Just before Christmas, the Amara case – already one of the most famous cases in American pension history – reached a new stage. The US Court of Appeals for the Second Circuit upheld the district court’s decision “reforming” the employer’s cash balance plan. Amara v. CIGNA Corporation, 2014 U.S. App. LEXIS 24265 (2d Cir. Dec. 23, 2014), affirming 925 F. Supp. 2d 242 (D. Conn. 2012). This is the second time that the Second Circuit has ruled on this case. The first time it issued a one paragraph opinion affirming two district court decisions for the plaintiffs, 348 Fed. Appx. 627 (2d Cir. 2006), which the US Supreme Court vacated and remanded in CIGNA Corporation v. Amara, 131 S. Ct. 2900 (2011). The latest opinion is notable for its holding that contract, as well as trust, principles could inform the availability of reformation, and for its conclusion that the elements of contract reformation had been established based on generalized circumstantial evidence of a unilateral mistake by the entire plaintiff class. Background In 1997 the employer revamped its pension and 401(k) plans. The pension plan, a traditional defined benefit plan, was frozen as of December 31, 1997, for participants who had been hired after 1988 or whose age and service totaled less than 45 years. (Nothing was changed for other participants.) The plan was then amended as of January 1, 1998, to provide the frozen participants with benefits under a cash balance formula. They were credited with initial cash balances equal to the present value of their frozen accrued benefits, calculated using a standard mortality table and an interest rate of 6.05% (5.05% for some older participants), which was not atypical in 1998. In subsequent years, the cash balance accounts would accrue additional “benefit credits” equal to a percentage of pay, varying with age and compensation. Account balances would also be credited with interest each quarter at a rate based on the yield on five-year US Treasury notes. At the same time, the employer amended its 401(k) plan to provide enhanced employer contributions for cash balance plan participants. The net effect was to decrease pension costs by an estimated $10 million a year and to increase 401(k) costs by the same amount, making the changes as a whole cost-neutral. To comply with ERISA’s prohibition against retroactively cutting accrued benefits, the amended pension plan provided that no participant could ever receive less than his accrued benefit as of the freeze date, December 31, 1997. Hence, he would not suffer if interest rate changes or other factors made the annuity equivalent of his cash balance account less than his frozen benefit. The technique of handling a transition between benefit formulas, known as “wear-away,” was commonplace in 1998 and still is in most situations, though Congress later eliminated its availability for conversions from traditional to cash balance formulas. An effect of wear-away is that, in some circumstances, participants may not accrue any additional benefits for some time after the change in formula. For many plan participants who severed from employment within a few years after the cash balance conversion, the frozen traditional pension benefit turned out to be greater than the cash balance benefit. There were three main reasons for this phenomenon:
The frozen pension formula included subsidized early retirement benefits, which were not reflected in the calculation of the opening balance, since there was no way to forecast participants’ retirement dates. Interest rates decreased after 1997, with the result that opening cash balances converted into smaller annuities than the originals from which they derived. If interest rates had gone up, the opposite would have occurred, and the wear-away period would have been shortened or nonexistent. A participant’s frozen pension benefit was forfeited if he died before his annuity starting date, except for a minimal qualified preretirement survivor annuity (available only to surviving spouses). In contrast, the cash balance benefit was payable upon death to any beneficiary. The opening cash balance account took this difference into account through a mortality discount. As a result, the value of the frozen benefit, calculated without taking mortality into account, was automatically greater than the opening account balance, thus generating a period of non-accrual for all participants, albeit a very short period in most instances. The employer explained its retirement plan changes in a series of communications to employees, including revised summary plan descriptions. These documents emphasized the positive side of the new benefit structure and were, in the eyes of the district court and the Second Circuit, misleading, particularly in their failure to make clear how wear-away operated. Also criticized by the courts was the employer’s representation that the changes would not decrease its total retirement plan expense, which was true for its plans as a whole, but not for the pension plan in isolation. To the Supreme Court In December 2001, a group of participants brought suit to overturn the cash balance conversion. The district court granted them class action status and, after a protracted proceeding, issued two decisions in their favor, the first finding liability on the employer’s part, 534 F. Supp. 2d 288 1011 (D. Conn., 2008), and the second ordering a remedy, 559 F. Supp. 2d 192 (D. Conn., 2008). The remedy devised by the court was not all that the plaintiffs had asked for. They wanted the cash balance formula nixed and the traditional formula reinstated for years after 1997. Instead, the district court ordered the plan to be “reformed” to accord with its interpretation of the summary plan description. Participants were to receive a “Part A” benefit – their accrued pensions as of December 31, 1997 – plus a “Part B” benefit – their cash balance credits earned in 1998 and later. This “A+B” formula would have no opening cash balance and no wear-away. The district court believed that it had the authority to remake the plan under section 502(a)(1)(B) of ERISA, which authorizes actions by a participant or beneficiary “to recover benefits due to him under the terms of the plan, to enforce his rights under the terms of the plan, or to clarify his right to future benefits under the terms of the plan.” The court also considered the possibility that reformation might be “appropriate equitable relief” that could be ordered under Yoga1smyjam!section 502(a)(3) of ERISA but expressed doubts about that approach, because the Supreme Court has interpreted section 502(a)(3) as authorizing only “those categories of relief that were typically available in equity (such as injunction, mandamus, and restitution, but not compensatory damages).” Mertens v. Hewitt Associates, 508 U.S. 248, 256 (1993) (emphasis in original). The district court was not confident that reforming the plan in the manner that it had directed would meet the Mertens standard. Nor did it feel a need to delve into the question, since, in its view, section 502(a)(1)(B) provided all the authority that was needed.
After a summary affirmance by the Second Circuit, Amara reached the Supreme Court, which held that the section 502(a)(1)(B) remedy was not available, because (contrary to a position widely adopted by the circuit courts) a summary plan description cannot alter the terms of an ERISA-covered plan. The court added, however, that relief under section 502(a)(3) was not necessarily foreclosed. Justice Breyer’s opinion for the court mused at some length on equity remedies that might be applicable, without relating any of them closely to the facts of the case. Specifically identified as possibly available were reformation, equitable estoppel and surcharge, all of which had been utilized by the equity courts. The Second Circuit’s decision was vacated and remanded for further proceedings. The court’s extended dictum on equitable remedies has proven far more influential than its actual holding and has formed the basis of a number of decisions that have applied section 502(a)(3) imaginatively to craft forms of relief that did not seem possible based on the court’s prior decisions. If Amara reaches the court again, we will learn to what extent the Justices truly meant to revolutionize the law of ERISA remedies. After the Supreme Court On remand, the district court granted the Amara plaintiffs the same “reformation” of the plan as before. Only the rationale changed. The court, in fact, weighed two rationales – contract reformation and surcharge – devoting roughly equal attention to each. The surcharge discussion, with its overtones (“In essence, the Court must imagine a counterfactual world. . . .,” 925 F. Supp. 2d at 258), is the more interesting, but the court ultimately did not rely on it, owing to uncertainty about whether, if it was employed, there were sufficient common questions of fact and law among the plaintiffs to sustain a class action. On appeal, the Second Circuit ignored that half of the district court opinion, and, for the sake of brevity, we do likewise. The appeal presented two main issues: whether the district court had properly applied the doctrine of contract reformation and whether, if it had, the case could continue as a class action in light of the Supreme Court’s decision in Wal-Mart Stores, Inc. v. Dukes, 131 S. Ct. 2541 (2011). That contract reformation was “typically available in equity” is indisputable. According to a standard treatise cited by both the district and the appellate courts, “[e]quity has jurisdiction to reform written instruments in  two well-defined cases,” one of which is – Where there has been a mistake of one party accompanied by fraud or other inequitable conduct of the remaining parties. In such cases the instrument may be made to conform to the agreement or transaction entered into according to the intention of the parties. [John Norton Pomeroy & Spencer W. Symons, 4 A Treatise on Equity Jurisprudence, §1376 (5th ed., 1941) (“Pomeroy”) (footnote omitted).] The courts concluded that the pension plan amendments exhibited “a mistake of one party” (the participants) “accompanied by fraud or other inequitable conduct of the remaining parties.” As the appellate court summarized: While no “single statement ... accurately define[s] the equitable conception of fraud,” it generally consists of “obtaining an undue advantage by means of some act or omission which is unconscientious or a violation of good faith.” 3 John N. Pomeroy, A Treatise on Equity Jurisprudence § 873 at 420-21 (5th ed. 1941). Here, defendants misrepresented the terms of CIGNA’s new pension plan and actively prevented employees from learning the truth about the plan. As Judge Kravitz put it in Amara I, “CIGNA employees suffered from the lack of accurate information in CIGNA’s disclosures, and CIGNA was aware of this fact.” Amara I, 534 F. Supp. 2d at 342; see also id. at 349 (deciding that CIGNA made “materially misleading statements” about wear away). CIGNA’s misbehavior was designed to “ease the transition to a less favorable retirement program.” Id. at 343. As a result, the district court did not err in finding that defendants obtained undue advantage through these actions by avoiding adverse employee reactions. [slip op. at 35-6] From there, the appellate panel drew a straight line to the conclusion that “reforming” the plan’s transition rule from wear-away to “A+B” was well within the discretion of a court of equity. Against that conclusion, the employer argued that, because ERISA derives from trust law, reformation could only be directed in accordance with the principles that the equity courts applied to trusts, where the objective of reformation was to reflect accurately the intent of the settlor. On that view, the plan amendments needed no reformation, as they embodied exactly what the settlor desired. (For an example of the application of trust reformation to ERISA plans, see Young v. Verizon’s Bell Atlantic Cash Balance Plan, 615 F.3d 808 (7th Cir.), cert. denied, 131 S. Ct. 2924 (2011), where the court invoked the doctrine to rectify a plan drafting error that, if left uncorrected, would have more than doubled some participants’ benefits.) The court rejected that argument, reasoning that CIGNA’s pension plan was part of a compensation package for its employees, and that trust reformation should be analyzed under contract principles “[w]here consideration is involved in the creation of the trust.” [citing Restatement (Third) of Trusts § 62 cmt. a (2003)] It did not look deeply, though, into the prerequisites for contract reformation where only one party is “mistaken.” The treatise quoted above notes some significant caveats: Another element of a fraudulent misrepresentation, without which there can be no remedy, legal or equitable, is, that it must be relied upon by the party to whom it is made, and must be an immediate cause of his conduct which alters his legal relations. Unless an untrue statement is believed and acted upon, it can occasion no legal injury. It is essential, therefore, that the party addressed should trust the representation, and be so thoroughly induced by it that, judging from the ordinary experience of mankind, in the absence of it he would not, in all reasonable probability, have entered into the contract or other transaction. [3 Pomeroy, §890 (footnotes omitted).] In this case, the pension plan already existed, and the employer had the right to amend it unilaterally, subject only to constraints imposed by law, such as the prohibition against retroactive reductions in accrued benefits. Since the participants’ consent wasn’t needed, they could not have “acted upon,” or been “induced” by, the employer’s allegedly misleading communications in a way that caused the adoption of the plan amendments. The most that can be said is that, by accentuating the positive, the employer ensured that its human resources department would field fewer questions, and hear fewer complaints, about the amended plan’s transition rules. The court acknowledged that the elements of contract reformation must be proven by “clear and convincing evidence,” but went on to hold that “generalized circumstantial evidence” can be used to establish a class-wide unilateral mistake where “defendants have made uniform misrepresentations about an agreement’s contents and have undertaken efforts to conceal its effect.” It then affirmed the district court’s finding of a class-wide unilateral mistake based on (a) evidence that the same misleading disclosures had been sent to the entire plaintiff class, (b) evidence that employees either read these disclosures looking for harmful changes or expected to hear through the “office grapevine” if the disclosures revealed any such harmful changes, (c) the employer’s failure to present evidence that any of its employees understood how they could be disadvantaged by the new plan, and (d) an inference drawn by the district court that “informed employees, aware that their pension benefits were less valuable, would have protested the change, requested a higher salary, filed a lawsuit, or left for another employer.” It is not at all clear that the traditional equity courts would have considered this evidence (or lack thereof) sufficient to establish a unilateral mistake on the part of the entire plaintiff class. The court’s conclusion that this evidence sufficed to establish a class-wide unilateral mistake also had the effect of undermining one of the employer’s principal arguments for decertifying the class – i.e., that reformation required individualized proof of a unilateral mistake that could not be shown on a class-wide basis. With that argument dead in the water, the court disposed of the remaining arguments for decertification in short order. It summarily rejected the employer’s argument that the “A+B” remedy would harm some class members, finding no evidence that any current employee would be harmed by the “A+B” formula. The court also rejected the employer’s argument that “A+B” would frustrate class members’ ability to take the “Part B” benefit as a lump sum. The pre-1998 plan generally didn’t offer lump sum distributions, while cash balance benefits, including the opening account balance, were available in that form. Under the “A+B” formula, a participant would be able to elect a lump sum only for the “B” portion of his benefit; the “A” portion would have to be taken as an annuity. The employer therefore argued that the “A+B” remedy deprived many participants of their preferred distribution option, so that all participants could not form a single class under Dukes. The court rejected that argument in a footnote, stating that “defendants have offered no reason, legal or practical, why any class member seeking a lump sum payment of the whole A+B benefit could not simply convert the Part A portion of the benefit into a lump sum through an ordinary commercial transaction on the open market.” Slip op. at 24, fn. 8. The court was evidently unaware that such a conversion is precluded by ERISA’s prohibition against the assignment or alienation of benefits, ERISA, §206(d); I.R.C., §401(a)(13). Participants cannot legally convert an annuity into a lump sum, unless the plan provides that option. The court mentioned only in passing another feature of “A+B” that would be detrimental to some class members. The plan’s cash balance benefits, unlike its traditional pension benefits, were not forfeited upon death. If a participant died before his pension commenced, the Part A portion was lost completely if he was unmarried, and in large part if he had a surviving spouse. To some participants, that feature would not matter much; to others, e. g., those who were in less than optimal health, it would be of critical importance. The employer further argued that the district court’s reformation remedy and award of monetary damages were improper under Rule 23(b)(2) of the Federal Rules of Civil Procedure, 1 under which the plaintiff class was certified. The court rejected these arguments, holding that the reformation remedy was appropriate under Rule 23(b)(2), since it was “properly understood as a declaration of the plaintiffs’ rights under the plan and an injunction ordering that the plan be reformed to reflect that declaration.” The monetary relief awarded to the plaintiff class was a “necessary consequence” of the injunction, and was therefore incidental to the “A+B” remedy. What Comes Next? Of the trio of equitable remedies discussed in the Supreme Court’s Amara opinion, reformation has received the least attention from the courts. The most intense focus has been on surcharge, which is often used by plaintiffs as a basis for seeking individualized tort-like damage awards under section 502(a)(3) of ERISA for any sort of fiduciary misconduct. While reformation is a narrower remedy, there are significant questions about how the reformation remedy “typically available in equity” should be applied in the ERISA context. If Amara reaches the Supreme Court again, we may find out how seriously 1 Rule 23(b)(2) authorizes class actions where “the party opposing the class has acted or refused to act on grounds that apply generally to the class, so that final injunctive or corresponding declaratory relief is appropriate respecting the class as a whole.”the current Court takes the word “typically”.