In Conkright v. Frommert (4/21/10), a divided court (5 to 3, with Justice Sotomayor recusing herself) refused to narrow further the circumstances under which interpretations of plan terms are entitled to deferential review. The majority held that the mere fact that the plan administrator's initial position was "arbitrary or capricious" did not change the standard of review for its second attempt to find the right answer. The plaintiffs and the US Department of Labor (DOL) had argued for a different rule; namely, that deference was discretionary with the court if the administrator was wrong the first time.

The case arose from Xerox Corporation's (Xerox) floor/offset pension plan. This relatively rare design consists of a defined benefit pension plan operating in conjunction with an individual account plan. Participants accrue benefits under the pension plan and contributions under the individual account plan. Upon retirement, they receive their account balances in the latter plan. To that is added a pension benefit equal to the actuarial value of the accrued benefit minus the account balance. If the individual account is more valuable than the pension, there is no pension benefit. In effect, then, the pension plan provides a floor benefit to protect participants against investment losses, while leaving them free to reap any gains.

The Xerox plan had a drafting flaw: It did not describe fully what would happen if a participant separated from service, received a distribution, and then returned to employment. It was clear that a person's prior service would be taken into account in calculating his/her pension benefit and that he/she would begin accumulating a new individual account. What the plan did not cover was how the prior distribution would affect future benefits. Ignoring it would give reemployed participants a windfall, since they would receive a pension based on total service while the individual account offset would reflect only contributions during the most recent period of employment.

The plan administrator's solution was to add to the actual balance in the individual account a "phantom account" equal to the amount distributed in the past plus the earnings (or losses) that would have accrued if the distribution had remained in the plan. The plan document had once stated this rule explicitly, but it had been dropped from the text, perhaps inadvertently, by the time the plaintiffs retired. They brought suit, seeking to have the phantom accounts removed from their benefit calculations.

A district court applied the "Firestone rule": A plan administrator's (in this instance, the employer's) interpretation of the terms of the plan must be upheld unless it is "arbitrary or capricious." The court granted summary judgment for Xerox but was reversed by the 2nd Circuit, which held that absence from the plan document made the phantom account an arbitrary or capricious device. An offset was permissible, the court said, but not one so unintuitive and potentially dramatic in its impact.

The case went back to the district court, where Xerox offered a new method of determining the offset. In place of the phantom account, it would add interest at a market rate to the prior distribution. It is hard to imagine that any judge could regard that methodology as arbitrary, capricious or otherwise objectionable, but the court decided that it no longer had to be deferential. Instead, it fashioned its own remedy, under which the offset would be the amount of the prior distribution without interest. The appeals court subsequently upheld the decision, saying that the plan administrator was entitled to no deference and the lower court had not abused its discretion.

As a further complication, a different appeals court, hearing a case involving the same plan and the same issue but different plaintiffs (Miller v. Xerox Corporation Retirement Income Guarantee Plan (9th Cir., 2006), rejected the "phantom account" in favor of a methodology very like that proposed by Xerox the second time around in Conkright.

Thus the case reached the Supreme Court, where the Chief Justice for the majority and Justice Breyer for the dissenters batted back and forth cases and treatises regarding the powers of private trustees. The majority's conclusion was that there is no solid support in trust law (on which this aspect of ERISA is modeled) for a rule that removes deference after a single error and much reason for not "creating ad hoc exceptions to Firestone deference." The court found little merit in apprehension "that continued deference would encourage plan administrators to adopt unreasonable interpretations of plans in the first instance, as administrators would anticipate a second chance to interpret their plans if their first interpretations were rejected," an argument that assumes bad faith in the private sector. Rather, "efficiency, predictability, and uniformity" weighed in favor of giving the plan administrator another chance, with the caveat that repeated capriciousness would justify bringing deference to an end.

The disposition by the lower courts should have made this an easy decision. Whatever the intricacies of trust law, requiring the plan to ignore the time value of money was without basis. The dissent makes an attempt to defend it, but the case should have been remanded to fashion a more reasonable remedy.