The SOA's 2026 Mortality Improvement Model extends historical data through 2023, adding four years of post-pandemic experience to the underlying datasets, and leaves the MP-2021 projection scale unchanged. The Retirement Plan Experience Committee concluded that "there is insufficient post-pandemic data to support the development of a new MP scale."

That is a defensible call. It is also load-bearing, because pension risk transfer is clearing at 99.7% of ABO and VM-20 prescribed mortality runs off the same scale.

Key Takeaways

  • Four years of new data, zero change to the projection scale. MIM 2026 carries mortality through December 2023 but the prospective framework still excludes everything from 2020 onward.
  • MP-2021 has not been revised since October 2021, so every VM-20 valuation and PRT quote written today is anchored to a pre-pandemic calibration that is now four years older.
  • 99.7% of ABO was competitive PRT cost in May 2026, down from 100.9% in March, which leaves the longevity margin close to nothing.
  • Each additional year of life expectancy at 65 raises pension liabilities by 4% to 5%, so at that price point the improvement scale is the pricing residual rather than one assumption among many.
  • COVID-19 deaths fell 73.2% to 49,932 in 2023 from 186,552, and life expectancy recovered to 78.4 years, but excess deaths of roughly 705,331 remained well above the pre-pandemic baseline.

What MIM 2026 Changed and What It Held

Three changes are material. The model now runs through December 2023, covering the acute peak in 2020 and 2021, the partial normalization in 2022 and the further recovery in 2023. The National Center for Health Statistics dataset for 2011 through 2019 has been revised, bringing improvement estimates closer to Social Security Administration experience and narrowing a discrepancy that had complicated interpretation. The retrospective workbook is now the Mortality Trends Explorer, with better graphing and downloadable Excel support.

None of that touches MP-2021, the scale that governs prospective improvement in pension funding, annuity pricing and principle-based reserves. The prospective side of MIM 2026 deliberately excludes data from 2020 onward when estimating trend. Practitioners can study 2020 through 2023 in the Explorer, but those observations land in the retrospective workbook.

The boundary is stated plainly: the post-pandemic pattern is observable in the model and does not yet revise what actuaries project forward.

Reading the extension against the record shows why. US life expectancy fell to 76.1 years in 2021, the lowest since 1996, then recovered to 77.5 in 2022 and 78.4 in 2023. COVID-19 deaths fell 73.2%, from 186,552 in 2022 to 49,932 in 2023. Excess deaths peaked near 1,098,808 in 2021 and fell to roughly 705,331 by 2023, a 36% reduction in a year and still above baseline.

At 99.7% of ABO the Scale Is the Margin

Competitive PRT cost fell from 100.9% of ABO in March 2026 to 100.1% in April and 99.7% in May. The duration-7 annuity purchase rate for May was 4.94% and duration-15 was 5.02%, among the highest since June 2025.

At 99.7% the price sits essentially at the accounting liability. ABO assumes no future salary growth and discounts at a corporate bond rate; it carries none of the longevity loading a carrier builds into pricing. The spread between a carrier's pricing mortality and the ABO basis is being competed away, which changes what the improvement assumption is doing in the quote.

The sensitivities are the point. October Three puts each additional year of unanticipated life expectancy at 65, roughly a 1% increase in improvement rates, at 4% to 5% on the liability. Using portfolio-specific mortality rather than generic tables moves present value by 3% to 5% either way, comparable to 30 to 40 basis points of annual asset return. A carrier pricing at 99.7% with MP-2021 embedded and then seeing 25 basis points of additional longevity has no other margin to absorb it.

The same scale sits inside VM-20, where prescribed mortality is the 2015 Valuation Basic Table projected with the individual life improvement scale the NAIC's Life Actuarial Task Force adopts annually, itself built on the MP-2021 framework. The valuation standard requires no projection of improvement beyond the valuation date, and that absence functions as an implicit margin in the Deterministic Reserve. When the scale eventually revises, that margin compresses or widens with it, which matters most on universal life with secondary guarantees, where the liability tail runs 30 to 50 years.

The 2025 PRT market closed at $49 billion across 700 transactions, a 6% decline on 2024, with 22 carriers offering group annuity contracts. For every one of those clearing at or under 100% of ABO, the improvement scale is the residual.

The Signal RPEC Is Waiting For Cuts Both Ways

RPEC has published no threshold for what data would be sufficient. Its 2025 update said the worst mortality effects of the pandemic have subsided while post-pandemic data remains insufficient, and MIM 2026 carries the same language forward. Read together, the committee appears to want both the acute effects gone and a stable stretch long enough to separate rebound from trend.

The 2024 and 2025 experience is the next input. CDC preliminary 2024 estimates should arrive in late 2026 and SSA will fold that into its intermediate-cost projections, which puts a new scale in a 2027 or 2028 cycle at the earliest.

What makes the wait genuinely hard is that the evidence points both directions. One reading argues for lower long-term improvement: US excess deaths stayed elevated after the acute phase, with more than 1.5 million deaths across 2022 and 2023 that peer-country rates would have avoided, and the socioeconomic gradients in pandemic mortality may have shifted who survives into older ages. The other argues for higher near-term improvement: the 2022 and 2023 recovery was fast, COVID-19 deaths now sit below pre-pandemic flu mortality, and a relatively healthy boomer cohort is aging into its seventies.

MP-2021 was calibrated to the pre-pandemic trend with modest long-term convergence, so it may prove too optimistic or too conservative depending on which dynamic dominates the 2024 and 2025 data. That is the part the deferral does not remove. An actuary using MP-2021 in a reserve or a PRT quote today is making the directional judgment RPEC declined to make: that the pre-pandemic trend is a reasonable proxy for the post-pandemic one.

Further Reading

Sources