Bonus To Reward States for High Performance Under the TANF Program

AGENCY:

Administration for Children and Families, HHS.

ACTION:

Final rule.

SUMMARY:

The Administration for Children and Families is issuing final regulations to implement section 403(a)(4) of the Social Security Act. This provision authorizes bonuses to high performing States in meeting the purposes of the Temporary Assistance for Needy Families Block Grant (the TANF program). We will base the bonus awards in FY 2002 and beyond on work measures (substantially the same work measures currently in effect for the FY 1999-2001 awards); measures that support work and self-sufficiency related to: participation by low-income working families in the Food Stamp Program, participation of former TANF recipients in the Medicaid and State Children's Health Insurance Programs (SCHIP), and receipt of child care subsidies; and a measure related to family formation and stability (increase in the number of children in the State who reside in married couple families).

Bonus funds of up to $200 million each year were authorized for awards in fiscal years 1999 through 2003. This rule specifies a formula for allocating these funds in FY 2002 and FY 2003. The amount awarded to each high performing State may not exceed five percent of the State's family assistance grant.

Earlier, we issued program guidance covering bonus awards in FY 1999, FY 2000, and FY 2001. We published a Notice of Proposed Rulemaking to cover awards beginning in FY 2002 on December 6, 1999 (64 FR 68202).

In a related regulatory action, we are amending 45 CFR Part 265, the TANF Data Collection and Reporting Requirements, to reduce the burden of reporting data on Separate State Program-Maintenance of Effort (SSP-MOE) programs. This amendment will allow waivers of certain reporting requirements under limited circumstances.

EFFECTIVE DATES:

These regulations are effective Ocotber 30, 2000 except for Section 270.4(e)(2)(ii), which requires an information collection that is not yet approved by OMB. We will publish an announcement in the Federal Register regarding the effective date of the additional data collection.

FOR FURTHER INFORMATION CONTACT:

Sean Hurley, Director, Division of Data Collection and Analysis, Office of Planning, Research and Evaluation, ACF, at 202-401-9297. Mr. Hurley's e-mail address is: shurley@acf.dhhs.gov.

This rule is accessible electronically via the Internet from the ACF Welfare Reform Home Page at http://www.acf.dhhs.gov/news/welfare.

SUPPLEMENTARY INFORMATION:

Table of Contents

I. Legislative Background.

A. The Temporary Assistance for Needy Families Program.

B. Summary of the Statutory Provisions Related to the High Performance Bonus Awards.

II. High Performance Bonus Awards in FY 1999, FY 2000, and FY 2001.

III. Summary of the Notice of Proposed Rulemaking.

IV. Overview of the Public Comments.

A. Overview of Comments on the Work Measures.

B. Overview of Comments on the Food Stamp and Medicaid/SCHIP Measures.

C. Overview of Comments on the Family Formation Measure.

D. Recommendations for the Inclusion of New Measures.

E. Other Recommendations and Suggestions.

V. Summary of the Final Rule.

VI. Section-by-Section Discussion of the Rule and the Public Comments.

VII. Amendment to 45 CFR Part 265.

VIII. Regulatory Impact Analyses.

A. Executive Order 12866.

B. Regulatory Flexibility Analysis.

C. Assessment of the Impact on Family Well-Being.

D. Paperwork Reduction Act.

E. Unfunded Mandates Reform Act of 1995.

F. Congressional Review.

G. Executive Order 13132.

I. Legislative Background

A. The Temporary Assistance for Needy Families Program

Title I of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996, Pub. L. 104-193, established the Temporary Assistance for Needy Families (TANF) program at title IV-A of the Social Security Act (the Act). TANF is a block grant program designed to make dramatic reforms in the nation's welfare system. Its focus is on moving recipients into work and turning welfare into a program of temporary assistance, preventing and reducing the incidence of out-of-wedlock births, and promoting stable two-parent families. Other key features of TANF include provisions that emphasize program accountability through financial penalties and rewards for high performance.

Title I also “de-linked” the eligibility for cash assistance and Medicaid benefits. Under the Medicaid amendments, a family's eligibility for Medicaid is based on whether the family would have been eligible under the State's prior AFDC plan.

TANF replaced the national welfare program known as Aid to Families with Dependent Children (AFDC) that provided cash assistance to needy families on an entitlement basis. It also replaced the related programs known as the Job Opportunities and Basic Skills Training (JOBS) program and the Emergency Assistance (EA) program.

The new TANF program went into effect on July 1, 1997, except in States that elected to submit a complete plan and implement the program at an earlier date. We published final regulations to implement the work, penalties, and data collection provisions of the TANF program in the Federal Register on April 12, 1999 (64 FR 17720). These rules became effective October 1, 1999. We also published a number of other related regulations, including rules covering annual reports of State child poverty rates in relation to the TANF program (Notice of Proposed Rulemaking published September 23, 1998 (63 FR 50837)) and bonuses to reward States for decreases in out-of-wedlock births (final rule published April 14, 1999 (64 FR 18484)).

The 1996 welfare reform law reflected widespread, bipartisan agreement on a number of key principles:

• Welfare reform should help move people from welfare to work.
• Welfare should be a short-term, transitional experience, not a way of life.
• Parents should receive the child care, health care, and other supports that they need to protect their children as they move from welfare to work.
• Child support enforcement programs should become tougher and more effective in securing support from noncustodial parents.
• Because many factors contribute to poverty and dependency, solutions to these problems should not be “one size fits all.” The system should allow States, Indian tribes, and localities to develop diverse and creative responses to these problems.
• The Federal government should place more emphasis on program results.

Under section 401(a)(1) of the Act, States (and certain Indian tribes) have the authority to use Federal welfare funds “in any manner that is reasonably calculated to accomplish the purpose” of the new program. They have broad flexibility to set eligibility rules and decide what benefits are most appropriate.

In short, they have the opportunity to try new, far-reaching changes that enable them to respond more effectively to the needs of families within their own unique environments.

B. Summary of the Statutory Provisions Related to the High Performance Bonus

Section 403(a)(4) of the Act requires the Secretary to award bonuses to “high performing States.” (Indian tribes are not eligible for these bonuses.) The term “high performing State” is defined in section 403(a)(4) of the Act to mean a State that is most successful in achieving the purposes of the TANF program as specified in section 401(a) of the Act. These purposes are to—

(1) provide assistance to needy families so that children may be cared for in their own homes or in the homes of relatives;

(2) end the dependence of needy parents on government benefits by promoting job preparation, work, and marriage;

(3) prevent and reduce the incidence of out-of-wedlock pregnancies and establish annual numerical goals for preventing and reducing the incidence of these pregnancies; and

(4) encourage the formation and maintenance of two-parent families.

Section 403(a)(4)(B) of the Act specifies that the bonus award for a fiscal year will be based on a State's performance in the previous fiscal year and may not exceed five percent of the State's family assistance grant.

Section 403(a)(4)(C) of the Act requires the Department to develop a formula for measuring State performance in consultation with the National Governors' Association (NGA) and the American Public Welfare Association, now known as the American Public Human Services Association (APHSA).

Section 403(a)(4)(D) of the Act requires the Secretary to use the formula developed to assign a score to each eligible State for the fiscal year preceding the bonus year and prescribe a performance threshold as the basis for awarding the bonus. Section 403(a)(4)(D) of the Act also specifies that $1 billion (or an average total of $200 million each year) will be awarded over five years, beginning in FY 1999.

II. High Performance Bonus Awards in FY 1999, FY 2000, and FY 2001

As we have done with all regulations related to the TANF program, we implemented a broad consultation strategy prior to our rulemaking. In addition, as required by section 403(a)(4)(C) of the Act, we consulted intensively with representatives of the NGA and the APHSA. We met with staff of these two national organizations as well as staff of the National Conference of State Legislatures (NCSL) and approximately 30 State representatives who participated by regularly scheduled conference calls over a period of approximately nine months.

We also consulted with a number of other audiences: researchers, data experts, and academics; other Federal and non-Federal agencies that had developed or were in the process of developing performance measures for their programs; and representatives of a broad range of non-profit, advocacy, and community-based programs.

We would have preferred to set the formula for all years through rulemaking. However, we were not able to conduct adequate consultations and complete a formal rulemaking process in order to advise States, in a timely way, how we would be assessing their performance (for both the performance year and the comparison year used to measure improvement) in FYs 1997-1998, FYs 1998-1999, and FYs 1999-2000, in order to make awards in FY 1999, FY 2000, and FY 2001. Therefore, we issued program guidance covering the first three award years without the benefit of a formal rulemaking process. (For the program guidance for the awards in FY 1999, see TANF-ACF-PI-98-1 and TANF-ACF-PI-98-5 (Form ACF 200, OMB #1970-0180); for the guidance for the FY 2000 awards, see TANF-ACF-PI-99-1; and for the guidance for the FY 2001 awards, see TANF-ACF-PI-99-5.)

The FY 1999 program guidance based the first-year bonus awards on four work measures, i.e., the job entry rate, the success in the work force rate (this is a combination of the job retention rate and the earnings gain rate), and improvement in each of these measures. We have based the FY 2000 and FY 2001 bonus awards on similar work measures.

On December 4, 1999, the President announced three actions relating to the high performance bonus:

• The award of $200 million for the first-year bonus awards to 27 States with the best records in moving parents on welfare into jobs and subsequent success in the work force;
• The program guidance for the FY 2001 awards; and
• The publication of a Notice of Proposed Rulemaking (NPRM) covering awards in FY 2002 and beyond.

According to the reports filed by the 46 States competing for the first-year bonus, nationwide more than 1.3 million adults on welfare went to work in the one-year period between October 1, 1997, and September 30, 1998. Retention rates were also promising: 80 percent of those who had gotten jobs were still working three months later. The States also reported an average earnings increase of 23 percent for welfare recipients (some of whom were now former recipients) from $2,088 in the first quarter of employment to $2,571 in the third quarter.

The States ranked the highest in each category were Indiana (job entry), Minnesota (success in the work force, i.e., job retention and earnings), Washington (biggest improvement in job entry), and Florida (biggest improvement in success in the work force).

The other States that received bonuses were: Arizona, California, Connecticut, Delaware, Hawaii, Illinois, Iowa, Louisiana, Massachusetts, Michigan, Nevada, New York, North Dakota, Oklahoma, Pennsylvania, Rhode Island, South Carolina, South Dakota, Tennessee, Texas, Utah, West Virginia, and Wyoming. Eleven States received bonuses in two categories, and one State, Minnesota, was successful in three.

In announcing these awards for FY 1999, we recognized that the award criteria did not necessarily identify all States that have implemented successful welfare reform strategies. For example, some States may have implemented exceptionally strong programs whose success was not captured by this award because of timing or the specific measures we used. In addition, although we awarded bonuses to the ten States with the highest scores in each measure, the performance scores for many other States were also high.

III. Summary of the Notice of Proposed Rulemaking

We faced a significant challenge in developing a performance measurement system for the new TANF program. Although there is considerable activity underway in both the public and private sectors, performance measurement is a field that is still evolving. Our aim in developing the bonus award system was to reflect outcomes based on the purposes of the Act, propose a system as simple as possible to understand and administer, and incorporate the best information available.

To provide context, in the NPRM, we included a discussion of some of the difficult and inter-related questions and issues with which we, and the groups with which we had consulted, had struggled, e.g., general approach questions, short-term versus long-term strategies, formula and distribution issues, and issues relating to the design of measures and the availability of data sources.

We also included a discussion of more specific issues related to TANF performance measurement, including issues around absolute performance and performance improvement and concerns about achieving a level playing field among States, and we discussed measures that we had considered and rejected. We also spoke about the difficulty of identifying appropriate measures without incurring new data collection responsibilities while relying, to the extent possible, on uniform, objective, and reliable State data; rewarding positive performance; and producing no unintended consequences.

Finally, as an additional encouragement to focus public comment on specific alternative approaches, we raised a series of questions on major sections of the proposed rule.

The consultations with NGA, APHSA, and others were very useful in helping us identify key issues, evaluate policy options, and develop both the program guidance for FY 1999, FY 2000, and FY 2001 and the Notice of Proposed Rulemaking. As a part of our consultations, NGA and APHSA developed a set of principles they believed should apply to a high performance bonus system. We found that these principles offered a positive framework for developing such a system and avoiding some major pitfalls. We also found these principles helpful as we addressed specific issues in developing the NPRM. The NGA/APHSA principles stated that a high performance bonus system should:

• Be simple, credible, quantifiable, understandable to the public, and consistent with the goals of the law;
• Focus on outcomes rather than process;
• Take varying State economic circumstances and policies into account and not impede the flexibility provided to States under Pub.L. 104-193;
• Minimize double jeopardy or reward. (For example, the law already provides bonuses for reducing out-of-wedlock births, a caseload reduction credit, and penalties and incentives related to child support enforcement and paternity establishment.);
• Avoid additional data collection requirements and costs and build on existing systems;
• Avoid unintended consequences;
• Focus on positive rather than negative measures; and
• Reflect the strong emphasis on employment and self-sufficiency in the Federal law and in the States' implementation of the law. This emphasis should influence the measures included in the system and the distribution of bonus funds.

We published a Notice of Proposed Rulemaking (NPRM) on December 6, 1999 (64 FR 68202). Since our initial consultations, we have held several additional formal and less formal discussions about TANF performance measures with States, State groups, and others. For example, on July 21, 1999, we invited States, advocates, researchers, and others to a day-long consultation on issues related to outcome and performance measurement related to the preparation of a “Study and Report to Congress on Alternative Outcome Measures” (section 107 of the Personal Responsibility and Work Opportunities Reconciliation Act). In addition, the core provisions of the NPRM were very similar to the measures, issues, and principles discussed in earlier consultations. Finally, we knew that the NPRM would provide an additional opportunity for public comment and believed it was important to move the regulations process forward.

In summary, the NPRM proposed to:

• Award bonuses beginning in FY 2002 based on four work measures (substantially the same work measures currently in use for FY 1999 and FY 2000 and specified for use in FY 2001);
• Award bonuses beginning in FY 2002 based on three non-work measures: one measure on family formation and family stability (increase in the number of children below 200 percent of poverty who reside in married couple families) and two measures that support work and self-sufficiency, i.e., participation by low-income working families in the Food Stamp Program and participation of former TANF recipients in the Medicaid program and the State Children's Health Insurance Program (SCHIP);
• Use one of two possible alternative sets of data for the four work measures, including the National Directory of New Hires;
• Use the Census Bureau's decennial and annual demographic programs as the data sources for two of the three work support measures, i.e., the measure on family formation and stability and the measure on participation in the Food Stamp Program;
• Measure performance on Medicaid/SCHIP participation, through State matches of TANF data with data on Medicaid/SCHIP enrollment;
• Award bonuses to the ten States with the highest scores in each measure;
• Specify an allocation of funds for each measure in FYs 2002 and FY 2003 (and beyond, if high performance bonus awards are subsequently authorized), under which we would award $140 million to the work measures and $60 million to the work support measures:
• Reiterate the requirement in § 265.3(d) of this chapter that, if a State wishes to receive a high performance bonus, it must file the information in Sections One and Three of the SSP-MOE Data Report; and
• Create an annual review process, as needed, if future modifications and technical changes are necessary.

We took this approach for several reasons. First, we believed that, given the primary focus of the TANF program on work, we should continue to focus the rewards to States for their efforts in this area. The funds allocation we proposed reflected the importance we placed on State performance directed towards work, i.e., $140 million for work and $60 million for work support measures.

Second, potential new data sources appeared to be available with respect to both the proposed work measures and the work support measures: i.e., the National Directory of New Hires would serve as a research data source and provide more comparable and reliable national work data; and data from the Census Bureau's decennial and annual demographic programs (e.g., the Supplementary Census 2000 Survey and the American Community Survey) would provide data sources for two of the three proposed non-work measures.

In developing both the program guidance and the NPRM, the Department has been interested in utilizing a broad set of measures (i.e., other than direct work measures) that more fully reflect other purposes of the TANF program. States, Congress, national organizations, and experts have also recommended the inclusion of other measures. During 1997 and 1998, we worked to develop other measures, but we were unable to identify measures for which we had reliable data sources.

Given the potential availability of the two new data sources, we proposed both work and work support measures. We strongly believe that Medicaid/SCHIP and Food Stamps are critical supports for many working families as they move towards self-sufficiency through employment. State performance to ensure that eligible families receive Food Stamps and Medicaid/SCHIP addresses two of the statutory purposes of the TANF program: Providing assistance to needy families so that children may be cared for in their own homes and ending the dependence of needy parents on government benefits by promoting job preparation and work. Receipt of Medicaid/SCHIP and Food Stamps supports purpose two by helping make it possible for families to move off of welfare into employment, sustain that employment, and progress on the job to eventual full economic independence.

In addition, the non-work measures reflected our concern that the lives of children and families, particularly low-income children and families, should be a focus of attention within the TANF program. We also believe that strong families are one of the key factors in developing and sustaining high levels of individual competence and functioning in our complex society. Thus, we concluded that States should be rewarded for their efforts in addressing family formation and the other purposes of the Act noted above.

IV. Overview of the Public Comments

We received 130 comment letters, some with multiple signatures, from a wide range of national, State, and local entities and organizations, including: City and county governments; State human service agencies, and national organizations representing States, State legislatures, and State human service organizations; national and State children, family, and domestic violence advocacy and service organizations; national and local faith-based organizations; national, State, and local employment, housing, and legal advocacy organizations; national labor unions and a State labor agency; food and nutrition service and advocacy organizations; Members of Congress; a national foundation; and others.

Some of the 130 individual comment letters were similar or identical to the more than 300 identical notecards we received as a result of a letter-writing campaign organized by a broad-based national coalition monitoring the effects of welfare reform.

The major themes of the comments included the following:

• Most commenters supported the work measures, but a number made recommendations for substantive and technical changes.
• There was a division of opinion on the inclusion of the Food Stamp and Medicaid/SCHIP measures. For a number of reasons, States objected to the inclusion of these measures. Advocacy, service, and faith-based organizations strongly supported these measures, as did all of the Members of Congress who commented on the NPRM.
• Almost all commenters objected to the family formation and stability measure, although a few suggested modifications.
• A large number of commenters, primarily national advocacy organizations and three Members of Congress, recommended the addition of a new measure on child care.
• To a lesser extent, a number of organizations also recommended other new measures, including domestic violence measures and worker protection measures.
• Some commenters made recommendations for changes in the allocation of funds, although these comments did not present a consistent view. Many who supported the Medicaid/SCHIP and Food Stamp measures suggested substantial increases in the dollars for these measures and decreases in the dollars for the work measures, while national organizations representing States and State human service agencies recommended that all dollars go to the work measures.

A. Overview of Comments on the Work Measures

With a few exceptions, commenters considered the work measures of job entry, job retention, and earnings gain to be the appropriate measures for assessing State performance in moving TANF recipients from welfare to work and self-sufficiency. At the same time, we received a number of substantive and technical suggestions on how we should modify these measures, e.g., establish a minimum level of earnings that would constitute employment; measure job retention and earnings gain over a longer time period; establish a separate measure of earnings gain (proposed as a combined job retention/earnings gain measure); measure performance improvement by percentage point change rather than percentage change; adjust performance scores by economic and demographic factors; and establish other threshold requirements, such as job placements above the poverty level. We address these comments in the section-by-section discussion below.

The States, their representative organizations, and other commenters expressed strong support for the proposed work measures (substantially the same work measures that are used for the high performance bonus awards for FYs 1999-2001). We considered a range of suggested changes, both substantive and technical, but, given the level of support for the proposed work measures, we made only a few technical changes in the final rule. We have changed the way we calculate performance improvement, i.e., we will use the percentage point change rather than the percentage change. We have also removed the distinction on what kinds of subsidized jobs count under the work measures. In addition, we have added clarifying definitions in ¶270.2 and incorporated other technical changes in ¶270.5. We will consider adding an earnings threshold in the future, after further analysis and consultation with States and other interested individuals.

In the NPRM, we also proposed that States report one of two alternative sets of data—either a minimal set of identifying information on adult TANF recipients, which we would match against data from the National Directory of New Hires (NDNH) at the Federal level, or a more extensive set of work performance data. We proposed the use of the NDNH in response to concerns that States raised about access to out-of-State and Federal employment data during our initial consultations and implementation of the FY 1999 bonus awards. States and other commenters were strongly supportive of the use of the NDNH.

We agree that the use of the NDNH, matched with State data, will result in reduced burden for States and greater accuracy in implementing bonus awards. Therefore, in the final rule, we require States to report identifying information on adult TANF recipients that we will match with the NDNH data. We address these changes later in the section-by-section discussion of the rule.

B. Overview of Comments on the Food Stamp and Medicaid/SCHIP Measures

The proposed rules contained two measures that focused on State efforts to provide critical supports needed by low-income working families. One measure looked at improvements in the percentage of families leaving TANF who were enrolled in Medicaid or SCHIP six months later. The second measure looked at improvement in the rate of food stamp participation for certain low-income working families. These two proposals generated extensive comments, which were highly diverse in nature.

Because many commenters addressed these proposals together, and the comments on the two provisions were somewhat similar, this overview will address both provisions. However, there were also a variety of comments that spoke more directly to the separate proposals. You will find the discussion of these detailed and distinct comments in the section-by-section analysis for §§ 270.4(d) and 270.4(e).

Comments: A significant majority of all commenters supported inclusion of the Medicaid/SCHIP and food stamp measures. Among the reasons cited were the importance of these benefits as work supports, particularly for families with entry-level employment; the negative consequences of the recent declines in these program caseloads; the ability of States to operate TANF in ways that facilitate food stamp and Medicaid participation by low-income families; and the value of encouraging States to take steps necessary to improve access. At the same time, a number of these commenters had suggestions for modifications to the proposals.

These two proposals also drew a significant negative response, primarily from State agencies and organizations representing States. While generally agreeing that these programs provide important supports for low-income families, commenters raised a variety of philosophical, programmatic, administrative, and equity objections to including these measures as part of the high performance bonus. Philosophically, and particularly for the food stamp measures, some commenters indicated that the measures were inconsistent with TANF purposes, promoting dependency rather than self-sufficiency. In addition, State agencies objected to being held accountable, under a TANF provision, for serving families that were beyond the reach of the TANF program and for complying with requirements in other Federal programs. In the case of food stamp participation, in particular, they also objected to being held accountable when they lacked control over many program rules, and they could not spend TANF funds to pay for activities that are reimbursable under the Food Stamp Act. They expressed concern about the adequacy of national data, the equity of looking at annual improvement only from FY 2000 forward, the equity of applying annual improvement measures when some States had made significant efforts to improve access prior to the measurement period, and a variety of other issues.

Response: We have decided to retain measures of Medicaid/SCHIP and food stamp participation in the final rule because we are committed to a high performance bonus system that rewards States not just for employment successes, but also for their efforts to support low-income families during their transitions. We believe these measures are consistent with, and support the statutory purposes of, TANF. By participating in Food Stamps and Medicaid or SCHIP, needy families receive the assistance they need to care for children in their homes (purpose one) and improve their chances of ending dependence on government benefits through work (purpose two). In fact, the bipartisan comments we received from Members of Congress on these measures uniformly supported their inclusion.

In response to the technical and substantive concerns raised by the States and others, we very carefully considered all the suggestions for how to improve the measures and looked for ways to address the States' concerns. As you will find in the section-by-section analysis, we have made a number of changes that respond to the concerns. For example, we have made it more explicit that States may choose whether to compete on the Food Stamp measure (consistent with our approach for all the measures), dropped the “qualifying conditions” for both the Food Stamp and Medicaid/SCHIP measures (i.e., the threshold conditions that States had to meet in order to compete on these measures), added awards for absolute performance (not just improvement), and modified the improvement measure so that it is less biased towards States starting with a low level of performance in the comparison year.

Also, we recognize State concerns about being held accountable for activities that are outside of TANF. However:

• Unlike prior law, under TANF, all the key statutory provisions regarding goals and responsibilities refer to the “State” rather than the “State agency”; the concept of “single State agency” is gone; and all notifications go to the chief executive officer of the State, not the State agency. Thus, the statutory language suggests that it is appropriate for the high performance bonus to look more broadly at State performance rather than TANF State agency performance.
• The legislative history suggests that Congress intended that Food Stamps and Medicaid remain as part of the safety net for needy families affected by the TANF changes and that Congress was referring to welfare benefits when it included statutory language about reducing dependency on government benefits. More specifically, Congress did not modify the entitlement nature of Food Stamps and Medicaid when it repealed the entitlement to cash assistance. Further, in enacting sections 1925 and 1931 of the Act, Congress clearly intended that needy families would maintain eligibility for Medicaid benefits on the same basis as prior law (or a less restrictive basis). Indeed, the fact that Congress did not budget any savings for either the Medicaid or Food Stamp programs as the result of TANF indicates that it did not anticipate the declines in program participation that occurred in both programs, and it suggests that Congress did not intend for the declines to happen.

Congressional interest in maintaining Food Stamps and Medicaid as part of the safety net is also suggested by the managers' statement which: (1) Refers to changes in the Food Stamp program, but does not suggest any TANF-related effects; (2) Refers to PRWORA as a “fundamental reform of welfare” that “promotes work over welfare” [emphasis added]; and (3) speaks to not abandoning “those Americans who truly need a helping hand” and guaranteeing that children “will continue to receive the support they need.” This interpretation of Congressional intent also corresponds with the consistent bipartisan support we received in comments from Members of Congress on this issue.

• The statutory purposes of the TANF program reflect a broad view of the program that goes beyond families that are needy and receiving cash assistance.
• In most cases, the same State and local agencies are administering the TANF, Medicaid, and Food Stamp programs (or the TANF agency is making Medicaid eligibility determinations on behalf of the Medicaid agency), and a single caseworker is often responsible for determining eligibility and benefits in the three programs. Thus, in the course of administering the TANF program, TANF program managers often have the opportunity to work on eliminating barriers that may be deterring clients from seeking or retaining Medicaid or food stamp benefits.

For example, they can work on clearing up client misunderstandings about the applicability of TANF requirements to other program benefits (e.g., believing there are food stamp and Medicaid time limits); ensuring that families served by TANF diversion programs have the opportunity to apply immediately for other benefits to which they are entitled; and ensuring that applications and notices are clear about the expectations of each program, the reasons why particular benefits are denied or terminated, and an individual's rights to pursue other benefits. They can also work to provide office hours, office locations, and cultural and language accommodations that are responsive to client needs and to minimize administrative requirements, such as reporting and face-to-face interviews, that might discourage participation by eligible families.

C. Overview of Comments on the Family Formation Measure

The proposed rules contained one non-work measure directed at the second and fourth statutory purposes of TANF—i.e., to end the dependency of needy parents by promoting marriage and encouraging the formation and maintenance of two-parent families. More specifically, based on Census Bureau data, the NPRM proposed to allocate $20 million of the annual high performance bonus award to the 10 States with the largest increase in the percent of children below 200 percent of poverty who reside in two-parent families.

Comments: This proposal generated a significant number of comments and a largely negative reaction. While a few commenters commended our efforts to encourage State initiatives in this area, almost all who commented on this section expressed serious methodological and substantive concerns. Commenters noted that:

• States could earn awards based on bad outcomes, and thus the measure could have unintended negative effects.
• The measure fails to reward increases in marriage rates among families with higher incomes.
• Success in increasing marriage among single parents could inadvertently diminish a State's chances of receiving a bonus.
• This measure might also disadvantage those 10 or more States with State or local EITC programs.

Among the philosophical objections were:

• The measure's focus on marriage as the one acceptable form of “two-parent” families, noting that TANF purpose four refers to two-parent families, not marriage;
• The measure's failure to recognize noncustodial parents and a variety of less traditional family structures or to recognize the value of strengthening families through means other than marriage;
• The appropriateness of promoting marriage, e.g., when there are contraindications such as domestic violence and substance abuse; and
• The appropriateness of engaging the government in decisions that are essentially personal and private.

In addition, some commenters questioned our preamble justification of the measure by referring to research findings that being raised in a single-parent family did not, in and of itself, negatively affect children.

Commenters also raised concerns about: (1) States being measured on something that seemed beyond their jurisdiction and control; (2) double jeopardy, e.g., based on the proposed measure's similarity to the out-of-wedlock birth bonus; (3) the adequacy of Census data; and (4) the lack of a State option on whether to compete or not.

We received some suggestions for changes to this measure or for alternative measures related to family formation. Two organizations suggested we might establish a competition and award bonuses based on innovative policy initiatives and program demonstrations, and one State suggested we evaluate individual State descriptions of their own initiatives in this area. Commenters also suggested that we consider marriage rates for the entire State population and reward only “noncoercive public education campaigns”; reward States for increasing the percentage of families receiving TANF cash assistance that are two-parent families; and add domestic violence provisions (either as threshold qualifying conditions or adjustments). A few commenters suggested, alternatively, that we could encourage States to reduce teen pregnancy.

Response: Since our earliest consultations with NGA, APHSA, NCSL, and the State representatives, we have actively explored the best means for incorporating non-work measures in order to encompass the broad statutory purposes of TANF. We also have had a number of conversations with Congressional staff, advocates, academics, and others to seek suggestions for such measures.

The proposed family formation measure in the NPRM reflected our best attempt to synthesize what we had heard and develop a measure that was feasible in light of the data that were available to us. While we recognized some of the measure's flaws, we hoped that proposal might either generate some useful suggestions for modifications that would improve it or present us with some viable alternatives.

We seriously considered the suggestion to establish a panel-based competitive process that would reward innovative initiatives or demonstrations. However, we did not include it in the final rule because the approach is inconsistent with the statutory language at section 403(a)(4)(C)-(E), which clearly expects us to employ quantitative measures. Also, this approach seemed to move us away from focusing on outcomes. We also thought that, without specific quantitative standards, it would be extremely difficult to implement a system that was sufficiently objective and fair to serve as the basis for awarding millions of dollars a year.

We are committed to the marriage and family formation purposes of the Act and believe it is important that these purposes, in addition to the work and work-related purposes, be addressed in the high performance bonus system. We also believe that it is important for us to help States focus on the non-work related purposes of the TANF statute. This measure is intended to provide an additional incentive for State activity and creativity in this area.

Thus, we have retained a family formation measure in the final rule similar, but not identical, to the measure proposed in the NPRM. We agree with commenters who recommended a broader population measure, i.e., that we measure the increase in the percent of all children in each State who reside in married couple families, not just low-income children, and we have made this change in the final rule. We believe that this will address commenters' concerns that including a ceiling could produce unintended consequences. However, because the measurement issues associated with family formation are more significant than those for the work and work support measures, we have reduced the funding allocation for this measure to $10 million. The final rule specifies that, in FY 2002 and beyond, we will award $10 million to the ten States with the greatest percentage point improvement in this measure. We have also made clear that States may choose to compete on this measure (we will rank only those States that indicate that they wish to compete), emphasizing our overall policy that participation in the high performance bonus system is voluntary.

We address comments more specifically in Part VI of the preamble regarding new § 270.4(f).

D. Recommendations for the Addition of New Measures

In the NPRM, we proposed not only specific measures for FY 2002 and beyond, but we discussed a number of other measures and data sources that we had considered but elected not to include for various reasons. We actively encouraged comments on all aspects of these measures and data sources and solicited recommendations for other measures and data sources that we might not have considered.

Over one-half of the letters we received and all of the notecards offered suggestions for the inclusion of a range of new bonus measures, either as a substitute for the family formation measure or as additional measures. Commenters discussed some measures in detail; others were mentioned as suggestions for future development. Some of the recommendations for new measures, e.g., child care, domestic violence, and child poverty, were among the measures we had discussed in the preamble to the NPRM, but had not included in the proposed rule.

We appreciate the number of thoughtful, well-reasoned comments we received regarding new measures, as well as the detailed analysis and other information provided in support of the commenters' recommendations. We also appreciate commenters' commitment to the success of welfare reform, the focus on work and self-sufficiency, and the importance of the well-being of families and children.

We gave considerable thought and attention to all recommendations for new measures, particularly where commenters had provided suggestions for further exploration and analysis. In evaluating measures and data sources, we based our deliberations on the NPRM and the final rule on the principles for a high performance bonus system developed by NGA and APHSA. We were at all times aware of the availability and sufficiency of data sources and wanted to avoid new data collection requirements and costs. We have been particularly aware of the issue of diversity among States and how that diversity might impact the design and implementation of a fair bonus system. Finally, we wanted the bonus system to remain as simple as possible to understand and administer and focus on (1) positive, not negative goals; and (2) outcomes, not processes.

In light of the comments we received, we have added a child care measure in the final rule. We strongly agree with commenters that child care subsidies or assistance are essential supports for low-income families and a critical part of a successful welfare reform program. A child care measure was the one measure that received the strongest and most consistent support from commenters. It was also the one for which commenters offered the most concrete suggestions about how we might specify the measure. Supporters included a broad array of national, State, and local advocacy and service organizations, Members of Congress, and a number of individual commenters.

We discuss the specific child care measure and respond to comments in Part VI of the preamble, “Section-by-Section Discussion of the Rule and the Public Comments,” § 270.4(e). Following the discussion of the child care measure, we also respond to commenters' recommendations for other new measures.

V. Summary of the Final Rule

We continue to be committed to a high performance bonus system that meets statutory requirements; reflects the principles developed by NGA and APHSA; is based on measurable outcomes using the most uniform, objective, and reliable data available; and offers States an opportunity to be recognized for their achievements in several areas.

In making changes to the final rule, we seriously considered all concerns and recommendations of the commenters. We appreciate the thoughtful and detailed letters we received, and we particularly appreciate the sense of common goals, expressed directly or indirectly in the letters, focusing on both effective implementation of the TANF program and the economic self-sufficiency and well-being of families and children.

We also paid attention to the concerns of States and State representative organizations, given the statutory provision on consultation with NGA and APHSA and the diversity of views on certain issues between States and a number of other commenters. We believe that the final rule takes a balanced approach to this diversity. We believe we have been responsive to, and incorporated a number of, State recommendations regarding ways of making the measures less burdensome and more workable; at the same time, we incorporated other provisions that were not generally supported by States but were supported by a very broad range of other commenters, e.g., retaining the Food Stamp and the Medicaid/SCHIP measures and adding a measure on receipt of child care. We discuss these changes and respond to specific comments in the detailed section-by-section discussion below. Briefly, however, the final rule:

1. Awards bonuses to the ten States with the highest scores in the four work measures proposed in the NPRM, with minor modifications;

2. Awards bonuses to the three States with the highest scores on a new absolute measure and the seven States with the highest scores on the proposed improvement measure related to the participation by low-income working families in the Food Stamp Program;

3. Awards bonuses to the three States with the highest scores on a new absolute measure and to the seven States with the highest scores on the proposed improvement measure related to the participation of former TANF recipients in the Medicaid and SCHIP programs;

4. Awards bonuses to the ten States with the highest scores on a new child care measure and the family formation and stability improvement measure;

5. Bases competition on the family formation and stability measure on a universal population, i.e., the increase in the percent of children in each State who reside in married couple families;

6. Makes more explicit that States may choose any of the measures on which they wish to compete in order to conform the language of the proposed Food Stamp and family formation measures to the overall policy that participation is voluntary;

7. Eliminates the qualifying conditions and qualifying options proposed in the NPRM for the Food Stamp and the Medicaid/SCHIP measures;

8. Allots $140 million to the work measures, $20 million each to the Food Stamp and Medicaid/SCHIP measures, and $10 million each to the child care and family formation measures;

9. Reduces the reporting burden on States by allowing waivers of the reporting requirements for SSP-MOE data under certain limited circumstances;

10. Reduces the reporting burden on those States competing on the work measures by requiring only minimal identifying information on adult TANF recipients that we will use to match with NDNH data at the Federal level;

11. Bases competition on the Food Stamp measure and the family formation and stability measure initially on the Census Bureau's Census 2000 Supplementary Survey and the Census Long-Form Transitional Database and, later, on data from the American Community Survey;

12. Bases competition on the Medicaid/SCHIP measure on State Medicaid/SCHIP data, matched with TANF data at the State level;

13. Bases competition in FY 2002 on the child care measure, which focuses on child care accessibility (the percent of CCDF-eligible children receiving services), affordability (assessed family co-payments), and child care quality (based on State reimbursement rates) using data States currently report to us under the CCDF program;

14. Specifies the dates by which States must report data and other information to us;

15. Clarifies the use of the bonus funds; and

16. Makes technical and clarifying changes in the work measures, e.g., changes the way we calculate the improvement measures from percentage change to percentage point change and drops the requirement that States identify those persons whose jobs are fully subsidized.

VI. Section-by-Section Discussion of the Final Rule and the Public Comments

Section 270.1 What Does This Part Cover?

We received no comments on this section and have made no changes to it.

Section 270.2 What Definitions Apply to This Part?

This section of the NPRM proposed a number of definitions used in this part.

We have made several changes in this section: (1) We have updated the acronym and name of the CHIP (Children's Health Insurance Program) to SCHIP (State Children's Health Insurance Program); (2) we have defined the acronym “CCDF” as the Child Care and Development Fund; (3) we have added the words “or the calendar year” in the definition of “performance year” to indicate that, for the Food Stamp measure and the family formation measure, we will be comparing State performance based on a calendar year rather than a fiscal year; (4) we have moved the definition of “improvement rate” as proposed in § 270.5(c) of the NPRM to this section; and (5) we have added a definition of “absolute rate.” We have added these last two definitions in this section for clarity and because these terms now apply to both the work measures and the work support measures.

We received no comments on the definition of “improvement rate,” but we want to call attention to one change we have made in this definition and explain how it affects our ranking of States and making bonus awards. In the final rule, “Improvement rate” means the positive percentage point change between the absolute rate of performance in the performance year and the comparison year, except for the calculation and ranking of States on the increase in success in the work force measure in § 270.5(a)(4). The definition proposed in the NPRM did not include an exception and would have prohibited us from considering a State with a negative score in one sub-measure in the increase in the success in the work force measure in the ranking process. For example, a State may have a negative score on one sub-measure (e.g., job retention) and a positive score on the other sub-measure (e.g., earnings gain). We did not want to exclude that State from the competition for a bonus. We have made corresponding changes in § 270.5.

We received the following comments on this section:

Comment: One State asked that we add definitions for the terms “TANF eligible,” “employed recipient,” and “leaving TANF assistance,” as these terms have different meanings across States.

Response: We have not added definitions of these terms for several reasons. First, the term “TANF eligible” was used in the NPRM to describe qualifying conditions for the Food Stamp measure. These conditions have been dropped in the final rule. Second, the term “leaving TANF assistance” is used in the description of the Medicaid/SCHIP measure, but it is clear in the language of § 270.4(d) that this term refers to persons no longer receiving TANF assistance. Finally, the term “employed recipient” is used in describing components of several of the work measures. We believe it is clear that employment connotes earnings or wages. Since we have not established a minimum earnings threshold, we believe it is not necessary to define this term.

Comment: In commenting specifically on the definition of the terms “comparison year,” “fiscal year,” and “performance year,” one commenter was concerned that these definitions, combined with the proposed work measures, resulted in a bonus system that penalizes those States that may have focused on these activities well before the first comparison year. For example, these definitions and our other proposals would penalize States that have achieved significant increases in health care coverage between the beginning of their welfare program and the comparison year, while providing an advantage to States that have started more slowly. (This is an example of the “level playing field” issue on which we received a number of comments.)

This commenter recommended that we should base the health coverage measure on the States' overall efforts beginning with the effective date of the TANF program.

Response: The “level playing field” issue is one that we and others have struggled with since the beginning of our consultations on establishing a high performance bonus system. We agree that the system in place for the awards in FYs 1999 through 2001 and specified in this final rule would not completely address the concerns of, and may disadvantage, some strong performers who initiated their welfare reform programs prior to FY 1997.

However, we have made no change in the definitions in response to this comment. The statute specifies the “bonus years” for purposes of these awards as FYs 1999 through 2003, and we based bonus awards in FY 1999 on a State's performance in FYs 1997 and 1998. We did not believe that measuring performance in earlier years was responsive to the requirement that awards reflect a State's performance under, and following the establishment of, the TANF program.

Nevertheless, we have made two changes in the final rule that may help address concerns regarding a “level playing field,” i.e., we have added an absolute outcome measure in both the Food Stamp and the Medicaid/SCHIP measures and we have changed the way we calculate the improvement measure from percentage to percentage point change. (See § 270.4(c) and (d).)

Section 270.3 What Is the Annual Maximum Amount We Will Award and the Maximum Amount That a State Can Receive Each Year?

In accordance with section 403(a)(4)(B)(ii) of the Act, we proposed that the amount payable to a State in a given bonus year will not exceed five percent of the State's family assistance grant (SFAG). We also published, as an Appendix to the NPRM, a list of the total amount of each State's SFAG and the amount equal to five percent of each State's SFAG.

Comment: One State asked that we clarify whether the SFAG is the “present grant amount” or the grant amount when the bonuses are awarded.

Response: The statute and the TANF regulations (45 CFR 260.30) define the State family assistance grant (SFAG) as the amount of the basic block grant allocated to each eligible State under the formula at section 403(a)(1) of the Act. Thus, other TANF funds that a State may receive under section 403, e.g., bonus funds, contingency funds, and supplemental funds, are not a part of the SFAG. Neither would we reduce a State's bonus award based on reductions to the “SFAG payable” due to a penalty against the State. The amount of the State's SFAG as published in the Appendix to the NPRM is accurate and remains in effect until the statute changes.

Section 270.4 On What Measures Will We Base the Bonus Awards?

In the NPRM, we proposed in paragraph (a) of this section to award bonuses based on four work measures and three “non-work” measures. We proposed the work measures in paragraph (b) of this section. As we said in the overview of comments on the work measures above, there was strong support for these measures, although we received a number of suggestions for substantive modifications and technical changes. We address these suggestions in the discussion of § 270.5 and § 270.6 below.

Section 270.4(c) Measure of Participation by Low-Income Working Households in the Food Stamp Program

Under the proposed food stamp outcome measure, we would measure the improvement in the number of low-income working families ( i.e., families with children under the age of 18 who have an income of less than 130 percent of poverty and earnings equal to at least half-time, full-year employment at the minimum wage) receiving food stamps as a percentage of the number of low-income working families in the State, using the same definition. For any given year, we would compare a State's performance on the measure with its performance in the previous year, beginning with a comparison of calendar year (CY) 2000 to CY 2001. We would rank all States and would award bonuses to the 10 States with the greatest percentage improvement in this measure. We proposed to allocate $20 million annually for the food stamp measure.

We also proposed that, in order to compete on the food stamp outcome measure, a State must be in compliance with four qualifying conditions. The qualifying conditions proposed in the rule were the following:

(1) The State agency has issued policy instructions or regulations clearly specifying that, at first contact with the State agency which administers the Food Stamp Program, individuals must be informed of the opportunity to apply for food stamps in accordance with 7 CFR 273.2(c)(1).

(2) The State agency has issued policy instructions or regulations clearly specifying that food stamp application forms are to be readily accessible and available upon request, in accordance with 7 CFR 273.2(c)(3).

(3) As evidenced through policy instructions, regulations, and administrative reviews, the State agency is complying with application processing time frames and expedited service rules, as required by 7 CFR 273.2(g).

(4) As evidenced through policy instructions, regulations, and administrative reviews, the State agency has taken steps to prevent inappropriate denials and terminations of eligible food stamp participants who have lost TANF eligibility, in accordance with 7 CFR 273.12(f). Since food stamp eligibility is not based on TANF eligibility, States may not deny food stamp eligibility to a family or family member simply because the family is ineligible for TANF.

We proposed that the Food and Nutrition Service (FNS) of the U.S. Department of Agriculture would determine each State's compliance with the qualifying conditions, as a part of its ongoing oversight of the Food Stamp Program.

As noted earlier in this preamble, the majority of total comments received on the food stamp outcome measure supported the proposed measure. However, for a number of reasons, almost all of the State commenters opposed the inclusion of the food stamp outcome measure. We have seriously considered all comments, particularly the concerns of States. We believe that we have addressed many, though not all, of their concerns in the final rule. We have also accepted recommendations made by other commenters.

Briefly, we have made the following changes in § 270.4(c) of the final rule:

(1) Added an absolute performance measure;

(2) Changed the award structure to grant bonuses to the three States that rank the highest on the absolute performance measure and the seven States that rank the highest on the improvement measure;

(3) Changed the measured unit from “families” to “households with children”;

(4) Revised the improvement component to measure the percentage point improvement, rather than the percentage improvement, in the participation of low-income working households with children;

(5) Dropped the qualifying conditions;

(6) Made more explicit that competition on the measure is optional for States, to conform to the overall bonus policy that participation is voluntary; and

(7) Clarified how we will deal with tie scores.

We address the specific comments below.

Comments: Some commenters claimed that awarding TANF high performance bonus funds based on a measure of food stamp performance exceeds the statutory authority of TANF. Others argued that the food stamp measure encourages continued dependence on government benefits and, thus, runs contrary to the second goal of the TANF program, which is to end the dependence of needy parents on government benefits by promoting job preparation, work, and marriage.

Response: We disagree with the commenters who believe that awarding TANF bonus funds based on State performance in the Food Stamp Program exceeds the statutory authority of TANF. Section 403(a)(4) of the Act requires the Secretary of the Department of Health and Human Services to award bonuses to those States that are most successful in achieving the goals and purposes of the TANF program. As noted earlier in the preamble, we believe that State performance to ensure that eligible working families receive food stamps addresses two of the statutory goals of the TANF program: providing assistance to needy families so that children may be cared for in their own homes; and ending the dependence of needy parents on government benefits by promoting job preparation and work.

We recognize that a number of commenters felt that, far from ending the dependence of needy parents on government benefits, the food stamp outcome measure encourages dependence by encouraging States to assist working families to participate in the Food Stamp Program. We strongly disagree with this viewpoint. Ending the dependence of needy parents on government assistance requires successfully transitioning parents from welfare to work. Key to that successful transition is the Food Stamp Program. Food stamps provide needed nutritional benefits during that period when families are working but are not earning at the level that will enable them to achieve full self-sufficiency. In some cases, working parents may only be able to keep their jobs and feed their families because food stamps help them make ends meet.

Comments: Some commenters opposed the food stamp outcome measure on the grounds that it does not take into account many factors that have contributed to the decline in food stamp participation, including policy changes that have affected the eligibility of single adults and non-citizens.

Response: We recognize that many factors combined to cause the significant decrease in program participation experienced since 1996, not the least of which were a strong economy and new food stamp requirements that barred many non-citizens from participating in the program and imposed work requirements on able-bodied, childless adults. However, other factors also appear to be at work. Between 1995 and 1998, food stamp participation fell three times as much as the fall in the number of poor people, suggesting that many poor families have left the program despite their continuing eligibility. In 1999, participation continued to decline, although the rate of decline has slowed.

Traditionally, the program has had lower participation rates among eligible low-income families who are not receiving cash assistance. This means that as more families move from cash assistance to work, we have begun to see a dramatic decline in the food stamp rolls even though many of these low-income families remain eligible for food stamps. The food stamp outcome measure is designed to provide States with an incentive to implement policies and procedures necessary to improve access to the program among working families.

Comments: Some commenters felt that the food stamp measure effectively holds States responsible for overcoming obstacles to program participation that are established in Federal law and regulation. The commenters noted that strict eligibility requirements in the Food Stamp Program and Federal policies in effect restrict the number of families who can receive food stamps. The commenters believe that if the Administration is committed to expanding food stamp participation, it should take the necessary steps to amend the law and relax Federal regulations. They recommended relaxing reporting and verification requirements for working families, improving conformity between food stamp and TANF rules, and simplifying rules related to self-employment.

Response: We recognize that complex Federal laws and regulations, as well as State policies and procedures, can prove to be a barrier to Food Stamp Program participation among working families. For their part, the United States Department of Agriculture (USDA) and the Food and Nutrition Service (FNS) have taken steps to simplify program rules and reduce administrative burdens on working families. In July 1999, the President announced a series of actions to help ensure working families' access to food stamps. These actions included: (1) Expanding categorical eligibility rules to make it easier for working families to own a car and still be eligible for food stamps; (2) so long as the household's eligibility is redetermined at least every six months, providing States the option to allow working households to report changes in their circumstances on a quarterly basis, report only changes in income of $100 or more a month, and report only when there is a change in a job, hours of work, or wage rate; and (3) raising the quality control threshold that establishes when a case is considered to be in error from $5 to $25.

In addition, in a recently published proposed rule, Noncitizen Eligibility and Certification Provisions of Pub. L. 104-193, as amended by Public Laws 104-208, 105-33, and 105-185, (65 FR 10855), FNS proposed a number of provisions for further simplifying program rules and expanding State flexibility. The rule proposed the following: (1) Simplifying current verification requirements by removing overly prescriptive requirements for use of specific documents for verification; (2) allowing for the use of a simplified method of calculating self-employment expenses for certain specified types of businesses; and (3) establishing the ground rules for implementing the Simplified Food Stamp Program, under which States may determine food stamp benefit levels for households receiving TANF by using food stamp requirements, TANF rules, or a combination of the two.

In regard to achieving better conformity between TANF and food stamps, FNS has tried to provide States with as much flexibility as possible in conforming food stamp rules to TANF requirements without compromising the food security of the low-income population the program serves. State efforts to conform food stamp rules with TANF rules need to recognize that the Food Stamp Program serves a large and diverse range of people, two-thirds of whom do not receive TANF assistance, i.e., primarily cash assistance.

Comments: Some commenters believed that food stamp participation is not the appropriate variable for measuring a State's performance, given the fact that TANF benefit amounts and income disregards vary by State. In States with liberal disregards, a family's earnings plus TANF benefits may cause ineligibility for food stamps or reduce the food stamp benefit level to such a low amount that the family may conclude that it is not worth the effort to comply with certification requirements. Other commenters felt that the measure would reward States that place clients in low paying jobs or otherwise keep families below 130 percent of the Federal poverty level so that they may continue to qualify for food stamps.

Response: We do not believe that the food stamp outcome measure disadvantages States with more liberal TANF programs. First, most State TANF assistance programs do not have eligibility standards that exceed 130 percent of poverty. Second, if a State has more liberal disregards, food stamp eligible working households with children are more likely to continue receiving TANF assistance, and thus are more likely than other working households to be participating on the Food Stamp Program.

Also, States should be focused on improving the food stamp participation rate among all low-income, working households with children, not just those receiving TANF assistance. There are many more low-income working households with children who are eligible for food stamps than there are TANF participants. Those States that will do the best in the improvement measure are not those who improve the food stamp participation rate the most among current or former recipients of TANF assistance. Rather, it will be those States that increase the food stamp participation rate the most among all low-income working households with children. Similarly, the absolute measure will reward States that serve the greatest percentage of low-income working households with children overall, not the most current or former TANF recipients.

Comments: Some commenters felt that the food stamp outcome measure failed to take into account the restrictive rules of the Food Stamp Program. They noted that the only measure being used is income below 130 percent of poverty, but income is not the only factor that must be measured in actually determining eligibility for the Food Stamp Program. The asset rules alone will make many low-income families ineligible.

Response: Limitations in the Census Bureau data that we will use to measure States' performances on the food stamp outcome measure make it difficult to screen households for food stamp eligibility factors other than income. However, we do not believe that using income below 130 percent of poverty as a proxy for food stamp eligibility disadvantages any State in the bonus competition. While it is true that, because of the food stamp asset test and non-financial eligibility tests, a State's ratio of working families participating in the Food Stamp Program to working families that are income eligible for the program may appear lower than it, in fact, is, this will be true for every State because the Food Stamp Program employs national eligibility criteria. Thus, no State should be disadvantaged in comparison to other States or to itself over time.

Also, States can close the gap between the number of households that are only income eligible for food stamps and those that are actually eligible for the program by taking advantage of the expanded categorical eligibility rules announced as part of the President's July 14, 1999 food stamp initiative. The new policy allows States to use their more generous TANF assets tests, including their vehicle tests, rather than the Food Stamp Program asset limits, in determining food stamp eligibility for families receiving or authorized to receive TANF benefits.

Comments: Several commenters noted that while the food stamp outcome measure gauges the TANF program's effectiveness in enrolling working poor families in the Food Stamp Program, States are prohibited by law from spending TANF and MOE money for food stamp outreach. These commenters felt that it is unreasonable to hold a TANF program accountable for increases or decreases in the food stamp caseload when States cannot use TANF funds for food stamp outreach.

Response: Section 16(k)(5) of the Food Stamp Act of 1977, as amended, prohibits States from using TANF or MOE funds to pay for food stamp costs that are eligible for reimbursement under the Food Stamp Act. This includes the cost of activities to inform low-income households about the availability, eligibility requirements, application procedures and benefits of the Food Stamp Program. However, although States may not spend TANF, or MOE, money on these activities, they may use other State money to fund these activities, and FNS will match the expenditures at the 50:50 rate. In addition, there are certain activities related to increasing food stamp participation that are not reimbursable under the Food Stamp Act and for which States can use TANF or MOE funds. These activities include recruiting individuals to participate in the Food Stamp Program, providing transportation to certification and issuance offices, and acting as an authorized representative.

Comments: Several commenters noted that while the food stamp measure refers to “families,” food stamp receipt is by household, which may or may not match the conventional (TANF) definition of family.

Response: We recognize that looking at families in the food stamp measures makes the measures somewhat incongruous with the Food Stamp Program, in which receipt is based on “household.” A family, defined as parent and child, may not match the food stamp household, which would include anyone that lives with the family and purchases and prepares meals with them.

In the proposed food stamp outcome measure, a family that is included in the count of working families in a State that are income eligible for food stamps may not, because of the presence of another person in the home who purchases and prepares meals with the family, be in fact eligible for food stamps. This incongruity could cause the ratio of working families participating in the Food Stamp Program to families that are income eligible for the program in a State to appear lower than it in fact is.

Because this would be true in all States, we do not believe that this incongruity creates a bias in favor of any State in the competition or affects over time comparisons within States. However, in the final rule, we have changed the measured unit in the food stamp measures from families to households in order to better align the measure with the Food Stamp Program. We have revised the proposed regulations at § 270.4(c) to indicate that we will measure the number of low-income working households with children participating in the Food Stamp Program as a percentage of the number of low-income working households with children in the State.

Comments: One commenter objected that the food stamp outcome measure effectively restores repealed Food Stamp Program client service requirements. The commenter noted that to effectively compete for a high-performance bonus under the Food Stamp Program measure, States must restore many client service requirements that were repealed by PRWORA. The commenter believed that HHS was using financial incentives as a trade-off for the flexibility and independence to operate local food stamp offices that was granted States under PRWORA.

Response: In replacing specific client service requirements with the broad requirement that States establish procedures that best serve households, PRWORA directed States to take into account households with special needs. Included in this special needs category are working families. Therefore, beyond any desire to compete for TANF bonus funds, States have a responsibility to make the Food Stamp Program accessible to working families by implementing practices such as holding evening office hours and increasing the availability of application sites. Awarding TANF bonus funds based on State performance in serving working families, while primarily a recognition of the importance of food stamps to the overall success of welfare reform, is a means of providing States with an additional incentive to implement practices that will improve enrollment among a needy, yet difficult-to-serve, population.

Comments: Some commenters believed that the food stamp measure was improperly designed and suggested alternative measures. A number of commenters felt that the proposed measure did not address the real issue—that families leaving the TANF rolls are not properly referred to and assisted in accessing food stamps, even though they may still be eligible. These commenters suggested that we re-design the measure to track food stamp receipt among former TANF recipients for the month following the end of TANF receipt to ensure continual access to the Food Stamp Program. Other commenters criticized the measure for not giving States credit for cases in which a family leaving TANF earns too much to qualify for food stamps.

Response: Our interest in improving participation in the Food Stamp Program extends to all low-income, working families, not just those served by the TANF program. The majority of low-income working families that are eligible for food stamps have never participated in TANF. Also, many States refer eligible TANF recipients into diversion programs that provide them needed services and keep them off of the TANF program. The ability of these households to support themselves is vital to the success of welfare reform. Without food stamps, many of these families are in danger of going hungry; this could impact their ability to hold a job and to remain off of government cash assistance. The food stamp outcome measure provides States with an incentive to ensure that eligible former TANF recipients are properly referred to and assisted in obtaining food stamps. At the same time, it provides States with an incentive to improve access to the program for low-income, working families who have never been served by the TANF program, but whose ability to achieve and sustain self-sufficiency is nevertheless critical to the success of welfare reform.

Comments: Several commenters suggested that we should expand the food stamp measure to evaluate the improvement in participation among all low-income families in a State, not just those who are working.

Response: While we recognize the importance of food stamps as a support for all low-income households, we believe that we should continue to focus the food stamp outcome measure on working poor families, given the focus on work in the TANF statute, including the second purpose of the program. Participation in the Food Stamp Program remains especially low among the working poor; in 1997, only 59 percent of individuals in households with employed adults who were eligible for food stamp benefits participated in the Food Stamp Program, compared to a participation rate of 63 percent overall. If welfare reform is to be a lasting success, States must increase the participation rate of low-income working families significantly. By restricting the food stamp outcome measure to working households with children, we can help States focus on improving access to food stamps for this hard-to-serve population. Thus, we have not made a change to the proposed regulation as a result of these comments.

Comments: One commenter felt that it was inappropriate to limit bonuses to the top 10 States. The commenter recommended that we expand the number of States who could benefit from the performance bonus.

Response: The bonuses are intended to reward those States that are the most successful in achieving the goals and purposes of the TANF program. We chose to limit the bonuses to the top 10 States in each performance measure in order to emphasize that the awards recognize the highest performance among States. Increasing the number of States eligible for a bonus under each measure would dilute the significance of the awards. For this reason, we have not made changes to the proposed rule.

Comments: A number of commenters noted that the proposed food stamp performance measure, because it is a measure of improvement only, disadvantages States that are already doing a good job of encouraging Food Stamp Program participation among low-income working families. Some commenters requested that we expand the measure to recognize the progress made by States prior to the first year of the bonus awards and the progress made in prior years as the bonus moves from year to year. Other commenters suggested that we include a measure of absolute performance as well as an improvement measure. These commenters further suggested that we rank States separately on both the absolute and improvement measures and award bonuses to the top five States in each category.

Response: We recognize the importance of rewarding States for both absolute performance and improvement in each high performance bonus category. Awarding bonuses for both absolute performance and improvement provides a way to ensure a more objective and fair competition, by allowing States that start from different baselines a reasonable chance to compete successfully for bonus money. Each of the four work measures has an absolute and improvement component. However, in the case of the food stamp outcome measure, because only $20 million is being allocated for the measure, we felt that dividing the bonus among 20 winners, 10 for the best performance and 10 for the most improved, would too greatly diminish the incentive the bonus would provide. We opted in the proposed rule, therefore, to make the food stamp outcome measure only a measure of improvement. Given the low participation rate of poor working families on the Food Stamp Program, we felt that it was more important to reward States that improve program access to this group than to reward States who are already doing a good job of serving them.

Based on the comments we have received on the provision, however, we have decided to modify the food stamp outcome measure by adding a measure of absolute performance. This measure is designed to reward those States that, in a given year, demonstrate the very best performance in serving low-income working families. Under the outcome measure in the final rule, we will award $6 million in bonus funds to the three States that serve the highest percentage of low-income working households with children in the current year (the absolute measure) and award $14 million to the seven States that show the most improvement in performance from the previous year to the current year. We chose to reward more States for improving performance than for maintaining high overall performance because we wish to keep the emphasis of the bonus on improving service to low-income working households with children. We believe that this provision offers an effective compromise between rewarding States that currently do the best job of serving low-income working families and providing an incentive for other States to improve their performance. We have revised the proposed regulations at § 270.4(c) to reflect these changes.

Comments: Several commenters suggested that we revise the food stamp measure to measure the percentage point improvement, rather than the percentage improvement, in the participation of low-income working families. They noted that under the proposed measure, we would rank a State that increases food stamp participation from 5 percent to 10 percent (100 percent improvement) higher than a State that increases participation from 30 to 45 percent (50 percent improvement).

Response: We agree with the commenters that a fairer measure of improvement would be to measure the percentage point improvement rather than the percentage improvement in the participation of low-income working families. Therefore, we are modifying the food stamp improvement measure at § 270.4(c) to reflect this change. This change is consistent with the change we made in the work improvement measure in § 270.6.

Comments: A commenter noted that the food stamp performance measure needs to have a method for dealing with tie scores similar to the method for the work measures.

Response: We agree with the commenters and are revising the food stamp outcome measures in paragraphs (c)(1) and (c)(2) of this section to include a method for dealing with tie scores. We will use the same method for resolving tie scores for the food stamp measures as we use for the work and the Medicaid/SCHIP measures. We will calculate the percentage rate for the absolute performance measure to two decimal points. If two or more States have the same percentage rate for this measure, we will calculate the rates for these States to as many decimal points as necessary to eliminate the tie. Likewise, we will calculate the percentage rate for the improvement measure to two decimal points. If two or more States have the same percentage rate for this measure, we will calculate the rates for these States to as many decimal points as necessary to eliminate the tie.

Comments: We received a number of comments related to the proposed qualifying conditions. Several commenters suggested that we strengthen the conditions. One commenter recommended that we require States to affirmatively demonstrate that their computer systems have been programmed so that when any TANF case closes, the food stamp case remains open until the worker makes an independent determination as to whether the household is still eligible for food stamps. Another commenter requested that States be required to notify the public that they are competing for a bonus related to food stamp participation and solicit comments on the extent to which agency practices are inconsistent with the qualifying conditions. The same commenter also recommended that HHS publish the preliminary determinations as to States' compliance and the basis for such conclusions, and seek comments from the public as to whether the determinations are accurate.

A number of other commenters, however, recommended that we eliminate the qualifying conditions. One commenter noted that FNS will not have the resources to undertake the new determinations and, as a result, will likely certify that States are in compliance based on incomplete information. If the agency did discover noncompliance with these policies at a later date, the earlier certification could interfere with administrative or legal actions the agency might wish to take.

Other commenters noted that the conditions proposed in the rule are already requirements in the Food Stamp Act of 1977, as amended, and therefore are among the many factors already monitored for compliance by both the States and FNS. These commenters recommended that FNS allow States to self-certify their compliance with the conditions. Otherwise, they argue, FNS would need to redirect its limited staff resources and focus on the qualifying conditions, to the exclusion of other important State assistance and monitoring activities.

Response: After carefully considering all of the comments received on the qualifying conditions, we have decided to remove these conditions from the food stamp outcome measure. While HHS and FNS both firmly believe that it would be inappropriate for a State to win bonus money related to improving food stamp participation among working poor families if they are not in compliance with the most basic rules and regulations that are designed to provide program access, FNS' ongoing compliance activities will not necessarily be compatible with the timing of the high performance bonus awards. FNS already monitors State compliance with the four qualifying conditions, and the Food Stamp Program already contains appropriate remedies for addressing compliance issues. In addition, the qualifying conditions are so basic to maintaining good program access for working families that States that fail to meet them will likely not perform well in the bonus competition.

Although we are removing the qualifying conditions from the food stamp outcome measure, State compliance with those requirements, and with other legal and regulatory provisions related to program access, remains a high priority with FNS. For example, FNS has released two program access guides, one for working families and another for elderly and disabled households, that are designed to assist State policy makers and others in understanding what the food stamp statute and regulations require of States in terms of food stamp eligibility application processing, recertification, notice and appeal rights, among other matters. In addition, FNS is conducting customer service access reviews in every State that are designed to identify barriers to program participation, including problems stemming from noncompliance with the program's legal and regulatory requirements. By the end of FY 2000, FNS will have completed between one to three access reviews in every State. Beyond FY 2000, FNS intends to make customer service access reviews a permanent part of its oversight of the Food Stamp Program.

Comments: We received a number of comments on our proposal to use Census Bureau decennial and annual demographic program data in ranking State performance on the food stamp measure. Many commenters expressed concern as to the reliability of Census Bureau data. They noted that, in the past, Census Bureau data have provided misleading information regarding food stamp participation when compared to actual State data. Also, they felt that, while using Census Bureau data simplifies setting the baseline, it could rapidly become outdated based on population growth in States, resulting in an inability to award State bonus funds accurately and appropriately. Many commenters wondered why we did not simply use State administrative data, which is more reliable and would match the method for tracking Medicaid and SCHIP enrollment.

Response: The food stamp outcome measure examines changes in the ratio of the number of working households with children in a State that participate on the Food Stamp Program to the number of working households with children in the State that are income eligible for the program. State administrative data can only provide us with the number of working families in a State that are participating in the Food Stamp Program. They cannot tell us the total number of families in the State who are income eligible for the program. The only data source that can provide us that information is Census Bureau data.

We recognize there are problems inherent in using existing Census Bureau data sources for awarding TANF bonus funds. However, we hope to avoid many of the pitfalls identified by commenters by using new Census Bureau surveys. We will use the annual State estimates produced by the Census Bureau from its annual household survey program, beginning with the Census 2000 Supplementary Survey and transitioning to the Census Bureau's American Community Survey by 2004.

Comment: One commenter asked that we publicize baseline information from the Census data used to determine State performance on the food stamp measure on all States so States will know what current data show and how they stand in relation to other States.

Response: We intend to release the baseline Census data, as well as other data relevant to the performance and rankings of competing States.

Comments: Some commenters noted that Census data would identify noncitizens as part of the low-income population potentially eligible for food stamps. However, they may not in fact be eligible for the program. The commenters noted that this would disadvantage States with significant noncitizen populations and suggested that we factor such noncitizen groups out of the outcome measure calculation or add a provision to the measure to count State-funded food stamp recipients toward a State's overall percentage of low-income working families receiving food stamp benefits.

Response: Based on our most recent available data, almost 85 percent of households participating in the Food Stamp Program in 1995 that contained a noncitizen also contained at least one citizen child. Thus, the majority of the noncitizen households identified as part of the low-income working population eligible for food stamps in a State would contain at least one member who is eligible for food stamps. We recognize that households containing eligible children, but ineligible parents, can be an extremely difficult population to serve. However, if we were to exclude these households from the food stamp outcome measure, we would be providing States no incentive for improving access to these needy children and families.

Also, we have not included State-funded food stamp recipients in the count of a State's number of low-income working families receiving food stamps. Many of the individuals served in the State-funded program who have citizen children will already be included in the count of food stamp participating households. The majority of the remaining participants in the State-funded programs will be individuals without children who will not be included in the count of the number of low-income working households eligible for food stamps in the State.

Comments: Several commenters felt that, regardless of whether they intend to compete in the non-work measures, States should be required to provide data on their progress in the food stamp outcome measures as a prerequisite to competing in the work measures.

Response: As noted above, the data source for the food stamp outcome measure will be Census data, not administrative data submitted by States. ACF and FNS, therefore, will have data on every State's performance on the food stamp outcome measure, regardless of whether or not a State chooses to compete on the measure. However, this information will not be used to restrict a State's ability to compete on the work measures.

Comments: One commenter felt that the food stamp measure should include both quantitative and qualitative components, especially worker-client relationship evaluations and customer satisfaction. The commenter believed that many of the barriers to participation in the food stamp and Medicaid programs are attributable to caseworker attitudes. More training and encouragement from the State agency could reverse this trend, thus increasing enrollment.

Response: Including qualitative components in the food stamp outcome measure, such as worker evaluations and reports on client satisfaction, would diminish our ability to rank States quickly and objectively. In addition, we are concerned that increasing administrative burdens on States by requiring them to collect and report such data would likely deter them from competing on the measures. We also believe the recommended new components would be process, not outcome, measures.

Finally, improving customer service is a vital component to increasing participation among low-income working families in the Food Stamp Program. States that wish to realistically compete for the food stamp related bonus will have to improve their customer service standards along the lines discussed in USDA's food stamp access guide, “The Nutrition Safety Net at Work for Families: A Primer for Enhancing the Nutrition Safety Net for Workers and Their Children,” published in 1999. Therefore, we are making no changes to the proposed rule.

Comments: Two commenters noted that the Economic Research Service (ERS) of the USDA is conducting research into the reasons families may not participate in the Food Stamp Program. These commenters felt that participation by low-income families in the Food Stamp Program should not be part of the TANF high performance bonus system until ERS completed this research and specific barriers are identified and resolved at the national program level.

Response: ERS is funding a study on Food Stamp Program access and declining participation. The study will examine the impact of local food stamp office policies and practices on food stamp participation. However, data collection for the study will not begin until Fall 2000, and a final report is not due until Winter 2001. While we expect the report to provide us with greater insight into the practices and policies of local offices that may deter individuals from applying for food stamps, we see no reason to wait two years to provide States with a fiscal incentive to begin removing barriers to participation. There are steps that States can take today to improve program access. We included a listing of best practices for serving working families in the proposed rule. They are also contained in USDA's publication “The Nutrition Safety Net at Work for Families: A Primer for Enhancing the Nutrition Safety Net for Workers and Their Children.”

Section 270.4(d) Measure of Participation By Low-Income Families in the Medicaid/SCHIP Programs

In the proposed rule, we included a non-work measure related to Medicaid/SCHIP that would reward State efforts to support work, self-sufficiency and the well-being of low-income families. This measure looked at improvement in the percentage of TANF families who were enrolled in Medicaid or SCHIP at the time they lost TANF and who are enrolled in Medicaid or SCHIP six months later. We chose this approach because nearly all of these families leaving TANF are likely to be eligible for a minimum of six months of transitional Medicaid under section 1925 of the Act or to qualify for Medicaid under other eligibility groups. In addition, there have been reports from consumer advocates and State and national studies indicating that many eligible families are losing Medicaid benefits when they leave TANF. While there may be a number of outside forces contributing to the decline in Medicaid enrollment, e.g., a strong economy, changes in public attitude toward welfare, we believe the challenges presented States by the delinking of cash assistance from Medicaid have also contributed to the decline. This proposed measure focused on how well States are providing Medicaid to eligible families who lose TANF. We believe that continued health insurance coverage is crucial to families making the transition from welfare to self-sufficiency, and we expect States to achieve a high rate of Medicaid and SCHIP participation among this population in order to be considered high performers.

We considered an outcome measure that would capture State performance in enrolling and retaining all eligible families and children in Medicaid and SCHIP, regardless of their former or current welfare status. However, we limited the outcome measure to individuals leaving TANF assistance because:

(1) States have a clear responsibility for serving these families under PRWORA; and

(2) welfare “leaver” studies and other studies on program participation indicated that these families frequently were not being served.

While a broader population measure would be consistent with a goal of expanding health coverage and have the positive effect of encouraging States to enroll eligible individuals who are diverted from TANF assistance or who do not apply for TANF assistance, the proposed measure was more directly related to the goals and purposes of TANF, as well as title I of PRWORA. Also, with no national data source on health coverage for low-income families, we believed that the focus on TANF “leavers” would result in a smaller reporting burden and in the collection of more accurate and consistent information by States. It, thus, should produce fairer comparisons in assessing State performance.

In the NPRM, we also proposed certain qualifying conditions, based on requirements in Medicaid law and regulations, that States must meet before competing for an award in the Medicaid/SCHIP measure. Those qualifying conditions were:

(1) The State has issued policy instructions or regulations clearly specifying that, at first contact with the TANF agency (when the TANF agency is also the Medicaid agency), an individual must be given the opportunity to apply for Medicaid in accordance with 42 CFR 435.906;

(2) When eligibility under section 1931 of the Act is lost due to hours of, or earnings from employment or loss of time-limited earning disregards, the State issues to the affected family a written notice that meets the requirements of section 1925(a)(2)(A) of the Act and a card or other evidence of the family's entitlement to assistance as required under section 1925(a)(2)(B) of the Act;

(3) The State has issued policy instructions or regulations clearly specifying that family members may not be terminated from Medicaid until it has been determined that they are not eligible under any other Medicaid group; and

(4) The State has fulfilled all data requirements under the law, including being up to date on all Medicaid and SCHIP data submissions, and having the MSIS on-line and operating properly.

We proposed these qualifying conditions because we did not believe that a State that is out of compliance with basic program requirements should be eligible for a bonus related to Medicaid and SCHIP participation.

In addition to complying with the qualifying conditions, we proposed that applicant States must meet at least two qualifying State options. We believe that States exercising these options are likely to increase enrollment of eligible families and, therefore, would perform better on the outcome measure. The proposed programmatic options were:

(1) The State accepts mail-in or phone-in applications for Medicaid for families and children, which can be completed without a face-to-face interview;

(2) State Medicaid workers have been outstationed at locations in addition to the locations required under 42 CFR 435.904(c)(1) and (c)(2);

(3) The State has expanded Medicaid eligibility for recipient and applicant families through the use of less restrictive methodologies, authorized by section 1931(b)(2)(B) and (C) of the Act;

(4) The State uses a definition of “unemployed parent” that includes parents who are employed more than 100 hours per month, as authorized under 45 CFR 233.101 and section 1931(b) of the Act.

(5) The State provides continuous Medicaid eligibility for children for a period of time without regard to changes in circumstances, as authorized by section 1902(e)(12) of the Act;

(6) The State provides a period of presumptive Medicaid eligibility for children as authorized by section 1920A of the Act; or

(7) The State has simplified the enrollment and re-enrollment processes for children and low-income families by implementing such improvements as shortened application forms.

We proposed that those States that met the qualifying conditions and options would be eligible to compete for the bonus award based on their performance under the outcome measure. Specifically, the outcome measure would assess Medicaid and SCHIP participation among persons whose TANF assistance cases were closed in the calendar year who also were enrolled in Medicaid or SCHIP at the time of case closure. The measure of State performance would be the percentage of such individuals who are enrolled in Medicaid or SCHIP six months after leaving TANF and who are not currently receiving TANF assistance in that month. We proposed to compare a State's performance to its performance in the previous year, beginning with a comparison of CY 2000 to CY 2001, and to award bonuses to the 10 States with the greatest percentage improvement in this measure. We proposed to allocate $20 million annually for this measure.

We received a significant number of comments from States objecting to the Medicaid measure. We received a larger number of comments from other individuals and organizations, including national advocacy organizations and Members of Congress, in support of the Medicaid measure. Some States objected based on philosophical grounds while others objected for programmatic, administrative and equity reasons. Those commenters supporting the inclusion of a Medicaid measure cited the importance of Medicaid to low-income working families and referred to several recent studies on the declines in Medicaid caseloads where individuals, particularly children, were eligible for Medicaid or SCHIP benefits. (See “Overview of Comments in the Food Stamp and Medicaid/SCHIP Measures” in Part IV above.)

Briefly, in response to the comments, we have made the following changes in the final rule:

• Added an absolute performance measure;
• Changed the award structure to grant bonuses to the three States that rank the highest on the absolute performance measure and the seven States that rank the highest on the improvement measure;
• Revised the improvement measure to measure the percentage point improvement, rather than the percentage improvement, in participation;
• Changed the six-month time frame to a four-month time frame;
• Dropped the qualifying conditions and qualifying options;
• Required that States competing on these measures submit data on a fiscal year, rather than a calendar year, basis; and
• Clarified how we would deal with tie scores.

In addition, based on our own review and analysis, we have revised the regulatory text at § 270.4(d) to clarify that the denominators of the Medicaid/SCHIP measures exclude individuals who are receiving TANF at the time of follow-up (i.e., the fourth month after leaving).

Below, we summarize the comments we received and our responses.

Comment: Most commenting States objected to the inclusion of Medicaid as a performance measure. They stated that including the Medicaid measure is at odds with the TANF goal to decrease dependence on Government benefits and is not specific to the TANF program. Specifically, they argued that:

• The measure is inappropriate because it unfairly rates the success of TANF on the State performance in other programs;
• The high performance bonus awards should not be used to enforce Medicaid law;
• Including Medicaid shifts the focus away from work and the TANF population; and
• It is not an outcome measure of the number of former TANF customers who are better off, but instead a process measure of the number of enrollees in another government program.

Commenters supporting inclusion of the Medicaid measure viewed Medicaid as a critical support to low-income working families; in view of the declines in the Medicaid rolls after passage of welfare reform, they noted that the measure looks to reward State improvements in increasing Medicaid and SCHIP participation.

Response: We believe that Medicaid is a vital support to low-income working families and the provisions in this regulation will measure overall State performance in achieving the TANF goal of promoting job preparation and work. In addressing this comment earlier in the section entitled “Overview of Comments on the Food Stamp and Medicaid/SCHIP Measures,” we gave many reasons why we believe the inclusion of the Food Stamp and Medicaid/SCHIP measures is appropriate. In this response, we expand on those thoughts, particularly as they relate to the Medicaid measure.

The commenters are correct that one of the goals under section 401(a) of the Act is to end welfare dependence by promoting job preparation, work, and marriage. However, States can best promote self-sufficiency through job preparation and work by providing the support systems, such as health insurance coverage, that are essential to families during their transition from welfare to work.

As noted by several commenters, there have been many studies that indicate the need for Medicaid coverage while families make this transition. A January 2000 Urban Institute study found that more than one-third of women and nearly one out of five children are uninsured within the first six months of leaving welfare. State studies of families that have left TANF are also finding that at least 20% of children and the majority of parents are no longer receiving Medicaid (see “Participation in Welfare and Medicaid Enrollment,” Kaiser Family Foundation, 1998). A May 1999 Families USA study found that over two-thirds of a million low-income individuals lost Medicaid coverage and became uninsured as of 1997 due to welfare reform.

In enacting section 114 of PRWORA, Congress clearly intended to preserve Medicaid coverage for low-income families whose parents left welfare and went to work if they needed health care coverage and otherwise qualified for Medicaid. Congress preserved the health care safety net because it considers Medicaid a critical support for working families who might otherwise have no health insurance.

We do not believe that the fact that Medicaid may be administered by an agency other than the agency principally responsible for TANF is a reason for not including Medicaid enrollment as a measure in this high performance bonus regulation. As we stated earlier, it is more appropriate to view the high performance bonus as an award for State, not State agency, performance. TANF funds are used by many State and local agencies to accomplish the goals of the TANF legislation; indeed, the TANF block grant opens up new opportunities for additional agencies and nongovernmental organizations to get involved in the administration of the TANF program and the delivery of TANF benefits and services. It also provides new incentives for improved State and local interagency cooperation and cross-program efforts to encourage work and self-sufficiency.

TANF and Medicaid are closely related whether or not the programs are administered jointly by the State. Inclusion of a Medicaid outcome measure as part of the high performance bonus award is not an attempt to enforce Medicaid law, but rather to measure a State's overall success in serving low-income families leaving welfare. We believe that we should use the high performance bonus to encourage and recognize State efforts to effectively coordinate TANF and Medicaid program operations and reduce or eliminate barriers to ongoing Medicaid coverage for eligible families leaving TANF. Inclusion of a Medicaid performance measure provides focus on how well a State is achieving the goals of TANF and further meets congressional intent to provide support services while ending dependence on cash assistance.

Comment: Several commenters objected to including the qualifying conditions and the qualifying options in the NPRM. The commenters argued that these conditions appeared too controlling and that the high performance bonus does not provide an appropriate vehicle for HCFA to evaluate whether a State is in compliance with the qualifying conditions. Other commenters also questioned whether the high performance bonus was an appropriate vehicle for evaluating or verifying State compliance with HCFA requirements. One commenter recommended that we offer programmatic options to States as suggestions for improving their performance.

A number of other commenters supported inclusion of the qualifying conditions and options, but recommended modifications. The specific suggestions included one to strengthen the qualifying conditions by requiring States to “affirmatively demonstrate compliance” and others to strengthen the qualifying options by requiring that States adopt a higher number of the seven qualifying options.

Response: We proposed the qualifying conditions based on the philosophy that States out of compliance with related Federal requirements should not be eligible for a bonus. We also believed that States meeting the qualifying options would perform better on the outcome measure. However, we recognize that the inclusion of the qualifying conditions and options conflicts with the NGA/APHSA principle that a high performance bonus system should focus on outcomes rather than process.

In addition, we have concluded that the bonus award system is not the appropriate vehicle by which to evaluate or certify State compliance with Federal Medicaid requirements. For example, at the time we are making the high performance awards, we might not have completed a recent assessment of all State programs or there might be a potential compliance issue pending with one State that cannot be resolved in a short enough timeframe. Thus, we agree that it would be more appropriate to address such issues through ongoing Federal oversight of State Medicaid programs and a vigorous agenda of technical assistance and guidance. Therefore, we are dropping the qualifying conditions and qualifying options from the final rule.

Among the significant activities in the Department's agenda to resolve Medicaid enrollment issues are the following:

• Reviews of all State Medicaid programs, primarily during the summer and fall of 1999, to assess compliance with Medicaid requirements and to advise States when corrective actions are necessary;
• Issuance of additional program guidance to State Medicaid Directors clarifying the expectations that apply (e.g., the April 7, 2000, letter that addressed expectations with respect to reinstatements, redeterminations, and computer systems modifications);
• Development and distribution of thousands of copies of a guide entitled Supporting Families in Transition: A Guide to Expanding Health Coverage in the Post-Welfare Reform World, which explains the basic rules for Medicaid eligibility under the PRWORA amendments;
• Development and distribution of special guides for State and local partners in the child care and Head Start communities to promote their participation in enrollment efforts;
• Issuance of guidance encouraging States to use the $500 million made available to help them provide outreach and address administrative changes related to delinking development and distribution;
• Issuance of a TANF guidance (in the form of a guide entitled “Helping Families Achieve Self-Sufficiency: A Guide of Funding Services for Children and Families through the TANF Program”) making clear that States may use TANF funds to “provide outreach activities that will improve access of needy families to medical benefits under the Medicaid and [S]CHIP programs”;
• In cooperation with the Robert Wood Johnson Foundation, interdepartmental support for the “Supporting Families” initiative to assist 22 sites in assessing and resolving barriers to initial and continuous participation in Medicaid and SCHIP. (Six of these sites will look at food stamp issues as well);
• Related contract support for development of a literature review and “promising practices” report to provide background information and technical assistance for all States; and
• Meetings with State agencies to discuss access issues of general concern.

Comment: Several States disagreed with the six-month time frame in the outcome measure, primarily because tracking families who leave TANF for six months would impose a significant burden on States. Also, data collection is problematic because SCHIP is a stand-alone agency in many States; States cannot always match the Medicaid records to TANF records (e.g., because the case composition may be different under the two programs); and some States do not have social security numbers for all SCHIP participants to match with TANF records. Commenters generally suggested limiting the time frame to the month following the month that families leave TANF. One other commenter suggested that States also demonstrate that families accessed health care services.

Response: We had proposed the six-month time frame because most families who leave TANF are eligible for six months of transitional Medicaid or for ongoing Medicaid under other eligibility categories. We also believed that States could easily identify these cases. However, a time frame shorter than six months may reduce the tracking burden on States because families will presumably have undergone fewer changes in this shorter time period, and case management information may be more useful. For example, there should be fewer families that have moved out of State or that have experienced significant changes in family composition. At the same time, we believe that the recommended one-month time frame is too short. Our concern is that some States may carry Medicaid coverage for one month after TANF benefits are terminated for systems reasons; thus, a one-month coverage period would not fairly assess whether policies and systems were in place to ensure ongoing Medicaid coverage for eligible families.

We are revising the final regulation at § 270.4(d) to reduce the measurement period in both the absolute measure and the improvement measure from six months to four months. We believe that a four-month time frame better accommodates States' concerns about tracking and the availability of case management information while still providing a reasonable time frame for assessing Medicaid or SCHIP participation by individuals in families who leave TANF. Also, the four-month time frame accommodates families that receive Medicaid extensions based on increased child support collections since this form of transitional benefit only lasts four months.

Most families who leave TANF are eligible for Medicaid through transitional Medicaid, under section 1931 of the Act, or under the medically needy or poverty level groups. Because families eligible under Medicaid must enroll in Medicaid rather than SCHIP, the instances under which children will be eligible for coverage under a separate SCHIP program are greatly limited. States can use methods such as case identifiers to match SCHIP and TANF cases in those instances.

We have also required that States competing on these measures submit their information on a fiscal year, rather than a calendar year basis as we proposed in the NPRM. We are changing to reporting semi-annually on a fiscal year basis for ease of processing the information and to parallel the requirements for reporting information for the work measures.

Comment: Several commenters responded to our invitation to comment on our decision to limit the outcome measure to individuals leaving TANF assistance, rather than all eligible families and children. Most of these commenters recommended using the larger population of Medicaid/SCHIP eligibles to assess overall State performance since these programs provide critical supports to all low-income families and children. They believed that the proposed measure merely rewards States for complying with section 1925 of the Act, by providing six months of transitional Medicaid to certain families who lose TANF assistance.

Response: In view of the decline in the Medicaid rolls nationwide since 1995, continued Medicaid for families losing TANF is of particular concern. In the NPRM, we proposed to concentrate the performance measure on States' efforts to provide continued Medicaid for eligible families leaving TANF since this is an area of program administration that has been identified by consumer advocate groups and local and national studies as needing improvement. In the final rule, we have aligned the Medicaid provisions with the food stamp provisions to allow for consistency to the extent possible. However, unlike the Food Stamp Program, there are many variables, as discussed below, that affect Medicaid participation among populations other than TANF leavers. For this reason, and other reasons that we also discuss below, the Medicaid outcome measure differs from the food stamp outcome measure in that it does not assess State performance based on participation of all Medicaid populations.

In response to widespread concerns that PRWORA's delinkage of Medicaid and cash assistance had negatively affected access of low-income families to medical benefits, HCFA conducted on-site reviews in all the States and Territories from September to December 1999 to examine State TANF and Medicaid application and enrollment policies and procedures to ensure that eligible families learn about, receive, and maintain Medicaid coverage. A particular focus of these reviews was on how the TANF application, denial, diversion, and termination processes affect application for and receipt of Medicaid. Based on these reviews, HCFA is in the process of identifying areas where States need to improve their Medicaid application, enrollment, and re-enrollment processes either solely in their Medicaid programs or in conjunction with the administration of their TANF programs. HCFA released policy guidance in some of these areas by way of an April 7, 2000, State Medicaid Directors Letter. This guidance directs States to identify individuals who have been improperly terminated from Medicaid and to reinstate their coverage; clarifies the proper procedures for eligibility redeterminations, and reviews the obligations imposed by Federal law with regard to operation of computerized eligibility systems. In view of the need for continued improvement in these areas and the purpose of the high performance bonus, we believe that the high performance bonus system should include a Medicaid/SCHIP measure that focuses on how well States are meeting the TANF goals of work preparation, work and self-sufficiency.

As stated earlier, we limited the outcome measure to individuals leaving TANF assistance because: (1) States have a clear responsibility for serving these families under title I of PRWORA, i.e., under the amendments to section 193l of the Social Security Act; and (2) welfare “leaver” studies and other studies on program participation indicate that these families frequently are not receiving Medicaid/SCHIP. Furthermore, we believe this type of measure will result in a significantly smaller reporting burden for States, as well more accurate and consistent reporting.

We do not agree with the comment that this measure merely rewards States for complying with the law. The transitional Medicaid provision under section 1925 of the Act covers only those families who were eligible and received Medicaid under section 1931 of the Act. (This is the PRWORA provision covering families who meet AFDC-related eligibility standards for three of the six months prior to losing Medicaid because of income or employment.) Some families leaving TANF because of work are not eligible for transitional Medicaid because they were not receiving Medicaid under section 1931 for three of the six months before losing Medicaid due to earnings or income. However, under section 1931, States may adopt less restrictive income methodologies to ensure that families seeking TANF benefits, but moving quickly to work, can qualify for transitional Medicaid benefits.

In cases not covered by this transitional Medicaid provision, the dependent children generally continue to receive Medicaid under the State's poverty levels groups or may qualify for benefits under the SCHIP program, and we want to ensure that these children are receiving these benefits. A different provision in section 1931 of the Act provides transitional Medicaid for families losing benefits as the result of child support collections. Other families might be eligible under a State's medically needy group.

In addition, many of the factors affecting enrollment in Medicaid are not compliance matters. Two States that fully comply with all Federal requirements could have vastly different participation rates because of differences in how they operate their programs. While there are potential compliance issues with respect to matters such as ex parte redeterminations and proper notice, States have discretion in numerous areas of policy and administrative practice. For example, in the policy area, States have flexibility to expand eligibility coverage through the use of more liberal income and resource standards and methodologies. In the area of administrative practice, States have broad flexibility with regard to variables such as the location of eligibility offices, office hours, length of application, amount of verification required, outstationing, and use of mail-in and phone-in applications to eliminate barriers to and simplify the application process and reduce procedural requirements. Limiting the outcome measure to families who leave TANF but remained enrolled in Medicaid focuses on the only group for which there are data easily accessible to all States on a uniform basis.

In view of the reasons stated above addressing the responsibilities of States to provide Medicaid to eligible families, particularly those leaving TANF, and the flexibility afforded States to meet these responsibilities, we are retaining in the final rule at § 270.4(d) the outcome measure limited to those families who leave TANF, but are enrolled in Medicaid after leaving TANF.

Comment: One commenter observed that the September 30, 2001, sunset date of section 1925 of the Act complicates the measurement since eligibility for transitional benefits will change after that date and one would expect a number of families would not be entitled to be on the rolls by month six.

Response: The commenter is correct. Unless Congress acts to change this provision, beginning in FY 2002, families who lose Medicaid eligibility because of hours of, or income from, employment or because of loss of earned income disregards will be eligible for a minimum four-month period of extended Medicaid eligibility. Since we are reducing the measurement time frame from six months to four months in this final rule, we do not expect States will be adversely affected in competing for an award.

Comment: A number of States commented that there are other factors—such as moves out-of-State, death, changes in family formation, increased earnings, and enrollment in private health insurance—that affect participation in Medicaid and SCHIP. Of particular concern was that the proposed measure did not allow any adjustment for families who obtain private health insurance.

Response: We agree that there are numerous factors that affect families' participation in the Medicaid and SCHIP programs. However, if we were to provide an adjustment for all circumstances that can affect States' caseloads, the outcome measure would conflict with the NGA/APHSA principles that the bonus system should: (1) Be simple, credible and quantifiable; and (2) rely on existing data where available. Also, if all States could not identify or exclude cases based on deaths, moves out-of-State, or other circumstances, then allowing adjustments would disadvantage the States that could not do so, thus creating an uneven playing field.

Similarly, we have decided not to make adjustments for families who obtain private health insurance after leaving TANF. While we applaud State efforts to get individuals into jobs that provide health coverage and related benefits, and we would like to be able to credit States somehow for success in this area, adequate data on private health insurance coverage do not exist. Further, the costs that would be associated with collecting comparable, adequate data for all States would be prohibitive. One underlying issue is that private insurance coverage varies substantially across employers and individual employee circumstances. This variability suggests that, in order to treat States fairly, we should somehow measure the quality and level of coverage under the private plans and include appropriate adjustments.

In summary, since private plans seldom offer benefits comparable to Medicaid, we would not necessarily want to give States the same credit for private coverage. Furthermore, participation in employer-sponsored insurance does not affect an individual's entitlement to transitional medical assistance.

Finally, although we recognize that some States have given a high priority to job placements that provide health coverage and have achieved some success, the national statistics suggest that it would be quite rare for those entering low-wage jobs to obtain private health insurance that is affordable and comparable to the benefits provided by Medicaid. The types of employment situations these families generally access, especially in the short run, mitigate against adequate health coverage. The types of industries and the types of occupations, union status, the size of establishments, length of time on the job, and the use of part-time or temporary employment all increase the chances that a family would not have adequate coverage.

For example, according to a study in the June 1995 Monthly Labor Review, while six of ten workers had employer-based coverage, only one-fourth of service workers in service-producing industries had such coverage. The study also noted that, in general, industries where coverage has traditionally been more prevalent have been in decline as a portion of the U.S. economy. It cited service jobs such as waitressing, cosmetology, and cleaning as occupations with less access to employer-based coverage. According to more recent data (1996-1997) from the Bureau of Labor Statistics (BLS), while 70 percent of those employed full-time in the private sector participate in employer-based plans, only 11 percent of part-time private-sector employees do.

In the event that private insurance is available, it is apt to be under a plan with a limited benefits package. Data from the 1996 and 1997 BLS Employee Benefits Surveys show that 78 percent of those participating in an employer-based plan had to contribute to the cost for family coverage (“Compensation and Working Conditions,” Winter 1999). Earlier BLS data showed a 200 percent increase in the average family premium between 1983 and 1993. Accordingly, “premiums may be difficult for some workers to afford, causing them to decline coverage.” Also, about three-fifths of all participants have coverage that is subject to a pre-existing condition clause, limiting the care they can receive. Thus, even where individuals have access to employee benefits, we are concerned that families receive the wrap-around services that Medicaid can provide and to which they are entitled.

Comment: A number of States commented that the performance measure disadvantages States that have achieved high levels of Medicaid/SCHIP participation since low-performing States can make greater improvements. A few of the commenters recommended that we use a percentage point improvement measurement rather than percentage improvement measurement to even the playing field.

Response: We agree that high performing States may be disadvantaged by the performance measure as proposed. Therefore, in the final regulations at § 270.4(d), we have substituted a measure of percentage point improvement for the proposed percentage improvement measure. We believe the measure in the final regulation puts States on a more even playing field, regardless of their baseline level of performance. At the same time, States whose performance was relatively low in the base year are still in an excellent position to be rewarded for their efforts toward increasing enrollment.

Under the proposed measure, a State that increased Medicaid participation from 5 percent to 10 percent (a 100 percent improvement) would be ranked higher than a State that increases participation from 30 percent to 45 percent (a 50 percent improvement). In the final rule, a State that increases participation from five percent to ten percent would achieve a five percentage point improvement while a State that increases participation from 30 percent to 45 percent would achieve a 15 percentage point improvement and would be ranked higher than the first State.

Comment: Some commenters recommended an absolute measure rather than a performance improvement measure. Other commenters recommended expanding the measure to have both an improvement measure and an absolute measure so that States could be ranked separately on both measures. They also suggested that we divide the bonus funds among the top five States achieving high performance in both measures.

Response: We agree that adding an absolute measure provides for a fairer system of awards and have made this change in the final rule. Under the NPRM, States that engaged in early and successful efforts to increase enrollment could have found it extremely difficult to compete for these awards, and their efforts might never have been recognized. The final rule sets aside some of the bonus awards for those States that have the highest overall success in enrolling individuals in eligible families. For the absolute performance measure, we will rank the States in order of the highest percentage of Medicaid/SCHIP participation rates by individuals in families four months after leaving TANF and who are not receiving TANF in the fourth month.

In awarding bonuses to the top performing States under both performance measures, we will divide the $20 million in bonus funds, as specified in § 270.8, among the three States that have the highest percentage of Medicaid/SCHIP participation rates by individuals in families four months after leaving TANF and who are not receiving TANF in the fourth month (absolute performance measure) and the seven States that show the most improvement performance from the previous year to the current year (improvement performance measure). This allocation of bonus funds among the States is consistent with the Food Stamp allocation of bonus funds among States that compete on the Food Stamps measure.

Comment: One commenter noted that the Medicaid measure should have a method for dealing with tied scores.

Response: We agree with this comment. In the interest of consistency, we will use the same method that we used for the work and food stamp measures. We specify in paragraphs (d)(1) and (d)(2) that we will calculate the percentage rate for the two measures to two decimal points. If two of more States have the same percentage rates, we will calculate the rates for these States to as many decimal points as necessary to eliminate the tie.

Also, since participation in the high performance bonus award system is voluntary, we will rank only those States that choose to compete, notify us by February 28 of the bonus year of their intent (§ 270.11), and provide the requisite data.

New Section 270.4(e) Child Care Subsidy Measure

A substantial number of commenters recommended a child care measure, either as an additional new measure or as a replacement for the family formation measure. We agree that child care is of critical importance to working families, and we share the commenters' view that access to affordable, high quality care is a necessary part of a welfare reform program. We believe that high quality care is of critical importance to children of TANF and other very low-income families. A growing body of research indicates that quality and stability of the child care setting influences outcomes for children as well as the ability of parents to retain employment.

In support of their recommendations, commenters observed that currently there is no reward for good performance in providing child care. A child care bonus, they believed, would be an excellent incentive toward better State performance, given that only a small percentage of income-eligible children are now receiving subsidies, according to published estimates from this Department. Many commenters based their rationale simply on the fact that child care is a critical work support needed in order for poor families to work. The suggestions for how we should construct a specific child care measure centered primarily on the number or proportion of children receiving child care subsidies and/or the amount of expenditures on eligible children.

In developing the NPRM, we had considered the possibility of including a child care measure, not only because of its importance to working families but also because the CCDF program and the TANF program are closely related. However, we did not propose a child care measure for various reasons. Our reasons included the lack of currently available data that would completely capture State performance on all of the crucial elements, including quality, affordability and accessibility of an effective child care subsidy delivery system. We also took into account the lack of data that would fully capture how well States are performing in serving the TANF/CCDF population given the considerable duplication in data sets and inconsistencies in statutory data collection between the two programs.

Based on the extensive support among commenters for a child care measure and upon further evaluation of the availability of data, we have devised a child care measure that we believe will reward States based on an appropriate range of important child care program elements. This new measure is located at § 270.4(e) of this final rule. The three components of the measure address child care accessibility, as indicated by the percent of CCDF-eligible children receiving services; affordability, as indicated by assessed family co-payments; and quality, as indicated by State reimbursement rates. We believe it is essential to include all three components in order to assess a State's performance in making high quality child care more accessible to low-income families. We will use Census Bureau and existing CCDF data on two elements for the FY 2002 bonus year and add an additional element for subsequent years that will require additional information to be reported by States choosing to compete for the measure. The top 10 performing States that choose to compete on this measure will each receive a portion of a total of $10 million in bonus funds awarded annually for this measure.

We intend to engage States and others, particularly data experts, in discussions regarding the technicalities of implementing key elements of the measure. While there were many comments in support of a child care measure, and many of these comments supported the areas that we have selected for measurement—accessibility, affordability, and quality of child care—there was no opportunity for detailed consultation or comment on the technical aspects of measurement within these areas, because a child care measure was not included in the NPRM. After consultation with States and others. We intend to issue details regarding the components of the measure by the end of the calendar year.

For the FY 2002 bonus year, the measure consists of two components: the percentage of eligible children served and the affordability of care for the families of the children served as indicated by the relationship between the State's reported family CCDF co-payments and reported family income. These components of the measure use existing data reported by the States on the ACF-801 and the ACF-800 as the source for the number of children served, family copayments, and family income. We will calculate the percentage of children served with “pooled” funds, i.e., CCDF funds (including transfers from TANF) and any other funds that are reported to us on the ACF-696 (CCDF financial reporting form) for the applicable year.

Each State's rank on the measure in the FY 2002 bonus year will be a composite weighted score of the two components, with the component on percent of population served having a weight of 6 and the component on affordability of family co-payment having a weight of 4.

We will use Census Bureau data (the Census 2000 Supplementary Survey and the Long-Form Transitional Database) as the data source for family income at 85% of the State's median income, i.e., the Federal eligibility limit set in statute for the CCDF, to determine income-eligibility in calculating the percentage of children served.

To determine affordability, we will compare family income with assessed state family co-payment as reported on the ACF-801. Because States have tremendous flexibility in setting sliding fee scales under the regulations governing the CCDF, in order that they can balance different needs and make child care affordable for families at a range of incomes, we will refine the technical details of this measure through additional consultation with States and data experts.

For FY the 2003 bonus year we will further strengthen the measure by adding a third component that compares actual rates paid by the State to the market rates applicable to the performance year. Each State's rank on the measure in the FY 2003 bonus year will be a composite weighted score of the three components, with the component on percent of population served having a weight of 5, the component on affordability of family co-payment having a weight of 3 and the third component on the comparison of rates paid to market rates having a weight of 2.

This third component cannot be implemented in the FY 2002 bonus year because States do not currently report the data collected in the CCDF-required market rate surveys, nor is there any consistency among States in how the surveys are conducted. However, we believe this additional component of the measure will strengthen our ability to assess a State's performance with respect to both affordability and quality, since access to higher quality, more stable care for families receiving subsidies is often linked to the rates paid to providers by the State.

This component will use existing data on actual rates paid for children receiving CCDF subsidies as reported on the ACF-801, and data on actual market rates that will be submitted by those States that choose to compete on the child care high performance measure. We will draw the necessary data from the market rate data collected by the State in the CCDF-required survey. Consistent with existing CCDF requirements, this survey must be completed no earlier than two years prior to the beginning of the performance year when the performance year is the first of the biennial State CCDF Plan cycle, or no earlier than three years prior to the beginning of the performance year when the performance year is the second year of the State CCDF Plan cycle. While States must complete their surveys within the specified time frame, CCDF regulations do not require submission of the survey data. A process for submission of this data by States choosing to compete on the child care measure and the precise methodology to be used in ranking States on the relationship between rates paid and market rates will be developed through additional consultation with States and data experts.

For all bonus years, we will distribute bonuses to the top 10 qualifying States that have both fully obligated their CCDF Matching Funds for the fiscal year corresponding to the performance year and fully expended their CCDF Matching Funds for the fiscal year preceding the performance year. The source of this financial information is the ACF-696 for the corresponding bonus performance period. This requirement contributes to the effective use of Federal funds and to a level playing field across States, by ensuring that no State can win the child care high performance bonus through substituting TANF or other 100% Federal funds for CCDF Matching Funds (although States may certainly add resources to the CCDF “pool” of funds). Thus, all States competing for the child care bonus must have committed all of their dedicated child care funds.

We address commenters' specific recommendations for the child care measure below. The approaches suggested by many of the commenters were similar. The commenters proposed a variety of options, some more detailed than others, while none provided an in-depth analysis of potential data sources. To avoid repetition, we have organized the comments by type, rather than by content of individual letters.

Comments: Commenters suggested that a child care measure might be based on various target populations, including:

(1) TANF recipients and former TANF recipients;

(2) Children eligible under the provisions of the Child Care and Development Fund (i.e., at or below 85 percent of State median income);

(3) Children at or below 200 percent of poverty;

(4) Children served under both the Child Care and Development Fund (CCDF) and, at State option, other subsidy programs funded by TANF or State sources; or

(5) Children in two-parent households receiving child care services (as a suggested substitute for the family formation measure).

Response: We concur that a child care measure should take into account the population served, i.e. a measure of accessibility to subsidies. Therefore, we have incorporated percentage of CCDF-eligible children served into the composite child care measure. We will include children served with “pooled” funds (all funds reported on the ACF-696 for the period corresponding to the performance year) in the percentage. We believe that the most appropriate denominator is the target population eligible under the Child Care and Development Block Grant Act, i.e., at or below 85% of the State Median Income.

We do not have an existing data source that would accurately capture child care subsidy services to all TANF recipients. Nor could we determine a method of eliminating duplicate counting of children served with TANF and CCDF maintenance-of-effort funds.

Since we are adopting a separate family formation measure, we did not see a rationale for focusing the child care measure solely on two-parent families. Child care is a critical support for one-parent as well as two-parent families when parents are working. Our measure does recognize that this is a TANF high performance bonus by capturing those children connected to TANF, transitioning from TANF or at risk of becoming eligible for TANF who are served in the CCDF system, including families served with “pooled” funds reported on the ACF-696.

Comments: A few commenters suggested that the bonus be based on a measure of State expenditures on child care subsidies divided by the estimated number of federally-eligible children under the age of 13. They suggested this simple measure in recognition that there are problems with consistency among programs in eligibility, payment levels, and other factors.

Response: While this measure would involve a minimal reporting burden, we do not believe that this is a meaningful measure, since it would not capture a measure of the extent of services provided to families, as our measure does. This suggested measure would be more process-based than the measure we adopted and would ignore critical elements of a State's child care performance.

Comments: A few commenters also suggested using both an absolute and an improvement measure for one or more of the suggested components of the child care bonus.

Response: We believe that it is most important to focus upon an absolute measure. First, we could not include quality until the third year of child care bonus if we were to seek to include an improvement measure, because there are no baseline data available. We would be very concerned about this consequence because, as we have noted, we believe that a balanced measure of accessibility, affordability, and quality is crucial to ensure beneficial outcomes for children. Second, we believed it to be especially important to reward those States that have already made considerable progress in improving the access, affordability, and quality of care for low-income families.

Comments: Some commenters suggested factors such as high payment rates and low family co-payments. One suggestion was to pay an increased bonus to States that adopted reimbursement rates at the 75th percentile of the local market rate.

Response: We concur that the child care performance measure should contain an indicator of affordability. Therefore, we have incorporated an indicator of affordability into the composite bonus, beginning with the first child care bonus year, by measuring the relationship of family copayments to family income.

In the second year of the child care bonus we will add a component that compares reimbursement rates to applicable market rates. These facets of the measure will also at least indirectly address quality of services, since families in top performing States would likely have access to a broader range of higher quality care, which often costs more than mediocre care. We cannot implement the payment component in the FY 2002 bonus year because States are not currently required to submit data on the relationship between their reimbursement rates and the market rates in the State. Nor do we have access to consistent information on what constitutes the 75th percentile in each State.

We believe that this approach is stronger than an approach that would link extra bonus funds to a specific level of rates, such as the 75th percentile. First, we believe that an approach that rewards States for rates that provide more access to the market for low-income families without setting a single standard is more consistent with the flexibility in the CCDF statute. The CCDF final rule (63 FR 39959) uses the 75th percentile as a benchmark, not a requirement. Second, we want to ensure that States have a continued incentive to improve the access of low-income families to the market beyond any specific marker. Third, we want to ensure that the bonus approach does not inadvertently inhibit the ability of States to try a variety of approaches to rate-setting that might enhance quality, including rate structures that offer incentives linked to provider qualifications, certification, or other quality measures. In the context of a shortage of dedicated Federal child care funds and the trade-offs that States could be forced to make as a result, we are concerned that a bonus linked to a single approach to payment could inadvertently be counter-productive. Our approach is intended to reward States that, by making the best use of all the resources and choices available to them, have established higher rates.

Comments: Some commenters suggested that the child care bonus incorporate certain measures of quality. The ideas forwarded by one or more commenters consisted of:

• Excluding children who are in informal or unlicensed care from the population measure;
• Measuring the use of licensed care by subsidized families; and
• The payment of higher rates to accredited centers.

Response: We concur that quality child care is important for the healthy development of all children and is especially crucial for children in low-income families who often are disadvantaged educationally as well as financially. Thus, we incorporated quality into the bonus measure by looking at factors that allow families to pay for better care, an approach that is consistent with the parental choice concept that is central to the statute governing the CCDF. Licensing and certification systems vary greatly, and we do not have the data to determine access to accredited care. Moreover, children often are in multiple arrangements or frequently are moved to different settings for various reasons. Available data would not support an unduplicated measure related to use of specific types of care.

Comments: Several commenters also suggested a child care threshold measure, either as a qualifying factor for a child care bonus or in order for a State to receive other bonus awards. They suggested:

1. Before a State could compete on a child care measure, the State must pay a child care rate equal to at least the 75th percentile of the local market rate based on a survey that is not more than two years old; or

2. To qualify to compete, a State must spend a required percentage of child care funding for care that meets State certification standards.

Other suggestions included granting high performance bonus awards only on the condition that a State based its child care payments on recent market surveys or served at least 70% of its income-eligible children.

Response: We agree with commenters on the critical importance of quality and access to child care services. However, we do not believe it is useful to include these factors as qualifying conditions. Instead, we have taken a stronger position by addressing payment rates and percentage of children served directly in the child care measure we developed. In our response regarding the suggestion that we include use of licensed or accredited care in the bonus itself, we explained why we believe that there are not data available to make a measurable determination on use of such care. As we pointed out earlier, we also sought to develop a measure that supports the central CCDF concept of parental choice.

Other Recommendations for New Measures in or Approaches to the Bonus System

In addition to recommendations for a new child care measure, we received many comments for other measures as well as suggestions for how the high performance bonus system might be improved. We summarize and respond to these comments below.

A. Domestic Violence

Comments: A substantial number of the letters and notecards recommended adding a measure of how well States address domestic violence. In support of this recommendation, some commenters provided detailed background information about the prevalence of domestic violence among women on TANF and how domestic violence can hinder an individual's ability to maintain work in a way that leads to self-sufficiency.

Most commenters on this issue recommended that the measure be designed to “look at the proportion of women who disclose they are victims of domestic violence who receive services or waivers under the [TANF] family violence option.” Alternatively, a few other commenters suggested that we add a threshold measure related to domestic violence, i.e., only States that adopted the TANF Family Violence Option and “meet the requirements of federally recognized good cause waivers” (45 CFR 260.55) could compete for other bonuses. One commenter suggested that the Department would have to adopt “a detailed statement on how to effectively implement a Family Violence Option”; another commenter suggested that competition for the bonus include interviewing domestic violence advocates in the State.

Response: We strongly agree that domestic violence services are important to the well-being of families and to support work and self-sufficiency. We are committed to efforts that both serve families who are victims of domestic violence and implement prevention programs. The Department has underway an on-going, coordinated, multi-agency initiative on Family and Intimate Partner Violence. This initiative is comprised of a wide range of activities whose purpose is to:

• Strengthen the health care system's ability to screen, treat, and prevent family and intimate partner violence;
• Provide education, training, and support for battered women and their families;
• Increase the ability of battered women, including those on welfare, to obtain and retain employment and obtain child support;
• Encourage greater linkages between child welfare, family and intimate partner violence, and criminal justice fields to better protect both children and parents in homes where violence occurs;
• Enhance community prevention and response systems by increasing collaboration between the Department's State and Tribal family violence grantees and the Department of Justice's State and community-based grantees and other community-based groups; and
• Increase the knowledge base about family and intimate partner violence, through data collection and research;

Specific examples of activities related to this initiative include:

• As the commenters recognized, the TANF final rule includes provisions pertaining to the Family Violence Option instituted under PRWORA. The TANF final rule also provides States penalty relief when they fail to meet the numerical standards for time limits and the work participation rates because they provide good cause domestic violence waivers to battered women. It also includes provisions for the reporting of the strategies and procedures the State has put into place to ensure that victims of domestic violence receive appropriate alternative services.
• ACF awarded grants to several States and localities to increase collaboration between domestic violence programs and welfare programs. These grants have been used for training, policy development, and joint intervention responses.
• ACF's Office of Child Support Enforcement (OCSE) has four ongoing grants examining child support cooperation/good cause and domestic violence, and a fifth cross-site evaluation of the projects. The grants will provide additional information about the incidence of domestic violence among child support recipients and ways the child support and domestic violence communities can work collaboratively to meet the needs of battered women.
• OCSE has also been working with States on implementation of the “Family Violence Indicator,” an automated flagging mechanism within OCSE's national database, the Federal Parent Locator Service, that will prevent the release of data on battered women.
• ACF has formed an interstate domestic violence working group that is examining a number of issues surrounding domestic violence and child support.
• The Department's Office of the Assistant Secretary for Planning and Evaluation is conducting assessments of State policies and practices regarding domestic violence in the TANF program.

ACF will issue periodic reports and technical assistance materials reflecting the results of these and other activities. For example, the Center for Law and Social Policy, in conjunction with OCSE, recently published “models of Safe Child Support Enforcement,” a guide for States and others.

Finally, through the Family Violence and Services Program, ACF also provides grants to all States, all State coalitions, and a number of Indian tribes to provide immediate shelter and related assistance to victims of family violence and their dependents. ACF also funds five national resource centers and the national Domestic Violence Hotline.

Although we are committed to addressing the problems and the often tragic consequences of domestic violence, our task with respect to the high performance bonus was to assess the appropriateness of such a measure in the context of our policy and evaluative framework. In the NPRM, we indicated that we considered a measure looking at the proportion of TANF recipients who received domestic violence services, but we noted that we had identified no objective and reliable data sources for this measure. Similarly, there is no existing source or uniform standards for determining whether a State is meeting federally recognized good cause waiver requirements (especially if it is not penalty-liable), and no existing Federal standards for qualitative measures of service.

After carefully considering the comments and verifying that no data were available to support an outcome rather than a process measure, we decided we would not include a domestic violence measure in the bonus system at this time. Because competition for the bonus is voluntary, we were concerned that additional data collection would be burdensome and would not generate competition—particularly since we had a fixed amount of bonus funds.

We also evaluated the recommendation for including domestic violence as a threshold measure, i.e., only States that adopted the Family Violence Option and met the federally recognized good cause provisions for domestic violence waivers would be allowed to compete for other bonuses. We have not accepted this suggestion. First, in this final rule, we have dropped the qualifying conditions and qualifying State options as threshold measures in the Food Stamp and Medicaid/SCHIP performance measures, and we decline to add a new threshold measure. Second, the suggested threshold is already used in the TANF penalty reduction process, and, thus, is a substantial incentive for States to adopt these practices in addressing domestic violence under welfare reform.

Nevertheless, we are continuing our efforts to encourage all States to plan pro-actively to meet the needs of victims of domestic violence and their dependents. We will be making technical assistance materials and research results available to States to enhance their efforts to prevent domestic violence and provide services to those in need.

B. Worker Displacement/Worker Protections

In the preamble to the TANF final rule (April 12, 1999, 64 FR 17748), we indicated that we would invite comments on whether to include a worker displacement/worker protection bonus measure during our rulemaking on the high performance bonus. We carried out this commitment in the NPRM when we specifically asked for comments on whether we should “consider State enforcement of the TANF non-displacement requirements in awarding bonuses, and, if so, how.”

Comments: We received four letters in support of new worker displacement and/or worker protection measures. Commenters recommended the following:

• Require States to provide evidence of anti-displacement measures, e.g., demonstrating the existence of a grievance procedure, either as a threshold measure in order to compete for any high performance bonus measure, or as a threshold measure in order to compete for the work measures.
• Add a new measure based on: (a) Evidence of the integration of TANF, the Work force Investment Act (WIA), and the Welfare-to-Work (WtW) worker protection procedures (e.g., a Memorandum of Understanding with appropriate agencies, providing worker information, and monitoring); and (b) submission of payroll records by employers with significant numbers of TANF employees or review of unemployment insurance records through which displacement might be detected.

These commenters believed in the importance of worker displacement protections in the TANF program and noted that they found that some States have been slow to put procedures into place. They provided one concrete suggestion for data sources. However, adopting this recommendation would have required additional State reporting—an approach that we said we wanted to avoid. One commenter specifically acknowledged that it is difficult to measure displacement accurately.

Additionally, as the commenters also observed, the statutory WIA and WtW requirements in the area of worker protection are more detailed than the statutory TANF requirements; these differences would appear to pose problems for establishing uniform standards.

Response: As we said in the preamble to the TANF final rule, it would not be consistent with the principle of State flexibility embodied in the TANF statute for us to regulate a State's administrative procedures and require States to adopt the more extensive WtW statutory provisions for the TANF program.

Worker displacement is a matter of concern to us, however, and we will be monitoring it through review of information each State provides to us in its TANF annual report. Specifically, a State must include a description of procedures that it has established and is maintaining to resolve displacement complaints. (See 45 CFR 265.9(a)(7).)

At the same time there are no standards available for us to objectively assess the extent and quality of State displacement procedures. Thus, we do not believe that we have adequate criteria or data upon which to base either a threshold or additional performance measure.

C. Child Poverty Measure

In the preamble to the proposed rule, we discussed our consideration of whether to include a child poverty measure in the high performance bonus system. We cited the importance of this matter, our belief that States had the flexibility and resources to make an impact on child poverty, and the connection of a child poverty measure to two of the purposes of TANF: (1) Promoting work and employment; and (2) strengthening child and family well-being by assisting needy children in their own homes or in the homes of relatives. We invited public comment on this issue.

On the other hand, child poverty is an area for which there are other mechanisms in the statute for monitoring and promoting positive State action. Section 413(i) of the Act requires that States report their child poverty rate annually and take corrective action when an increase in the child poverty rate is the result of the TANF program in the State. We published a final rule implementing these provisions on June 23, 2000 (65 FR 39234).

Comments: A number of commenters were concerned that a State could not be performing well if large numbers of already poor children and families were allowed to fall deeper into poverty. One commenter suggested that we add compliance with the child poverty requirements under section 413(i) of the Act as a threshold measure.

The commenters recommended the use of the official poverty measure (developed by the Census Bureau) and suggested possible measurement approaches for our future consideration. They also suggested that when there were economic conditions beyond the control of States, ACF could use rules for setting aside the measure.

Response: We continue to believe that poverty, and child poverty in particular, is an issue of great importance, but we are not convinced that the best way to address the issue is through the TANF high performance bonus award. We considered the recommendation that State compliance with section 413(i) of the Act be added as a threshold measure, but we believe that the several requirements States may need to meet under section 413(i) do not lend themselves to effective inclusion in the high performance bonus system.

D. Other Suggested New Measures

Comment: One national organization recommended the addition of a number of new measures directed at achieving economic independence and self-sufficiency, particularly for women and girls. As a part of their recommendations, they asked that we define the term “self-sufficiency” in the final rule, using the Wider Opportunities for Women's “Self-Sufficiency Standard.” They believed that this comprehensive standard, which describes how much money is needed to meet a family's basic needs (i.e., for housing, food, child care, transportation, clothing, and related work expenses, calculated by family size on a per county basis) without public assistance, is an accurate and sensible indicator of true self-sufficiency. Without such a measure, they believed that it would be difficult, if not impossible, to accurately and consistently define self-sufficiency for a given family in the United States.

Response: We agree that helping families achieve economic independence and self-sufficiency is one of the most important goals of the TANF program, and we believe that the recommended “Self-Sufficiency Standard” is a useful tool in evaluating State and local efforts towards achieving that goal. However, since none of our measures incorporate this term, and any measure based on this concept would entail substantial new data collection, we have not accepted this recommendation.

Comments: A number of commenters mentioned other topics on which new measures might be based, but the commenters' suggestions were general and, for the most part, undeveloped in terms of both design and data sources. One or two letters suggested each of the following topics:

• Diversion from TANF;
• Recidivism, i.e., returns to TANF;
• Savings and asset-building;
• Housing (with the measure to be developed by ACF in collaboration with the Department of Housing and Urban Development, e.g., a measure of the use of TANF funds to provide housing assistance for families moving from welfare to work);
• Transportation, e.g., demonstrating cooperation between the TANF and State transportation agencies;
• Education or training, i.e., the number of teens on TANF attending or completing high school or an equivalency program, or, using performance data under the Carl D. Perkins Vocational and Technical Education Act to determine the skill attainment success rate of public assistance recipients who enroll in advanced education and training programs following job entry;
• Reduction in the incidence of teenage pregnancy through abstinence education and other programs that encourage children to postpone starting a family until married;
• Availability of and utilization of various transitional services, e.g., Medicaid, and transportation;
• Programs to enhance family relationships and reduce family violence;
• Decreases in the number of children in foster care; and
• Child support, i.e., measurement of how many families in transition from welfare to work are receiving payments on child support or payments on child support arrearages; and medical child support, i.e., measurement of the number of children in families that are in transition [from TANF] that have medical support orders as part of their child support order and who are receiving benefits from those orders.

Response: Many of these suggested areas are viable strategies for helping families move toward self-sufficiency. However, most of these suggestions do not lend themselves to the construction of a quantifiable outcome measure; lack an objective, uniform, and reliable data source; or have other problems when viewed in the context of our policy and evaluative framework. In the case of the suggestion regarding child support and associated medical orders, the proposed measure would duplicate the existing incentive and penalty system that is already part of the child support enforcement program.

Comment: Another suggestion, rewarding States for the enactment of a State EITC, had support from a substantial number of commenters. However, none of the commenters was specific about how we might construct an outcome measure. One commenter noted that the existence of a State and local EITC program worked against the State's being able to compete successfully in the proposed family formation measure. Another commenter noted that enrollment in EITC is a difficult measure to document.

Response: We agree that an EITC program can be a major support to families working towards self-sufficiency. At the present time, however, approximately ten States administer State EITC programs. We believe that, to be a meaningful incentive, the high performance bonus system should offer all States the opportunity to compete on all of the measures. Thus, we have not accepted this suggestion.

E. Recommendations for Other Approaches to the Design of the Bonus System

We received a few additional recommendations and suggestions to which we wish to respond.

Comment: A commenter recommended that DHHS adopt an overall high performance award that would be given to the top one to three States that exhibited overall outstanding success in meeting the TANF goals. They argued that such a bonus award was needed to overcome various shortcomings in the proposed system, e.g., rewarding States for one area of performance when they fell short overall, giving awards in areas of performance where all States were experiencing only mediocre success, and having the unintended effect of encouraging States to take a narrowly focused approach to welfare reform in order to achieve the specified rewards. Such a measure was needed, they believed, to measure performance with respect to all poor families in the State, not just the segments of the population covered by the proposed bonuses.

The commenter recognized the lack of currently available data that could be used to support such a measure, but suggested that we could develop such data. They gave examples of factors that we could measure in an overall bonus but did not go into detail regarding the data that would need to be developed. Rather, they suggested an alternate approach of letting individual States choose multiple areas in which they wished to compete.

Response: We have not accepted this recommendation for a number of reasons. As the commenter pointed out, data are limited and would have to be developed in order for the measure to be workable. From a logistical standpoint alone, new information sources could not be developed in time to be implemented for the 2002 award year.

We believe the measures in the final rule permit an overall high performing State to compete successfully on several of the measures and to receive a commensurate monetary award. At the same time, a statutory cap on the total award a State can receive in a year provides a deterrent to any misplaced incentive to over-focus on certain aspects of welfare reform that might be generated by the proposed system of bonuses.

In addition, having States compete for an overall bonus that encompasses items of their own choosing would not only be exceedingly complex, but would be a challenge to objectivity. Also, based on our experience in awarding the FY 1999 bonuses, we did not identify a large number of States winning on certain measures that exhibited only mediocre success on other measures.

Comment: Two commenters suggested new bonuses based on innovative or extraordinary practices. One commenter encouraged bonus awards for “extraordinary practices that lead to quality achievement in advancing the goals of TANF.” Similar to the previous comment, the commenter noted that this approach would recognize individual programs that show promising results but do not directly contribute to a State's performance in the specific bonus categories. The commenter also noted that this approach would serve to reward individual efforts in States “that, for various reasons, may not be able to compete well in the specific categories.” One commenter also suggested a separate measure that “rewards states for implementing policies or programs that address a particular need in their state.” The State would identify a specific category of client need and submit an explanation of how the State responded to the need.

Response: While we agree that innovative or extraordinary practices deserve to be recognized and promoted, a high performance bonus award for such practices would not meet the statutory requirements for comparing and ranking States and awarding bonuses based on a formula. In addition, such a bonus would focus on process rather than outcomes.

ACF has in place a strong system of peer technical assistance that serves to recognize promising practices. The Welfare Peer Technical Assistance Network provides guidance and instruction to States on promising practices for moving welfare and low-income families to self-sufficiency. It provides information to States about TANF program resources and increases communication among the States about promising practices. Access to this Network is available at: http://www.calib.com/peerta/.

Comment: In response to a question we posed in the NPRM regarding whether we should consider thresholds (such as denying a bonus to a State that was subject to a work participation or other noncompliance penalty), one commenter suggested that a competing State should not be in a penalty situation of any kind. The commenter stated that the purpose of the bonus is to reward high performing States and that penalties indicate a failure to perform in some way.

Response: Although we specifically solicited comments on whether we should consider such a threshold measure, we received only one recommendation that we do so. We gave this proposal serious consideration and concluded that potential for delay in the due process provisions for TANF penalties would make the recommendation difficult to implement, i.e., to match the penalty year with the performance year for the bonus. We also foresaw other implementation issues, such as how to treat States under corrective compliance plans and whether to take into account the severity of a penalty.

Comment: A national organization recommended that the Department “collect and evaluate data broken down by race and ethnicity for all components of the high performance measure, and move towards the ultimate goal of assessing states' efforts to eliminate racial, ethnic, or other disparities in their welfare programs.” To support this proposal, the commenter cited various studies that reported disparity in treatment of various population groups and in their success in leaving the welfare rolls. The commenter said that “states should not be rewarded if their programs treat certain groups of clients differently,” but should be rewarded “for working proactively to address the different needs of different communities and to correct problems that may occur in their programs.”

The commenter recommended that ACF could use the data sources proposed in the NPRM since they either contained race/ethnicity variables or could be matched with a data source, such as TANF data, that contained the variables. The commenter acknowledged, however, that there could be additional reporting burden and that “states are the only entities in a position to perform the matching of data to determine whether certain groups are faring worse than others.” The commenter urged DHHS to explore ways to incorporate the results of the various studies and analyses into the bonus system.

In addition, the commenter suggested that DHHS “use any statistically significant racial or ethnic disparity of outcomes as a factor in evaluating the state's performance the following year,” i.e., to show improvement in the disparity as a threshold measure for competing in the measure in which the disparity occurred.

Response: We agree with the importance of equitable treatment of various welfare populations. The Department has demonstrated its strong commitment to fair and equitable treatment of individuals in the welfare population. For example, in August, 1999, we provided governors and State TANF administrators with comprehensive written technical assistance to help them understand the application of Federal civil rights laws in the implementation of welfare reform. We have also worked on developing technical assistance to clarify the responsibilities of health and social agencies that receive Federal funding in fulfilling their responsibilities to persons of limited English proficiency, pursuant to title VI of the Civil Rights Act of 1964.

We also believe that, if it were feasible to secure adequate data, there may be factors beyond a State's control when there are disparities in the outcomes among welfare sub-populations. Immigration and refugee resettlement patterns are examples of factors that are beyond State control. Thus, it would be difficult to construct a measure that would be fair to States. The commenter did not assess the availability and interaction of racial/ethnic data for the various measures. Without doing a full analysis, we were also concerned that the suggestion might require that States conduct additional data collection and analysis. For these reasons, we have decided not to add a performance or a threshold measure based on racial/ethnic disparities.

Former Section 270.4(e), Now New Section 270.4(f) Measure of Family Formation and Stability

In the overview section above (IV.C. Overview of Comments on the Family Formation Measure), we described our continuing efforts to develop an outcome measure related to family formation and our commitment to including in the high performance bonus system measures to address the non-work purposes of the statute. In addition, we summarized the public comments on, and the objections to, this proposed measure and our rationale for retaining the measure in the final rule.

By including this measure in the final rule, we want to emphasize that our primary focus is on the second statutory purpose of TANF, i.e., “* * * to promote marriage.” At the same time, including this measure in the high performance bonus does not preclude State efforts to support two-parent families or responsible fatherhood activities for parents who are not married. Nor does the focus on this measure preclude parents making responsible choices that best meet their needs and the needs of their children.

We have made one substantive change in this measure. In response to comments, we will base this measure on a universal population, i.e., the increase in the percentage of all children in the State who reside in married couple families, regardless of income. Given the remaining issues, we have also reduced the allocation for this measure to $10 million. In addition, we have made one clarifying and one technical change in this section: We have clarified that we will rank only those States who wish to compete on this measure and added the provision that we will measure performance based on percentage point change rather than the percentage change.

We want to respond more specifically to some of the commenters' major concerns.

Comment: A number of commenters objected to the proposed use of a quantitative measure (promoting marriage) while ignoring the more qualitative goal of “encouraging the formation of two-parent families,” i.e., purpose four in the statute. They noted that, under the proposed measure, stable but less traditional families, such as separated families, common-law families, same sex families, or two related adults living together would not be counted for bonus purposes. In addition, they pointed out that DHHS had supported the qualitative position in policy guidance by encouraging both parents to meet parental responsibilities whether they are married and live together or not.

Response: Although we explored a possible measure based on “two-parent families,” we were constrained by the lack of available data. The Census Bureau, our source of data for this measure, does not collect information on two-parent families as a category but does collect information on married couple families. We recognize the diversity of views on this issue, but point out that the second purpose of TANF includes the promotion of marriage. In using this measure in the bonus system, we are not intending to diminish our strong support for responsible parenthood, regardless of parental living arrangements. We also reward States that support and encourage responsible parenthood by non-custodial parents through the child support program and its incentives.

Comment: Some commenters objected to the design of the proposed measure. They observed that the measure appears to reward States for increasing the number of children in poverty in married couple families; it will not reward States when a married couple's income exceeds 200 percent of poverty. Further, States that are successful in encouraging single parents to marry might be less likely to receive a bonus since single mothers usually improve their economic situation when they marry.

Response: We agree with commenters that the proposed measure, based on the percent of children below 200 percent of poverty who reside in married couple families, raises programmatic and measurement issues. We believe the change we have made in the final rule, i.e., to measure the percent of all children in the State who reside in married couple families, will address these concerns. It also makes this measure more consistent with the out-of-wedlock birth bonus which, by statute, is based on all out-of-wedlock births and not just those to low-income mothers.

Comment: A number of commenters believed that the proposed family formation measure, like the proposed Food Stamp measure, was dissociated from the TANF population and State efforts to help TANF families become self-sufficient. Rather, they believed that it would merely reward States for changes in State demographics or fluctuations in the State economy. In addition, some commenters envisioned negative consequences to families as a result of coercive actions a State might take in order to compete for the bonus.

Response: We believe this bonus provision may be an incentive to States to increase their attention to some of the non-work purposes of the TANF program, i.e., purposes that go beyond the population receiving TANF assistance. We agree that this is a new program area for many States, and we believe that States will respond by reducing barriers and developing new activities to support marriage and strengthen families. The universal measure in the final rule addresses the concern about fluctuations in the States' economy.

Some of the recent steps ACF has taken will help us track State efforts to meet the family formation goals of TANF. First, under the provisions in the final TANF rule, published April 12, 1999 (64 FR 17720), the new fiscal reporting form (ACF-196) will capture expenditures on “Two-parent family formation and maintenance activities.” This will help us determine which States are making the biggest investments in this area.

Second, the new annual report from States will include summaries of State programs and activities directed at the third and fourth purposes of the TANF program (see 45 CFR 265.9(a)(8)). We anticipate that this information will provide leads about promising practices that we can share among States to encourage innovation and increase efforts in these areas. It might also provide leads about policies or practices that merit further review.

Third, in the summer of 1999, we issued a TANF funding guidance document entitled “Helping Families Achieve Self-Sufficiency.” This document provides several examples of family formation activities that States could undertake to encourage and reduce barriers to marriage, funded with Federal TANF or State MOE funds, e.g., providing premarital and marriage counseling and mediation services, and changing State TANF eligibility rules to provide incentives for single parents to marry and/or for two-parent families to stay together.

Finally, recent results from a rigorous evaluation of the Minnesota Family Investment Plan found that a program combining generous work incentives with work requirements significantly increased the proportion of married families for both two-parent recipients and single-parent, long-term recipient families.

Comment: A few commenters objected to this proposed measure because it did not meet the principles outlined by NGA and APHSA, i.e., that it did not minimize double jeopardy or reward. They believed that there was a very narrow distinction between the proposed measure of the out-of-wedlock birth bonus in section 403(a)(2) of the Act. Some commenters urged us to develop a measure to reward reductions in teen pregnancies instead of the family formation measure.

Response: We believe the out-of-wedlock birth bonus and the family formation measure are sufficiently different in focus so as not to violate the principle of non-duplication. For example, the out-of-wedlock birth bonus focuses on all births, in conjunction with the abortion rate in the State, while the family formation measure focuses on children of all ages, including newborns. In addition, the out-of-wedlock birth bonus addresses the third purpose of TANF (preventing and reducing the incidence of out-of-wedlock pregnancies) while the family formation measure addresses the broader goal of promoting job preparation, work, and marriage.

We had previously devoted considerable thought to the question of whether we should include a teen pregnancy measure because this question came up in our earlier consultation with the States. While the Centers for Disease Control and Prevention and the Alan Guttmacher Institute both publish State-level teen pregnancy data, some of these data are based on estimates because teen abortion data are not available for all States, and we did not consider these figures sufficiently reliable for the purpose of awarding bonuses.

In the areas of teen pregnancy prevention and out-of-wedlock childbearing, we have sought to focus more State attention on the pregnancy prevention goals of TANF. First, we believe the new reporting form (ACF-196) and the annual reports from States will provide information on promising State practices that we can share to encourage innovation in this area. The TANF funding guidance (“Helping Families Achieve Self-Sufficiency”) also provided several examples of pregnancy prevention activities that States could undertake with Federal TANF or State MOE funds.

In addition, the Department has continued its efforts to improve State and national outcomes in these areas. The latest national data show continued success in reducing the teen birth rate (which dropped 20 percent between 1991 and 1999). The 1999 rate of births to all unmarried women has dropped six percent from its peak in 1994, and the birth rate for unmarried teens dropped 11 percent from its peak in 1994 to 1998, based on the most recent available data.

On April 14, 1999, we issued the final regulations on the bonus to reward States with the largest decreases in their out-of-wedlock childbearing. On September 13, 1999, Secretary Shalala announced the first five winners of that bonus: Alabama, California, the District of Columbia, Massachusetts, and Michigan. Each received $20 million. The reductions in their out-of-wedlock birth ratios ranged from 1.5 to 5.7 percent.

Prior to these issuances, the Department had undertaken a number of initiatives directed at reducing out-of-wedlock and teen pregnancies.

• In 1995, the Department produced the Report to Congress on Out-of-Wedlock Childbearing, and Beginning Too Soon: Adolescent Sexual Behavior, Pregnancy and Parenthood—both of which contained valuable information about the occurrence of out-of-wedlock and teen pregnancy as well as strategies for addressing these concerns.
• In 1997, the Department developed the National Strategy to Prevent Teen Pregnancy, as required in section 905 of PRWORA. The Department has released three annual reports to Congress since then. Among other things, the 2000 report noted that HHS has funded teen pregnancy prevention programs in at least 35 percent of communities across the country and listed more than 20 Departmental programs aimed at educating teens and preventing pregnancy (including Girl Neighborhood Power! and demonstration grants to 13 communities in 11 States funded through the Centers for Disease Control and Prevention Community Coalition Partnership Programs).
• The Department, in partnership with the National Campaign to Prevent Teen Pregnancy and Johnson and Johnson, has developed “Get Organized: A Guide to Preventing Teen Pregnancy.” This publication stresses a localized approach, a long-term commitment, and careful evaluation. The Department will be disseminating additional information to communities regarding programs that specifically target boys and young men.
• The Department has been administering the State Abstinence Education Program, as authorized by section 912 of the PRWORA. This program authorizes $50 million per year beginning in FY 1998. Every State applied for this money to build on its efforts to prevent teen pregnancy in FY 1998 and FY 1999 (although New Hampshire declined its funding for FY 1998 and California did not draw down its 1998 or 1999 grant). For FY 2000, all States applied for an abstinence education grant except California. As mandated in the Balanced Budget Act of 1997, the Department is conducting an evaluation of a selected number of sites receiving funding under this provision.
• The Department is actively supporting expanding pregnancy prevention efforts to include a focus on boys and young men.
• The Department's Regional Offices have awarded $2 million in small grants to Title X Family Planning Clinics to develop pilot programs designed to prevent premature fatherhood. These projects employ male high school students as interns to provide them with on-the-job training in clinic operations and allied health occupations and provide education about male responsibility, family planning and reproductive health.
• In addition to these initiatives, the Department supports other research efforts, including the National Study of Adolescent Health, the National Survey of Family Growth, and the National Survey of Adolescent Males, which have all provided important insight into adolescent risk behaviors including sexual activity and response to pregnancy.

New Section 270.4(g) Option to Compete

Under the NPRM, we proposed to rank only the competing States for the work measures, but to rank all States that met the qualifying conditions for the Food Stamp and the family formation measure, based on Census Bureau data.

Comment: A number of commenters objected to not being given the option whether to compete on all measures.

Response: In the final rule, we have added new language in §§ 270.4 (c), (d), (e), (f), and (g) and § 270.6(c) to make it clear that States have the option whether to compete on all measures. Under § 270.11, each State must submit to us a list of the measures on which it is competing by February 28 of each bonus year.

Section 270.5 What Factors Will We Use To Determine a State's Score on the Work Measures?

In § 270.5 of the NPRM, we proposed definitions of the four work measures and a description of the factors that we would use to determine a State's score on the work measures. We also proposed that States could compete on one, any number of, or none of these work measures. We would score and rank competing States and award bonuses to the ten States with the highest scores in each measure.

The four work measures are: Job Entry; Success in the Work Force (Job Retention and Earnings Gain); and improvement from the prior fiscal year in each of these measures. We would use the proposed measures to measure State performance along three parameters of employment: the extent to which States are moving recipients into the work force; the degree to which recipients are able to remain in the work force; and the degree to which their earnings increase over time.

The comments were strongly supportive of our proposed work measures, although a number of commenters suggested substantive and technical modifications or recommended the addition of threshold measures. A few commenters opposed some of the measures. Briefly, we made the following changes in the final rule in response to comments:

(1) Revised the calculation of the improvement measures to measure percentage point improvement rather than percentage improvement in § 270.5(a);

(2) Revised the calculation of the State rankings in § 270.5(b) to drop the proposed double-weighting of the job retention sub-measure—thus giving both sub-measures equal weight;

(3) Dropped the distinction on what kinds of subsidized jobs count under the work measures;

(4) Streamlined the description of the ranking procedures in § 270.5(b);

(5) Clarified that we will award bonuses only to States with positive scores; and

(6) Clarified how we will rank States on the increase in success in the work force improvement measure.

Following is a discussion of the comments we received, by issue.

A. Establishing a Performance Threshold

The NPRM did not specify a level of earnings or other threshold factor that a TANF recipient would need to achieve in order for the State to count the individual in the job entry, job retention, or earnings gain measures. Thus, as little as one dollar of earnings in a quarter would count in determining who entered employment, how long they remained employed, and how much their earnings increased.

Comment: Some commenters suggested that we tie work measures to a minimum threshold, e.g., that we count only those persons whose wages are at the poverty level or above the minimum wage. These commenters reasoned that the proposed measures reward States that place recipients in jobs without regard to how long they will last or whether they move the family towards self-sufficiency. One commenter suggested a specific threshold factor, i.e., that at least 50 percent of those who leave TANF be employed at wage levels above poverty one year after leaving TANF.

Two commenters recommended that we link job entry with success in the work place. They asserted that a State could receive bonus grants for moving TANF recipients into work and at the same time have a dismal record regarding recipients' remaining employed and/or earning a livable wage that would propel them out of poverty.

Several commenters recommended that we establish a threshold requiring a minimum percentage of expenditures to be spent on education and training or requiring the full use of Federal programs, e.g., the Work Investment Act (WIA) and Welfare-to-Work (WtW) programs, before we would consider a State for a bonus on a work measure.

Response: It is our intention, and many commenters agree, to reward the statutory work purpose of TANF across a wide range of part-time and full-time work. We take account of how good the jobs are with the success in the work force measure. We have not established an earnings threshold in the final rule because there is currently insufficient baseline information for selecting a threshold. We believe a threshold should be high enough to foster placement in substantive jobs, but not so high as to disadvantage States with large numbers of recipients with significant employment barriers or substantially more difficult labor markets. We may consider adding an earnings threshold (e.g., a minimum amount of quarterly earnings based on the NDNH reporting), after further analysis and consultation with States and other interested parties.

We do not agree with the commenters' recommendations that we establish a threshold requiring a minimum percentage of expenditures to be spent on education and training or requiring the full use of Federal programs, e.g., the Workforce Investment Act (WIA) and Welfare-to-Work (WtW) programs, before we would consider a State for a bonus on a work measure. Such a threshold would hinder a State's flexibility in designing its TANF program and would not be a measure or indicator of outcomes.

B. A More Rigorous Job Retention Measure

In the definition of “job retention rate,” we proposed a job retention period of six months, i.e., States could count those individuals employed in one quarter who remain employed in the next two consecutive quarters.

Comment: There were a number of comments in support of extending the retention measurement period; only one State commenter recommended a shorter retention period. Most of these commenters supported extending the retention measure to one year, but other recommendations included an 18 month period or a longer period, if possible. In general, they believed that six months is too short a period of time to demonstrate that an individual has achieved job stability. Other specific suggestions ranged from extending retention to the sixth tracking quarter, or measuring retention only in the same job unless the change is to a job that pays a higher wage, offers enhanced benefits, or promotes job stability or career growth. The latter suggestion did not include a specified period of time for measuring retention.

Response: After careful consideration of the comments, we did not adjust the retention period. We proposed in the NPRM that job retention be measured in the initial quarter and the two consecutive subsequent quarters, in part, because this is similar, though not identical, to measures of job retention in the WIA and WtW programs. We continue to believe that job retention over six months is a reasonable indication of stable employment. In addition, an extended retention period would delay when critical performance data would be available. Given the lag in data availability, the longer time frame would not allow us to make the bonus awards in the bonus year to which they apply. Further, if we did not issue the FY 2003 bonus funds by September 30, 2003, they would return to the Treasury, unless Congress reauthorizes the bonus and appropriates funds.

We also considered the suggestion that we measure retention in the same job, but decided against it because, moving to another job can, in fact, represent moving up to a better job. In addition, tracking job characteristics, other than wages, e.g., benefits, would be extremely burdensome, if not impossible.

C. Success in the Work Force Measures

In the NPRM, we proposed a “success in the work force measure” that combined the performance scores of two sub-measures, i.e., job retention and earnings gain. We proposed a combined measure because of the linkage between these two outcomes, i.e., with earnings gain viewed as dependent on an individual's success at retaining employment. In ranking and combining the scores of these measures, we proposed a double weight for the job retention measure. We also proposed to measure the change in the earnings for those employed in one quarter who were also employed in the second subsequent quarter.

Comment: Several commenters thought that we had not given sufficient weight to earnings gain. Commenters also recommended that we:

• Treat earnings gain as a separate measure instead of incorporating it into the success in the work force measure;
• Measure earnings gain at both six and 12 months;
• Adjust the earnings gain for the cost of living in each State using the HUD fair market rent amount;
• Measure only full quarters' earnings; or
• Allow administrative data on a voluntary basis as a source of earnings.

Response: Several commenters questioned why we weighted the job retention submeasure at twice the rate of the earnings gain submeasure and recommended that we give each measure equal weight. They believed earnings gain is a better indicator of stable employment and a key component to achieving economic independence.

For the reasons expressed by the commenters and, after further analysis of the available data, we agree that job retention and earnings gain should be weighted equally in developing State rankings. We have made this change in § 270.5(b).

We agree that separate measures for both job retention and earnings gain would provide a discrete focus and reward for each measure. Job retention is an important measure of success in transitioning from welfare to work. It is correlated with long-term employment stability, e.g., the longer an individual remains employed the more opportunities there are to acquire specific job skills and refine successful work habits. However, we also believe that job retention and performance in earnings gain are directly linked in the work world. Research in this area has shown a correlation between persistent work force attachment and earnings. We also believe that de-linking these measures could create perverse incentives for States, e.g., to focus on placement and retention of recipients in any job without regard to the quality of the job or to “cream” in order to show larger earnings gains. Finally, creating a separate earnings gain measure would result in fewer bonus dollars for any one work measure.

Therefore, we have retained the link between job retention and earnings gain in order to focus attention and resources on both. Helping people rapidly get employed, remain employed, or enter higher paying jobs after training and progress to even higher paying jobs are all important strategies. States may choose to emphasize different strategies at different times, depending on their populations, goals, and economies.

Several commenters recommended that we measure earnings gain at both six and twelve months. We did not want to make this change for essentially the same reasons we did not want to extend the retention rate to twelve months, i.e., it would delay obtaining critical data for analysis and impede our ability to award the bonus in the bonus year.

In addition, we have not accepted the comment to adjust the earnings gain sub-measure based on the cost of living in each State using the HUD fair market rate amount as this would add a level of complexity that would counter one of our basic principles, i.e., to maintain simplicity. First, it is not clear that differences in earnings directly track differences in fair market rent. More importantly, it is not clear that differences in the cost of living would make any difference in the comparison of earnings gains between States. Because we are measuring an increase in earnings in a single State from the reporting quarter to the second subsequent quarter, differences in the cost of living among different States would not be relevant. It is not the absolute difference between earnings in different States that is relevant, but the percentage point difference of the change in earnings amount from one year to the following year in the same State. Thus, a State would not be advantaged or disadvantaged by having a higher or lower cost of living than other States.

We also did not accept the recommendation to measure only full quarters' earnings because that would require a substantially increased data collection and reporting burden on the States. Because Unemployment Insurance (UI) and NDNH records do not indicate whether earnings reflect a full quarter of employment or not, we would have to establish a proxy measure, require States to collect and submit administrative data, or conduct new surveys. Also, allowing States to submit administrative data on a voluntary basis would not be helpful because we believe that the bonus awards must be based on data that are uniformly available and comparable across States.

Finally, although we received no comments on this matter, we have clarified how we will calculate scores, rank States, and award bonuses for the increase in success in the work force measure. First, we will award bonuses only to States with a positive improvement score on at least one of the sub-measures. Second, we will not exclude a State with a negative score in calculating a rate and ranking States on this measure. For example, a State may have a negative score on one sub-measure (e.g., job retention) and a positive score on the other sub-measure (e.g., earnings gain). We have added language in § 270.5(a)(4) to specify that we will award bonuses only to States that achieve a positive percentage point difference on at least one sub-measure between the rate for the performance year and the comparison year.

D. Sustained Employment Rate

Comment: Two organizations suggested a somewhat different approach to the proposed work measures. Their approach would create a “sustained employment rate,” a separate earnings gain rate, and a new measure of the earnings gap for poor families in a State.

They recommended that we develop a single sustained employment rate in lieu of the job entry and job retention rates. The base (denominator) for the sustained employment rate would be all TANF adult recipients except those engaged in employment with earnings equal to or greater than an average of twenty hours per week for the quarter multiplied by the Federal minimum wage. A recipient would be counted as a “sustained employment recipient” if the recipient has at least two consecutive quarters with earnings at or above the threshold noted above during the next year. The earnings gain rate would be measured over a one-year period beginning after the initial quarter of employment where the earnings are at or above the earnings threshold in the measurement quarter, e.g., quarter two and quarter six.

Under the earnings gap measure, the difference between the amount of earnings for poor families with children in a State and the official poverty level would be determined. We would rank States on their success (improvement) in closing that difference. We would also make adjustments based on changes in unemployment rates and changes in the number of families in the measurement period.

Response: These recommendations would result in a significant change in the work performance measures, add a high degree of complexity, and substantially increase the data collection burden. Further, because the “sustained employment rate” recommendation does not differentiate between the unemployed and the underemployed, it would result in the loss of information on the number of adult recipients entering employment for the first time in a year. The suggested time frames for these recommendations would result in the issuance of the bonus awards well after the end of the bonus year. In addition, most commenters supported staying within the existing framework for the work measures. For these reasons, we are not adopting this proposal.

E. Subsidized Work

In the NPRM, we proposed to count only jobs that were not fully subsidized in order to focus on jobs that were likely to lead to self-sufficiency.

Comment: Several commenters suggested that we count fully subsidized jobs, or at least exclude them from the denominator as well as the numerator. They contended that, unlike community experience and work experience, wage-paying subsidized jobs resemble unsubsidized employment in every aspect, except for the subsidy paid to the employer. Another commenter recommended that we link work experience and subsidized work with the provision of ancillary services such as education and training. One commenter suggested that we clarify that, while subsidized work would not be counted, “supported” work would be, since ongoing funding is provided to the job coaches rather than to the individuals' wages.

Response: First, “supported” work currently is counted in the calculation of the work measure. Second, we have decided to count wage-paying, fully subsidized jobs in the numerator and the denominator of the work measures in the final rule. We have made this change because we believe the numbers are small, and we recognize that the distinction between partially and fully subsidized employment is somewhat artificial and is governed by changes based on the wage rate of the individual TANF recipient. We also believe that this change will reduce the burden on States as they will not need to separate out and report different types of subsidized jobs. We also note that community service and work experience “jobs” would also count to the extent that wages are actually paid and reported to the State Employment Service agency.

F. Measuring Improvement

In the NPRM, we proposed to award bonuses to the States demonstrating the greatest improvement in job entry and success in the work force. We proposed to measure the percentage increase from the comparison year to the performance year.

Comment: There was strong support for the improvement measures, including one commenter who expressed the view that we should use only improvement measures, in order to give States more incentive for continuous progress, and to give all States an equal chance of earning a bonus, regardless of their starting point.

Another commenter, however, recommended that we should use only absolute measures or reduce the amount of the awards for the improvement measures because States that have been high performers in the past would have little room for improvement. They believed that the proposed method of calculating the performance improvement rate for job entry and the success in the work force measure (combination of job retention and earnings gain) did not take into consideration the past performance of States that had implemented welfare reform early. These commenters indicated that States were disadvantaged if they had achieved earlier success in moving recipients from welfare to work and self-sufficiency that would not be measured. In addition, such States' remaining caseloads have a higher proportion of recipients that have significant employment barriers. Commenters made the following recommendations:

• Adjust the proposed method of calculating the improvement rate, i.e., average the comparison and performance year rate scores and add the percent change between the two years to the average score.
• Set a minimum target for improvement, e.g., ten percent, and make awards to all States that met the target without limiting the number of States that could receive an award.

Response: The bonus awards are based on current performance. Even for the improvement measures, we assess the level of current performance in relation to the immediately preceding year. Because the statute specifies the performance year for each of the bonus years, it is not clear that the recommendations are consistent with either that statute or Congressional intent. Further, even if we agreed with the argument that we should consider past performance (prior to the FY 1997 comparison year), there is no objective way to provide for such an adjustment, because we do not have access to data for prior-year periods.

Finally, we recalculated the FY 1999 bonus improvement scores based on the recommended adjustment formula and found that it had the effect of narrowing the range of the improvement scores and had some limited impact on bonus winners. There was only a modest change in the improvement rate ranking of “early reform” States, based on this recommended approach. For these reasons, we have not adopted the suggestion to consider past performance earlier than the comparison year.

We also have not accepted the recommendation that we establish a minimum target for improvement largely because caseloads are changing. With the available baseline data it would be difficult to determine a reasonable goal, especially over more than one bonus year.

However, in § 270.5(a) and (b)(3), we have changed the method of calculating the improvement rate. Specifically, we have decided to use the percentage point change instead of the percentage change. For example, under the NPRM, a State that went from a 50 percent to a 60 percent job entry rate would have an improvement rate of 20 percent ((60%−;50%)/50%=20%), while a State that went from a lower base of 40 percent to a 50 percent job entry rate would have a bigger increase, i.e., 25 percent ((50%−40%)/40%=25%). Under the final rule, these same States would have identical improvement rates of ten percent (60%−50%=10% and 50%−40%=10%). We believe this new approach will reward substantial improvement rather than relative improvement and will raise expectations about the level of improvement required to receive a bonus, particularly among States whose baseline performance was low.

G. Leveling the Playing Field

We stated in the NPRM that we believed that competition for the high performance bonus should primarily reflect a State's welfare to work strategies and should be a competition among States that is objective and fair. We indicated that there are factors over which the State has little control, such as the health of the State's economy, the demographics of its TANF caseload and its resident population, and State population growth. We asked if we should attempt to develop adjustment factors in order to ensure an objective and fair competition.

Comment: There were two comments suggesting that we level the playing field to take into account such factors as economic, demographic, and cultural differences. We did not receive any specific proposals.

Response: We believe that by incorporating both absolute performance and performance improvement measures, we have helped produce a more level playing field for States competing for the high performance bonus. Further, by changing from a percentage increase to a percentage point increase to measure improvement, we have struck a balance between recognizing past performance and encouraging improvement from a low base.

The adjustment of performance scores by external factors would be a complex and difficult task involving the establishment of a correlation between external factors and the performance being measured and, if a strong correlation was detected, determining the scope of the adjustment. Also, such adjustment(s) could add a level of subjectivity and contentiousness to the performance system beyond the value of potentially leveling the playing field.

Nevertheless, in order to test whether we could detect a correlation between certain external factors and the work measures, we performed regression analysis using job entry rate performance and such factors as the unemployment rate, recipient characteristics, and TANF payment levels. (We chose job entry rate performance because we thought State performance under this work measure would more likely be influenced by external factors.) Our analysis showed that none of these factors was highly correlated with job entry rate. In fact, the highest correlation coefficient was −0.28 for the unemployment rate. Thus, the implication is that these specific factors do not determine the job entry rate to any significant degree. In addition, adding adjustment factors makes it much more difficult to explain performance and for States to set meaningful targets. Therefore, we have decided not to make any adjustments to the way we calculate job entry or any of the other work measures based on economic, demographic, or other factors.

H. Other Comments and Recommendations

Comment: One commenter recommended that we should measure all those served, not just those receiving TANF cash assistance.

Response: We have not limited the population we are measuring with the work measures to just those currently receiving cash assistance. We also include all individuals receiving TANF “assistance” as specified in § 260.31 of the TANF final rule, i.e., primarily cash assistance. We also look at work-related outcomes for those who received cash assistance in a prior period. There is no practical way that we could include all those served, regardless of whether they are receiving assistance, e.g., diverted individuals. Such a change would involve a major new data collection effort and impose a substantial burden on the States.

Comment: In response to a series of questions we posed, one commenter recommended that we should include core measures, while another suggested that States should have to compete on a universal set of measures.

Response: We have not required any core measures or a universal set of measures because we want to allow States maximum flexibility to elect the areas in which they will compete. We believe this is consistent with the flexibility provided States in the operation of their TANF programs and the voluntary nature of the high performance bonus.

Comment: Two commenters indicated that some States have expressed concern over the current method of calculating the job entry rate. They suggested instead that the measure should be the percent of the total recipients in the current quarter employed for the first time. No reason was given for wanting this change.

Response: It is unclear what the commenters would change in the proposed method for calculating the job entry rate. Under our proposed method, the denominator is the unduplicated number of adult recipients who were unemployed at some point during the year, and the numerator is the unduplicated number of job entries. We believe the commenters are proposing to change the denominator to either employed adult recipients or to all adult recipients. This suggested change would significantly alter the job entry measure. The purpose of this measure is to determine the extent to which unemployed adult recipients enter employment for the first time in the year as a percentage of those who are unemployed. The suggestion, if adopted, would result in a different focus, i.e., of the adult recipients who are employed, how many are new job entries or, of all adult TANF recipients (employed or unemployed), how many are new job entries. We continue to believe that the appropriate focus for this measure should be on the impact of States' efforts on its unemployed caseload. For this reason, we have not made this change.

Comment: One commenter suggested that because research has shown that many TANF recipients attain and lose jobs several times, that we should reward States for job entries over a longer period of time, such as a two-year performance period. Specifically, for the job entry measure, the commenter indicated that we should count only job entries that are the first job entry in a two-year period. They proposed that retention would then be measured from the point of job entry to a time period one year later.

Response: Given the lag in data availability, the longer time frame would not allow us to make the bonus awards in the bonus year to which they apply. If we did not issue the FY 2003 bonus funds by September 30, 2003, they would return to the Treasury, unless Congress reauthorizes the bonus program and appropriates funds. In addition, the statute requires that we award bonuses for a single performance year, not over a two-year period.

Section 270.6 What Data for the Work Measures Must a State Report to Us? (Title of This Section in the NPRM.)

Section 270.6 What Data and Other Information Must a State Report to Us? (Title of This Section in the Final Rule.)

In the NPRM, we proposed that the State, if it chose to compete on any or all of the work measures, must report one of two alternative sets of data, as specified by the Secretary. In one alternative, the State would provide three items of identifying information on its adult TANF recipients that we would match against the NDNH data. In the second alternative, the State would provide actual performance data for the work measures based on data matches with State UI records or other records. We also specified the SSP-MOE reporting requirements.

We have broadened the content of this section in the final rule. In paragraph (a), we specify the data a State must report if it wishes to compete on the work measures. In paragraph (b), we specify that a State must report data on SSP-MOE programs in order to compete on any high performance bonus measure. In new paragraphs (c) and (d), we specify the data a State must report if it wishes to compete on the Medicaid/SCHIP measures and/or the child care measure. Finally, in paragraph (e), we have retained the requirement (paragraph (d) in the NPRM) that each State must notify us regarding which measures it will compete on in each bonus year.

Comments: Regarding paragraph (a) and the data States must report on the work measures, most commenters supported reporting minimal information on recipients and the use of the NDNH. A number of States believed strongly that no other source could provide Federal employment and out-of-State employment. However, a number of commenters raised issues about the use of NDNH data. Several questioned whether the NDNH contains all the necessary information to calculate the performance scores for all the work measures. One raised concerns about privacy protection of the NDNH data. Two wanted to use State administrative data to supplement the NDNH data, since we could not detect a period of unemployment within a quarter through the use of NDNH quarterly wage data.

Response: After considering the comments and other factors relating to our operational use of the NDNH, we agree with commenters that State reporting of minimal identifying information on all adult TANF recipients and a match of this information by ACF at the Federal level would result in the least burden to States and the maximum accuracy in implementing the bonus awards.

We have not specified the identifying information that States must submit in § 270.6(a). Rather, because some operational factors are unclear in our use of the NDNH, we will specify these data in program guidance. The data that will be required will be limited to some or all of the information proposed in the NPRM, i.e., the name, date of birth, and social security number of all adult TANF recipients.

In addition, we want to clarify what data are available through the NDNH. The NDNH contains not only a national database of new hires, but also national wage data compiled by the State Unemployment Insurance agencies and Federal employment and wages. (This addresses commenters' concerns for information on out-of-State and Federal employment.) Matching adult TANF recipient data with quarterly wage data on the NDNH data base will provide the necessary work performance information to rank States on all the work measures.

With respect to the privacy concerns, the match of adult TANF recipients with the NDNH database will not reflect information about individuals, but will produce only aggregate information for use in calculating the State rankings on the work measures.

Prior to these regulations, we provided States flexibility in what data source(s) they could use for compiling performance data for the work measures. We note that States competing for the FY 1999 bonus awards compiled their work measurement performance data based on their State unemployment wage data system. Our proposal builds on and strengthens that system. The use of the NDNH will ensure that we rank States based on the most uniform, objective, and reliable data available. In addition, States will have the benefit of the employment data on their TANF recipients, including out-of-State employment and Federal employment.

We specify in new paragraph (c) that we will issue program guidance on the data a State must submit if it wishes to compete on the Medicaid/SCHIP measure.

We specify in new paragraph (d) that if a State wishes to compete on the child care measure in FY 2002, it must report the data required by the CCDF program. These data are found in ACF Forms 800, 801, and 696. In addition, after external consultation, we will issue program guidance to specify the additional data on child care market rates that States must submit in order to compete in FY 2003.

Section 270.7 What Data Will We Use To Measure Performance on the Work Support and Other Measures?

In the NPRM, we proposed to use Census Bureau data to rank State performance on the Food Stamp and the Family Formation measures. We also proposed to rank State performance on the Medicaid/SCHIP measure based on the data submitted by States following their match of individuals no longer receiving TANF assistance with Medicaid/SCHIP enrollment data.

In the final rule, we specify in paragraph (c) that we will use data from the ACF Forms 800, 801, and 696 to rank State performance on the child care measure. Also, after external consultation, we will issue program guidance specifying the other information States must submit in order to compete on this measure.

In addition, we have made the following editorial and technical changes in this section: we have substituted the acronym SCHIP for the acronym CHIP and clarified that we will rank only those States that choose to compete on these measures. In addition, we received a number of comments on this section.

Comment: Several commenters expressed concern about the proposed use of Census Bureau data. Commenters based their concerns on perceived problems with the Current Population Survey (CPS) or the decennial census, e.g., the CPS data could not produce reliable State-level estimates for all States, given the sample sizes. They cited the lack of reliability of the CPS data and recommended, alternatively, that we base Food Stamp and the Medicaid/SCHIP measures on State administrative data that are more current. They also believed that the decennial census data, despite periodic updates, under-reports many low-income populations, focuses on married households, and undercounts households where two adults may be responsible for parenting and child-rearing. Undercounting the increasing number of grandparents raising children was also of concern to commenters.

Response: We agree that State administrative data, in some cases, are more current than decennial census data or the CPS estimates. Our aim in the NPRM, however, was to propose to use the most uniform, reliable, and objective data available with as little burden to States as possible.

We want to clarify that we will use data from the Census Bureau's Census 2000 Supplementary Survey and the Census Long-Form Transitional Database in awarding bonuses for FY 2002 and FY 2003. The Supplementary Survey and the Transitional Database will provide reliable State-level data based on a sample of 700,000 cases. If high performance bonus awards are authorized in subsequent years, we plan to use data from the Census Bureau's American Community Survey (ACS), which will provide annual reliable State and county-level data, starting in 2004 for areas with populations of 65,000 or more.

Comment: A number of commenters recommended that the final rule require States to report data separately on the number of adults and the number of children no longer receiving TANF assistance who are enrolled in Medicaid or SCHIP. They commended our proposal to evaluate States on the percentage of individuals enrolled in Medicaid or SCHIP, rather than the percentage of families enrolled. They based their recommendation on the need for better data and findings from recent studies that indicated that often only selected family members retain health care coverage after leaving welfare. In particular, parents appear to be at much greater risk than children of losing out on health care coverage for which they are eligible.

Response: We agree that information on the participation of adults and children in these programs could be useful in monitoring program outcomes. However, we do not believe that the value of this information justifies the additional data collection effort that would be required.

Section 270.8 How Will We Allocate the Bonus Award?

This section of the NPRM proposed to:

• Make awards to the ten States with the highest scores in each measure;
• Allocate a total of $140 million to the four work measures as follows:
• Job entry rate—$56 million
• Success in the work force—$35 million
• Increase in job entry rate—$28 million
• Increase in success in the work force—$21 million
• Allocate a total of $60 million to the non-work measures with $20 million each to the Food Stamp, Medicaid/SCHIP, and family formations measures; and
• Within each measure, distribute the bonus money based on each State's percentage of the total amount of the SFAG of the ten States that will receive a bonus.

There were no major objections to this section of the NPRM, but several commenters made alternate recommendations. In general, most States and their representative organizations recommended that the full amount of the annual bonus ($200 million) be awarded based on the work measures. On the other hand, a large number of the other commenters recommended various increases in the amount of funding allocated to the work support measures.

In addition, although there was general support for awarding bonuses to the top ten States with the highest scores in each measure, we received several suggestions for alternative approaches to the Food Stamp and Medicaid/SCHIP measures: (1) Award bonuses in each measure for both absolute performance and performance improvement; (2) award five bonuses for absolute performance and five for improved performance; (3) alternatively, award three bonuses for absolute performance and seven for performance improvement; and (4) award the same amount of money for each measure.

Comment: Several commenters recommended specific changes in the amounts proposed for the four work measures that would emphasize “success in the workplace.” They believed that the proposed allocations seem unduly weighted towards job entry (i.e., $56 million of the $140 million) at a time when the focus of the TANF program has shifted to an emphasis on earnings gain (measured by success in the work force and increased success in the work force). They indicated that job entry is no longer the critical measure that job success is. One commenter, for example, suggested that we alter the bonus amounts as follows:

• Job entry—$28 million
• Success in work force—$56 million
• Increase in job entry—$21 million
• Increase in success in work force—$35 million

Response: We considered this recommendation in light of our knowledge of and experience with State TANF programs to date. At the present time, we believe it is important to retain a priority focus on job entry, not only because of the emphasis throughout the TANF statute on work and economic self-sufficiency, but also because initial employment is a prerequisite for the rest of the work measures. It also remains a primary focus for most State programs in leading to self-sufficiency. However, the bonuses also recognize that job entry is not sufficient for meeting the work-related objectives of TANF and that TANF goals encompass more than work.

Comment: Commenters were concerned that, in competing for the bonus, a State with a relatively small SFAG might be likely to lose money if the cost of competing for the bonus was greater than the amount the State might receive as a winner. They recommended that States should at least be able to recover the costs of applying for the bonus in a category where they made the top ten winners list. They also recommended that we award any State receiving a bonus some minimum amount based on an estimated cost of applying for the bonus in that category. We could allocate the remainder of the award money in that category as proposed.

Response: We are aware that, for a few States, it may be difficult to compile data in order to compete for the bonuses. In the final rule, we have addressed this difficulty, in part, by requiring no information from States for the Food Stamp and the family formation measures and minimal information from States for the work measures and matching these minimal data with the NDNH data at the Federal level. For the child care measure in FY 2002, we have relied on data States report to us under the CCDF program. With respect to the commenter's specific suggestions, we could find no support for specifying minimum bonuses in section 403(a)(4) of the Act and believe it could be very problematic to administer. Thus, we have not accepted these recommendations.

Comment: A number of commenters, including Members of Congress, recommended increases in the amounts allocated to each of the work support measures, ranging from increasing the allocation for the Food Stamp and Medicaid/SCHIP measures to $30, $35, or $40 million; allotting amounts ranging from $20 to $40 million for a new child care measure; and allotting $20 million for a new measure on domestic violence and $20 million each for two measures on worker protections. Commenters offered the following opinions:

• States already have many incentives to help individuals enter the work force, e.g., the penalties if States do not meet the work participation requirements. Skewing performance bonuses further towards work is unnecessary. Giving equal weight to work and work support measures better reflects the reality that a job alone is not enough for a family to succeed.
• Families will not be able to get or maintain employment without these essential supports. Further, the work support measures (including the addition of any new measures related to child care or other measure) support two of the four purposes on TANF by providing assistance so that children can be cared for in their own homes as well as promoting job preparation and work.
• As proposed in the NPRM, the $20 million for each of the Food Stamp and Medicaid/SCHIP measures, divided among the top ten winning States, creates little incentive for States, compared to the more generously funded work measures.

Response: We seriously considered increasing the allocations for the Food Stamp measure and the Medicaid/SCHIP measure. However, we continue to believe that the work measures most directly address the overall focus of the TANF program, including purposes one, two, and four of the Act. Therefore, we continue our proposed allocation plan and will allot $140 million to the work measures, $20 million to the Food Stamp measures, and $20 million to the Medicaid/SCHIP measures. We also specify in new paragraph (c) that we will allot $10 million to the child care measure and $10 million to the family formation measure.

Comment: In commenting on one State's receipt of a multi-million dollar bonus in 1999 for 1.7 and 2.9 percent increases in two work measures, one organization objected to the lack of a benchmark or other threshold standard as a part of the bonus system. They believed such “minuscule changes” in State performance do not represent high performance. Rather than funding the ten States with the highest scores in each measure, they recommended that we set limits for what we consider acceptable and successful performance.

Response: In the preamble to the NPRM, we solicited public comment on some of the issues we had considered related to absolute performance, performance improvement, and threshold levels. (See Section IV. Discussion of Other Issues Related to Performance Measurement.) The issue of a threshold level, benchmark, or performance standard is one we struggled with, both internally and in our consultations with external groups. Given the lack of experience in establishing and implementing a bonus system, the general consensus on this matter was that an ordinal standard, such as rewarding the ten States with the highest scores, was initially most appropriate.

As we said in the NPRM, awarding bonuses to the top ten States not only provides a clear incentive to States, but also helps avoid problems associated with the need to re-allocate funds. More importantly, we did not want to set a numerical threshold based on an absolute level of performance, given the absence of baseline data. Another factor we considered is that what works now in terms of a bonus system may not be appropriate as the States' caseloads continue to change.

Finally, we believe that State programs are still evolving and continue to reflect the flexibility provided in the statute in their programs and services. Given that diversity, we have decided not to change our proposed standard.

Comment: This same commenter recommended that we invest most of the award funds in the work support measures on the grounds that the State Food Stamp and Medicaid/SCHIP enrollment data are the most solid, i.e., they follow individuals over time, and they are backed by quality control efforts. They recommended that we delay awarding bonuses to measure job retention and earnings until we have a valid source of data for these categories.

Response: We believe the minimal data States must provide to compete for the work measures, cross-referenced to or matched with the NDNH data, will provide the basis for national uniform, objective, and reliable job retention and earnings information on which we can make bonus awards on these measures with full confidence. As proposed, we will use Census Bureau data as the data source for the Food Stamp and the family formation measures; we will use State Medicaid and SCHIP data (matched with TANF data at the State level) as the data source for the Medicaid/SCHIP measure.

Comment: Rather than basing the amount of the award on the percentage of TANF funds a State is allotted, one commenter recommended that, in distributing funds among the top ten winners in each measure, we base the bonus amount on the percentage improvement a State achieves.

Response: We continue to believe that the allocation of bonus funds based on a percentage of a State's family assistance grant in relation to the total bonus award for that measure is the most appropriate allocation method. This approach recognizes the potential number of individuals affected by the State's performance, the State resource costs to achieve levels of high performance, and the five-percent funding limitation in the statute.

Section 270.9 How Will We Redistribute Funds If That Becomes Necessary?

In the NPRM, we proposed a two-step process to redistribute funds, if for some reason we cannot award the full annual amount of $200 million in any fiscal year. We proposed that, if we could not distribute the funds as specified in § 270.8, due to the statutory limit on each State's bonus award, we would reallocate the funds among the measures proposed in § 270.4. If the funds still could not be distributed within the bonus year, we proposed that they would be available for distribution in the following year.

We have made one editorial change in the regulatory text for clarity. In § 270.9(a), we deleted the phrase “due to the statutory limit on the amount of each State's bonus award,” as there may be additional reasons why funds could not be awarded, e.g., if insufficient numbers of States qualify for bonuses in a particular measure in a given year.

Comments: Two commenters concurred with the proposal that we re-allocate any undistributed funds among the top ten States. However, they also recommended that if we still cannot distribute funds within the bonus year, we should award the funds to the eleventh and twelfth State in rank order.

Response: We reconsidered these suggestions, which we had evaluated earlier, in the context of our external consultations. We find no compelling reason to accept the suggestion that funds be awarded to the eleventh or twelfth State or to change the process we proposed in the NPRM. As we indicated above, we believe ten awards in each measure offers a reasonable and clear incentive to States. We do not believe diluting this incentive would further the aims of the bonus system, particularly as we proposed another more efficient mechanism to ensure that funds that cannot be awarded in a fiscal year will remain available, i.e., will be rolled into the next fiscal year's awards.

However, we will reconsider this suggestion if: (1) We find that, in FY 2003, the bonus awards have not been re-authorized in the statute, and we cannot roll funds forward; and (2) we have a situation in which funds would remain unawarded.

Section 270.10 How Will We Annually Review the Award Process?

In this section of the NPRM, we proposed to annually review the measures, data sources, and funding allocations for the high performance bonus system to determine if modifications, adjustments, or technical changes were needed. We stated explicitly that we would not add new measures or change funding allocations except through regulations.

Also in this section, we proposed that we would consult with NGA and APHSA and other interested parties before we made our final decisions on performance components for bonus awards in FY 2002, FY 2003 and beyond; notify States through program guidance of our decisions; and post this information on the Internet.

Comments: Most commenters were supportive of these proposed provisions but provided additional comments and recommendations.

(1) Regarding the proposed annual review of data sources, one commenter urged us to use the review process to evaluate annually whether new data sources are available. They supported a broader measure for determining State success in assuring health coverage for low-income families that would require additional data sources.

(2) Regarding the proposed consultation activities, a State “welcomed the flexibility and input but only if changes produce little burden on States.”

(3) Other commenters urged us to include members of the advocacy community in the review process, i.e., “specify in the final rule that our consultation with interested parties will include, among others, local elected officials (such as mayors and county officials), labor unions, charitable organizations (such as those providing emergency food assistance and monitoring services to families with children), and low-income clients and their advocates.”

(4) One commenter recommended that we should release the findings of the annual review process.

Response: One of the purposes of the annual review is to identify and possibly implement certain changes, modifications, and technical corrections in the high performance bonus system, but not changes in the measures or funding allocations. We will change these latter items only through rulemaking.

We view the high performance bonus system, however, as one that is still new for both the States and the Federal government and one that will evolve and need refinements in the future. We agree with the first commenter and are committed to looking to identify new, more reliable data sources. We also agree that our external consultations, as they have done in the past, will include a broad range of “interested parties.” However, we have not added specific examples of such agencies and organizations in the regulatory text because we do not think it is necessary or appropriate. (We specified the National Governors' Association and the American Public Human Services Association in this section because they are specified in the statute.)

Finally, we have not accepted the suggestion to release the findings of the annual review. We believe such a release would be duplicative because the results will be shared through our consultation activities and/or through periodic guidance. If any changes result, the guidance containing these changes will explain the basis for the changes, including our rationale and the results of our review and consultation.

Comment: One commenter recommended that, as a part of the consultation process, we should post on the Internet the data submitted by States in competing for the bonus.

Response: We will make some of the data that are submitted by the States competing for the bonus available annually, after we have ranked the States and officially awarded the bonuses.

We will not make any social security numbers available.

Section 270.11 When Must the States Report the Adult Recipient Data and Other Information Related to the Work Measures? (Title of This Section in the NPRM)

Section 270.11 When Must the States Report the Data and Other Information in Order To Compete for Bonus Awards? (Title of This Section in the Final Rule)

In the NPRM, we proposed time frames for States to submit various information and data required to compete on the proposed work measures.

In the final rule, we have broadened the content of this section to include information on when the data for the Medicaid/SCHIP measures and the child care measure must be reported to us. In paragraph (a), we specify that each State must submit a list of measures on which it wishes to compete by February 28 of each bonus year. This provision is unchanged from the NPRM. In paragraphs (b) and (d), we specify that the dates for submitting data for the work measures and the Medicaid/SCHIP measures will be specified in program guidance. In paragraph (c), we specify when States must submit SSP-MOE data. This provision is unchanged from the NPRM. In new paragraph (d), we specify that States competing on the child care measure must report the child care information by the date specified by us.

In the NPRM, we proposed that States must submit data for the work measures by February 28 and August 31 of each bonus year. These time frames reflected the proposed option for States to match adult recipient data with Unemployment Insurance data. We have dropped that option in the final rule. The final rule provides that we will match State adult recipient identifying information provided by competing States with wage data in the NDNH. We are working closely with the Office of Child Support Enforcement to finalize all technical issues related to this match, and we will specify the reporting timeframes in program guidance at a later date.

Comment: One commenter recommended that the final data be reported 13 months after the last quarter of the Federal fiscal year. The commenter did not discuss the reason for the suggested revision.

Response: If we adopted the time frame suggested by the commenter, we could not issue the awards during the bonus year. Therefore, we have not accepted this comment.

Section 270.12 Must States File the Data Electronically?

In this section, we proposed that, in order to compete for the high performance bonus, each State must submit data electronically on the work measures and on the Medicaid/SCHIP outcome measure in a manner that we and HCFA will specify.

Comment: One commenter said that, if States are to submit data electronically, we must develop procedures well in advance, with adequate testing at the State and Federal levels.

Response: We agree that we should develop and test the data submission procedures in time for use in the FY 2002 bonus year. We intend to have further discussions on these matters with Federal and State staff following publication of the final rules and prior to implementation.

Section 270.13 What Do States Need To Know About the Use of Bonus Funds?

In the NPRM, we specified some of the requirements for the use of bonus funds, e.g., that funds must be used to carry out the purposes of the Act (section 401) and must meet the requirements of section 404 (Use of Funds) and 408 (Prohibitions; Requirements) of the Act.

Comments: The majority of commenters on this section urged us to provide as much flexibility as possible in the use of these funds, e.g., to “exempt these funds from the constraints of the regulations and administrative caps associated with the SFAG.” One commenter urged us to follow the model of the Department of Agriculture's Food and Nutrition Service, which rewards States for payment accuracy by providing additional administrative funds without restrictions. Without this flexibility, the bonus award would not provide an incentive to achieve but would become “just another quarterly advice of TANF funding availability.”

In emphasizing recommendations for flexibility, several commenters noted the lack of congruence between the proposed bonus awards tied to success in the Food Stamp and the Medicaid/SCHIP programs and a State TANF agency's inability to use TANF dollars to, for example, increase health care coverage for an additional portion of the low-income population, provide food stamp outreach, or directly fund food shelves. They recommended that we allow bonus dollars for expenditures that could positively affect the outcomes being measured, i.e., an increase in health care coverage or Food Stamp use.

Other commenters suggested that, if flexibility were provided, States could use bonus funds for a wide range of activities, such as economic development targeted at TANF families; supplemental wages for newly employed TANF workers; allowable medical services; supplements for physician reimbursement rates for Medicaid to stimulate more access to health care; State programs or initiatives that reduce poverty, such as EITC credits; and increases in the number and quality of child care slots.

One commenter, aware of the statutory restrictions on these funds, urged us to seek a legislative amendment to provide greater flexibility.

Response: First, we want to reiterate that a State has the same flexibility in the use of these bonus funds as it has in the use of other TANF block grant funds.

Second, despite a commenter's assertion that the statute does not require these limitations on the bonus funding, sections 404 and 408 of the Act limit the purposes for which high performance bonus funds can be used. Both sections refer to the use of a grant “under section 403.” Since high performance bonus awards (as well as bonus awards to reduce out-of-wedlock births, contingency funds, and supplemental funds) are grants under section 403, the limitations in sections 404 and 408 are statutory. Likewise, the restriction on the use of TANF funds for Food Stamp outreach is statutory, based on the Agricultural Research, Extension, and Education Reform Act of 1998, Pub. L. 105-185.

Comment: One commenter questioned whether 45 CFR 263.13(a)(i) applied the 15-percent administrative cap limitation to the State's SFAG or to all funds a State may receive under section 403 of the Act.

Response: Under 45 CFR 263.13(a)(i), the 15-percent cap applies to all funds a State receives under section 403 of the Act, except for Welfare-to-Work funds under section 403(a)(5). As explained in the preamble to the final TANF rule, this section provides for a consolidated administrative cap. Thus, it limits the total amount a State could spend on administrative costs based on the total amount of funding a State receives under section 403. We will not apply the 15-percent cap separately to each grant or award under section 403.

Comment: Other commenters asked that we clearly state that the high performance bonus funds, if not expended in the year of the award, would be available in future fiscal years until such time as they are spent.

Response: Under section 404 of the Act, there is no expenditure period for TANF funds awarded to States under section 403 of the Act. The “carryover” provision includes the bonus award funds. Therefore, bonus funds, once awarded to States, are available until such time as they are spent in accordance with TANF requirements (including the requirement that reserved funds must be spent on “assistance” or associated administrative costs under the TANF program).

In the NPRM, we also indicated that any expenditures for “assistance” are subject to the restrictions on the use of Federal funds at 45 CFR 263.11. We have revised the proposed regulatory language because we did not want it misinterpreted. Regardless of the purpose to which the bonus funds are put, a State's expenditure of bonus funds is subject to the requirements in § 263.11. We had referred specifically to “assistance” in the NPRM because we wanted to alert States to the special restrictions applying to the use of reserved (i.e., carryover) funds under section 404(e) of the Act. A State may spend reserved funds only on assistance and related administrative costs. (See 64 FR 17840 for further discussion.)

For clarity, we have added new paragraph (d) in this section to specify that States must report quarterly on the use of the bonus funds, along with other TANF funds, using the ACF-196.

VII. Amendment to 45 CFR Part 265

Comment: A few commenters objected to the requirement that States must submit Sections One and Three of the SSP-MOE reports in order to qualify for the high performance bonus. They also commented that a State should not be required to submit MOE data as a condition of competing for the food stamp and Medicaid/SCHIP awards since these measures do not use the SSP-MOE data.

Response: This first comment was similar to one we received on the NPRM for the general TANF rule. It reappeared because § 270.11(b) of the high performance bonus NPRM reiterated the requirement from the general TANF rule (at 45 CFR 265.3(d)) that States wishing to compete for a high performance bonus must submit data reports on their separate State programs (SSPs). The SSP-MOE reporting requirement in the general TANF rule covers all high performance bonus measures.

Last year, when we published the TANF final rule and the high performance bonus NPRM, we believed that submission of the SSP-MOE reports was critical to understanding State caseload changes and work performance. However, as TANF and SSPs have continued to evolve, and we have gained more experience with the caseload reduction credit and high performance bonus processes, we have encountered situations where a waiver of the SSP-MOE reporting requirement seemed appropriate. Most notably, as discussed in more detail later, States have raised questions about a very few SSPs that do not address basic needs; do not appear to be particularly germane to assessing State caseload reductions, work efforts, or performance; and are not amenable to TANF-like reporting. Our general TANF rules do not allow us any discretion to grant waivers of the SSP-MOE reporting requirements, even in such circumstances. Based on discussions with States, we are concerned that, in some of these cases where the SSP-MOE reporting requirements are particularly problematic, States might elect not to provide certain benefits that support the goals of TANF rather than to develop the data collection and reporting system that the SSP-MOE requirements would entail.

Since these problematic situations can arise with respect to either the high performance bonus or the caseload reduction credit, we wanted to provide an opportunity for waiver of the requirements in both circumstances. We believe the clearest way to make that change is in the general TANF rule. Thus, in this rulemaking, we have revised the TANF final rules at § 265.3(d)(2) to allow waivers of the SSP-MOE reporting requirements, under very limited circumstances. More specifically, we would allow waivers only if the benefits being provided in the SSPs were considered assistance under § 260.31(a)(3) and the State made a clear case that the cost and burden associated with collection and reporting of the data substantially outweighed any potential benefit.

The first condition means that waivers would be available only for benefits that are defined as “assistance” because they are “supportive services such as transportation and child care provided to families that are not employed.” We would not waive SSP-MOE reporting requirements for SSPs that provide assistance under paragraphs (a)(1) or (a)(2)—that is, for programs that: (1) provide assistance that meets the basic needs of a family (even if the family's receipt of such benefits is conditioned upon participation in work experience or community service); or (2) provide benefits that would have historically been considered “welfare” benefits. For example, we would never waive SSP-MOE reporting when the State had elected to meet the basic needs of its two-parent TANF cases through a SSP. We would only grant waivers for SSPs that provide supportive services, such as transportation and child care, to families that are unemployed. In fact, we anticipate that SSPs receiving reporting waivers would typically be serving a mix of employed and unemployed families, but mostly employed families. However, separating the families into employed and unemployed families and gathering detailed characteristics data on a monthly basis for only the employed families would be extremely burdensome.

In deciding whether a State had made its case, we would look at factors such as:

• The capacity of the SSP to provide the kind of information required in the SSP-MOE report;
• The size of the separate State program (e.g., the number of beneficiaries and the proportion of the TANF caseload that would represent); and
• Whether the data would be important to a full understanding of the State's work efforts, caseload changes, or performance.

An example of a situation where we might waive SSP-MOE reporting would be the following: a State provides funds to a local transportation initiative that provides shuttle bus service between a low-income, inner-city neighborhood and suburban jobs. While most of the shuttle passengers are employed, unemployed neighborhood residents can use the shuttle to get to a pre-employment training program. Thus, there may be families receiving benefits that meet the definition of “assistance” under § 260.31(a)(3). The State agency, working with the transportation program, can determine the proportion of “eligible families” using this service through a simple survey. However, the program does not collect detailed case-specific and monthly information on the families they serve, and it has no mechanism for collecting such data or submitting data electronically to the State. In addition, the number of families receiving “assistance” in the program is thought to be negligible, and they would have little chance of showing up in, or affecting, the SSP sample.

VIII. Regulatory Impact Analyses

A. Executive Order 12866

Executive Order 12866 requires that regulations be drafted to ensure that they are consistent with the priorities and principles set forth in the Executive Order. The Department has determined that this final rule is consistent with these priorities and principles.

The Executive Order encourages agencies, as appropriate, to provide the public with meaningful participation in the regulatory process. This rulemaking implements statutory authority based on broad consultation and coordination. Section 403(a)(4) of the Act requires the Department to consult with the National Governors' Association and the American Public Human Services Association in the development of a system for awarding high performance bonuses. As described earlier in the preamble and in section G. of this Regulatory Impact Analysis, ACF consulted with States, their representative organizations, and a broad range of advocacy groups, researchers, and others to obtain their views. This rule reflects the discussions with and the concerns of the groups with whom we consulted.

This rule is a significant regulatory action that will have an annual effect on the economy of $100 million or more, according to section 3(F)(1) of the Executive Order. It will determine how $200 million will be awarded annually to high performing States to be used to carry out the purposes of the TANF program. It will also have the additional effect of improving State efforts to implement welfare reform. High performing States could see their funding increase by as much as five percent of their State family assistance grant. We believe the cost of competing for a high performance bonus award in FY 2002 should be minimal since competition for these awards will be based, to the extent possible, on existing data sources. After consultation with States, advocates, and others, we will specify the data States must submit in order to compete on the child care measure in FY 2003.

B. Regulatory Flexibility Analysis

The Regulatory Flexibility Act (5 U.S.C. Ch. 6) requires the Federal government to anticipate and reduce the impact of rules and paperwork requirements on small businesses and other small entities. Small entities are defined in the Act to include small businesses, small non-profit organizations, and small governmental entities. This rule will affect only the 50 States, the District of Columbia, and certain territories. Therefore, the Secretary certifies that this rule will not have a significant impact on small entities.

C. Assessment of the Impact on Family Well-Being

We certify that we have made an assessment of this rule's impact on the well-being of families, as required under section 654 of The Treasury and General Appropriations Act of 1999. The high performance bonus awards are a statutory part of the TANF program and are designed to reward State efforts in strengthening the economic and social stability of families and carrying out other purposes in the statute. The final rule does not limit State flexibility to design programs to serve these purposes.

D. Paperwork Reduction Act

Under the Paperwork Reduction Act of 1995 (PRA), no persons are required to respond to a collection of information unless it displays a valid OMB control number. As required by the PRA, we have submitted the data collection requirements to OMB for review and approval. We used the NPRM as a vehicle for seeking comment from the public on these and any additional information collection activities that they believe should be added as a part of the bonus award process.

Based on this final rule, we will award bonuses, in FY 2002 and beyond, on four work measures, five work support measures, and one measure on family formation and stability in § 270.4. We have computed the burden based only on the work measures and the measure of Medicaid/SCHIP participation. No reporting burden would fall on the States in competing on the Food Stamp measure or the family formation measure as we will use the Census Bureau's Census 2000 Supplementary Survey and the Transitional Long-Form Database as the data source for this measure. In FY 2002, no additional reporting burden will fall on the States in competing on the child care measure as States will be ranked based on data they currently report under the Child Care and Development Fund program (ACF Forms 800, 801, and 696). After external consultation on the child care measure, we will specify, by the end of the calendar year, the additional information States must submit in order to compete on this measure in FY 2003 and submit any additional paperwork burden requirements to OMB for approval. These requirements would not become effective until approved by OMB.

Burden Estimate for the Work Measures

In § 270.6 of the NPRM, we proposed the use of two alternative sets of data. In the first alternative, States would collect quarterly and report semi-annually a minimal set of identifying information on adult TANF recipients that we would match against the information in the National Directory of New Hires (NDNH) to determine the State's scores for the work measures. In the second alternative, the State would submit more detailed work performance data based on its matching of adult recipient data with its UI data. Commenters strongly supported the first alternative and the use of the NDNH whenever possible.

In the final rule, we specify that we will use the first reporting alternative. We estimate the reporting burden for the first reporting alternative in § 270.6(a) to be 1,728 hours, based on the requirement that States report some or all of the following three data elements: the name, birth date, and social security number of all adult TANF and SSP-MOE recipients. (The specific data elements will be issued in program guidance.) Our estimate of the burden is as follows: 16 hours per response, times 54 respondents, times two (semi-annual reporting), for a total annual burden of 1,728 hours.

In addition, if a State wishes to receive a high performance bonus, it must report the data in Sections One and Three of the SSP-MOE Data Report as required in § 265.3(d) of this chapter. The burden for this reporting requirement was previously estimated in the TANF final rule, published April 12, 1999 (64 FR 17720).) We have not revised our estimates, but we note that this burden may be reduced in view of the amendment to § 265.3(d) included in this rulemaking which waives SSP-MOE reporting requirements under certain circumstances.

We believe the burden of reporting the identifying information on work measures will be minimal for most States, particularly as we will be using the NDNH as a match at the Federal level. In addition, States already have experience in extracting case/individual identifying information from their electronic data bases for matching purposes, including the Income and Eligibility Verification System (IEVS) matches required by statute.

Burden Estimate for the Measures on Medicaid/SCHIP Participation

The Medicaid/SCHIP performance measures at § 270.4(d) are based on semi-annual reporting of the data from a match of TANF data and Medicaid/SCHIP enrollment data, using information from HCFA's MSIS system and the HCFA Form 21-E. Because this activity is similar to State activity in matching TANF data and UI data, as is currently done for the ACF-200 (OMB No. 0970-0180), we estimate that the burden will be approximately the same, i.e., 4,320 hours, excluding start-up costs. We understand that some States may not have social security numbers for SCHIP recipients. In that instance, there may be an additional burden.

The total annual burden estimate includes the development of a one-time extraction program (based on our specifications), computer run-time to execute the program, the creation of an extract data file, and transmitting the information.

We estimate that the 50 States, the District of Columbia, Guam, Puerto Rico, and the United States Virgin Islands will be potential respondents. (Currently, American Samoa has not applied to implement the TANF program.)

The annual burden estimate for this data collection is:

E. Unfunded Mandates Reform Act of 1995

Section 202 of the Unfunded Mandates Reform Act of 1995 (Unfunded Mandates Act) requires that a covered agency prepare a budgetary impact statement before promulgating a rule that includes any Federal mandate that may result in the expenditure by State, local, and Tribal governments, in the aggregate, or by the private sector, of $100 million or more in any one year.

If a covered agency must prepare a budgetary impact statement, section 205 further requires that it select the most cost-effective and least burdensome alternative that achieves the objectives of the rule and is consistent with the statutory requirements. In addition, section 203 requires a plan for informing and advising any small government that may be significantly or uniquely impacted by the proposed rule.

We have determined that this final rule will not result in the expenditure by State, local, and Tribal governments, in the aggregate, or by the private sector, of more than $100 million in any one year. Competition for a high performance bonus is entirely at State option. Accordingly, we have not prepared a budgetary impact statement, specifically addressed the regulatory alternatives considered, or prepared a plan for informing and advising any significantly or uniquely impacted State or small government.

F. Congressional Review

This final rule is a major rule as defined in 5 U.S.C., Chapter 8.

G. Executive Order 13132

On August 4, 1999, the President issued Executive Order 13132, “federalism.” The purposes of the Order are: “to guarantee the division of governmental responsibilities between the national government and the States that was intended by the Framers of the Constitution, to ensure that the principles of federalism established by the Framers guide the executive departments and agencies in the formulation and implementation of policies, and to further the policies of the Unfunded Mandates Reform Act * * * *”

We certify that this final rule does not have a substantial direct effect on States, on the relationship between the Federal government and the States, or on the distribution of power and responsibilities among the various levels of government. The final rule does not pre-empt State law and does not impose unfunded mandates.

This rule does not contain regulatory policies with federalism implications that would require specific consultations with State or local elected officials. The statute, however, requires consultations with the National Governors' Association and the American Public Human Services Association in the development of a high performance bonus system.

During the development of the NPRM, we held two types of consultations. First, we raised issues related to this provision in the general TANF consultation meetings with a broad range of representatives of State and local governments; nonprofit, advocacy, and community organizations; foundations; and others. Second, we consulted intensively with representatives of the National Governors' Association, the American Public Human Services Association, the National Conference of State Legislatures, and approximately 30 State representatives who participated by regularly scheduled conference calls over a period of approximately nine months.

List of Subjects in 45 CFR Parts 265 and 270

• Grant programs—social programs; Public assistance programs; Reporting and Recordkeeping Requirements; Poverty

For the reasons set forth in the preamble, we are amending 45 CFR Chapter II as follows:

PART 265—DATA COLLECTION AND REPORTING REQUIREMENTS

1. The authority citation for part 265 continues to read as follows:

Authority: 42 U.S.C. 603, 605, 607, 609, 611, and 613.

2. We are amending § 265.3 by redesignating paragraph (d)(2) as paragraph (d)(2)(i) and adding a new paragraph (d)(2)(ii) to read as follows:

PART 270—HIGH PERFORMANCE BONUS AWARDS

3. We are adding a new part 270 to read as follows:

Authority: 42 U.S.C. 603(a)(4).