PBGC's New Overpayment Rule Drops the Actuarial Reduction Method

PBGC's July 7, 2026 proposed rule would scrap the actuarially-equivalent formula that has governed benefit overpayment recoupment since 1985, replacing a reduction generally capped at 10% of a participant's monthly benefit with a flat 5% rate.

It would also end recoupment from surviving spouses and from most revised benefit determinations. For actuaries working trusteed plan valuations that is a liability-modeling change, not an administrative one.

Key Takeaways

• A flat 5% rate replaces a fraction equal to the net overpayment divided by the present value of the participant's remaining Title IV benefit, computed on PBGC's own interest and mortality assumptions.
• $12,000 of overpayment against a $2,000 monthly benefit computes under 2% for a participant in their mid-40s and 8% or 9% for one in their late 70s. Flattening to 5% inverts which cohort repays fastest.
• 926,000 retirees were already in pay status across PBGC's trusteed book in FY2025, the cohort for whom the current formula computes a rate above 5%.
• Two lives become one. Ending spousal recoupment collapses the collection horizon on a joint-and-survivor election to the participant's own remaining lifetime, compounding the flat-rate effect rather than offsetting it.
• September 4, 2026 is the comment deadline on Docket 2026-13639, filed against $6.4 billion of annual PBGC benefit payments.

What the Rule Replaces

Under 29 CFR 4022.82, PBGC reduces each future benefit payment by a fraction equal to the net overpayment divided by the present value of the participant's remaining benefit under Title IV of ERISA, with that present value computed on PBGC's own interest and mortality assumptions as of the plan's termination date.

The resulting percentage is capped at the greater of 10% per month or the portion of the monthly benefit exceeding the maximum guaranteeable amount under ERISA Section 4022(b)(3)(B), a ceiling PBGC set at $93,477 a year for a straight-life annuity starting at 65 in 2026, up from $89,181 in 2025.

That is a genuine actuarial equivalence calculation rather than an arbitrary percentage. A given dollar overpayment converts into a smaller monthly percentage for a participant with a long expected remaining payment stream, because the present-value denominator is larger, and into a larger percentage for a shorter stream.

PBGC's stated rationale is not that the formula is inaccurate. It is that participants cannot verify or follow it. A reduction percentage depending on an interest assumption and a mortality table the retiree has never seen is the complexity the agency wants out of a letter explaining why a monthly check got smaller. The proposal pairs the flat rate with a $250 de minimis waiver, the end of spousal recoupment, and the end of recoupment on revised determinations except where a qualified domestic relations order is submitted and qualified after PBGC has already issued a benefit determination. Comments on Docket 2026-13639 are due September 4, 2026.

A Flat Rate Redistributes the Recovery

Trading an age-and-mortality-sensitive formula for one flat percentage does not shrink total expected recoupment uniformly. It moves the recovery timeline across the participant population, and at the older end a share of expected recovery converts into an unrecovered residual.

Take two participants in the same trusteed plan, both corrected to a $2,000 monthly benefit, each carrying the same $12,000 net overpayment accumulated during the estimated-benefit period that follows a takeover. A participant in their mid-40s has several hundred expected future monthly payments ahead, so the present-value denominator is large and the fraction needed to amortize $12,000 might run well under 2%. A participant in their late 70s has a compressed remaining stream, and the same $12,000 against a much smaller present value computes to 8% or 9%, close to the current ceiling.

Flatten both to 5% and the arithmetic inverts. The younger participant repays faster than actuarial equivalence required, a timing shift at no real cost to the fund. The older participant repays more slowly, and part of that overpayment becomes an expected loss realized whenever the benefit stops before recoupment finishes.

The spousal provision compounds it rather than offsetting it. A correction against a participant electing a joint-and-100%-survivor form currently has two lives over which recovery can occur, and that combined stream is what makes the actuarial-equivalence percentage lower than a single-life calculation would produce. Remove spousal recoupment and the effective collection horizon collapses to the participant's own remaining lifetime, the shorter of the two lives in the pairing.

Neither effect is evenly spread across PBGC's book. The agency trusteed 31 single-employer plans in FY2025, covering roughly 20,000 current and future retirees, and now pays benefits tied to an estimated 1.4 million current and future retirees, disbursing more than $6.4 billion in FY2025 to nearly 926,000 retirees already in pay status. Plans arrive at PBGC after years of frozen accruals, acquisitions, formula amendments and recordkeeping turnover, so the retirees affected by the resulting errors skew older and already in pay, exactly the cohort for whom the current formula computes above 5%.

What It Does to PBGC's Own Trusteeship Valuation

The participant-communication rationale sits alongside a second effect, on the agency's own books rather than the retiree's.

When PBGC assumes trusteeship it performs an ERISA Section 4044 valuation to establish benefit liabilities and available assets, and anticipated overpayment corrections enter that reconciliation as a partial offset. A dollar PBGC expects to recoup from future payments is, in present-value terms, a dollar it does not carry as a realized cost of trusteeship. Lowering expected recovery through all three provisions at once raises the net liability booked at the point of trusteeship for every affected plan.

The revised-determination carve-out is the sharpest of the three, because it removes recovery by rule rather than by circumstance. A misapplied benefit formula, a missed offset, a data-entry error carried over from a predecessor recordkeeping system: each loses its recoupment mechanism entirely once the determination is revised, with only the post-determination QDRO case surviving. Benefit-determination errors surface routinely in plans that ran through multiple mergers, spinoffs and formula changes before termination, which describes a large share of the trusteed book.

The increment is small against $6.4 billion of annual payments, and the roughly 31 plans trusteed in FY2025 are a modest cohort. It accumulates the way any systematic reduction in expected recovery accumulates in an insurance reserve: across every affected cohort, for as long as the flat rate and the two carve-outs stand. The concentration is the problem rather than the size. The plans carrying the highest baseline rate of administrative error are the same legacy, formula-amended plans whose participants are furthest through their payment streams, so the rule reduces recovery most where errors are most likely to be found.

Further Reading on actuary.info

• PBGC's $62B Surplus Resets the Actuarial Case for Premium Reform
• When PRT Competition Stalls: Bid Economics for Disabled Lives and Complex Benefit Forms
• The DC Plan In-Plan Annuity Adoption Gap Under SECURE 2.0
• SOA's 2026 Mortality Improvement Model and What It Means for Pension De-Risking
• Retirement and Pension Actuarial Outlook 2026