finding that a negative multiplier suggests that "the requested percentage[-]based fee is fair and reasonable"Summary of this case from Carlin v. DairyAmerica, Inc.
No C-03-5138 VRW.
November 26, 2007
As described in this court's preliminary approval order of June 30, 2007 (Doc #175), plaintiffs in this class action securities litigation reached a settlement with the defendant, Portal Software, Inc ("Portal"), on March 9, 2007. Doc #168. In that order the court preliminarily approved the proposed settlement, certified the settlement class pursuant to FRCP 23 and approved notice by publication. Notice having been mailed to 17,937 potential class members, posted on lead counsel's website, posted on the claims administrator's website and a summary notice published in Investor's Business Daily on July 5, 2007 (Doc #176 at 2), the parties now move for final approval of the settlement. In particular plaintiffs seek: (1) final approval of the proposed settlement; (2) final approval of the plan of allocation; and (3) final approval of plaintiffs' attorneys' request for fees and expenses.
The court held a final settlement approval hearing on October 25, 2007. For the reasons that follow, the court GRANTS final approval of the proposed settlement, GRANTS approval of the plan of allocation and GRANTS an award of expenses and attorneys' fees. The court will discuss each of the above issues in turn.
Portal provides billing and subscriber management solutions to its clients primarily through its "Infranet" software, for which Portal charges companies "license fees." Doc #135 at ¶ 68. Portal also charges customers "service fees" for system implementation, consulting, maintenance and training. Doc #135 at ¶ 68. Following the dot-com market crash of 2001, Portal lost many of its dot-com startup customers and incurred financial losses that wiped out more than ninety-six percent of its equity. Doc #135 at ¶ 69.
Portal subsequently began to market its Infranet product to more established and sophisticated business customers, including telecommunications providers. Doc #135 at ¶ 69. Portal's new clients required greater software customization than had the dot-com startups, which in turn affected Portal's license fee revenue recognition. Doc #135 at ¶ 69. Plaintiffs contend that under GAAP, a software provider cannot recognize licensing revenues for software that requires customization for a client until a substantial portion of the modification has been completed. Doc #135 at ¶¶ 4, 44(e), 69. Although Portal could recognize revenue when it delivered its Infranet product to its dot-com clients, the greater customization required by Portal's new, more established clients required the company to defer recognizing revenue from many of its contracts until customization was complete. Doc #135 at ¶ 153. Plaintiffs allege that during the class period, Portal began to manipulate its license fees to recognize more revenue "up-front." Doc #135 at ¶¶ 70-71.
On September 12, 2003, Portal completed a secondary offering to the public at a price of $13.25 per share, thereby generating $60 million in net proceeds. Doc #135 at ¶ 9. On November 13, 2003, defendants announced that due to contract delays, revenue recognition deferrals and service execution issues, Portal expected net losses of $0.36 to $0.40 per share for the third quarter of fiscal year (FY) 2004. Doc #135 at ¶ 10. These losses contrasted with the $0.04 net profits per share that Portal projected previously for the quarter. Doc #135 at ¶ 10. After this announcement, Portal's common share price plummeted more than 42.5% to $8.77 in afterhours trading. Doc #135 at ¶ 113.
Plaintiffs' complaint alleges that Portal made the accounting misstatements described above to inflate reported revenue numbers. Doc #135. Portal then used those numbers to create false and misleading statements regarding its financial health and future business prospects. Doc #135 at ¶ 3. According to plaintiffs, these false and misleading statements artificially inflated Portal's stock price and allowed Portal to complete the $60 million secondary offering on September 12, 2003. Doc #135 at ¶¶ 2, 9.
Plaintiffs' claims for violations of the 1933 Securities Act (`33 Act) are based on the alleged false and misleading statements made in the registration statement and prospectus issued in connection with the secondary offering. Doc #135 at ¶¶ 142-65. Plaintiffs' claims for violations of the 1934 Exchange Act (`34 Act) are based on alleged false and misleading statements disseminated to the investing public via SEC filings and press releases. Doc #135 at ¶¶ 166-81.
On August 17, 2006, the court denied defendants' motion to dismiss plaintiffs' claims under sections 11, 12(a) and 15 of the `33 Act and granted defendants' motion to dismiss plaintiffs' claims under sections 10b and 20(a) of the `34 Act. Doc #155. Additionally, because plaintiffs amended their complaint four times but still had not satisfied the Private Securities Litigation Reform Act's (PSLRA) heightened pleading requirements, the court dismissed plaintiffs' claims under the `34 Act with prejudice. Doc #155. Accordingly, the only remaining claims are under sections 11, 12(a) and 15 of the `33 Act.
The court first addresses the fairness of the settlement, consisting of $3.25 million in cash. In making its assessment, the court must adopt the point of view of the class members, who can no longer depend on their attorneys' rigorous adherence to their fiduciary duties. See Court Awarded Attorney Fees, Report of the Third Circuit Task Force, Oct 8, 1985, 108 FRD 237, 255 (1985) ("[T]he court now must monitor the disbursement of the fund and act as a fiduciary for those who are supposed to benefit from it, since typically no one else is available to perform that function — the defendant has no interest in how the fund is distributed and the plaintiff class members rarely become involved."). In assessing whether a settlement is "fair, reasonable and adequate" under FRCP 23(e)(1)(C), the court is to consider several factors:
(1) the strength of the plaintiffs' case; (2) the risk, expense, complexity, and likely duration of further litigation; (3) the risk of maintaining class action status throughout the trial; (4) the amount offered in settlement [presumably in comparison to comparable cases]; (5) the extent of discovery completed and the stage of the proceedings; (6) the experience and views of counsel; (7) the presence of a governmental participant; and (8) the reaction of class members to the proposed settlement.Churchill Village v General Electric, 361 F3d 566, 575 (9th Cir 2004), citing Hanlon v Chrysler Corp, 150 F3d 1011, 1026 (9th Cir 1998). To these factors, the court adds (9) the procedure by which the settlement was arrived at, see Manual for Complex Litigation (Fourth) § 21.6 (2004), and (10) the role taken by the lead plaintiff in that process, a factor somewhat unique to the PSLRA.
Factor (1), the strength of plaintiffs' case, somewhat favors settlement because plaintiffs' remaining claims are tenuous. Plaintiffs assert that establishing liability and damages at trial would be difficult because of the uncertainties associated with proving its claims, which are "exacerbated by the unpredictability of a lengthy and complex jury trial." Doc #176 at 6. Additionally, the heightened pleading requirement of the PSLRA and the application of Dura Pharms, Inc v Broudo, 544 US 336 (2005), which poses significant risks to plaintiffs' ability to survive the pending summary judgment and prevailing at trial, suggest that settlement here is prudent. Further, because the `34 Act claims were dismissed with prejudice, plaintiffs' case is relatively weak.
Similarly, the risk, expense, complexity and likely duration of further litigation, factor (2), favor settlement because further litigation would entail substantial risk to the class of recovering nothing. Any further litigation would likely be complex, expensive and a favorable outcome improbable. Additional consideration of increased expenses of fact and expert discovery and the inherent risks of proceeding to summary judgment, trial and appeal also support the settlement. See Doc #178 at 10. Also, the court's ruling on the `34 Act claims effectively barred class members with `34 Act claims from recovery if the case proceeded to trial. Accordingly, the risk, expense, complexity and likely duration of further litigation favor settlement.
Factor (3) does not weigh in favor of settlement because class treatment is generally appropriate in such litigation and any risk in maintaining the class is low.
The amount offered in settlement, addressed in factor (4), also favors the settlement. Plaintiffs contend the potential aggregate damages at trial for the `33 Act claims is approximately $13 million — a four-fold increase from the $3.25 million settlement. Doc #176 at 7-8. The settlement therefore represents twenty-five percent of plaintiffs' estimated maximum recovery at trial. The Ninth Circuit has held that "a cash settlement amounting to only a fraction of the potential recovery will notper se render the settlement inadequate or unfair." Officers for Justice, 688 F2d 615, 628 (9th Cir 1982); see In re Mego Financial Corp, 213 F3d 454, 459 (9th Cir 2000). And the Ninth Circuit has upheld percentage recoveries lower than the twenty-five percent recovery here. See In re Mego Financial Sec Litig, 213 F3d at 459 (noting that a settlement of $2 million out of a potential $12 million, or 16.66%, was fair and adequate). The settlement also exceeds median percentages for all securities class actions in 2005 and 2006, which were approximately 3.1% and 2.4% respectively. See Laura E Simmons Ellen M Ryan, Securities Class Action Settlements: 2006 Review and Analysis, at 5 (Cornerstone 2007). For securities class action settlements below $50 million, median settlements as a percentage of estimated damages were 10.5% through year end 2005 and 8.8% in 2006. Simmons Ryan, supra, at fig 5. Accordingly, the amount here favors settlement.
The extent of discovery completed and the stage of the proceedings, factor (5), also supports settlement. By the time the settlement was reached, the litigation had proceeded to a point in which both plaintiffs and defendants "ha[d] a clear view of the strengths and weaknesses of their cases." In re Warner Communications Sec Litig, 618 F Supp 735, 745 (SDNY 1985) aff'd 798 F2d 35 (2d Cir 1986). The settlement reflects three and a half years of completed work including pre-filing investigation, locating and interviewing over twenty-one witnesses, four amended complaints, substantial research, opposing defendant's motion to dismiss and plaintiff's analysis of defendant's motion for summary judgment. Doc #176 at 5. As a result, the true value of the class's claims was well-known. Plaintiffs' counsel possessed an accurate and sufficient understanding of the strengths and weaknesses of the case.
The views of counsel, factor (6), support settlement as well. While some courts have indicated that such views are entitled to deference, see, for example, Williams v Vukovich, 720 F2d 909, 922-23 (6th Cir 1983), this court is reluctant to put much stock in counsel's pronouncements, given their pecuniary interest in seeing the settlement approved. Here, plaintiffs' counsel are experienced in complex class action and securities litigation. The court is well acquainted with the views of plaintiffs' counsel and considered their memoranda in support of the settlement approval. See Doc ##176, 178. Further, the court knows of no objections by defendant's counsel which would indicate anything other than support for the settlement. While the court is disinclined to give counsels' views too much weight, those views do support settlement.
Factor (7) does not support settlement, inasmuch as there is no government participant present.
Factor (8) considers the reaction of class members to the proposed settlement and supports the settlement because no objections and only one opt-out was made from the class of roughly 17,937 members. The Ninth Circuit has approved settlements over objections if the settlement otherwise meets the fairness requirements. See, for example, Churchill Village, 361 F3d at 577 (500 opt-outs and 45 objections out of approximately 90,000 notified class members); In re Mego Financial Sec Litig, 213 F3d at 459 (one objection out of a potential class of 5400);Marshall v. Holiday Magic, Inc, 550 F2d 1173, 1178 (9th Cir 1977) (one percent of the class disapproved). Accordingly, the reaction of the class members as a whole supports the settlement.
Factor (9), the procedure by which the settlement was arrived, supports the settlement. The arms-length, contentious negotiations that culminated in the settlement agreement indicate that the settlement was reached in a procedurally sound manner. Doc #176 at 4. There is nothing in the record indicating any collusion or bad faith by the parties. Additionally, in its preliminary approval order the court found that "the procedure for reaching this settlement was fair and reasonable * * *. Experienced counsel on both sides, each with a comprehensive understanding of the strengths and weaknesses of each party's respective claims and defense, negotiated this settlement over an extended period of time in early 2007." Doc #175 at 10.
Finally, factor (10), the role taken by the lead plaintiff in the settlement process, supports settlement because lead plaintiff was intimately involved in the settlement negotiations. See Doc #178 at ¶ 57. Congress sought to foster such involvement through the PSLRA's lead plaintiff provisions.
For the reasons discussed above and in its preliminary approval order, the court finds that, on balance, the settlement is fair, reasonable and adequate to the class within the meaning of FRCP 23(e)(1)(C). Accordingly, the court GRANTS the motion for final approval of the settlement.
The court turns next to the proposed plan of allocation, which must be fair, reasonable and adequate. Class Plaintiffs v Seattle, 955 F2d 1268, 1284 (9th Cir 1992).
Plaintiffs' counsel submitted a plan of allocation which they developed from consultations with lead plaintiffs, representative plaintiffs and representative plaintiffs' outside counsel. Plaintiffs also consulted a damages expert named Bjorn Steinholt. Doc ##170, 176 at 13. Plaintiffs hired Steinholt to draft a plan of allocation to ensure a fair distribution of the available settlement proceeds. Steinholt's proposed plan distinguishes between class members asserting `34 Act claims, which were dismissed with prejudice, and those asserting `33 Act claims, which remain active. Doc #170.
The proposed plan provides that each authorized claimant shall be allocated a percentage of the net settlement fund according to their recognized claim. The plan allocates ninety-five percent of the net settlement fund to the `33 Act class members and five percent of the net settlement to the `34 Act class members. Doc #178 at 14-15.
The `33 Act claims, representing ninety-five percent of the net settlement fund, will take as follows:
For shares of Portal securities that were purchased or acquired in the September 12, 2003, Secondary Offering, and
(a) sold on or before November 13, 2003, the claim per share is $0;
(b) retained at the close of trading on November 13, 2003, the claim per share is the difference between the $13.25 offer price per share and the $8.40 per share (November 14, 2003 closing price).
Doc #178 at ¶ 67.
The `34 Act claims, representing five percent of the net settlement fund, will take as follows:
For shares of Portal securities that were purchased or acquired from May 20, 2003 through November 13, 2003, and
(a) sold on or before November 13, 2003, the claim per share is $0;
(b) retained at the close of trading on November 13, 2003, the claim per share is equal to the lesser of: (i) the purchase price per share less $8.40 per share (November 14, 2003 closing price), or (ii) $6.86 per share (November 14, 2003 price decline).
Doc #178 at ¶ 67.
Under this plan, the estimated average payment to the `34 Act class members is approximately $0.01 per share before deduction of fees and expenses. Doc #178 at ¶ 68. The estimated average payment to the `33 Act class members is $1.10 for each share before deduction of fees and expenses. Doc #178 at ¶ 68.
Plaintiffs' counsel reasons that because the `34 Act claims were dismissed with prejudice and are likely to fail on appeal, those claims had little weight during settlement negotiations and should, therefore, be limited to five percent of the net settlement proceeds. Doc #176 at ¶ 69. Counsel further asserts that it was in the best interest of the `34 Act class members to pool their claims with the `33 Act class members, and the plan was designed to allocate fairly and rationally the proceeds of the settlement among the class. Doc #176 at ¶ 69.
Courts endorse distributing settlement proceeds according to the relative strengths and weaknesses of the various claims. SeeIn re Warner Communications Sec Litig, 618 F Supp 735, 745 (SDNY 1985), aff'd, 798 F2d 35 (2d Cir 1986); In re Agent Orange Prod Liab Litig, 611 F Supp 1396, 1411 (EDNY 1985) ("[I]f one set of claims had a greater likelihood of ultimate success than another set of claims, it is appropriate to weigh `distribution of the settlement * * * in favor of plaintiffs whose claims comprise the set' that was more likely to succeed."), quoting In re Corrugated Container Antitrust Litigation, 643 F2d 195, 220 (5th Cir 1981);Petrovic v AMOCO Oil Co, 200 F3d 1140, 1152 (8th Cir 1999) (upholding distribution plan where class members received different levels of compensation and finding that no subgroup was treated unfairly). Distinguishing between the `33 and `34 Act claims seems appropriate here, because the court dismissed the `34 Act claims with prejudice. Accordingly, the court cannot conclude that the plan of allocation is obviously deficient or is not "within the range of possible approval." Schwartz v Dallas Cowboys Football Club, Ltd, 157 F Supp 2d 561, 570 n12 (ED Pa 2001).
While the plan of allocation is likely fair and reasonable, the court finds it appropriate in light of the Ninth Circuit's recent opinion In re Veritas Software Corp Sec Litig, 496 F3d 962 (9th Cir 2007), to examine whether the class notice properly disclosed the amount of recovery and the amount of expenses for each claimant.
Veritas involved a class action lawsuit alleging that Veritas Software Corporation ("Veritas") "falsely represented that it had entered a $50 million deal with AOL, structured to appear as if Veritas had sold $50 million in software and services to AOL and had purchased $20 million in online advertising from AOL. This `round-trip' transaction allowed both companies to artificially inflate their revenues and earnings." In re Veritas, 496 F3d at 965. The class, consisting of individuals and entities who purchased securities of Veritas between January 3, 2001 and January 16, 2003, claimed that because of the false representations the price of Veritas securities was inflated artificially and the individuals purchasing Veritas securities were therefore harmed. In re Veritas, 496 F3d at 965.
The parties settled, and lead counsel sent notice of the proposed settlement to the class members in March 2005, stating that the "estimated average recovery per share will be approximately $0.25." In re Veritas, 496 F3d at 965. Michael Malone, a class member, objected to the initial notice because it unfairly excluded four classes of Veritas securities from the settlement. In re Veritas, 496 F3d at 965-66. Accordingly, the court required a revised notice to be sent out, which again stated that "estimated average recovery per share of common stock will be approximately $0.25." In re Veritas, 496 F3d at 966.
Malone objected to the revised notice, this time questioning the calculation of the $0.25 in damages. In re Veritas, 496 F3d at 966. As a result, the district court ordered subsequent briefing to explain how the estimate of $0.25 was calculated. Lead plaintiff then explained that $0.25 stemmed from an assumption, based on a NERA Economic Consulting study, that only 43% of the class members would actually file claims. In re Veritas, 496 F3d at 966. Malone responded that if all eligible shares filed claims, the recovery per share would be $0.085. In re Veritas, 496 F3d at 966. Despite the objections, the district court approved the settlement and plan of allocation. In re Veritas, 496 F3d at 967-68.
On appeal, Malone argued that the "revised notice of proposed settlement was inadequate because it failed to meet the PSLRA requirement that it provide a calculation of the amount of settlement proposed to be distributed on a per share basis." In re Veritas, 496 F3d at 969. Additionally, Malone asserted that "because [the revised notice] calculated the estimated average recovery per share of common stock at $0.25 based on an undisclosed assumption that only 43% of class members would file a claim," the notice failed the notice requirements under the PSLRA. In re Veritas, 496 F3d at 969.
The PSLRA requires disclosure of "[t]he amount of settlement proposed to be distributed to the parties to the action, determined in aggregate and on an average per share basis." 15 USC § 78u-4(a)(7)(A). Additionally, the PSLRA requires a statement of "the average amount of damages per share that would be recoverable if the plaintiff prevailed on each claim alleged." 15 USC § 78u-4(a)(7)(B). The purpose of the PSLRA's notice requirement is "to allow class members to evaluate a proposed settlement" because "[w]ith sufficient notice, class members can * * * weigh the risks and rewards of proceeding to trial or participating in the proposed settlement." In re Veritas, 496 F3d at 969.
The Ninth Circuit noted the PSLRA's policy of providing transparency to class members and, interpreting the PSLRA's notice requirement under 15 USC § 78u-4(a)(7)(A)-(B) as "a question of first instance in this and all other circuits," held that "the estimated average recovery per share must be based on all of the shares in the class." In re Veritas, 496 F3d at 971. The notice may also include an "estimated average per share recovery," but it must be "based on explicitly disclosed projections of how many class members are likely to file claims." In re Veritas, 496 F3d at 971. Accordingly, because the notice in Veritas implied that the calculations were based on 100% of the class filing claims but were actually based on undisclosed assumptions "significantly less than 100%," the notice "clearly did not satisfy the requirements of the PSLRA." In re Veritas, 496 F3d at 970-71. The Ninth Circuit vacated and remanded to the district court to issue a new class notice.
Considering the overarching policy concern of fair, accurate disclosure to class members, the court turns to: (a) whether plaintiffs' counsel complied adequately with the disclosure requirement of Veritas; and (b) whether the estimated settlement amounts per share for the `33 Act ($1.10/share) and `34 Act ($0.01/share) sufficiently disclose attorneys' fees and expenses.
At first glance, it is unclear whether the notice in this case complies with the disclosure requirement in Veritas. The notice is silent on the key fact that warranted vacatur in Veritas — the precise assumptions behind the estimated average recovery per share. The notice is ambiguous, stating that claimants "could get more money" if fewer than the "anticipated" class members send in claim forms. Doc #179 Ex A at 4. The notice does not define what the "anticipated" number is, failing the Ninth Circuit's requirement that estimates must be based on "explicitly disclosed projections of how many class members are likely to file claims."In re Veritas, 496 F3d at 971.
After examining the matter in greater detail, the court concludes that the notice substantially satisfies the rule inVeritas. To illustrate counsel's substantial compliance withVeritas, some simple calculations are necessary.
Taking the `34 Act claims first, the notice appears to comply with Veritas. The total recovery allocated to those claims before deducting fees and expenses is five percent of $3,250,000, or $162,500. Counsel declared that the total number of damaged shares in the `34 Act claims is 13,155,060. See Doc #176 at 14. Thus the gross recovery per share is $0.012. That amount rounds down to $0.01 per share, which is consistent with the recovery stated in the notice. These calculations demonstrate that counsel assumed that one hundred percent of the 13,155,060 damaged shares would submit a claim.
162500 / 13155060 = 0.012
For the `33 Act claims, the picture is less clear. The total recovery allocated to those claims before fees and expenses is ninety-five percent of $3,250,000, or $3,087,500. Counsel declared that the total number of damaged shares in the `33 Act claims is 2,881,701. See Doc #176 at 14. Thus the gross recovery per share is $1.07. That amount directly contradicts the stated recovery per share of $1.10 that was advertised to the class.
3087500 / 2881701 = 1.07
The $1.07 amount suggests that counsel did not assume that one hundred percent of eligible shares would submit a claim; if fewer than one hundred percent of the 2,881,701 shares submit a claim, then the per share recovery would rise from $1.07 to a higher number, such as $1.10. To illustrate this point, one must calculate backwards from the stated recovery of $1.10. If the stated recovery of $1.10 is correct, and if the recovery for the `33 Act claims is $3,087,500, then the number of damaged shares should be 2,806,818. That number of damaged shares is 74,883 less than counsel's stated number of damaged shares. Counsel offers no explanation for that discrepancy.
3087500 / 2806818 = 1.10.
The following table summarizes the above calculations. Counsel's proffered arithmetic (3087500 / 2881701 = 1.10) does not compute.Recovery Amount $3,087,500 $3,087,500 Number of Damaged 2,881,701 2,806,818 Shares Recovery per Share $1.07 $1.10 Difference from 0 74,883 Stated Number of Damaged Shares Rounding does not account for the discrepancy. A recovery per share of $1.095 (which would round up to 1.10) implies a damaged number of shares of 2,819,635, for a discrepancy of 62,066. A recovery per share of $1.104999 (which would round down to 1.10) implies a damaged number of shares of 2,794,118, for a discrepancy of 87,583. Accordingly, if the $1.10 per share recovery stated in the notice is correct, then the actual number of damaged `33 Act shares is somewhere between 2,794,118 and 2,819,635, rather than counsel's declared number of 2,881,701. The following table summarizes the range of possible damaged shares and recoveries per share. Recovery Amount $3,087,500 $3,087,500 $3,087,500 Number of 2,819,635 2,806,818 2,794,118 Damaged Shares Recovery per $1.095 $1.10 $1.104999 Share Difference from 62,066 74,883 87,583 Stated Number of Damaged Shares Percentage 2% 2.5% 3% Difference from Stated Number of Damaged Shares The foregoing calculations imply that counsel did not assume that one hundred percent of the stated number of damaged shares would submit a claim. Instead, the numbers reflect an implicit assumption that ninety-seven to ninety-eight percent of the stated number of damaged shares would submit a claim.
Nonetheless, a small discrepancy of this scale does not run afoul of Veritas, which stated that "what matters is that the assumption [of class members who would submit claims] was undisclosed and significantly less than 100%." In re Veritas, 496 F3d at 970. The two to three percent difference is not a significant departure, and there is no indication that counsel made a specific assumption that less than one hundred percent of damaged shares would submit a claim, as did counsel in Veritas. The more likely explanation for the discrepancy is that counsel should be taken literally when they claim that the number of damaged `33 Act shares is "approximately 2,881,701." Doc #176 at 14. A differential of two or three percent satisfies any common sense definition of "approximately." The three-cent difference between $1.10 and $1.07, while not insignificant, is not so large that the notice "overstated the likely recovery per share [in a way that] may have discouraged other objectors from speaking up."In re Veritas, 496 F3d at 972. The notice here was in line with the purpose of the PSLRA's notice requirement, which is to allow class members to evaluate a proposed settlement effectively and fairly. In re Veritas, 496 F3d at 969-70. Accordingly, the notice here substantially satisfies the rule in Veritas.
In future motions to approve final settlements, counsel should provide to the court spreadsheets containing the above calculations. Such calculations, supported by declarations, are the best way for district courts to ensure compliance withVeritas.
The court next turns to the estimated recovery amounts listed in the notice for the `33 Act ($1.10/share) and `34 Act ($0.01/share). Specifically, the court notes the failure to include estimated attorneys' fees and expenses in the recovery estimates. Considering the goals of providing transparency and accurate information to the class, the court finds it appropriate to analyze whether failing to include attorneys fees in the calculation of estimated recovery amounts complies with the text and structure of the PSLRA as well as the spirit ofVeritas.
The "Notice of Pendency and Proposed Settlement of Class Action" published and mailed to the class states:
[T]he estimated average recovery per share of securities purchased or acquired in the Secondary Offering will be approximately $1.10 before deduction of Court-approved fees and expenses and the average recovery per share of securities not purchased or acquired in the Secondary Offering will be approximately $0.01 before deduction of Court-approved fees and expenses.
Doc #179 Ex A. Three paragraphs later the notice states that if the requested expenses and attorneys fees are approved by the court, those funds will total "20% of the Settlement Fund," and "the average cost per share of securities will be $0.04."
The $1.10 and $0.01 amounts fail to include costs and expenses, creating inflated estimates. The notice discloses that those estimates were calculated "before deduction of Court-approved expenses," but under the PSLRA after Veritas, it is questionable whether the notice provides clarity to the potential class members so that they could accurately weigh the risks of accepting or rejecting the settlement.
The notice may technically violate the PSLRA, which requires disclosure of "[t]he amount of the settlement proposed to be distributed to the parties to the action." 15 USC § 78u-4(a)(7)(A). Here, for example, "the amount" to be "distributed" to the `33 Act claimants is only $0.88 instead of $1.10, but the notice does not say that explicitly. The ordinary meaning of the statutory language is that the notice must disclose a single amount that the class member can multiply by his number of shares and then take to the bank, as it were. While the notice discloses that expenses and fees are not included in the per share amount, the PSLRA requires more. The amounts disclosed — $1.10 per share and $0.01 per share — do not reflect the net distribution to the class members, which should account for attorneys' fee and expenses. 15 USC § 78u-4(a)(7)(A).
$1.10 minus twenty percent ($0.22) is $0.88.
Even though the expenses deduction comes only three paragraphs later and the class members can calculate for themselves, the concern is with the initial false impression of the amount of recovery. And as a rule of law, due to the PSLRA's concern with transparency and clarity, a per se rule requiring a single disclosure of the final amount of recovery is more desirable than a flexible inquiry into the complexity of the algebraic calculations that settlement notices might impose upon class members in the future.
Moreover, the notice is not exactly clear what the amount of expenses is. The notice states that costs will total "20% of the Settlement Fund," while the next sentence states that "the average cost per share of securities will be $0.04." The latter number is grossly misleading because it does not explain that the cost per share of securities for the `33 Act claims is over one hundred times larger than the cost per share for the `34 Act claims. The cost per share for the `33 Act claims is $0.22, whereas the cost per share for the `34 Act claims is $0.002. Because that information is not spelled out, a `33 Act claimant might read the notice and conclude that she will receive $1.06 per share ($1.10 minus $0.04), while she will in fact receive only $0.88. A difference that large might affect a class member's decision whether to accept or reject the settlement.
Even if the notice tries to avoid that problem by listing expenses as "20% of the Settlement Fund," the $0.04 figure is listed in the same unit as the average recovery — dollars per share — and more easily invites calculation. After looking at the notice, the understandable if not natural reaction would be to subtract 0.04 from 1.10 instead of multiply 1.10 by (100% minus 20%). The notice is confusing because it lists multiple different numbers each requiring a different calculation, even though the PSLRA requires a single, bottom line number.
Counsel's failure to list separately the cost per share of securities for each specific class of claims (i e $0.22 and $0.002) may itself violate the PSLRA. The statute requires the class notice to disclose "the amount of [attorneys'] fees and costs determined on an average per share basis." 15 USC § 78u-4(a)(7)(C). At first glance, the notice complies with that provision perfectly by stating that "the average cost per share of securities will be $0.04." Doc #179 Ex A. That seeming compliance is inadequate in this case because the class was bifurcated into two classes asserting different claims, whereas the text and structure of the PSLRA contemplate only a single class. See 15 USC § 78u-4(a)(7) (referring to "the class" in the singular); see also 15 USC § 78u-4(a)(5)-(8) (same). Because the statute was written with only a single class in mind, the provision requiring disclosure of costs "on an average per share basis" is unobjectionable on its face. It makes little sense, however, to apply that provision unthinkingly to a settlement with more than one discrete group of claimants. In such an instance, the best reading of the statute is that costs must be disclosed on an average per share basis for each group. A requirement along those lines would give each shareholder a clear idea of what her individual claim is worth. Particularly in this case, where the cost per share for one group is one hundred times the cost per share for the other group, the increased clarity would have been appropriate. Accordingly, because the notice included a useless and misleading figure of $0.04 per share, the notice may have run afoul of the PSLRA's requirement that cost per share information be "clearly summarized." 15 USC § 78u-4(a)(7)(C). Instead, a separate statement of the cost per share for the `33 Act claimants and the cost per share for the `34 Act claimants would have been more in line with the statute. The `33 Act claimants could have seen more clearly that their net recovery was not $1.06 but $0.88.
Overall, the ambiguity in the class notice here fails to satisfy the "clear purpose of the notice requirement" which is "to allow class members to evaluate [the] proposed settlement."In re Veritas, 496 F3d at 969. Providing estimated per share recovery in one location and fees and expenses in a separate location fails the notice requirement because seemingly the only purpose of breaking up those two pieces of information is to make the per share recovery appear higher than it actually is. The PSLRA requires a statement of the net amount to be distributed. Listing the cost per share for all claims together also fails the notice requirement because it does not properly differentiate between distinct claims that may have substantially different costs per share.
The Ninth Circuit did not address the cost disclosure issue despite similar language in the disputed notice in Veritas. There, the notice stated that "the estimated average payment for common stock will be approximately $0.25 for each share before deduction of court approved fees and expenses." In re Veritas, 496 F3d at 970 (emphasis added). Veritas confined its discussion to "undisclosed assumptions" only, but the notice there "overstated the likely recovery per share" and "may have discouraged other objectors from speaking up," thus adversely impacting the overall fairness of the settlement. In re Veritas, 496 F3d at 972.
Despite the possibility that the notice here fails to satisfy the PSLRA, the court does not find the notice utterly deficient. Because the Ninth Circuit did not address this issue and because guidance is lacking on the matter, the court does not reject the notice. Rather the court suggests that because the PSLRA andVeritas require the notice to be as clear and accurate as possible, including attorneys' fees and expenses in the estimated recovery amounts is not only a more appropriate way of attaining clarity, but is a reasonable step for plaintiffs' attorneys to take. The court understands that the costs of finalizing a class action recovery cannot be known with certainty at the time notice is given. Attorney fees may have to be incurred in shepherding the settlement through the approval process subsequent to the notice and certain out-of-pocket expenses incurred in the administration of the fund. But plaintiffs' attorneys, not the class members, understand the nuances of the settlement and are the best providers of reasonable estimates of these costs and can be expected to make cautious estimates of what class members should expect lest they give rise to disappointments that might further complicate resolution of the litigation. Accordingly, future class notices should provide a single number showing the actual amount of take-home recovery for each share, using reasonable estimates of any future expenses of class administration.
Although the settlement amount is fair, the court must scrutinize class counsel's request for expenses and attorneys' fees. Under the proposed fee award, plaintiffs' counsel would receive an award of 20% of the common fund inclusive of expenses, with an expenses cap of $95,000. Doc #179 Ex A. This translates to a total of $650,000, reflecting $83,274.81 for expenses and $566,725.19 for attorneys' fees. Doc #177 at 1. After the award, $2.6 million will remain for the class members.
The court first turns to the request for expenses in the amount of $83,274.81, accounting for 2.56% of the total fund. Doc ##181-83. The expenses represent costs incurred by plaintiffs' counsel over three and a half years of litigation, and the largest expenses are attributed to "experts, consultants and investigators" (44%), "meals, hotels and transportation" (22%) and "shareholder notice" (21%). Doc ##181-83. Facially the expenses raise no indication of unreasonableness. Class action securities litigation is complex. Considering the 1700-plus hours spent on the litigation, awarding expenses in the amount of $83,274.81 seems reasonable. Further, the current proposal bundles expenses with fees, which is consistent with the court's preference. SeeWenderhold v. Cylink Corp, 189 FRD 570, 573 (ND Cal 1999).
In addition to expenses, plaintiffs' counsel are entitled to compensation for bringing the case and obtaining a fair settlement. See Boeing Co v Van Gemert, 444 US 472 (1980). Plaintiffs' counsel requests fees of 17.44% of the common fund, or $566,725.19.
Plaintiffs' counsel assert that the attorneys' fee request is reasonable because of the extent of the litigation; the risks of litigation and the need to ensure the availability of competent counsel in high-risk, contingent securities cases; the standing and expertise of plaintiffs' counsel; and the standing and caliber of oppositions' counsel. See Doc #178 at 17, 20. In addition, plaintiffs' counsel note that the requested 17.44% fee is below the Ninth Circuit's 25% benchmark, and the fee request represents a reduction from an ex ante negotiated fee agreement of 25% of recovery plus expenses. See Doc #177 at 5, 10 n8.
The 17.44% fee appears at first impression reasonable. See In re HPL, 366 F Supp 2d 912, 918 (ND Cal 2005) (noting that a fee of 15% of the common fund appeared reasonable). In assessing a percentage-based fee, the court looks to the Ninth Circuit's benchmark of 25% as a starting point. See Paul, Johnson, Alston Hunt v Graulty, 886 F2d 268, 272 (9th Cir 1989). This percentage may then be adjusted based on a number of circumstances, including the size of the settlement, the risk undertaken by counsel in pursuing the case and the extent to which this case precluded counsel from taking other matters. See Vizcaino v Microsoft Corp, 290 F3d 1043, 1048-50 (9th Cir 2002); Six (6) Mexican Workers v Ariz Citrus Growers, 904 F2d 1301, 1311 (9th Cir 1990).
But "[r]elying on percentages without reference to * * * other factors can be, like blind reliance on benchmarks, an `all too tempting substitute for the searching assessment that should properly be performed.'" In re HPL, 366 F Supp 2d at 918, quotingGoldberger v Integrated Resources, Inc, 209 F3d 43, 52 (2d Cir 2000). Accordingly, it is the court's practice to assess a percentage fee award not only by using the usual litany of factors bearing on the reasonableness of a fee (see, for example,Vizcaino, 290 F3d at 1047-50), but also by cross-checking the percentage fee award against a rough fee computation under the lodestar method. See, for example, In re HPL, 366 F Supp 2d at 919-25. The court's view that a lodestar cross-check is appropriate in setting a fee award mirrors its concern that even a modest percentage fee recovery may represent a windfall to lead counsel. Even though it is unlikely a lodestar calculation here would suggest an award far below the percentage fee, the court believes a cross-check is appropriate.
In conducting a lodestar cross-check, the court must first determine the dollar value of the proposed percentage-based fee award. Here that amount is $566,725.19, representing a requested 17.44% fee out of the $3.25 million settlement.
The second step is to cross-check the proposed percentage fee against the lodestar figure. "Three figures are salient in a lodestar calculation: (1) counsel's reasonable hours, (2) counsel's reasonable hourly rate and (3) a multiplier thought to compensate for various factors" such as unusual skill or experience of counsel, or ex ante risk of nonrecovery. In re HPL, 366 F Supp 2d at 919. The multiplier is computed from the ratio of the proposed percentage fee to the computed lodestar fee and will be assessed for reasonableness. In re HPL, 366 F Supp 2d at 919. "Accordingly, the court need only consider counsel's reasonable hours and counsel's reasonable hourly rate in computing the lodestar." In re HPL, 366 F Supp 2d at 919.
Plaintiffs' counsel provided calculations of lodestar figures (Doc ##181-83) and sworn declarations from the lead attorneys in charge of billing records for the case attesting to (1) the experience and qualifications of the attorneys who worked on the case (Doc #177); (2) those attorneys' customary billing rates during the pendency of the case; and (3) the hours reasonably expended by those attorneys in pursuing the case. Doc ##181-83. But plaintiffs' counsel failed to provide detailed experience levels sufficient for proper analysis. Plaintiffs' counsel merely distinguished between partners and associates, with no specification of the respective years of experience each attorney possesses. Accordingly, in computing its own lodestar, the court assumes that attorney experience is measured from the year of law school graduation or the year admitted to the bar. See In re Rite Aid Corp Sec Litig, 396 F3d 294, 306 (3d Cir 2005) ("The lodestar cross-check calculation need entail neither mathematical precision nor bean-counting.").
The court will conduct its own assessment based on reasonable billing rates. See In re HPL, 388 F Supp 2d at 919; Vaughn R Walker Ben Horwich, The Ethical Imperative of a Lodestar Cross-Check: Judicial Misgivings About "Reasonable Percentage" Fees in Common Fund Class Actions, 18 Georgetown J Legal Ethics 1453, 1455 (2005). To determine appropriate hourly rates, it is the practice of the court to rely on official data. One reliable official source for rates that vary by experience levels (and used by the court in In re HPL), is the Laffey matrix used in the District of Columbia. See http://www.usdoj.gov/usao/dc/Divisions/Civil_Division/Laffey_Matrix_6.html, citing Laffey v Northwest Airlines, Inc, 572 F Supp 354 (D DC 1983), aff'd in part, rev'd in part on other grounds, 746 F2d 4 (DC Cir 1984).
Under the 2007 Laffey matrix, attorneys bill at the following rates according to experience:Experience Rate Per Hour 20+ Years $440 11-19 Years $390 8-10 Years $315 4-7 Years $255 1-3 Years $215 Paralegals Law Clerks $125 These figures are tailored for the District of Columbia, which has a lower cost of living than both San Diego (the city in which lead counsel for lead plaintiffs firm operates) and the San Francisco Bay area, but a higher cost of living than Crestwood, Kentucky (a suburb of Louisville, where representative plaintiffs' counsel's firm operates). The court will adjust theLaffey figures accordingly based on the locality pay differentials within the federal courts. See 2007 General Schedule (GS) Locality Pay Tables, at http://www.opm.gov/oca/07tables/indexGS.asp; see also In re HPL, 366 F Supp 2d at 921 (adjusting locality pay differentials based on the geographical region in which lead counsel's firm operated). The Washington-Baltimore area has a +18.59% locality pay differential; the San Diego area has a +20.34% locality pay differential; the San Francisco-Oakland-San Jose area has a +30.33% locality pay differential; and the Louisville, Kentucky area has a +17.38% locality pay differential. Adjusting theLaffey matrix figures accordingly will yield appropriate rates for the respective geographical regions: +1.5% for San Diego and +10% for San Francisco-Oakland-San Jose. No adjustment is made for Louisville, which, for simplicity, will be compensated at theLaffey DC rates.
Applying these adjustments the court obtains the following rates:San Diego: San Francisco — Oakland — + 1.5% Adjustment San Jose: +10% Adjustment 20+ Years 446.60 484.00 11-19 Years 395.85 429.00 8-10 Years 319.72 346.50 4-7 Years 258.82 270.50 1-3 Years 218.22 236.50 Paralegals Law Clerks 126.87 137.50 Substituting the values above and correlating them to the appropriate attorney or paralegal, the following table reflects the court's adjusted lodestar calculations:Attorney Experience 2007 Billing Total Total Rate (per hr) Hours Lodestar Alba (SD) 5 years 258.82 14.75 $3,817.60 Alexander 20+ years 446.60 11.00 $4,912.60 Bull 19 years 395.85 110.00 $43,543.50 Cauley 13 years 395.85 4.00 $1,583.40 Daley 11 years 395.85 5.0 $1,979.25 Geller 14 years 395.85 0.25 $98.96 Gunyan 12 years 395.85 111.00 $43,939.35 Reid Avallone 5 years 258.82 1.75 $452.94 Reise 20+ years 446.60 2.75 $1,228.15 Robbins 5 years 258.82 0.25 $64.71 Rosenfeld 7 years 258.82 2.75 $711.76 Rothman 13 years 395.85 463.75 $183,575.44 Rudman 15 years 395.85 10.25 $4,057.47 Stein 13 years 395.85 13.00 $5,146.06 Svetcov 20+ years 446.60 354.75 $158,431.35 Wilens 11 years 395.85 250.25 $99,061.46 Green (SF) 20+ years 484.00 37.05 $17,932.20 Pillette unknown unknown 0.50 unknown Jigarjian 20+ years 484.00 12.00 $5,808.00 Sharma 3 years 236.50 18.21 $4,306.67 Welling 7 years 270.50 0.50 $135.25 Roelandt (KY) 20+ years 435.60 31.70 $13,808.52 Stewart 20+ years 435.60 45.09 $19,641.20 Law Clerk (SF) — 137.50 1.00 $137.50 Paralegals(KY) — 123.75 0.25 $30.94 Paralegals(SD) — 126.87 206.50 $26,198.67 Paralegals(SF) — 137.50 4.98 $684.75 Total 1713.28 $684,287.65 Only categories covered by the Laffey matrix are included in the above tabulation, and consolidated paralegal work is based on geographical region.
The court now turns to the lodestar cross-check, which entails evaluation of the multiplier implied by lead counsel's requested fee (17.44% percent of a $3.25 million settlement, or $566,725.19) and lead counsel's lodestar fee (computed above as $684,287.65). These data imply a multiplier of 0.83.
Here, the lodestar amount ($684,397.65) exceeds the percentage-based award ($566,725.19), reflecting a somewhat unusual situation. See Walker and Horwich, supra, at 1470 n71. The resulting so-called negative multiplier suggests that the percentage-based amount is reasonable and fair based on the time and effort expended by class counsel. Although multipliers are frequently greater than one and often on the order of two to four, see, for example, Van Vranken v Atlantic Richfield Corp, 901 F Supp 294, 298 (ND Cal 1995) ("Multipliers in the 3-4 range are common in lodestar awards for lengthy and complex class action litigation."); Behrens v Wometco Enters, Inc, 118 FRD 534, 549 (SD Fla 1988) ("[The] range of lodestar multiples in large complicated class actions [varies from] a low of 2.26 to a high of 4.5."), in a case such as this, a negative multiplier is appropriate to ensure at least some recovery for the class and to reflect the very modest achievement of counsel in this case. Even if the court accepted the unadjusted lodestar from plaintiffs' counsel ($922,884.75), the correlating multiplier of 0.74 would still reflect a negative multiplier, further suggesting that the requested percentage based fee is fair and reasonable.
Although the court's role is to make a fair ex post allocation of the settlement fund, the essential inquiry is: What kind of fee and expenses arrangement would the class have struck with class counsel ex ante? See, for example, In re Wells Fargo Sec Litig, 156 FRD 223, 225-26 (ND Cal 1994). Such an approach admittedly has its difficulties in a class action "because no member of the class has a sufficient stake to drive a hard — or any — bargain with the lawyer." Matter of Continental Ill Sec Litig, 962 F2d 566, 572 (7th Cir 1992); see also Goldberger v Integrated Resources, Inc, 209 F3d 43, 53 (2d Cir 2000). Nevertheless, "some guides are available," including "data from large common-pool cases where fees were privately negotiated; and information on class-counsel auctions, where judges have entertained bids from different attorneys seeking the right to represent a class." In re Synthroid Marketing Litig, 264 F3d 712, 719 (7th Cir 2001).
Here, plaintiffs' counsel asserts that an ex ante negotiated agreement was reached with plaintiffs, which would entitle plaintiffs' counsel to a fee award of 25% plus expenses. Doc #177 at 10 n8. If such an agreement does exist the court will give it weight assuming the agreement was the product of procedurally sound, arms-length negotiations. While the negotiated fee is instructive, the parties failed to provide any information to ensure the fee agreement's reasonableness, fairness and accuracy. Because of the lack of evidentiary support and the possibility that the 25% figure may have been boilerplate rather than an economically sound, reasonably calculated fee, the court is reluctant to give the agreement too much weight. On the other hand, if the fee agreement resulted from arms-length bargaining, the court might well find this persuasive in determining the reasonableness of the attorneys' fees, because of the reduction from 25% to 17.44%. But no such documentation was provided here, and the resolution of this litigation should not be further delayed.
The court GRANTS the motion for final approval of the settlement. The court GRANTS approval of the plan of allocation and GRANTS an award of costs and attorneys' fees.
IT IS SO ORDERED.