It has been nearly seven years since we last wrote about “actuarial equivalence” litigation—a new flavor of ERISA class action lawsuits that emerged in 2018. (See Defined Benefit Plan Actuarial Equivalence Litigation – A Formidable Threat or An Unfounded Theory?) More than three dozen such lawsuits have been filed to date, but we still do not have definitive answers to the fundamental questions at the heart of this litigation: what does ERISA §205(d)(2)(A), which requires “actuarial equivalence” between single-life and joint and survivor annuity options,[1] actually mean—and what must plan sponsors do to comply? A spate of recent court decisions—mostly favorable to defendants—shows that the federal courts are still struggling with these questions, and ultimately they will likely need to be decided by the Supreme Court.
Most of the actuarial equivalence cases have been brought against defined benefit pension plans sponsored by large corporations and their fiduciaries: American Airlines, PepsiCo, Kellogg Company, Intel, U.S. Bancorp and FedEx are among the notable defendants. A few have been brought against university plans and multiemployer pension plans. Almost all of the lawsuits share a common allegation: that the actuarial assumptions—in particular, the mortality table used to estimate how long a participant is expected to live—used to calculate “joint and survivor” (and sometimes other) benefit options are outdated and unreasonable, resulting in reduced pension payments to retirees and their spouses.[2] These reduced payments, the lawsuits allege, do not satisfy ERISA’s “actuarial equivalence” requirement and result in an impermissible forfeiture of a participant’s accrued benefit, in violation of both Section 1055(d) as well as 29 U.S.C. 1053(a).
A little actuarial background may be helpful. A single-life annuity, or “SLA,” is typically a monthly benefit amount paid out solely over the life of the plan participant, while a joint and survivor annuity, or “JSA,” is paid out over two lifetimes—those of the participant and their spouse. Because of the likelihood that JSA payments will last longer than SLA payments (i.e., if the spouse survives the participant), plans typically pay smaller amounts to participants who elect JSAs, so that the two types of benefits are “actuarially equivalent.” To achieve this actuarial equivalence, plans employ either a set of assumptions—a specific mortality table along with an interest rate, which is necessary to calculate the present value of the expected stream of benefit payments—or a table of percentages (sometimes a single percentage) that in turn was derived from mortality and interest assumptions. So, for example, a participant who has earned an SLA benefit of $1,000 a month might have the option under the plan terms of choosing a 50% JSA that pays him $900 per month for life and his spouse $450 per month for her life if she survives him.[3] Under the plan’s assumptions, the JSA and SLA are said to be “actuarially equivalent.”
In fact, ERISA requires that the primary types of JSAs offered to married participants—known as the qualified joint and survivor annuity and qualified optional survivor annuity—be the “actuarial equivalent” of an SLA “for the life of the participant.”[4] But the statute provides no further elaboration on what “actuarial equivalent” means, or how it must be calculated, leaving plan sponsors to their own devices to choose the method by which actuarial equivalence is determined. Many pension plans still use mortality tables from the 1970s or 1980s, which may have been the most recently available mortality tables at the time the provisions were put in place, and ERISA contains no express requirement that the tables must be updated to reflect more current mortality. Using a more recent mortality table (with longer life expectancy) will generally result in higher benefits for most participants, as a lower reduction is then needed to cover the cost of the joint annuitant’s expected benefit. In the example above, the retiree’s $1,000 monthly benefit might only be reduced to $950 per month, and his spouse’s to $475, if a more recent mortality table is used.
All of that said, courts have now wrestled for eight years with the question of whether the ERISA provision requiring actuarial equivalence in this context includes an implicit requirement that the underlying actuarial assumptions be “reasonable.” The word does not appear in the relevant statutory provision. Nevertheless, a number of federal district courts have ruled that there is an implicit “reasonableness” requirement and allowed actuarial equivalence cases to proceed beyond the pleadings stage.[5] A number of other courts, however, have hewn more strictly to the law’s plain language and ruled that there is no such requirement, dismissing cases that relied on an allegation that a plan’s mortality assumptions were outdated and therefore “unreasonable.”[6]
This year has started with three rulings for defendants and one for plaintiffs, and one that had a little something for both sides. In April, Judge Edward Davila of the Northern District of California granted summary judgment for Intel Corporation, ruling that actuarial equivalence “merely reflects a mathematical principle, devoid of an implied ‘reasonableness’ requirement.”[7] The court noted that Congress used the word “reasonable” in many other sections of ERISA, and its absence in the actuarial equivalence provision at issue indicated that Congress did not intend to include it there.[8] A month earlier, Judge Cristian Stevens of the Eastern District of Missouri reached the same conclusion (on a motion to dismiss) and dismissed a case brought against the Olin Corporation.[9] And in January, Judge Nancy Brasel of the District of Minnesota ruled similarly in favor of U.S. Bancorp, granting summary judgment for the plan and its fiduciaries in a case challenging early retirement factors.[10]
One thing notable about these three decisions is that they were issued by judges with seemingly quite different philosophical bents, perhaps demonstrating some broadening appeal of the strict constructionist “let’s hew to the plain language of the statute” argument. Judge Davila was appointed by President Obama, while Judge Stevens and Judge Brasel were both appointed by President Trump—Stevens (a Federalist Society member) in 2025, and Brasel (a moderate put forward by Democratic Senator Amy Klobuchar) in 2018. Yet these three judges were able to find common ground in three distinct cases with a straightforward, literal analysis of the statute.
In yet another case out of Minnesota, Judge Jeffrey Bryan (appointed by President Biden in 2023) allowed an actuarial equivalence claim to proceed beyond the pleadings stage—without clearly deciding whether actuarial assumptions must be reasonable or continuously updated — but threw out other claims based on statute of limitations grounds.[11]
Plaintiffs are not without their own recent victory—at least for now. In March, a divided panel of the Sixth Circuit Court of Appeals reversed two district court rulings that had dismissed cases filed against the Kellogg Co. and FedEx Corporation, holding that actuarial equivalence does in fact require up-to-date, professionally reasonable actuarial assumptions.[12] Majority opinion author Judge Jane Stranch echoed similar rulings and noted that without a “reasonableness” requirement, plan sponsors would be free to calculate JSAs using outrageously outdated mortality assumptions—say, from 500 years ago (or more).[13] Judge John Nalbandian, the third judge on the panel, wrote a strongly-worded dissent, taking the majority to task for relying on peripheral sources and straining to go beyond the simple language of the statute.[14] As for the 500-year-old mortality table bugaboo, Judge Nalbandian noted that even if ERISA’s actuarial equivalence provision contains no reasonableness requirement, the IRS has promulgated regulations that do impose such a requirement,[15] and therefore any plan that employed a mediaeval mortality table ran the risk of plan disqualification—a much more severe penalty than an ERISA lawsuit might impose.[16] Thus, he wrote, “the majority needn’t rush in to do the work of the legislature.”[17]
Whether Judge Stranch’s majority opinion or Judge Nalbandian’s dissent eventually carries the day remains to be seen. Shortly after the decision was issued, the defendants/appellees petitioned for a rehearing “en banc” (i.e., by the entire 6th Circuit bench), and the court thought enough of their arguments that they took the uncommon step of ordering the plaintiffs to respond to the petition, indicating that an en banc rehearing is at least somewhat likely.
Meanwhile, down in Georgia, the parties in Drummond v. The Southern Company await a decision from the 11th Circuit, which heard oral arguments on the actuarial equivalence issues in September 2025 (after a ruling for defendants in the district court). Parties in other cases are awaiting the 11th Circuit ruling before proceeding further. Roughly one-third of the cases filed have been settled, with no rulings yet in any case on what the bounds of reasonableness actually are (assuming there is even such a requirement).
For those defendants that have chosen not to settle and instead fight on, there is still much uncertainty. That uncertainty will likely remain until a clear pattern emerges from the circuit courts or the issues reach the Supreme Court, leaving these defendants the difficult decision of settling or litigating on an unknown landscape. In the meantime, plan sponsors may want to consult with their legal and actuarial professionals and undertake an evaluation of their plans’ actuarial equivalence provisions and the litigation risk they may pose—that would seem to be both a prudent and reasonable step.
[1] 29 U.S.C. 1055(d)(1)(b).
[2] Some of the actuarial equivalence lawsuits challenge the assumptions used to calculate other forms of benefits, such as early or late retirements, but the argument is essentially the same—that the assumptions, in particular the mortality table, are outdated and unreasonable.
[3] In this example, the “factor” was .9. Typically, table factors vary with the ages of the participant and their spouse.
[4] See 29 U.S.C. 1055(d).
[5] E.g., Belknap v. Partners Healthcare Sys., Inc., 588 F. Supp. 3d 161 (D. Mass 2022); Drummond v. S. Co. Servcs., Inc., No. 2:23-CV-00174-SCJ, 2024 WL 4005945 (N.D. Ga. July 30, 2024).
[6] E.g., Urlaub v. CITGO Petroleum Corp., No. 21 C 4133, 2022 WL 523129, at *6 (N.D. Ill. Feb. 22, 2022); Smith v. Rockwell Automation, Inc., 438 F. Supp. 3d 912, 924 (E.D. Wis. 2020).
[7] Berkeley v. Intel, No. 5:23-cv-00343-EJD, 2026 WL 948725, at *4 (N.D. Ca., April 8, 2026).
[8] Id. at *6.
[9] Landel v. Olin Corp., No. 4:25-cv-00096-CMS, 2026 WL 785044 (E.D. Mo., Mar. 20, 2026).
[10] Adams v. U.S. Bancorp, No. 22-CV-509 (NEB/LIB), 2026 WL 151825 (D. Minn., Jan. 16, 2026). Plaintiffs have filed a notice of appeal in the Eighth Circuit.
[11] Bennett v. Ecolab, Inc., No. 24-CV-0546 (JMB/SGE), 2026 WL 810758 (D. Minn. Mar. 24, 2026)
[12] Reichert v. Kellogg Co., 170 F.4th 473, 480-84 (6th Cir. 2026).
[13] Id. at 485. Here, too, seemingly strange bedfellows joined forces, this time in reading “reasonableness” into the statute. Judge Stranch was appointed by President Obama, while Judge Bush, who joined her opinion, is another Federalist Society member appointed by President Trump.
[14] Id. at 488 (Nalbandian J., dissenting)
[15] See Treasury Reg. 1.401(a)-11(b)(2).
[16] Id. at *496. Indeed, it is likely that every pension plan that has been sued in these actuarial equivalence cases was reviewed by the IRS for compliance with the Code’s qualification requirements in conjunction with determination letter applications (and perhaps on other occasions as well), and yet allegations have not emerged that the IRS ever found their actuarial equivalence assumptions to be defective.
[17] Id.
