests comments on whether the CAT NMS Plan should specify an approach for user management. How would the various alternatives affect the benefits of CAT, such as accessibility? How would the various alternatives affect the implementation costs of CAT? How would the various alternatives affect the ongoing costs of CAT for CAT Reporters, Participants, and the Central Repository? How would the various alternatives affect the risk of a security breach or misuse of the CAT Data? Please explain and provide estimates, if available.

See CAT NMS Plan, supra note 3, at Appendix C, Section D.12(k). User management is a business function that grants, controls, and maintains user access to a system. Id. at n.253.

See id. for more specific information on the relative strengths and weaknesses of each approach.

449. Required Reportable Events. The CAT NMS Plan states that the Participants considered requiring the reporting of multiple additional order event types, such as the “results order event” and the “CAT feedback order event.” According to the CAT NMS Plan, a “results order event” type would not provide additional value over a “daisy chain” linkage method and a “CAT feedback order event” can be generated by the Plan Processor, making reporting by others unnecessary. The Commission requests comments on these additional order event types and any other order event types that the Plan might require. Should the CAT NMS Plan require additional order event types? What are these order event types and what distinguishes them from the required order event types? What would be the purpose of these order event types? Would they make the CAT Data more complete or more accurate? How would regulators use these event types? How much would these additional order event types cost to report, to validate, and/or to store? Are there any other costs associated with these additional order event types? Please provide estimates, if available.

See CAT NMS Plan, supra note 3, at Appendix C, Section D.12(l).

Id.

450. Data Feed Connectivity. The Plan discusses requiring the collection of SIP data in real-time as opposed to through an end-of-day batch process. According to the Plan, real-time data would provide for more rapid access to SIP Data, but may require additional processing support to deal with out-of- sequence or missing records. Because CAT Reporters are only required to report order information on a next-day basis, the Plan does not require the Plan Processor to have real-time SIP connectivity. The Commission requests comments on whether the Plan should require a particular SIP connectivity. The Commission requests comment on the costs and benefits of requiring real-time SIP connectivity, or conversely, the costs and benefits of requiring end-of-day batch SIP connectivity (and not allow real-time). What would the Plan Processor do with real-time SIP data? Would the real-time SIP data be available to regulators, and if so, what would regulators do with that data? Do all regulators currently receive real-time SIP data? How much would the various SIP connectivity alternatives cost? How much processing would each alternative require to be of use to the Plan Processor or regulators?

See CAT NMS Plan, supra note 3, at Appendix C, Section D.12(n).

See id.

I. Request for Comment on the Economic Analysis

The Commission has identified above the economic effects associated with the proposed CAT NMS Plan and requests comment on all aspects of its preliminary economic analysis. The Commission encourages Commenters to identify, discuss, analyze, and supply relevant data, information, or statistics regarding any such economic effects. Commenters should, when possible, provide the Commission with data to support their views. Commenters suggesting alternative approaches should provide comprehensive proposals, including any conditions or limitations that they believe should apply, the reasons for the suggested approaches, their analysis of the cost-benefit trade-offs of suggested approaches compared to the Plan, and their analysis regarding why their suggested approaches would satisfy the objectives of Rule 613. In particular, the Commission seeks comment on the following:

451. Do Commenters agree with the Commission's analysis of the potential economic effects of the Plan? Why or why not?

452. Has the Commission considered all relevant economic effects? If not, what other economic effects should the Commission consider?

453. Do Commenters have information that could help the Commission fill in gaps in the economic analysis related to a lack of information on details in the plan that could significantly affect the economic analysis? If so, please provide this information and explain how it could affect the economic analysis.

454. Do Commenters have data that could help the Commission fill in gaps in the economic analysis related to a lack of available data? If so, please provide this information and explain how it could affect the economic analysis.

455. Do Commenters believe that there are additional categories of benefits or costs that could be quantified or otherwise monetized? If so, please identify these categories and, if possible, provide specific estimates or data.

456. Do Commenters believe that the CAT NMS Plan would change the behavior of any market participant in such a way as to create unintended effects? For example, would requirements to report certain data elements or events change the activities of market participants in ways other than deterrence but that create second-order economic effects? If so, please explain. Would such effects be economic benefits or economic costs? Please explain.

V. Paperwork Reduction Act

Certain provisions of Rule 613 contain “collection of information requirements” within the meaning of the Paperwork Reduction Act of 1995 (“PRA”) and the Commission has submitted them to the Office of Management and Budget (“OMB”) for review in accordance with 44 U.S.C. 3507 and 5 CFR 1320.11. An agency may not conduct or sponsor, and a person is not required to respond to, a collection of information unless it displays a currently valid OMB control number. The title of the collection of information is “Creation of a Consolidated Audit Trail Pursuant to Section 11A of the Securities Exchange Act of 1934 and Rules Thereunder.”

44 U.S.C. 3501 et. seq.

As noted above, Rule 613 of Regulation NMS (17 CFR part 242) requires the Participants to jointly submit to the Commission the CAT NMS Plan to govern the creation, implementation, and maintenance of the consolidated audit trail and Central Repository for the collection of information for NMS securities. The CAT NMS Plan must require each Participant and its respective members to provide certain data to the Central Repository in compliance with Rule 613. When it adopted Rule 613, the Commission discussed the burden hours associated with the development and submission of the CAT NMS Plan. In doing so, the Commission noted that the development and submission of the CAT NMS Plan that would govern the creation, implementation and maintenance of a consolidated audit trail is a multi-step process and accordingly that the Commission was deferring its discussion of the burden hours associated with the other paperwork requirements required by Rule 613 and ongoing burdens since they would only be incurred if the Commission approves the CAT NMS Plan.

See Adopting Release, supra note 9, at 45804. On September 25, 2015, the Commission submitted to OMB a request for approval of an extension of the collection of information related to the development and submission of the CAT NMS Plan. The Commission stated that, although that collection of information pertained to the development and submission of an NMS plan, and that such NMS plan had already been developed and submitted, the Commission believed it was prudent to extend the collection of information during the pendency of the Commission's review of the NMS plan. The Commission provided estimates for 19 SROs, stating that they would spend a total of 2,760 burden hours of internal legal, compliance, information technology, and business operations time to comply with the existing collection of information, calculated as follows: (880 programmer analyst hours) + (880 business analyst hours) + (700 attorney hours) + (300 compliance manager hours) = 2,760 burden hours to prepare and file an NMS plan, or approximately 52,440 burden hours in the aggregate, calculated as follows: (2,760 burden hours per SRO) × (19 SROs) = 52,440 burden hours. Amortized over three years, the annualized burden hours would be 920 hours per SRO, or a total of 17,480 for all 19 SROs. The Commission further estimated that the aggregate one-time reporting burden for preparing and filing an NMS plan would be approximately $20,000 in external legal costs per SRO, calculated as follows: 50 legal hours × $400 per hour = $20,000, for an aggregate burden of $380,000, calculated as follows: ($20,000 in external legal costs per SRO) × (19 SROs) = $380,000. Amortized over three years, the annualized capital external cost would be $6,667 per SRO, or a total of $126,667 for the 19 SROs. See Submission for OMB Review; Comment Request for Extension of Rule 613; SEC File No. 270-616, OMB Control No. 3235-0671 (September 25, 2015), 80 FR 59209 (October 1, 2015).

Id.

The Commission is now publishing its preliminary estimates of the paperwork burdens of the CAT NMS Plan. These estimates are based on the requirements of Rule 613 and take into account the Exemption Order discussed above. Information and estimates contained in the CAT NMS Plan that was submitted by the Participants also informed these estimates because they provide a useful, quantified point of reference regarding potential burdens and costs. The Commission acknowledges that the CAT NMS Plan as filed contains provisions in addition to those required by Rule 613 (e.g., requiring the inclusion of OTC Equity Securities; the availability of historical data for not less than six years in a manner that is directly available and searchable without manual intervention from the Plan Processor; a complete symbology database to be maintained by the Plan Processor, including the historical symbology; as well as issue symbol information and data using the listing exchange symbology format ).

See Exemption Order, supra note 18.

See CAT NMS Plan, supra note 3, at Section 1.1 (defining “Eligible Security” as all NMS securities and all OTC Equity Securities); Appendix C, Section A.1(a).

See id. at Section 6.5(b)(i).

See CAT NMS Plan, supra note 3 at Appendix C, Section A.1(a); Appendix D, Section 2.

A. Summary of Collection of Information Under Rule 613

Rule 613 requires that the CAT NMS Plan must provide for an accurate, time-sequenced record of an order's life, from receipt or origination, through the process of routing, modification, cancellation and execution. The Central Repository, created by the Participants, would be required to receive, consolidate and retain the data required under the Rule. Such data must be accessible to each Participant, as well as the Commission, for purposes of performing regulatory and oversight responsibilities.

See 17 CFR 242.613(e)(1), (e)(2).

Rule 613 provides that the CAT NMS Plan must require that all Participants that are exchanges, and their members, record and report to the Central Repository certain data for each NMS security registered or listed on a national securities exchange, or admitted to unlisted trading privileges on such exchange, and each Participant that is a national securities association, and its members, record and report for each NMS security for which transaction reports are required to be submitted to the national securities association in a uniform electronic format or in a manner that would allow the Central Repository to convert the data to a uniform electronic format for consolidation and storage. This data must be recorded contemporaneously with the Reportable Event and reported to the Central Repository in no event later than 8:00 a.m. Eastern Time on the trading day following the day such information has been recorded by the national securities exchange, national securities association, or member.

Rule 613 also provides that the CAT NMS Plan must require each member of a Participant to record and report to the Central Repository other information which may not be available until later in the clearing process no later than 8:00 a.m. Eastern Time on the trading day following the day the member receives such information. The CAT NMS Plan also requires the Participants to provide to the Commission, at least every two years after the effectiveness of the CAT NMS Plan, a written assessment of the operation of the consolidated audit trail.

Rule 613 requires all Participants to make use of the consolidated information, either by each developing and implementing new surveillance systems, or by enhancing existing surveillance systems. The Rule also requires the CAT NMS Plan to require Participants to submit to the Commission a document outlining the manner in which non-NMS securities and primary market transactions in NMS and non-NMS securities can be incorporated into the consolidated audit trail.

1. Central Repository

Rule 613 provides that the CAT NMS Plan must require the creation and maintenance of a Central Repository that would be responsible for the receipt, consolidation, and retention of all data submitted by the Participants and their members. The Rule also requires that the CAT NMS Plan require the Central Repository to retain the information reported pursuant to subparagraphs (c)(7) and (e)(7) of the Rule for a period of not less than five years in a convenient and usable standard electronic data format that is directly available and searchable electronically without any manual intervention. The Plan Processor is responsible for operating the Central Repository in compliance with the Rule and the CAT NMS Plan. In addition, the Rule provides that the CAT NMS Plan must include: Policies and procedures to ensure the security and confidentiality of all information submitted to the Central Repository, including safeguards to ensure the confidentiality of data; information barriers between regulatory and non-regulatory staff with regard to access and use of data; a mechanism to confirm the identity of all persons permitted to use the data; a comprehensive information security program for the Central Repository that is subject to regular reviews by the CCO; and penalties for non-compliance with policies and procedures of the Participants or the Central Repository with respect to information security. Further, the Rule provides that the CAT NMS Plan must include policies and procedures to be used by the Plan Processor to ensure the timeliness, accuracy, integrity, and completeness of the data submitted to the Central Repository, as well as policies and procedures to ensure the accuracy of the consolidation by the Plan Processor of the data.

See 17 CFR 242.613(e)(8). The Commission notes that the CAT NMS Plan proposes to require that the Central Repository retain data reported in a convenient and usable standard electronic data format that is directly available and searchable electronically without any manual intervention for six years. See CAT NMS Plan, supra note 3, at Section 6.5(b)(i).

Id.

2. Data Collection and Reporting

Rule 613 provides that the CAT NMS Plan must require each Participant, and any member of such Participant, to record and electronically report to the Central Repository details for each order and Reportable Event documenting the life of an order through the process of original receipt or origination, routing, modification, cancellation, and execution (in whole or part) for each NMS security. For national securities exchanges, Rule 613 requires the CAT NMS Plan to require each national securities exchange and its members to record and report to the Central Repository the information required by Rule 613(c)(7) for each NMS security registered or listed for trading on an exchange, or admitted to unlisted trading privileges on such exchange. Rule 613 provides that the CAT NMS Plan must require each Participant that is a national securities association, and its members, to record and report to the Central Repository the information required by Rule 613(c)(7) for each NMS security for which transaction reports are required to be submitted to the Participant. The Rule requires each Participant and any member of a Participant to record the information required by Rule 613(c)(7)(i) through (v) contemporaneously with the Reportable Event, and to report this information to the Central Repository by 8:00 a.m. Eastern Time on the trading day following the day such information has been recorded by the Participant or member of the Participant. The Rule requires each Participant and any member of a Participant to record and report the information required by Rule 613(c)(7)(vi) through (viii) to the Central Repository by 8:00 a.m. Eastern Time on the trading day following the day the Participant or member receives such information. The Rule requires each Participant and any member of such Participant to report information required by Rule 613(c)(7) in a uniform electronic format or in a manner that would allow the Central Repository to convert the data to a uniform electronic format for consolidation and storage.

See 17 CFR 242.613(c)(1), (c)(5), (c)(6), (c)(7).

See 17 CFR 242.613(c)(1), (c)(5).

See 17 CFR 242.613(c)(1), (c)(6).

Such information must also be reported to the Central Repository with a time stamp of a granularity that is at least to the millisecond or less to the extent that the order handling and execution systems of a Participant or a member utilize time stamps in finer increments. The Commission understands that any changes to broker-dealer recording and reporting systems to comply with Rule 613 may also include changes to comply with the millisecond time stamp requirement.

3. Collection and Retention of NBBO, Last Sale Data and Transaction Reports

Rule 613(e)(7) provides that the CAT NMS Plan must require the Central Repository to collect and retain on a current and continuing basis: (i) Information on the NBBO for each NMS Security; (ii) transaction reports reported pursuant to a transaction reporting plan filed with the Commission pursuant to, and meeting the requirements of, Rule 601 of Regulation NMS; and (iii) Last Sale Reports reported pursuant to the OPRA Plan. The Central Repository must retain this information for no less than five years.

See 17 CFR 242.613(e)(7); 17 CFR 242.601.

4. Surveillance

Rule 613(f) provides that the CAT NMS Plan must require that every Participant develop and implement a surveillance system, or enhance existing surveillance systems, reasonably designed to make use of the consolidated information contained in the consolidated audit trail. Rule 613(a)(3)(iv) provides that the CAT NMS Plan must require that the surveillance systems be implemented within fourteen months after effectiveness of the CAT NMS Plan.

5. Participant Rule Filings

Rule 613(g)(1) requires each Participant to file with the Commission, pursuant to Section 19(b)(2) of the Exchange Act and Rule 19b-4 thereunder, a proposed rule change to require its members to comply with the requirements of Rule 613 and the CAT NMS Plan approved by the Commission. The burden of filing such a proposed rule change is already included under the collection of information requirements contained in Rule 19b-4 under the Exchange Act.

15 U.S.C. 78s(b)(2) and 17 CFR 240.19b-4.

See Securities Exchange Act Release No. 50486 (October 5, 2004), 69 FR 60287, 60293 (October 8, 2004) (File No. S7-18-04) (describing the collection of information requirements contained in Rule 19b-4 under the Exchange Act). The Commission has submitted revisions to the current collection of information titled “Rule 19b-4 Filings with Respect to Proposed Rule Changes by Self-Regulatory Organizations” (OMB Control No. 3235-0045). According to the last submitted revision, for Fiscal Year 2012 SROs submitted 1,688 Rule 19b-4 proposed rule changes.

6. Written Assessment of Operation of the Consolidated Audit Trail

Rule 613(b)(6) provides that the CAT NMS Plan must require the Participants to provide the Commission a written assessment of the consolidated audit trail's operation at least every two years, once the CAT NMS Plan is effective. Such written assessment shall include, at a minimum, with respect to the CAT: (i) An evaluation of its performance; (ii) a detailed plan for any potential improvements to its performance; (iii) an estimate of the costs associated with any such potential improvements; and (iv) an estimated implementation timeline for any such potential improvements, if applicable.

See id.

7. Document on Expansion to Other Securities

Rule 613(i) provides that the CAT NMS Plan must require the Participants to jointly provide to the Commission, within six months after the CAT NMS Plan is effective, a document outlining how the Participants could incorporate into the CAT information regarding: (1) Equity securities that are not NMS securities; (2) debt securities; and market transactions in equity securities that are not NMS securities and debt securities.

As noted above, the CAT NMS Plan would require the inclusion of OTC Equity Securities, while Rule 613 does not include such a requirement. See supra note 1373.

B. Proposed Use of Information

1. Central Repository

Rule 613 states that the Central Repository is required to receive, consolidate and retain the data required to be submitted by the Participants and their members. Participant and Commission Staff would have access to the data for regulatory purposes.

2. Data Collection and Reporting

The Commission believes that the data collected and reported pursuant to the requirements of Rule 613 would be used by regulators to monitor and surveil the securities markets and detect and investigate activity, whether on one market or across markets. The data collected and reported pursuant to Rule 613 would also be used by regulators for the evaluation of tips and complaints and for complex enforcement inquiries or investigations, as well as inspections and examinations. Further, the Commission believes that regulators would use the data collected and reported to conduct timely and accurate analysis of market activity for reconstruction of broad-based market events in support of regulatory decisions.

See Section IV.E.2, supra.

3. Collection and Retention of NBBO, Last Sale Data and Transaction Reports

The CAT NMS Plan must require the Central Repository to collect and retain NBBO information, transaction reports, and Last Sale Reports in a format compatible with the order and event information collected pursuant to Rule 613(c)(7). Participant and Commission Staff could use this data to easily search across order, NBBO, and transaction databases. The Commission believes that having the NBBO information in a uniform electronic format compatible with order and event information would assist Participants in enforcing compliance with federal securities laws, rules, and regulations, as well as their own rules. The Commission also believes that a CAT NMS Plan requiring the Central Repository to collect and retain the transaction reports and Last Sale Reports in a format compatible with the order execution information would aid regulators in monitoring for certain market manipulations.

The Commission and Participants use the NBBO to, among other things, evaluate members for compliance with numerous regulatory requirements, such as the duty of best execution or Rule 611 of Regulation NMS. See 17 CFR 242.611; see also, e.g., ISE Rule 1901 and Phlx Rule 1084.

Rules 613(e)(7)(ii) and (iii) require that transaction reports reported pursuant to an effective transaction reporting plan and Last Sale Reports reported pursuant to the OPRA Plan be reported to the Central Repository. This requirement should allow regulators to evaluate certain trading activity. For example, trading patterns of reported and unreported trades may cause Participant or Commission staff to make further inquiries into the nature of the trading to ensure that the public was receiving accurate and timely information regarding executions and that market participants were continuing to comply with trade reporting obligations under Participant rules. Similarly, patterns in the transactions that are reported and unreported to the consolidated tape could be indicia of market abuse, including failure to obtain best execution for customer orders or possible market manipulation. The Commission and the Participants would be able to review information on trades not reported to the tape to determine whether they should have been reported, whether Section 31 fees should have been paid, and/or whether the trades are part of a manipulative scheme.

4. Surveillance

The requirement in Rule 613(f) that the Participants develop and implement a surveillance system, or enhance existing surveillance systems, reasonably designed to make use of the consolidated information in the consolidated audit trail, is intended to position regulators to make full use of the consolidated audit trail data in order to carry out their regulatory obligations. In addition, because trading and potentially manipulative activities could take place across multiple markets, and the consolidated audit trail data would trace the entire lifecycle of an order from origination to execution or cancellation, new or enhanced surveillance systems may also enable regulators to investigate potentially illegal activity that spans multiple markets more efficiently.

5. Written Assessment of Operation of the Consolidated Audit Trail

Rule 613(b)(6) requires the CAT NMS Plan to require the Participants to provide the Commission a written assessment of the CAT's operation at least every two years, once the CAT NMS Plan is effective. These assessments would aid Participant and Commission Staff in understanding and evaluating any deficiencies in the operation of the consolidated audit trail and to propose potential improvements to the CAT NMS Plan. The Commission believes the written assessments would allow Participants and Commission Staff to periodically assess whether such potential improvements would enhance market oversight. Moreover, the Commission believes these assessments would help inform the Commission regarding the likely feasibility, costs, and impact of, and the Participants' approach to, the consolidated audit trail evolving over time.

6. Document on Expansion to Other Securities

Rule 613(i) requires the CAT NMS Plan to require the Participants to jointly provide to the Commission, within six months after the CAT NMS Plan is effective, a document outlining how the SROs could incorporate into the CAT information regarding certain products that are not NMS securities. A document outlining a possible expansion of the consolidated audit trail could help inform the Commission about the SROs' strategy for potentially accomplishing such an expansion over a reasonable period of time. Moreover, such document would aid the Commission in assessing the feasibility and impact of possible future proposals by the SROs to include such additional securities and transactions in the consolidated audit trail.

See 17 CFR 242.613(i). See also supra note 1408.

C. Respondents

1. National Securities Exchanges and National Securities Associations

Rule 613 applies to the 20 Participants (the 19 national securities exchanges and the one national securities association (FINRA)) currently registered with the Commission.

The Participants are: BATS Exchange, Inc., BATS-Y Exchange, Inc., BOX Options Exchange LLC, C2 Options Exchange, Incorporated, Chicago Board Options Exchange, Incorporated, Chicago Stock Exchange, Inc., EDGA Exchange, Inc., EDGX Exchange, Inc., Financial Industry Regulatory Authority, Inc., International Securities Exchange, LLC, ISE Gemini, LLC, ISE Mercury, LLC, Miami International Securities Exchange LLC, NASDAQ OMX BX, Inc., NASDAQ OMX PHLX LLC, The NASDAQ Stock Market LLC, National Stock Exchange, Inc., New York Stock Exchange LLC, NYSE MKT LLC, and NYSE Arca, Inc. The Commission understands that ISE Mercury, LLC will become a Participant in the CAT NMS Plan and thus is accounted for as a Participant for purposes of this Section. See supra note 3.

2. Members of National Securities Exchanges and National Securities Association

Rule 613 also applies to the Participants' members, that is, broker-dealers. The Commission believes that Rule 613 applies to 1,800 broker-dealers. The Commission understands that there are currently 4,138 broker-dealers; however, not all broker-dealers are expected to have CAT reporting obligations. The Participants report that approximately 1,800 broker-dealers currently quote or execute transactions in NMS Securities, Listed Options or OTC Equity Securities and would likely have CAT reporting obligations.

The Commission understands that the remaining 2,338 registered broker-dealers either trade in asset classes not currently included in the definition of Eligible Security or do not trade at all (e.g., broker-dealers for the purposes of underwriting, advising, private placements). See supra note 864.

D. Total Initial and Annual Reporting and Recordkeeping Burden

1. Burden on National Securities Exchanges and National Securities Associations

a. Central Repository

Rule 613 requires the Participants to jointly establish a Central Repository tasked with the receipt, consolidation, and retention of the reported order and execution information. The Participants issued an RFP soliciting Bids from entities to act as the consolidated audit trail's Plan Processor. Bidders were asked to provide total one-year and annual recurring cost estimates to estimate the costs to the Participants for implementing and maintaining the Central Repository. There are currently three remaining Bidders, any of which could be selected to be the Plan Processor. The Plan Processor would be responsible for building, operating, administering and maintaining the Central Repository.

See Section III.A.1, supra.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(B). The CAT NMS Plan listed the following as primary drivers of Bid costs: (1) Reportable volumes of data ingested into the Central Repository; (2) number of technical environments that would be have to be built to report to the Central Repository; (3) likely future rate of increase of reportable volumes; (4) data archival requirements; and (5) user support and/or help desk resource requirements. See id. at Section B.7(b)(i)(B).

The Plan's Operating Committee, which consists of one voting representative of each Participant, would be responsible for the management of the LLC, including the Central Repository, acting by Majority or Supermajority Vote, depending on the issue. In managing the Central Repository, among other things, the Operating Committee would have the responsibility to authorize the following actions of the LLC: (1) Interpreting the Plan; (2) determining appropriate funding-related policies, procedures and practices consistent with Article XI of the CAT NMS Plan; (3) terminating the Plan Processor; (4) selecting a successor Plan Processor (including establishing a Plan Processor Selection Subcommittee to evaluate and review Bids and make a recommendation to the Operating Committee with respect to the selection of the successor Plan Processor); (5) entering into, modifying or terminating any Material Contract; (6) making any Material Systems Change; (7) approving the initial Technical Specifications or any Material Amendment to the Technical Specifications proposed by the Plan Processor; (8) amending the Technical Specifications on its own motion; (9) approving the Plan Processor's appointment or removal of the CCO, CISO, or any Independent Auditor in accordance with Section 6.1(b) of the CAT NMS Plan; (10) approving any recommendation by the CCO pursuant to Section 6.2(a)(v)(A); (11) selecting the members of the Advisory Committee; (12) selecting the Operating Committee chair; and (13) determining to hold an Executive Session of the Operating Committee.

See id. at Section 4.2(a).

See Section IV.E.3.d(1), supra.

See CAT NMS Plan, supra note 3, at Section 4.3(a)(iii).

See id. at Section 4.3(a)(vi).

See id. at Section 4.3(b)(i).

See id. at Section 4.3(b)(iv).

See id. at Section 4.3(b)(v).

See id. at Section 4.3(b)(vi).

See id. at Section 4.3(b)(vii).

See id. at Section 4.3(b)(iii).

See id. at Section 4.3(a)(iv).

See id. at Section 4.3(a)(ii).

See id. at Section 4.3(a)(i).

See id. at Section 4.3(a)(v).

Additionally, in managing the Central Repository, the Operating Committee would have the responsibility and authority, as appropriate, to: (1) Direct the LLC to enter into one or more agreements with the Plan Processor obligating the Plan Processor to perform the functions and duties contemplated by the Plan to be performed by the Plan Processor, as well as such other functions and duties the Operating Committee deems necessary or appropriate; (2) appoint as an Officer of the Company the individual who has direct management responsibility for the Plan Processor's performance of its obligations with respect to the CAT; (3) approve policies, procedures, and control structures related to the CAT System that are consistent with Rule 613(e)(4), Appendix C and Appendix D of the CAT NMS Plan that have been developed and will be implemented by the Plan Processor; (4) approve any policy, procedure or standard (and any material modification or amendment thereto) applicable primarily to the performance of the Plan Processor's duties as the Plan Processor; (5) for both the CCO and CISO, render their annual performance reviews and review and approve their compensation; (6) review the Plan Processor's performance under the Plan at least once each year, or more often than once each year upon the request of two Participants that are not Affiliated Participants; (7) in conjunction with the Plan Processor, approve and regularly review (and update as necessary) SLAs governing the performance of the Central Repository; (8) maintain a Compliance Subcommittee for the purpose of aiding the CCO as necessary; and (9) designate by resolution one or more Subcommittees it deems necessary or desirable in furtherance of the management of the business and affairs of the Company.

See id. at Section 6.1(a).

See id. at Section 4.6(b).

See id. at Section 6.1(c).

See id. at Section 6.1(e).

See id. at Section 6.2(a)(iv) and Section 6.2(b)(iv).

See id. at Section 6.1(n).

See id. at Section 6.1(h).

See id. at Section 4.12(b).

See id. at Section 4.12(a).

The CAT NMS Plan also proposes to establish a Selection Committee comprised of one Voting Senior Officer from each Participant, which is tasked with the review and evaluation of Bids and the selection of the initial Plan Processor. The Selection Committee would determine, by Majority Vote, whether Shortlisted Bidders will have the opportunity to revise their Bids. The Selection Committee would review and evaluate all Shortlisted Bids, including any permitted revisions submitted by Shortlisted Bidders, and in doing so, may consult with the Advisory Committee (or the DAG until the Advisory Committee is formed) and such other Persons as the Selection Committee deems appropriate. After receipt of any permitted revisions, the Selection Committee would select the Initial Plan Processor from the Shortlisted Bids in two rounds of voting where each Participant has one vote via its Voting Senior Officer in each round. Following the selection of the Initial Plan Processor, the Participants would file with the Commission a statement identifying the Initial Plan Processor and including the information required by Rule 608.

See id. at Section 5.1(a).

See id. at Section 5.1.

See id. at Section 5.1(d)(i).

See id. at Section 5.1(d)(ii).

See id. at Section 5.1(e).

See id. at Section 6.7(a)(i).

For its initial and ongoing internal burden and cost estimates associated with the management of the Central Repository, the Commission is relying on estimates provided in the CAT NMS Plan for the development of the CAT NMS Plan, which the Participants “have accrued, and will continue to accrue,” and have described in the CAT NMS Plan as “reasonably associated with creating, implementing, and maintaining the CAT upon the Commission's adoption of the CAT NMS Plan.”

See id. at Appendix C, Section B.7(b)(iii).

See id.

The Commission believes that the activities of the Operating Committee and the Selection Committee overlap with those undertaken by the Participants to develop the CAT NMS Plan. The CAT NMS Plan describes the costs incurred by the Participants to develop the CAT NMS Plan as including “staff time contributed by each Participant to, among other things, determine the technological requirements for the Central Repository, develop the RFP, evaluate Bids received, design and collect the data necessary to evaluate costs and other economic impacts, meet with Industry Members to solicit feedback, and complete the CAT NMS Plan submitted to the Commission for consideration.” For the building and management of the Central Repository, the Selection Committee and the Operating Committee would have comparable responsibilities. The Selection Committee would be required to review and evaluate all Shortlisted Bids, including any permitted revisions submitted by Shortlisted Bidders, and then to select the initial Plan Processor from those Bids. As part of its overall management of the Central Repository, the Operating Committee would have responsibility for decisions associated with the technical requirements of the Central Repository. Furthermore, the Operating Committee would be required to establish a Selection Subcommittee to evaluate Bids received to select a successor Plan Processor, and would also be required to authorize the selection of the members of the Advisory Committee, comprising members of the Industry, to advise the Participants on the implementation, operation, and administration of the Central Repository. Because the responsibilities of the Operating Committee and the Selection Committee are similar to those described in the CAT NMS Plan for the development of the CAT NMS Plan itself, the Commission believes that it is reasonable to use the CAT NMS Plan estimates as the basis for its burden and cost estimates for the initial and ongoing management of the Central Repository.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii).

For example, the Operating Committee would be required to authorize the following actions of the LLC: Entering into, modifying or terminating any Material Contract (see id. at Section 4.3(b)(iv)); making any Material Systems Change (see id. at Section 4.3(b)(v)); amending the Technical Specifications on its own motion (see id. at Section 4.3(b)(vii)); and approving the initial Technical Specifications or any Material Amendment to the Technical Specifications proposed by the Plan Processor (see id. at Section 4.3(b)(vi)). Further, the Operating Committee would be able to approve policies, procedures, and control structures related to the CAT System that are consistent with Rule 613(e)(4), Appendix C and Appendix D of the CAT NMS Plan that have been developed and will be implemented by the Plan Processor (see id. at Section 6.1(c)); and in conjunction with the Plan Processor, approve and regularly review (and update as necessary) SLAs governing the performance of the Central Repository (see id. at Section 6.1(h)).

See id. at Section 4.3(b)(i).

See id. at Section 4.3(a)(ii).

See id. at Section 4.13(d).

(1) Initial Burden and Costs To Build the Central Repository

As proposed, each Participant would contribute an employee and a substitute for the employee to serve on the Operating Committee that would oversee the Central Repository. Additionally, each Participant would select a Voting Senior Officer to represent the Participant as a member of the Selection Committee responsible for the selection of the Plan Processor of the Central Repository.

In the case of Affiliated Participants, one individual may be the primary representative for all or some of the Affiliated Participants, and another individual may be the substitute for all or some of the Affiliated Participants. See id. at Section 4.2(a).

In the case of Affiliated Participants, one individual may be (but is not required to be) the Voting Senior Officer for more than one or all of the Affiliated Participants. Where one individual serves as the Voting Senior Officer for more than one Affiliated Participant, such individual will have the right to vote on behalf of each such Affiliated Participant. See id. at Section 5.1(a).

The Commission preliminarily estimates that, over the 12-month period after the effectiveness of the CAT NMS Plan within which the Participants would be required to select an initial Plan Processor and begin reporting to the Central Repository, each Participant would incur an initial internal burden of 720 burden hours associated with the management of the creation of the Central Repository and the selection of the Plan Processor (including filing with the Commission the statement identifying the Initial Plan Processor and including the information required by Rule 608), for an aggregate initial estimate of 14,407 burden hours.

Rule 613(a)(3)(i) requires the selection of the Plan Processor within 2 months after effectiveness of the CAT NMS Plan. See 17 CFR 242.613(a)(3)(i).

Rule 613(a)(3)(iii) requires the Participants to provide to the Central Repository the data required by Rule 613(c) within one year after effectiveness of the CAT NMS Plan. See 17 CFR 242.613(a)(3)(iii).

The Commission is basing this estimate on the internal burden estimate provided in the CAT NMS Plan related to the development of the CAT NMS Plan. See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii) (stating “. . . the Participants have accrued, and will continue to accrue, direct costs associated with the development of the CAT NMS Plan. These costs include staff time contributed by each Participant to, among other things, determine the technological requirements for the Central Repository, develop the RFP, evaluate Bids received, design and collect the data necessary to evaluate costs and other economic impacts, meet with Industry Members to solicit feedback, and complete the CAT NMS Plan submitted to the Commission for consideration. The Participants estimate that they have collectively contributed 20 FTEs in the first 30 months of the CAT NMS Plan development process”). The Commission believes the staff time incurred for the development of the CAT NMS Plan would be comparable to the staff time incurred for the activities required of the Operating Committee and the Selection Committee for the creation and management of the Central Repository once the Plan is effective). (20 FTEs/30 months) = 0.667 FTEs per month for all of the Participants to develop the CAT NMS Plan. Converting this into burden hours, (0.667 FTEs) × (12 months) × (1,800 burden hours per year) =14,407 initial burden hours for all of the Participants to develop the CAT NMS Plan. (14,407 burden hours for all Participants/20 Participants) = 720 initial burden hours for each Participant to develop the CAT NMS Plan.

Additionally, the Commission preliminarily estimates that the Participants will collectively spend $2,400,000 on external public relations, legal and consulting costs associated with the building of the Central Repository and the selection of the Plan Processor for the Central Repository, or $120,000 per Participant. The Commission is basing this estimate on the estimate provided in the CAT NMS Plan for public relations, legal and consulting costs incurred in preparation of the CAT NMS Plan. Because the Participants described such costs as “reasonably associated with creating, implementing and maintaining the CAT,” the Commission preliminarily believes these external cost estimates should also be applied to the creation and implementation of the Central Repository.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii) (stating “the Participants have incurred public relations, legal and consulting costs in preparation of the CAT NMS Plan. The Participants estimate the costs of these services to be $8,800,000”). $2,400,000 for all Participants over 12 months = ($8,800,000/44 months between the adoption of Rule 613 and the filing of the CAT NMS Plan) × (12 months). ($2,400,000/20 Participants) = $120,000 per Participant over 12 months.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii).

The CAT NMS Plan provides the estimates given by the Shortlisted Bidders for the one-time total cost associated with the Plan Processor that would build the Central Repository. The CAT NMS Plan states that this includes internal technological, operational, administrative and “any other material costs.” Using the estimates in the CAT NMS Plan, which are based on the Bids of the six Shortlisted Bidders, the Commission preliminarily estimates that the initial one-time cost to develop the Central Repository would be an aggregate initial external cost to the Participants of $91.6 million, or $4.6 million per Participant. Therefore, the Commission preliminarily estimates that each Participant would incur initial one-time external costs of $7 million to build the Central Repository, or an aggregate initial one-time external cost across all Participants of $140 million.

The Selection Committee narrowed the list of Shortlisted Bidders from six to three Shortlisted Bidders. See “Participants, SROs Reduce Short List Bids from Six to Three for Consolidated Audit Trail” (November 16, 2015), available at http://www.catnmsplan.com/pastevents/catnms_release_downselect_111615.pdf . However, the costs provided by the SROs in the CAT NMS Plan are based on the Bids of the six Shortlisted Bidders.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(B). See also id. at Appendix C, Section B.7(b)(iv)(A)(1). The Commission notes that the cost associated with the build and maintenance of the Central Repository includes compliance with the requirement in Rule 613(e)(8) that the Central Repository retain information collected pursuant to Rule 613(c)(7) and (e)(7) in a convenient and usable standard electronic data format that is directly available and searchable electronically without any manual intervention for a period of not less than five years. See id. at Section 6.1(d)(i) (requiring the Plan Processor to comply with the recordkeeping requirements of Rule 613(e)(8)). See also id. at Appendix C, Section D.12(l) (stating that Rule 613(e)(8) requires data to be available and searchable for a period of not less than five years, that broker-dealers are currently required to retain data for six years under Rule 17a-4(a), and that the Participants are requiring CAT Data to be kept online in an easily accessible format for regulators for six years, though this may increase the cost to run the CAT). The Commission notes that a Shortlisted Bidder may be permitted to revise its Bid prior to approval of the CAT NMS Plan if the CAT Selection Committee determines by Majority Vote that such revisions are necessary or appropriate, so the estimates provided in the CAT NMS Plan may be subject to change. See id. at Section 5.2(c)(ii). In addition, changes in technology between the time the Bids were submitted and the time the Central Repository is built could result in changes to the costs to build and operate the Central Repository.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(B).

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(B) (describing the minimum, median, mean and maximum Bidder estimates for the build and maintenance costs of the Central Repository).

Id. The Bidders provided a range of estimates. For purposes of this Paperwork Burden Act analysis, the Commission is using the build cost of the maximum Bidder estimate. $4,580,000 = $91,600,000/20 SROs.

$7 million for each Participant to build the Central Repository = ($4.6 million per Participant in initial one-time costs to compensate the Plan Processor to build the Central Repository) + ($2.4 million per Participant in initial one-time public relations, legal and consulting costs associated with the building of the Central Repository and the selection of the initial Plan Processor).

$140 million for all of the Participants to build the Central Repository = $7 million per Participant to build the Central Repository) × (20 Participants). Id.

(2) Ongoing, Annual Burden Hours and Costs for the Central Repository

After the Central Repository has been developed and implemented, there would be ongoing costs for operating and maintaining the Central Repository, including the cost of systems and connectivity upgrades or changes necessary to receive, consolidate, and store the reported order and execution information from Participants and their members; the costs to store data, and make it available to regulators, in a uniform electronic format, and in a form in which all events pertaining to the same originating order are linked together in a manner that ensures timely and accurate retrieval of the information; the cost, including storage costs, of collecting and maintaining the NBBO and transaction data in a format compatible with the order and event information collected pursuant to the Rule; the cost of monitoring the required validation parameters, which would allow the Central Repository to automatically check the accuracy and completeness of the data submitted and reject data not conforming to these parameters consistent with the requirements of the proposed Rule; and the cost of paying the CCO. The CAT NMS Plan provides that the Plan Processor would be responsible for the ongoing operations of the Central Repository. The Operating Committee would continue to be responsible for the management of the Central Repository. In addition, the CAT NMS Plan states that the Participants would incur costs for public relations, legal, and consulting costs associated with maintaining the CAT upon approval of the CAT NMS Plan.

See supra note 1469.

See CAT NMS Plan, supra note 3, at Section 6.1.

See id. at Appendix C, Section B.7(b)(iii).

The Commission preliminarily estimates that the Participants would incur an ongoing annual internal burden of 720 burden hours associated with the continued management of the Central Repository, for an aggregate annual estimate of 14,407 burden hours across the Participants.

The Commission is basing this estimate on the internal burden estimate provided in the CAT NMS Plan for the development of the CAT NMS Plan. The Commission notes that the CAT NMS Plan describes the internal burden estimate for the development of the CAT NMS Plan as a cost the Participants will continue to accrue; therefore, the Commission preliminarily believes that it is reasonable to use this burden estimate as the basis for its ongoing internal burden estimate for the maintenance of the Central Repository, particularly as the Commission believes the reasons for the staff time incurred for the development of the CAT NMS Plan would be comparable to those of the staff time to be incurred by the Operating Committee and the Selection Committee for the continued management of the Central Repository. See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii) (stating “ . . . the Participants have accrued, and will continue to accrue, direct costs associated with the development of the CAT NMS Plan. These costs include staff time contributed by each Participant to, among other things, determine the technological requirements for the Central Repository, develop the RFP, evaluate Bids received, design and collect the data necessary to evaluate costs and other economic impacts, meet with Industry Members to solicit feedback, and complete the CAT NMS Plan submitted to the Commission for consideration. The Participants estimate that they have collectively contributed 20 FTEs in the first 30 months of the CAT NMS Plan development process”). (20 FTEs/30 Participants) = 0.667 FTEs per month for all of the Participants to continue management of the Central Repository. Converting this into burden hours, (0.667 FTEs) × (12 months) × (1,800 burden hours per year) = 14,407 ongoing annual burden hours for all of the Participants to continue management of the Central Repository. (14,407 ongoing annual burden hours for all Participants/20 Participants) = 720 ongoing annual burden hours for each Participant to continue management of the Central Repository.

Additionally, the Commission estimates that the Participants will collectively spend $800,000 annually on external public relations, legal and consulting costs associated with the continued management of the Central Repository, or $40,000 per Participant.

The Commission is basing this external cost estimate on the public relations, legal and consulting external cost estimate provided in the CAT NMS Plan associated with the preparation of the CAT NMS Plan (which the Participants consider “reasonably associated with creating, implementing, and maintaining the CAT upon the Commission's adoption of the CAT NMS Plan”). See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii) (stating “the Participants have incurred public relations, legal and consulting costs in preparation of the CAT NMS Plan. The Participants estimate the costs of these services to be $8,800,000”). $2,400,000 for all Participants over 12 months = ($8,800,000/44 months between the adoption of Rule 613 and the filing of the CAT NMS Plan) × (12 months). Because the Central Repository will have already been created, the Commission believes it is reasonable to assume that the Participants will have a lesser need for public relations, legal and consulting services. The Commission is estimating that the Participants will incur one-third of the external cost associated with development and implementation of the Central Repository to maintain the Central Repository. $800,000 = (0.333) × ($2,400,000). ($800,000/20 Participants) = $40,000 per Participant over 12 months.

The CAT NMS Plan includes the estimates the six Shortlisted Bidders provided for the annual ongoing costs to the Participants to operate the Central Repository. The CAT NMS Plan did not categorize the costs included in the ongoing costs, but the Commission believes they would comprise external technological, operational and administrative costs, as the Participants described the costs included in the initial one-time external cost to build the Central Repository. Using these estimates, the Commission preliminarily estimates that the annual ongoing cost to the Participants to compensate the Plan Processor for building, operating and maintaining the Central Repository would be an aggregate ongoing external cost of $93 million, or approximately $4.7 million per Participant. Therefore, the Commission preliminarily estimates that each Participant would incur ongoing annual external costs of $4,740,000 to maintain the Central Repository, or aggregate ongoing annual external costs across all Participants of $94,800,000.

See Section IV.F.1.a, supra, for a discussion of the total five-year operating costs for the Central Repository presented in the CAT NMS Plan. See also CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(B); supra note 840; supra note 1467.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(B).

See supra note 1469.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(B).

The Bidders provided a range of estimates. For purposes of this Paperwork Burden Act analysis, the Commission is using the maximum operation and maintenance cost estimate. $4,650,000 = $93,000,000/20 Participants. See also Section IV.F.1.a, supra. The Commission noted several uncertainties that may affect the Central Repository cost estimates, including (1) that the Participants have not yet selected a Plan Processor and the Shortlisted Bidders have submitted a wide range of cost estimates for building and operating the Central Repository; (2) the Bids submitted by the Shortlisted Bidders may not be final because they may be revised before the final selection of the CAT Processor; and (3) neither the Bidders nor the Commission can anticipate the evolution of technology and market activity with precision, as improvements in available technology may allow the Central Repository to be built and operated at a lower cost than is currently anticipated, but if levels of anticipated market activity are materially underestimated, the capacity of the Central Repository may need to be increased, resulting in an increase in costs.

$4,740,000 for each Participant to build the Central Repository = ($4.7 million per Participant in ongoing annual costs to build the Central Repository) + ($40,000 per Participant in ongoing annual public relations, legal and consulting costs associated with the maintenance of the Central Repository).

$94,800,000 for all of the Participants to maintain the Central Repository = ($4,740,000 per Participant to compensate the Plan Processor and for external public relations, legal and consulting costs associated with the maintenance of the Central Repository) × (20 Participants). Id.

b. Data Collection and Reporting

Rule 613(c)(1) requires the CAT NMS Plan to provide for an accurate, time-sequenced record of orders beginning with the receipt or origination of an order by a Participant, and further to document the life of the order through the process of routing, modification, cancellation and execution (in whole or in part) of the order. Rule 613(c) requires the CAT NMS Plan to impose requirements on Participants to record and report CAT information to the Central Repository in accordance with specified timelines.

Rule 613(c) would require the collection and reporting of some information that Participants already collect to operate their business and are required to maintain in compliance with Section 17(a) of the Exchange Act and Rule 17a-1 thereunder. For instance, the Commission believes that the national securities exchanges keep records pursuant to Section 17(a) of the Exchange Act and Rule 17a-1 thereunder in electronic form, of the receipt of all orders entered into their systems, as well as records of the routing, modification, cancellation, and execution of those orders. However, Rule 613 requires the Participants to collect and report additional and more detailed information, and to report the information to the Central Repository in a uniform electronic format, or in a manner that would allow the Central Repository to convert the data to a uniform electronic format for consolidation and storage.

The CAT NMS Plan provides estimated costs for the Participants to report CAT Data. These estimates are based on Participant responses to the Participants Study that the Participants collected to estimate CAT-related costs for hardware and software, FTE costs, and third-party providers, if the Commission approves the CAT NMS Plan. For these estimates, the Commission is relying on the cost data provided by the Participants because it believes that the Plan's estimates for Participants to report CAT Data are reliable since all of the Participants provided cost estimates, and most Participants have experience collecting audit trail data, as well as knowledge of both the requirements of Rule 613 as well as their current business practices. The Commission is providing below its paperwork burden estimates for the initial burden hours and external costs, and ongoing, annual burden hours and external costs to be incurred by the Participants to comply with the data reporting requirements of Rule 613.

Third-party provider costs are generally legal and consulting costs, but may include other outsourcing. The template used by respondents is available at http://catnmsplan.com/PastEvents/ under the Section titled “6/23/14” at the “Cost Study Working Template” link.

The Commission notes that throughout this Paperwork Reduction Act analysis, it is categorizing the FTE cost estimates for the Participants, as well as the broker-dealer respondents, that were provided in the CAT NMS Plan as an internal burden. To convert the FTE cost estimates into internal burden hours, the Commission: (1) Divided the FTE cost estimates by a divisor of $424,350, which is the Commission's estimated average salary for a full-time equivalent employee in the securities industry in a job category associated with regulatory data reporting; and then (2) multiplied the quotient by 1,800 (the number of hours a full-time equivalent employee is estimated to work per year).

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(ii)(C) at n.192. The Participants represented that the cost per FTE is $401,440. The $401,440 figure used in the CAT NMS plan was based on a Programmer Analyst's salary ($193 per hour) from SIFMA's Management & Professional Earnings in the Securities Industry 2008, multiplied by 40 hours per week, then multiplied by 52 weeks per year. The Commission has updated this number to include recent salary data for other job categories associated with regulatory data reporting in the securities industry, using the hour and multiple methodology used by the Commission in its paperwork burden analyses. The Commission is using $424,350 as its annual cost per FTE for purposes of its cost estimates. The $424,350 FTE cost = 25% Compliance Manager + 75% Programmer Analyst (0.25) × ($283 per hour × 1,800 working hours per year) + (0.75) × ($220 per hour × 1,800 working hours per year). The $282 per hour figure for a Compliance Manager and the $220 per hour figure for a Programmer Analyst are from SIFMA's Management & Professional Earnings in the Securities Industry 2013, modified by the Commission to account for an 1800-hour work-year and multiplied by 5.35 to account for bonuses, firm size, employee benefits and overhead.

(1) Initial Burden Hours and External Cost

The CAT NMS Plan provides the following average costs that the Participants would expect to incur to adopt the systems changes needed to comply with the data reporting requirements of the consolidated audit trail: $10,300,000 in aggregate FTE costs for internal operational, technical/development, and compliance functions; $770,000 in aggregate third party legal and consulting costs; and $17,900,000 in aggregate total costs.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(B)(2). Of the $17,900,000 in aggregate total costs, $11,070,000 is identified (subtotal of FTE costs and outsourcing), but the remaining $6,830,000 is not identified in the CAT NMS Plan. The Commission believes that the $6,830,000 may be attributed to hardware costs because the Participants have not provided any hardware costs associated with data reporting elsewhere and the Commission believes that the Participants will likely incur external costs to purchase upgraded hardware to report data to the Central Repository.

Based on estimates provided in the CAT NMS Plan, the Commission preliminarily estimates that the initial internal burden hours to develop and implement the needed systems changes to capture the required information and transmit it to the Central Repository in compliance with the Rule for each Participant would be approximately 2,185 burden hours. The Commission also estimates that each Participant would, on average, incur approximately $38,500 in initial third party legal and consulting costs for a total of $380,000 in initial external costs. Therefore, the Commission preliminarily estimates that, for all Participants, the estimated aggregate one-time burden would be 43,690 hours and the estimated aggregate initial external cost would be $7,600,000.

($10,300,000 anticipated initial FTE costs)/(20 SROs) = $515,000 in anticipated initial FTE costs per Participant. ($515,000 in anticipated initial FTE costs per Participant)/($424,350 FTE costs per Participant) = 1.214 anticipated FTEs per Participant for the implementation of data reporting. (1.214 FTEs) × (1,800 working hours per year) = 2,184.5 initial burden hours per Participant to implement CAT Data reporting.

($770,000 anticipated initial third party costs)/(20 Participants) = $38,500 in initial anticipated third party costs per Participant.

To determine the total initial external cost per Participant, the Commission subtracted the anticipated initial FTE cost estimates for the Participants as provided in the Plan from the total aggregate initial costs to the Participants and divided the remainder by 20 Participants. ($17,900,000 total aggregate initial cost to Participants) − ($10,300,000 initial FTE cost to Participants) = $7,600,000. ($7,600,000)/20 Participants = $380,000 in initial external costs per Participant. See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(B)(1) for the Participants' anticipated costs associated with the implementation of regulatory reporting to the Central Repository.

43,690 initial burden hours = (20 Participants) × (2,184.5 initial burden hours).

$7,600,000 = ($380,000 in initial external costs) × (20 Participants).

(2) Ongoing, Annual Burden Hours and External Cost

Once a Participant has established the appropriate systems and processes required for collection and transmission of the required information to the Central Repository, the Commission preliminarily estimates that Rule 613 would impose on each Participant ongoing annual burdens associated with, among other things, personnel time to monitor each Participant's reporting of the required data and the maintenance of the systems to report the required data; and implementing changes to trading systems that might result in additional reports to the Central Repository. The CAT NMS Plan provides the following average aggregate costs that the Participants would expect to incur to maintain data reporting systems to be in compliance with Rule 613: $7,300,000 in anticipated annual FTE costs for operational, technical/development, and compliance functions related to data reporting; $720,000 in annual third party legal, consulting, and other costs; and $14,700,000 total annual costs.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(B)(2). The CAT NMS Plan did not identify the other costs.

Of the $14,700,000 in aggregate total annual costs, $8,020,000 is identified (subtotal of FTE costs and outsourcing), but the remaining $6,680,000 is not identified in the CAT NMS Plan. The Commission believes that this amount may be attributed to hardware costs because the Participants have not provided any hardware costs associated with data reporting elsewhere and the Commission believes that the Participants will likely incur costs to upgrade their hardware to report data to the Central Repository.

Based on estimates provided in the CAT NMS Plan, the Commission believes that it would take each Participant 1,548 ongoing burden hours per year to continue compliance with Rule 613. The Commission preliminarily estimates that it would cost, on average, approximately $36,000 in ongoing third party legal and consulting and other costs and $370,000 in total ongoing external costs per Participant. Therefore, the Commission preliminarily estimates that the estimated aggregate ongoing burden for all Participants would be approximately 30,966 hours and an estimated aggregate ongoing external cost of $7,400,000.

($7,300,000 in anticipated Participant annual FTE costs)/(20 Participants) = $365,000 in anticipated per Participant annual FTE costs. ($365,000 in anticipated per Participant FTE costs)/($424,350 FTE cost per Participant) = 0.86 anticipated FTEs per Participant. (0.86 FTEs) × (1,800 working hours per year) = 1,548.3 burden hours per Participant to maintain CAT Data reporting.

($720,000 in annual third party costs)/(20 Participants) = $36,000 per Participant in anticipated annual third party costs.

To determine the total external annual cost per Participant, the Commission subtracted the anticipated annual FTE cost estimates for the Participants as provided in the Plan from the total aggregate annual costs to the Participants and divided the remainder by 20 Participants. ($14,700,000 total aggregate annual cost to Participants) − ($7,300,000 annual FTE cost to Participants) = $7,400,000. ($7,400,000)/20 Participants = $370,000 in annual external costs per Participant. See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(B)(1) for the Participants' anticipated maintenance costs associated with regulatory reporting to the Central Repository.

30,966 annual burden hours = (20 Participants) × (1,548.3 annual burden hours).

$7,400,000 = ($370,000 in total annual external costs) × (20 Participants).

c. Collection and Retention of NBBO, Last Sale Data and Transaction Reports

Rule 613(e)(7) provides that the CAT NMS Plan must require the Central Repository to collect and retain on a current and continuous basis NBBO information for each NMS security, transaction reports reported pursuant to an effective transaction reporting plan, and Last Sale Reports reported pursuant to the OPRA Plan. Additionally, the CAT NMS Plan must require the Central Repository to maintain this data in a format compatible with the order and event information consolidated and stored pursuant to Rule 613(c)(7). Further, the CAT NMS Plan must require the Central Repository to retain the information collected pursuant to paragraphs (c)(7) and (e)(7) of Rule 613 for a period of not less than five years in a convenient and usable uniform electronic format that is directly available and searchable electronically without any manual intervention. The Commission notes that the CAT NMS Plan includes these data as “SIP Data” to be collected by the Central Repository. The Commission believes the burden associated with SIP Data is included in the burden to the Participants associated with the implementation and maintenance of the Central Repository.

Id.

See CAT NMS Plan, supra note 3, at Section 6.5(a)(ii).

d. Surveillance

Rule 613(f) provides that the CAT NMS Plan must require that every national securities exchange and national securities association develop and implement a surveillance system, or enhance existing surveillance systems, reasonably designed to make use of the consolidated information contained in the consolidated audit trail. Rule 613(a)(3)(iv) provides that the CAT NMS Plan must require that the surveillance systems be implemented within fourteen months after effectiveness of the CAT NMS Plan.

(1) Initial Burden Hours and External Cost

The CAT NMS Plan states that the estimated total cost to the Participants to implement surveillance programs within the Central Repository is $23,200,000. This amount includes legal, consulting, and other costs of $560,000, as well as $17,500,000 in FTE costs for operational, technical/development, and compliance Staff to be engaged in the creation of surveillance programs.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(B)(2).

Id. The Commission also notes that based upon the data provided by the Participants, the source of the remaining $5,140,000 in initial costs to implement new or enhanced surveillance systems is unspecified. The Commission believes that this amount may be attributed to hardware costs because the Participants have not provided any hardware costs associated with surveillance elsewhere and the Commission believes that the Participants will likely incur costs to implement new or enhanced surveillance systems reasonably designed to make use of the consolidated audit trail data.

Based on the estimates provided in the CAT NMS Plan, the Commission preliminarily estimates that the initial internal burden hours to implement new or enhanced surveillance systems reasonably designed to make use of the consolidated audit trail data for each Participant would be approximately 3,711.6 burden hours, for an aggregate initial burden hour amount of 74,232 burden hours. The Commission also estimates that each Participant would, on average, incur an initial external cost of approximately $28,000 for outsourced legal, consulting and other costs in order to implement new or enhanced surveillance systems, for a total of $285,000 in initial external costs, for an aggregate one-time initial external cost of $5,700,000 across the 20 Participants to implement new or enhanced surveillance systems.

($17,500,000 in anticipated initial FTE costs)/(20 Participants) = $875,000 in anticipated FTE costs per Participant. ($875,000 in anticipated initial FTE costs per Participant)/($424,350 FTE cost per Participant) = 2.06 anticipated initial FTEs per Participant. (2.06 FTEs) × (1,800 working hours per year) = 3,711.6 initial burden hours per Participant to implement new or enhanced surveillance systems.

(3,711.6 initial burden hours per Participant to implement new or enhanced surveillance systems) × (20 Participants) = 74,232 aggregate initial burden hours.

$28,000 = $560,000/20 Participants.

$285,000 = ($23,200,000 in total initial surveillance costs—$17,500,000 in FTE costs)/(20 Participants).

$5,700,000 = $285,000 × 20 Participants.

(2) Ongoing, Annual Burden Hours and External Cost

The CAT NMS Plan states that the estimated total annual cost associated with the maintenance of surveillance programs for the Participants is $87,700,000. This amount includes annual legal, consulting, and other costs of $1,000,000, as well as $66,700,000 in annual FTE costs for internal operational, technical/development, and compliance Staff to be engaged in the maintenance of surveillance programs. Based on the estimates provided in the CAT NMS Plan, the Commission preliminarily estimates that the ongoing internal burden hours to maintain the new or enhanced surveillance systems reasonably designed to make use of the consolidated audit trail data for each Participant would be approximately 14,146 annual burden hours, for an aggregate annual burden hour amount of 282,920 burden hours. The Commission also estimates that each Participant would, on average, incur an annual external cost of approximately $50,000 for outsourced legal, consulting and other costs in order to maintain the new or enhanced surveillance systems, for a total estimated ongoing external cost of $1,050,000, for an estimated aggregate ongoing external cost of $21,000,000 across the 20 Participants to maintain the surveillance systems.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(B)(2).

Id. The Commission also notes that based upon the data provided by the Participants, the source of the remaining $21,000,000 in ongoing costs to maintain the new or enhanced surveillance systems is unspecified. The Commission believes that this amount may be attributed to hardware costs because the Participants have not provided any hardware costs associated with surveillance elsewhere and the Commission believes that the Participants would likely incur costs associated with maintaining the new or enhanced surveillance systems.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(B)(2).

($66,700,000 in anticipated ongoing FTE costs)/(20 Participants) = $3,335,000 in anticipated ongoing FTE costs per Participant. ($3,335,000 in anticipated ongoing FTE costs per Participant)/($424,350 FTE cost per Participant) = 7.86 anticipated FTEs per Participant. (7.86 FTEs) × (1,800 working hours per year) = 14,146 ongoing burden hours per Participant to maintain the new or enhanced surveillance systems.

(14,146 annual burden hours per Participant to maintain new or enhanced surveillance systems) × (20 Participants) = 282,920 aggregate annual burden hours.

$50,000 = $1,000,000 for ongoing legal, consulting and other costs associated with maintenance of surveillance programs/20 Participants.

$1,050,000 = ($87,700,000 in total ongoing surveillance costs−$66,700,000 in ongoing FTE costs)/20 Participants

$21,000,000 = $1,050,000 × 20 Participants.

e. Written Assessment of Operation of the Consolidated Audit Trail

Rule 613(b)(6) provides that the CAT NMS Plan must require the Participants to provide the Commission a written assessment of the CAT's operation at least every two years, once the CAT NMS Plan is effective. The assessment must address, at a minimum, with respect to the consolidated audit trail: (i) An evaluation of its performance; (ii) a detailed plan for any potential improvements to its performance; (iii) an estimate of the costs associated with any such potential improvements; and (iv) an estimated implementation timeline for any such potential improvements, if applicable. Thus, the Participants must, among other things, undertake an analysis of the consolidated audit trail's technological and computer system performance.

17 CFR 242.613(b)(6). See also Section IV.E.3.a, supra.

The CAT NMS Plan states that the CCO would oversee the assessment required by Rule 613(b)(6), and would allow the Participants to review and comment on the assessment before it is submitted to the Commission. The CCO would be an employee of the Plan Processor and would be compensated by the Plan Processor. The Commission assumes that the overall cost and associated burden on the Participants to implement and maintain the Central Repository includes both the compensation for the Plan Processor as well as its employees for the implementation and maintenance of the Central Repository.

See CAT NMS Plan, supra note 3, at Section 6.6.

Id. at Section 6.2(a).

The Commission preliminarily estimates that it would take each Participant approximately 45 annual burden hours of internal legal, compliance, business operations, and information technology staff time to review and comment on the assessment prepared by the CCO of the operation of the consolidated audit trail as required by Rule 613(b)(6). The Commission preliminarily estimates that on average, each Participant would outsource 1.25 hours of legal time annually to assist in the review of the assessment, for an ongoing annual external cost of approximately $500. Therefore, the Commission preliminarily estimates that the ongoing annual burden of submitting a written assessment at least every two years, as required by Rule 613(b)(6), would be 45 ongoing burden hours per SRO plus $500 of external costs for outsourced legal counsel per Participant per year, for an estimated aggregate annual ongoing burden of 900 hours and an estimated aggregate ongoing external cost of $10,000.

The Commission calculated the total estimated burden hours based on a similar formulation used for calculating the total estimated burden hours of Rule 613(i)'s requirement for a document addressing expansion of the CAT to other securities. See Section V.D.1.f., infra. The Commission assumes that the review and potential revision of the written assessment required by Rule 613(b)(6) would be approximately one-half as burdensome as the document required by Rule 613(i) as the Participants are delegating the responsibility to prepare the written assessment required by Rule 613(b)(6) to the CCO and the Participants would only need to review the written assessment and revise it as necessary. As noted in note 1530, infra, to estimate the Rule 613(i) burden, the Commission is applying the internal burden estimate provided in the CAT NMS Plan for Plan development over a 6-month period, and dividing the result in half. See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii). To estimate the Rule 613(b)(6) written assessment burden, the Commission is dividing the result further by half. 0.667 FTEs required for all Participants per month to develop the CAT NMS Plan = (20 FTEs/30 months). 0.667 FTEs × 6 months = 4 FTEs. 4 FTEs/2 = 2 FTEs needed for all of the Participants to create and submit the Rule 613(i) document. 2 FTEs/2 = 1 FTE needed for all of the Participants to review and comment on the written assessment. (1 FTE × 1,800 working hours per year) = 1,800 ongoing annual burden hours per year for all of the Participants to review and comment on the written assessment. (1,800 burden hours/20 Participants) = 90 ongoing annual burden hours per Participant to review and comment on the written assessment prepared by the CCO. The Commission notes that this assessment must be filed with the Commission every two years and is providing an annualized estimate of the burden associated with the assessment as required for its Paperwork Reduction Act analysis. To provide an estimate of the annual burden associated with the assessment as required for its Paperwork Reduction Act analysis, Commission is dividing the 90 ongoing burden hours in half (over two years) = 45 ongoing annual burden hours per Participant to review and comment on the written assessment prepared by the CCO.

$500 = ($400 per hour rate for outside legal services) × (1.25 hours). The Commission based this estimate on the assumption that the written assessment required by Rule 613(b)(6) would require approximately one-half the effort of drafting and submitting the document required by Rule 613(i) regarding the expansion of the CAT to other securities because the Participants have delegated the responsibility to draft the written assessment on the CCO, rather than having to draft it themselves (as with the expansion report), but would also have to review the written assessment and revise it as necessary. See Section V.D.1.f., infra. Because the written assessment is a biennial requirement, the Commission is further dividing the cost of the written assessment in half (over two years) to estimate the annual ongoing external cost per Participant for outside legal services to review and comment on the written assessment prepared by the CCO.

900 ongoing annual burden hours = (45 ongoing annual burden hours) × (20 Participants).

$10,000 = 20 Participants × ($400 per hour rate for outside legal services) × (1.25 hours).

f. Document on Expansion to Other Securities

Rule 613(i) provides that the CAT NMS Plan must require the Participants to jointly provide to the Commission, within six months after the CAT NMS Plan is effective, a document outlining how the Participants could incorporate into the consolidated audit trail information regarding: (1) Equity securities that are not NMS securities; (2) debt securities; and (3) primary market transactions in equity securities that are not NMS securities and debt securities. The document must also detail the order and Reportable Event data that each market participant may be required to provide, which market participants may be required to provide such data, an implementation timeline, and a cost estimate. Thus, the Participants must, among other things, undertake an analysis of technological and computer system acquisitions and upgrades that would be required to incorporate such an expansion.

As noted above, the CAT NMS Plan would require the inclusion of OTC Equity Securities, while Rule 613 does not include such a requirement. See supra note 1408.

The Commission preliminarily estimates that it would take each Participant approximately 180 burden hours of internal legal, compliance, business operations and information technology staff time to create a document addressing expansion of the consolidated audit trail to additional securities as required by Rule 613(i). The Commission preliminarily estimates that on average, each Participant would outsource 25 hours of external legal time to create the document, for an aggregate one-time external cost of approximately $10,000. Therefore, the Commission preliminarily estimates that the one-time initial burden of drafting the document required by Rule 613 would be 180 initial burden hours plus $10,000 in initial external costs for outsourced legal counsel per Participant, for an estimated aggregate initial burden of 3,600 hours and an estimated aggregate initial external cost of $200,000.

The Commission is basing this estimate on the internal burden provided in the CAT NMS Plan related to the development of the CAT NMS Plan. See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii) (stating “[t]he Participants estimate that they have collectively contributed 20 FTEs in the first 30 months of the CAT NMS Plan development process”). Because this document is much more limited in scope than the CAT NMS Plan, and because the Commission assumes that in drafting the CAT NMS Plan, the Participants have already contributed time toward considering how the CAT can be expected to be expanded in accordance with Rule 613(i), the Commission is applying the CAT NMS Plan development internal burden over a 6-month period (Rule 613(i) requires this document to be submitted to the Commission within six months after effectiveness of the CAT NMS Plan), divided by half. 0.667 FTEs required for all Participants per month to develop the CAT NMS Plan = (20 FTEs/30 months). 0.667 FTEs × 6 months = 4 FTEs. 4 FTEs/2 = 2 FTEs needed for all of the Participants to create and submit the document. 2 FTEs × 1,800 working hours per year = 3,600 burden hours. 3,600 burden hours/20 Participants = 180 burden hours per Participant to create and file the document.

$10,000 = (25 hours of outsourced legal time per Participant) × ($400 per hour rate for outside legal services). The Commission derived the total estimated cost for outsourced legal counsel based on the assumption that the report required by Rule 613 would require approximately fifteen percent of the Commission's approximated burden of drafting and filing the CAT NMS Plan. This assumption is based on the Participants leveraging their knowledge gained from their drafting and filing of the CAT NMS Plan and applying it to efficiently preparing the report required by Rule 613 with respect to other securities' order and Reportable Events, implementation timeline and cost estimates.

The initial burden hour estimate is based on: (20 Participants) × (180 initial burden hours to draft the report). The initial external cost estimate is based on: (20 Participants) × ($10,000 for outsourced legal counsel).

2. Burden on Members of National Securities Exchanges and National Securities Associations

a. Data Collection and Reporting

Rule 613(c)(1) requires the CAT NMS Plan to provide for an accurate, time-sequenced record of orders beginning with the receipt or origination of an order by a broker-dealer member of a Participant, and further documenting the life of the order through the process of routing, modification, cancellation and execution (in whole or in part) of the order. Rule 613(c) requires the CAT NMS Plan to impose requirements on broker-dealer members to record and report CAT information to the Central Repository in accordance with specified timelines.

The Commission acknowledges the inherent difficulty in establishing precise burden estimates because the Commission does not know the exact method of data reporting the Participants would decide for broker-dealers. For these estimates, the Commission is relying, in part, on the cost data provided by the Participants in the CAT NMS Plan, and, as noted earlier, on its own estimates of the costs that broker-dealers are likely to face for CAT implementation and ongoing reporting in compliance with Rule 613.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b).

See Sections IV.F.1.c(1) and IV.F.1.c(2), supra.

The Commission's estimates delineate broker-dealer firms by whether they insource or outsource, or are likely to insource or outsource, CAT Data reporting obligations. The Commission preliminarily believes that firms that currently report high numbers of OATS ROEs strategically would decide to either self-report their CAT Data or outsource their CAT Data reporting functions, while the firms with the lowest levels of activity would be unlikely to have the infrastructure and specialized employees necessary to insource CAT Data reporting and would almost certainly outsource their CAT Data reporting functions. The Commission recognizes that more active firms that will likely be CAT Reporters and insource regulatory data reporting functions may not have current OATS reporting obligations because they either are not FINRA members, or because they do not trade in NMS equity securities.

See Section IV.F.1.c(2)B, supra.

Id.

The Commission also preliminarily recognizes as discussed above that some broker-dealer firms may strategically choose to outsource despite the Plan's working assumption that these broker-dealers would insource their regulatory data reporting functions.

The Commission preliminarily estimates that there are 126 OATS-reporting Insourcers and 45 non-OATS reporting Insourcers. The Commission's estimation categorizes the remaining 1,629 broker-dealers that the Plan anticipates would have CAT Data reporting obligations as Outsourcers.

See Section IV.F.1.c(2)B, infra.

Id.

(1) Insourcers

A. Large Non-OATS Reporting Broker-Dealers

i. Initial Burden Hours and External Cost

The Commission relies on the Reporters Study's large broker-dealer cost estimates in estimating costs for large broker-dealers that can practicably decide between insourcing or outsourcing their regulatory data reporting functions. The Commission estimates that there are 14 large broker-dealers that are not OATS reporters currently in the business of electronic liquidity provision that would be classified as Insourcer firms.

These broker-dealers are not FINRA members and thus have no regular OATS reporting obligations. See supra note 937.

Additionally, the Commission estimates that there are 31 broker-dealers that may transact in options but not in equities that can be classified as Insourcer firms. Although the Exemptive Relief may relieve these firms of the obligation to report their option quoting activity to the Central Repository, these firms may have customer orders and other activity off-exchange that would cause them to incur a CAT reporting obligation.

See supra note 939.

The Commission assumes the 31 options firms and 14 ELPs would be typical of the Reporters Study's large, non-OATS reporting firms; for these firms, the Commission relies on the cost estimates provided under Approach 1 for large, non-OATS reporting firms in the CAT NMS Plan.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(i)(A)(2). The Reporters Study requested broker-dealer respondents to provide estimates to report to the Central Repository under two approaches. Approach 1 assumes CAT Reporters would submit CAT Data using their choice of industry protocols. Approach 2 assumes CAT Reporters would submit data using a pre-specified format. Approach 1's aggregate costs are higher than those for Approach 2 for all market participants except in one case where service bureaus have lower Approach 1 costs. See supra note 946. For purposes of this Paperwork Reduction Act analysis, the Commission is not relying on the cost estimates for Approach 2 because overall the Approach 1 aggregate estimates represent the higher of the proposed approaches. The Commission believes it would be more comprehensive to use the higher of the two estimates for its Paperwork Reduction Act analysis estimates.

The CAT NMS Plan provides the following average initial external cost and FTE count figures that a large non-OATS reporting broker-dealer would expect to incur to adopt the systems changes needed to comply with the data reporting requirements of Rule 613 under Approach 1: $450,000 in external hardware and software costs; 8.05 internal FTEs; and $9,500 in external third party/outsourcing costs. Based on this information, the Commission preliminarily estimates that the average initial burden associated with implementing regulatory data reporting to capture the required information and transmit it to the Central Repository in compliance with the Rule for each large, non-OATS reporting broker-dealer would be approximately 14,490 initial burden hours.

Approach 1 also provided $3,200,000 in initial internal FTE costs. The Commission believes the $3,200,000 in internal FTE costs is the Participants' estimated cost of the 8.05 FTEs. (8.05 FTEs) × ($401,440 Participants' assumed annual cost per FTE provided in the CAT NMS Plan) = $3,231,592. See CAT NMS Plan, supra note 3, at n. 192. See also supra note 1487.

See CAT NMS Plan, supra note 3, at Section B.7(b)(iii)(c)(2)(a). The Commission believes that the third party/outsourcing costs may be attributed to the use of service bureaus (potentially), technology consulting, and legal services.

14,490 initial burden hours = (8.05 FTEs for implementing CAT Data reporting systems) × (1,800 working hours per year).

The Commission also preliminarily estimates that these broker-dealers would, on average, incur approximately $450,000 in initial costs for hardware and software to implement the systems changes needed to capture the required information and transmit it to the Central Repository, and an additional $9,500 in initial third party/outsourcing costs. Therefore, the Commission preliminarily estimates that the average one-time initial burden per ELP and options market-making firm would be 14,490 internal burden hours and external costs of $459,500, for an estimated aggregate initial burden of 652,050 hours and an estimated aggregate initial external cost of $20,677,500.

See supra note 1544.

($450,000 in initial hardware and software costs) + ($9,500 initial third party/outsourcing costs) = $459,500 in initial external costs to implement data reporting systems.

The Commission preliminarily estimates that 45 large non-OATS reporting broker-dealers would be impacted by this information collection. (45 large non-OATS reporting broker-dealers) × (14,490 burden hours) = 652,050 initial burden hours to implement data reporting systems.

($450,000 in hardware and software costs) + ($9,500 third party/outsourcing costs) × 45 large, non-OATS reporting broker-dealers = $20,677,500 in initial external costs to implement data reporting systems.

ii. Ongoing, Annual Burden Hours in External Cost

Once a large non-OATS reporting broker-dealer has established the appropriate systems and processes required for collection and transmission of the required information to the Central Repository, the Commission preliminarily estimates that the Rule would impose ongoing annual burdens associated with, among other things, personnel time to monitor each large non-OATS reporting broker-dealer's reporting of the required data and the maintenance of the systems to report the required data; and implementing changes to trading systems that might result in additional reports to the Central Repository. The CAT NMS Plan provides the following average ongoing external cost and internal FTE count figures that a large non-OATS reporting broker-dealer would expect to incur to maintain data reporting systems to be in compliance with Rule 613: $80,000 in external hardware and software costs; 7.41 internal FTEs; and $1,300 in external third party/outsourcing costs. Based on this information, the Commission preliminarily believes that it would take a large non-OATS reporting broker-dealer 13,338 burden hours per year to continue to comply with the Rule. The Commission also preliminarily estimates that it would cost, on average, approximately $80,000 per year per large non-OATS reporting broker-dealer to maintain systems connectivity to the Central Repository and purchase any necessary hardware, software, and other materials, and an additional $1,300 in third party/outsourcing costs.

Approach 1 also provided $3,000,000 in internal FTE costs related to maintenance. The Commission believes the $3,000,000 in ongoing internal FTE costs is the Participants' estimated cost of the 7.41 FTEs. (7.41 FTEs) × ($401,440 Participants' assumed annual cost per FTE provided in the CAT NMS Plan) = $2,974,670. See CAT NMS Plan, supra note 3, at n.192. See also supra note 1487.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(C)(2)(b). The CAT NMS Plan did not break down these third party costs into categories.

13,338 ongoing burden hours = (7.41 ongoing FTEs to maintain CAT data reporting systems) × (1,800 working hours per year).

See supra note 1544; CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(C)(2)(b).

Therefore, the Commission preliminarily estimates that the average ongoing annual burden per large non-OATS reporting broker-dealer would be approximately 13,338 hours, plus $81,300 in external costs to maintain the systems necessary to collect and transmit information to the Central Repository, for an estimated aggregate ongoing burden of 600,210 hours and an estimated aggregate ongoing external cost of $3,658,500.

($80,000 in ongoing external hardware and software costs) + ($1,300 ongoing external third party/outsourcing costs) = $81,300 in ongoing external costs per large non-OATS reporting broker-dealer.

The Commission estimates that 45 large non-OATS reporting broker-dealers would be impacted by this information collection. (45 large non-OATS reporting broker-dealers) × (13,338 burden hours) = 600,210 aggregate ongoing burden hours.

($80,000 in ongoing external hardware and software costs) + ($1,300 ongoing external third party/outsourcing costs) × (45 large non-OATS reporting broker-dealers) = $3,658,500 in aggregate ongoing external costs.

B. Large OATS-Reporting Broker-Dealers

i. Initial Burden Hours and External Cost

Based on the Commission's analysis of data provided by FINRA and discussions with market participants, the Commission estimates that 126 broker-dealers, which reported more than 350,000 OATS ROEs between June 15 and July 10, 2015, would strategically decide to either self-report CAT Data or outsource their CAT data reporting functions. To conduct its Paperwork Burden Analysis for the 126 broker-dealers, the Commission is relying on the Reporters Study estimates used by the CAT NMS Plan of expected costs that a large OATS-reporting broker-dealer would incur as a result of the implementation of the consolidated audit trail under Approach 1.

See Section IV.F.1.c.2.B and Section IV.F.1.c(2)B.i, supra. See also supra note 901, stating that the Commission believes that broker-dealers that report fewer than 350,000 OATS ROEs per month are unlikely to be large enough to support the infrastructure required for insourcing data reporting activities.

See supra note 1544.

The CAT NMS Plan provides the following average initial external cost and internal FTE count figures that a large OATS-reporting broker-dealer would expect to incur as a result of the implementation of the consolidated audit trail under Approach 1: $750,000 in hardware and software costs; 14.92 internal FTEs; and $150,000 in external third party/outsourcing costs. Based on this information the Commission preliminarily estimates that the average initial burden to develop and implement the needed systems changes to capture the required information and transmit it to the Central Repository in compliance with the Rule for large OATS-reporting broker-dealers would be approximately 26,856 internal burden hours. The Commission also preliminarily estimates that these large OATS-reporting broker-dealers would, on average, incur approximately $750,000 in initial external costs for hardware and software to implement the systems changes needed to capture the required information and transmit it to the Central Repository, and an additional $150,000 in initial external third party/outsourcing costs.

Approach 1 also provided $6,000,000 in initial internal FTE costs. The Commission preliminarily believes the $6,000,000 in initial internal FTE costs is the Participants' estimated cost of the 14.92 FTEs. (14.92 FTEs) × ($401,440 Participants' assumed annual cost per FTE provided in the CAT NMS Plan) = $5,989,485. See CAT NMS Plan, supra note 3, at n. 192. See also supra note 1487.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(C)(2)(a). The CAT NMS Plan did not break down these third party costs into categories. The Commission preliminarily believes that these costs may be attributed to the use of service bureaus, technology consulting, and legal services.

26,856 initial burden hours per large OATS-reporting broker-dealer = (14.92 FTEs for implementation of CAT data reporting systems) × (1,800 working hours per year).

See CAT NMS Plan, supra note 3, at Section B.7(b)(iii)(C)(2)(a).

Therefore, the Commission preliminarily estimates that the average one-time initial burden per large OATS-reporting broker-dealer would be 26,856 burden hours and external costs of $900,000, for an estimated aggregate initial burden of 3,383,856 hours and an estimated aggregate initial external cost of $113,400,000.

($750,000 in initial external hardware and software costs) + ($150,000 initial external third party/outsourcing costs) = $900,000 in initial external costs per large OATS-reporting broker-dealer to implement CAT data reporting systems.

The Commission preliminarily estimates that 126 large OATS-reporting broker-dealers would be impacted by this information collection. 126 large OATS-reporting broker-dealers × 26,856 burden hours = 3,383,856 initial burden hours to implement data reporting systems.

($750,000 in initial external hardware and software costs) + ($150,000 initial external third party/outsourcing costs) × 126 large OATS-reporting broker-dealers = $113,400,000 in initial external costs to implement data reporting systems.

ii. Ongoing, Annual Burden Hours and External Cost

Once a large OATS-reporting broker-dealer has established the appropriate systems and processes required for collection and transmission of the required information to the Central Repository, the Commission preliminarily estimates that the Rule would impose on each broker-dealer ongoing annual burdens and costs associated with, among other things, personnel time to monitor each broker-dealer's reporting of the required data and the maintenance of the systems to report the required data; and implementing changes to trading systems which might result in additional reports to the Central Repository.

The CAT NMS Plan provides the following average ongoing external cost and FTE count figures that a large OATS-reporting broker-dealer would expect to incur to maintain data reporting systems to be in compliance with Rule 613: $380,000 in ongoing external hardware and software costs; 10.03 internal FTEs; and $120,000 in ongoing external third party/outsourcing costs. Based on this information the Commission preliminarily believes that it would take a large OATS-reporting broker-dealer 18,054 ongoing burden hours per year to continue compliance with the Rule. The Commission preliminarily estimates that it would cost, on average, approximately $380,000 per year per large OATS-reporting broker-dealer to maintain systems connectivity to the Central Repository and purchase any necessary hardware, software, and other materials, and an additional $120,000 in external ongoing third party/outsourcing costs.

Approach 1 also provided $4,000,000 in internal FTE costs related to maintenance. The Commission believes the $4,000,000 in ongoing internal FTE costs is the Participants' estimated cost of the 10.03 FTEs. (10.03 FTEs) × ($401,440 Participants' assumed annual cost per FTE provided in the CAT NMS Plan) = $4,026,443. See CAT NMS Plan, supra note 3, at n. 192. See also supra note 1487.

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(C)(2)(b). The CAT NMS Plan did not categorize these third party costs. The Commission preliminarily believes that these costs may be attributed to the use of service bureaus, technology consulting, and legal services.

18,054 ongoing burden hours = (10.03 ongoing FTEs for maintenance of CAT data reporting systems) × (1,800 working hours per year).

See CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(iii)(C)(2)(b).

Therefore, the Commission preliminarily estimates that the average ongoing annual burden per large OATS-reporting broker-dealer would be approximately 18,054 burden hours, plus $500,000 in external costs to maintain the systems necessary to collect and transmit information to the Central Repository, for an estimated aggregate burden of 2,274,804 hours and an estimated aggregate ongoing external cost of $63,000,000.

($380,000 in ongoing external hardware and software costs + $120,000 in ongoing external third party/outsourcing costs) = $500,000 in ongoing external costs per large OATS-reporting broker-dealer.

The Commission preliminarily estimates that 126 large OATS-reporting broker-dealers would be impacted by this information collection. (126 large OATS-reporting broker-dealers) × (18,054 burden hours) = 2,274,804 aggregate ongoing burden hours.

($380,000 in ongoing external hardware and software costs + $120,000 in ongoing external third party/outsourcing costs) × 126 large OATS-reporting broker-dealers = $63,000,000 in aggregate ongoing external costs.

(2) Outsourcing Firms

A. Small OATS-Reporting Broker-Dealers

i. Initial Burden Hours and External Cost

Based on data provided by FINRA, the Commission estimates that there are 806 broker-dealers that report fewer than 350,000 OATS ROEs monthly. The Commission preliminarily believes that these broker-dealers generally outsource their regulatory reporting obligations because during the period June 15-July 10, 2015, approximately 88.9% of their 350,000 OATS ROEs were reported through service bureaus, with 730 of these broker-dealers reporting more than 99% of their OATS ROEs through one or more service bureaus. The Commission estimates that these firms currently spend an aggregate of $100.1 million on annual outsourcing costs. The Commission estimates these 806 broker-dealers would spend $100.2 million in aggregate to outsource their regulatory data reporting to service bureaus to report in accordance with Rule 613, or $124,373 per broker-dealer. These external outsourcing cost estimates are calculated using the information from Staff discussions with service bureaus and other market participants, as applied to data provided by FINRA.

See Section IV.F.1.c(2)B.i, supra. Because of the extensive use of service bureaus in these categories of broker-dealers, the Commission assumes that these broker-dealers are likely to use service bureaus to accomplish their CAT data reporting.

The average broker-dealer in this category reported 15,185 OATS ROEs from June 15-July 10, 2015; the median reported 1,251 OATS ROEs. Of these broker-dealers, 39 reported more than 100,000 OATS ROEs during the sample period. See Section IV.F.1.c(2)B.ii, supra.

Id.

$124,373 = $100,200,000/806 broker-dealers. This amount is the average estimated annual outsourcing cost to firms that currently report fewer than 350,000 OATS ROEs per month. Id.

See Section IV.F.1.c(2)B.ii, supra.

Firms that outsource their regulatory data reporting still face internal staffing burdens associated with this activity. These employees perform activities such as answering inquiries from their service bureaus, and investigating reporting exceptions. Based on conversations with market participants, the Commission estimates that these firms currently have 0.5 full-time employees devoted to these activities. The Commission estimates that these firms would need to hire one additional full-time employee for one year to implement CAT reporting requirements.

Id.

Id.

Based on this information, the Commission preliminarily estimates that the average initial burden to implement the needed systems changes to capture the required information and transmit it to the Central Repository in compliance with the CAT NMS Plan for small OATS-reporting broker-dealers would be approximately 1,800 burden hours. The Commission believes the burden hours would be associated with work performed by internal technology, compliance and legal staff in connection with the implementation of CAT data reporting. The Commission also preliminarily estimates that each small OATS-reporting broker-dealer would incur approximately

This estimate assumes that, based on the expected FTE count provided, a small OATS-reporting broker-dealer would have to hire 1 new FTE for implementation. The salary attributed to the 1 FTE would be (1 × $424,350 FTE cost) = $424,350 per year. To determine the number of burden hours to be incurred by the current 0.5 FTE for implementation, multiply 0.5 FTE by 1,800 hours per year = 900 initial burden hours.

$124,373 in initial external outsourcing costs. Therefore, the Commission preliminarily estimates that the average one-time initial burden per small OATS-reporting broker-dealer would be 1,800 burden hours and external costs of $124,373, for an estimated aggregate initial burden of 1,450,800 hours and an estimated aggregate initial external cost of $100,244,638.

See Section IV.F.1.c(2)B.ii, supra. The Commission preliminarily believes the outsourcing cost would be the cost of the service bureau, which would include the compliance and legal costs associated with changing to CAT Data reporting. The Commission assumes these costs of changing to CAT would be included in the cost of the service bureau because the broker-dealers would be relying on the expertise of the service bureau to report their data to CAT on their behalf. See supra note 941.

The Commission preliminarily estimates that 806 small OATS-reporting broker-dealers would be impacted by this information collection. (806 small OATS-reporting broker-dealers × 1,800 burden hours) = 1,450,800 aggregate initial burden hours.

($124,373 in outsourcing costs) × (806 small OATS-reporting broker-dealers) = $100,244,638 in aggregate initial external costs.

ii. Ongoing, Annual Burden Hours and External Cost

Small OATS-reporting broker-dealers that outsource their regulatory data reporting would likely face internal staffing burdens and external costs associated with ongoing activity, such as maintaining any systems that transmit data to their service providers. Based on conversations with market participants, the Commission estimates these firms would need 0.75 FTEs on an ongoing basis to maintain CAT reporting.

See Section IV.F.1.c(2)B.ii, supra.

Based on this information the Commission preliminarily believes that it would take a small OATS-reporting broker-dealer 1,350 ongoing burden hours per year to continue compliance with the Rule. The Commission believes the burden hours would be associated with work performed by internal technology, compliance and legal staff in connection with the ongoing operation of CAT Data reporting. The Commission preliminarily estimates that it would cost, on average, approximately $124,373 in ongoing external outsourcing costs to ensure ongoing compliance with Rule 613.

1,350 ongoing burden hours = (0.75 FTE for maintenance of CAT Data reporting systems) × (1,800 working hours per year).

See Section IV.F.1.c(2)B.ii, supra. See supra note 1581.

Therefore, the Commission preliminarily estimates that the average ongoing annual burden per small OATS-reporting broker-dealer would be approximately 1,350 hours, plus $124,373 in external costs, for an estimated aggregate ongoing burden of 1,088,100 hours and an estimated aggregate ongoing external cost of $100,244,638.

The Commission preliminarily estimates that 806 small OATS-reporting broker-dealers would be impacted by this information collection. (806 small OATS-reporting broker-dealers × 1,350 burden hours) = 1,088,100 aggregate ongoing burden hours to ensure ongoing compliance with Rule 613.

$100,244,638 = $124,373 in ongoing outsourcing costs × 806 broker-dealers.

B. Non-OATS Reporters

i. Initial Burden Hours and External Cost

In addition to firms that currently report to OATS, the Commission estimates there are 799 broker-dealers that are currently exempt from OATS reporting rules due to firm size, or excluded because all of their order flow is routed to a single OATS reporter, such as a clearing firm, that would incur CAT reporting obligations. A further 24 broker-dealers have SRO memberships only with one Participant; the Commission believes this group is comprised mostly of floor brokers and further preliminarily believes these firms would experience CAT implementation and ongoing reporting costs similar in magnitude to small equity broker-dealers that currently have no OATS reporting responsibilities.

See Section IV.F.1.c(2)B.ii, supra. Rule 613 does not exclude from data reporting obligations SRO members that quote or execute transactions in NMS Securities and Listed Options that route to a single market participant. See also CAT NMS Plan, supra note 3, at Appendix C, Section B.7(b)(ii)(B)(2).

See Section IV.F.1.c(2)B.ii, supra.

Id.

The Commission assumes these broker-dealers would have very low levels of CAT reporting, similar to those of the lowest activity firms that currently report to OATS. For these firms, the Commission assumes that under CAT they would incur the average estimated service bureau cost of broker-dealers that currently report fewer than 350,000 OATS ROEs per month, which is $124,373 annually. Furthermore, because these firms have more limited data reporting requirements than other firms, the Commission assumes these firms currently have only 0.1 full-time employees currently dedicated to regulatory data reporting activities. The Commission assumes these firms would require 2 full-time employees for one year to implement CAT.

Id.

Id.

Id.

Based on this information, the Commission preliminarily estimates that the average initial burden to develop and implement the needed systems changes to capture the required information and transmit it to the Central Repository in compliance with the Rule for small, non-OATS-reporting broker-dealers would be approximately 3,600 initial burden hours. The Commission believes the burden hours would be associated with work performed by internal technology, compliance and legal staff in connection with the implementation of CAT Data reporting. The Commission also preliminarily estimates that each small non-OATS-reporting broker-dealer would incur approximately $124,373 in initial external outsourcing costs.

3,600 initial burden hours = (2 FTEs for implementation of CAT Data reporting systems) × (1,800 working hours per year).

See supra note 1590.

Therefore, the Commission preliminarily estimates that the average one-time initial burden per small OATS-reporting broker-dealer would be 3,600 burden hours and external costs of $124,373 for an estimated aggregate initial burden of 2,962,800 hours and an estimated aggregate initial external cost of $102,358,979.

The Commission preliminarily estimates that 823 small non-OATS-reporting broker-dealers would be impacted by this information collection. (823 small non-OATS-reporting broker-dealers × 3,600 burden hours) = 2,962,800 aggregate initial burden hours.

($124,373 in outsourcing costs) × (823 small non-OATS-reporting broker-dealers) = $102,358,979 in aggregate initial external costs.

ii. Ongoing, Annual Burden Hours and External Cost

Small non-OATS-reporting broker-dealers that outsource their regulatory data reporting would likely face internal staffing burdens and costs associated with ongoing activity, such as maintaining any systems that transmit data to their service providers. Based on conversations with market participants, the Commission estimates these firms would need 0.75 full-time employees annually to maintain CAT reporting.

Based on this information the Commission preliminarily believes that it would take a small non-OATS-reporting broker-dealer 1,350 ongoing burden hours per year to continue compliance with the Rule. The Commission preliminarily estimates that it would cost, on average, approximately $124,373 in ongoing external outsourcing costs to ensure ongoing compliance with Rule 613. Therefore, the Commission preliminarily estimates that the average ongoing annual burden per small non-OATS-reporting broker-dealer would be approximately 1,350 hours, plus $124,373 in external costs, for an estimated aggregate ongoing burden of 1,111,050 hours and an estimated aggregate ongoing external cost of $102,358,979.

1,350 ongoing burden hours = (0.75 FTEs for maintenance of CAT data reporting systems) × (1,800 working hours per year).

The Commission assumes these firms would have very low levels of CAT reporting, similar to those of the lowest activity firms that currently report to OATS. For these firms, the Commission assumes that under CAT they would incur the average estimated service bureau cost of firms that currently OATS report fewer than 350,000 OATS ROEs per month of $124,373 annually.

The Commission preliminarily estimates that 823 small non-OATS-reporting broker-dealers would be impacted by this information collection. (823 small non-OATS-reporting broker-dealers × 1,350 burden hours) = 1,111,050 aggregate ongoing burden hours to ensure ongoing compliance with Rule 613.

($124,373 in ongoing external outsourcing costs) × 823 = $102,358,979 in aggregate ongoing external costs to ensure ongoing compliance with Rule 613.

E. Collection of Information is Mandatory

Each collection of information discussed above would be a mandatory collection of information.

F. Confidentiality C

Rule 613 requires that the information to be collected and electronically provided to the Central Repository would only be available to the national securities exchanges, national securities association, and the Commission for the purpose of performing their respective regulatory and oversight responsibilities pursuant to the federal securities laws, rules and regulations. Further, the CAT NMS Plan is required to include policies and procedures to ensure the security and confidentiality of all information submitted to the Central Repository, and to ensure that all SROs and their employees, as well as all employees of the Central Repository, shall use appropriate safeguards to ensure the confidentiality of such data and shall agree not to use such data for any purpose other than surveillance and regulatory purposes. The Commission will receive confidential information. To the extent that the Commission does receive confidential information pursuant to this collection of information, such information will be kept confidential, subject to the provisions of applicable law.

G. Recordkeeping Requirements

National securities exchanges and national securities associations would be required to retain records and information pursuant to Rule 17a-1 under the Exchange Act. Broker-dealers would be required to retain records and information in accordance with Rule 17a-4 under the Exchange Act. The Plan Processor would be required to retain the information reported to Rule 613(c)(7) and (e)(6) for a period of not less than five years.

17 CFR 240.17a-1.

17 CFR 240.17a-4.

17 CFR 242.613(c)(7) and (e)(6).

H. Request for Comments

Pursuant to 44 U.S.C. 3506(c)(2)(A), the Commission solicits comment to:

(1) Evaluate whether the proposed collections are necessary for the proper performance of our functions, including whether the information shall have practical utility;

(2) Evaluate the accuracy of our estimate of the burden of each collection of information;

(3) Determine whether there are ways to enhance the quality, utility, and clarity of the information to be collected; and

(4) Evaluate whether there are ways to minimize the burden of each collection of information on those who are to respond, including through the use of automated collection techniques or other forms of information technology.

Persons submitting comments on the collection of information requirements should direct them to the Office of Management and Budget, Attention: Desk Officer for the Securities and Exchange Commission, Office of Information and Regulatory Affairs, Washington, DC 20503, and should also send a copy of their comments to Brent J. Fields, Secretary, Securities and Exchange Commission, 100 F Street NE., Washington, DC 20549-1090, with reference to File No. 4-698. Requests for materials submitted to OMB by the Commission with regard to these collections of information should be in writing, with reference to File No. 4-698, and be submitted to the Securities and Exchange Commission, Office of FOIA/PA Services, 100 F Street NE., Washington, DC 20549-2736. As OMB is required to make a decision concerning the collections of information between 30 and 60 days after publication in the Federal Register, a comment to OMB is best assured of having its full effect if OMB receives it within 30 days of publication.

VI. Solicitation of Comments

Interested persons are invited to submit written data, views, and arguments concerning the foregoing, including whether the CAT NMS Plan is consistent with the Act. Comments may be submitted by any of the following methods:

Electronic Comments

  • Use the Commission's Internet comment form ( http://www.sec.gov/rules/sro.shtml ); or
  • Send an email to rule-comments@sec.gov. Please include File Number 4-698 on the subject line.

Paper Comments

  • Send paper comments in triplicate to Brent J. Fields, Secretary, Securities and Exchange Commission, 100 F Street NE., Washington, DC 20549-1090.

All submissions should refer to File Number 4-698. This file number should be included on the subject line if email is used. To help the Commission process and review your comments more efficiently, please use only one method. The Commission will post all comments on the Commission's Internet Web site ( http://www.sec.gov/rules/sro.shtml ). Copies of the submission, all subsequent amendments, all written statements with respect to the CAT NMS Plan that are filed with the Commission, and all written communications relating to the CAT NMS Plan between the Commission and any person, other than those that may be withheld from the public in accordance with the provisions of 5 U.S.C. 552, will be available for Web site viewing and printing in the Commission's Public Reference Room, 100 F Street NE., Washington, DC 20549 on official business days between 10:00 a.m. and 3:00 p.m. Copies of the submission will also be available for inspection and copying at the Participants' principal offices. All comments received will be posted without change; the Commission does not edit personal identifying information from submissions. You should submit only information that you wish to make available publicly. All submissions should refer to File Number 4-698 and should be submitted on or before July 18, 2016.

By the Commission.

Brent J. Fields,

Secretary.

[FR Doc. 2016-10461 Filed 5-16-16; 8:45 am]

BILLING CODE 8011-01-P