I imagine that, for a federal judge, getting reversed is not pleasant, even though it’s part of the job. Well, pity poor Judge Larimer of the Western District of New York, who has now been reversed three times in the same case – twice by the Second Circuit and once by the Supreme Court.

The name of this case will sound familiar to ERISA practitioners for one of two reasons. The Supreme Court ruled in 2010 that a plan administrator does not lose its discretionary authority to interpret plan terms merely because its initial interpretation was unreasonable; what it referred to as a “one strike and you’re out” rule. Conkright v. Frommert, 559 U.S. 506, 130 S.Ct. 1640, 1651–52, 176 L.Ed.2d 469 (2010) (“The question here is whether a single honest mistake in plan interpretation justifies stripping the administrator of that deference for subsequent related interpretations of the plan. We hold that it does not.”). Alternately, practitioners who are actuarial wonks might relish the debate about how to address lump-sum distributions in calculating future benefits under a floor-offset retirement plan.

Frommert is still around, and shows no sign of ending anytime soon. This is thanks to a recent Second Circuit decision that it refers to as Frommert III (but may more accurately be called Frommert VII, counting the District Court and Supreme Court decisions).

Frommert involved a dispute between Xerox and employees who had left the company, received a lump-sum distribution of retirement benefits, and were subsequently rehired. The issue in the case is how the lump-sum distribution affects future benefits under the plan. The plan is called a “floor-offset plan,” which “uses a defined-benefit structure (with pension payments linked to years of work and high salary) to buffer the uncertainty of a defined-contribution system (where pension payments depend on the performance of investments in each employee’s account).” In essence, the Xerox plan has a standard defined contribution component, with contributions earning a market rate of interest, and also has a “Retirement Income Guarantee Plan formula,” which is a calculation based on years of service and salary. The RIGP does not vary with market interest rates, and provides a floor: the benefit payable to any employee cannot be lower than the RIGP amount, but might be higher, depending on how well the defined contribution component performs. Simple, right? Well, it gets better.

Xerox’s plan used something called a “phantom account” to determine how the lump-sum distribution to the rehired employees affected their future retirement benefits. This assumed that the prior distribution had been invested and earned the same interest as the funds remaining in the plan; this amount was used to reduce the RIGP for future benefits. The employees sued under 29 U.S.C. § 1132(a)(1)(B), arguing that the phantom account calculation short-changed them, and that Xerox violated ERISA’s notice requirements in not adequately disclosing the phantom account offset. In 2004, Judge Larimer granted summary judgment  for Xerox. Frommert v. Conkright, 328 F.Supp.2d 420 (W.D.N.Y. 2004). The Second Circuit reversed, in what it refers to as Fommert I, Frommert v. Conkright, 433 F.3d 254 (2d Cir.2006). Frommert I held that the Xerox plan did not contain the phantom account offset at its inception, and that the plan was never validly amended; therefore, Xerox’s use of that method to offset pension benefits was unreasonable. The Second Circuit instructed Judge Larimer to calculate retirement benefits based on “the appropriate pre-amendment calculation[.]”

On remand, Judge Larimer used as an offset the nominal value of the lump-sum distribution; in doing so, he refused to give deference to alternate methods of calculating the offset proposed by Xerox. Frommert v. Conkright, 472 F.Supp.2d 452 (W.D.N.Y.2007).The Second Circuit affirmed in Frommert II, Frommert v. Conkright, 535 F.3d 111 (2d Cir.2008). But the Supreme Court reversed, rejecting the “one strike and you’re out” rule.

On remand (again), Judge Larimer deferred to Xerox’s new offset calculation (which reduced the RIGP formula by an amount determined by assuming that the prior distribution had been converted to an annuity), found it reasonable, and also determined that the offset was not invalidated by lack of adequate notice to plan participants.  Frommert v. Conkright, 825 F.Supp.2d 433 (W.D.N.Y. 2011). That set the stage to come back to the Second Circuit, which just reversed Judge Larimer again, and remanded, again, in Frommert III. Frommert v. Conkright, — F.3d –, 2013 WL 6726965 (2d Cir. Dec. 23, 2013).

The Second Circuit found that Xerox’s annuity formula was unreasonable, apparently because the RIGP calculation (the guaranteed minimum benefit) would always be lower for a rehired employee than for a comparable employee who had not left and had not received a lump-sum payment. Of course, it would seem logical that an employee who already received a pension payment should receive less in the future than an employee who had received nothing. The Second Circuit did acknowledge that an ERISA plan could be written “to change the risk borne by rehired employees or reduce such employees’ benefits in a manner that treats them worse than newly hired employees[.]” But, the court held, the Xerox plan was not written to achieve that result. Which is sort-of expected, given that the court had previously found that the provision written to address rehired employees (the phantom account provision) was unenforceable.

Even if the annuity offset calculation had been reasonable, the court held that Xerox had not given adequate notice to participants. First, though the plan said that a participant’s lump sum distribution “shall” offset future benefits, the SPD said that it “may” reduce benefits. The Court found this improper, because a participant would have no reason to conclude that a lump sum distribution would reduce his RIGP calculated minimum benefit. One struggles to understand why it is would be unusual for an employee to understand that an early payment of retirement benefits would decrease future benefits. In any event, this whole debate is only about the reduction in the RIGP benefit, which is the benefit that would be payable only in the event that the defined contribution component performed poorly. Therefore, a reduction in the RIGP benefit might, or might not, affect the benefit payable, which would seem to confirm that the SPD was correct. But the Second Circuit did not see it that way.

The Second Circuit also held that Xerox was not saved by a previous Second Circuit decision declining to impose a “blanket rule” that an SPD must describe the method of calculating an actuarial reduction or provide clarifying examples. Frommert III held that it was not imposing a blanket rule here; it was merely saying that here, the SPD did need to describe the method of calculating the actuarial reduction or provide clarifying examples. But other plans might not need to do so. How can plan administrators tell if a description or examples are necessary? The court didn’t say.

The Second Circuit also rejected Xerox’s argument that requiring detailed calculations or examples would make SPDs unreadable, holding that Xerox had no problem describing the phantom-account offset calculation in an SPD. No matter that the phantom-account offset was improperly adopted and never effective; the fact that Xerox could describe it meant that it should be able to describe any other offset calculation.

The Second Circuit provided some guidance to Judge Larimer about what he should do on remand (again). He “should first consider equitable remedies[,]” evaluating what remedy is appropriate, and whether the employees have established the requisite level of harm as a result of Xerox’s failure to give notice of the offset. This is essentially the Cigna v. Amara analysis, where the level of harm required varies with the specific equitable remedy being imposed. If Judge Larimer finds that an equitable remedy is not available, he “should enforce a reasonable interpretation of the Plan, without again considering the issue of notice” but giving “the appropriate deference to the interpretation of the Plan Administrator[.]”Apparently the Second Circuit has now determined that even two honest mistakes do not justify stripping deference.

As a final side issue, the Second Circuit upheld Judge Larimer’s determination that the plaintiffs were not entitled to additional conflict discovery after the Supreme Court’s remand. The first two determinations in the District Court had been made before Metropolitan Life Ins. Co. v. Glenn, and the plaintiffs argued that Glenn justified discovery that they had not previously sought. Judge Larimer denied that request, and the Second Circuit held that the role of a conflict “had been established long before Glenn[;]” therefore, there was no basis to re-open discovery.