Late actuarial tweaks just turned a $1.8M bill into $6.2M – and the high court is fine with it
Employers just lost a unanimous Supreme Court fight over pension exit bills, and the math can swing by millions.
A unanimous US Supreme Court ruled on May 21, 2026 that multiemployer pension plans can use actuarial assumptions adopted after the official measurement date to calculate what a withdrawing employer owes. The decision was written by Justice Jackson. It is a clean win for plans and a hard loss for employers trying to exit them.
For HR and benefits leaders at companies that participate in union pension plans, the message is blunt. Withdrawal liability, the share of a plan's unfunded benefits an employer must cover on the way out, can grow significantly between the date a company decides to leave and the date the bill lands. The court has now confirmed that federal law allows it.
The case involved four employers who withdrew from the IAM National Pension Fund, an underfunded plan covering workers represented by the International Association of Machinists and Aerospace Workers, between April and December 2018.
In November 2017, the Fund's actuary, Cheiron, valued the plan's unfunded vested benefits at close to $500 million using a 7.50% discount rate. Two months later, in January 2018, the Fund and Cheiron settled on a lower rate of 6.50%. A lower discount rate makes future benefit obligations look bigger in today's dollars. When Cheiron published the 2017 plan year valuation in April 2019 using the new rate, unfunded vested benefits had jumped to just over $3 billion, about six times the previous year's figure.
The dollar impact on individual employers was steep. M&K Employee Solutions was assessed withdrawal liability of around $6.2 million. Under the old assumptions, it would have owed roughly $1.8 million.
The employers challenged their bills in arbitration and won. Arbitrators ruled the Fund should have used the assumptions in effect on the measurement date of December 31, 2017, meaning the 7.50% rate. Federal district courts disagreed with the arbitrators. The D.C. Circuit affirmed the district courts. The Supreme Court took the case to resolve a split with the Second Circuit, which had ruled the other way in a 2020 decision involving the National Retirement Fund.
The ruling turns on two small words in ERISA: as of. Section 1391 says withdrawal liability must be calculated based on a plan's unfunded vested benefits as of the measurement date. The employers argued that phrase locks in the actuarial assumptions on that date. Justice Jackson rejected that reading. Hard data about the plan, such as the number of beneficiaries and the value of the assets, has to be fixed on the measurement date. Actuarial assumptions, the court said, are not facts. They are predictive judgments and tools used to do the calculation, and the calculation itself can happen later.
The court also looked at ERISA section 1393, which governs how assumptions are used in this context. That section requires assumptions to be reasonable and to reflect the actuary's best estimate of what the plan will experience. It does not set a deadline. Congress wrote a deadline into another section of the statute, the court noted, but chose not to write one here. The court treated the silence as intentional.
The employers warned that the ruling invites manipulation, with plans picking assumptions after the fact to drive up bills. The court was unmoved. It pointed out that employers can still challenge unreasonable assumptions in arbitration, and that policy worries do not override the best reading of the statute.
For HR teams managing multiemployer plan participation, the ruling resolves a long-running circuit split and reinforces a structural reality. The cost of leaving these plans is not locked in on any date the employer controls. Plans and their actuaries keep meaningful flexibility on when to update key assumptions, and that flexibility can move withdrawal liability by millions of dollars.
The decision is final.
