After the Pandemic Signal: Resetting Mortality Assumptions in 2026 Life Pricing

The Society of Actuaries released its 2025 Individual Life Insurance Mortality Improvement Scale for use with AG-38 and VM-20 in September, and the NAIC's Life Actuarial Task Force adopted it, handing life actuaries the improvement factors they will project base mortality with for year-end reserving and 2026 product pricing. The backdrop is a mortality picture that has mostly, but not entirely, normalized: excess mortality has fallen steadily from its pandemic peak, yet emerging data through June 2025 still shows a small residual of excess deaths in the population aged 65 and older. A signal that is fading but has not cleared is the hard case to price. An actuary setting the mortality assumption for a twenty-year level term contract cannot wait for the data to fully settle, and cannot treat a transient pandemic distortion as a permanent change in the underlying trend.

Sept 2025

Release of the SOA Individual Life Mortality Improvement Scale for AG-38 and VM-20, since adopted by the NAIC LATF.

+10%

Year-over-year growth in 2025 US individual life new annualized premium, a record $17.5 billion.

65+

The age band where a small residual of pandemic excess mortality still showed in data through June 2025.

What the 2025 Improvement Scale Changes

The SOA's individual life mortality improvement scale is not a mortality table; it is the set of annual factors that project a base table forward in time. The 2025 recommendation, prepared by the Mortality Improvement Life Working Group, supplies improvement factors that vary by gender and attained age and are designed to be applied to the 2008 or 2015 Valuation Basic Table for AG-38 purposes. Because the NAIC's Life Actuarial Task Force has adopted it, the scale is not advisory background; it is the input a company uses to roll its base mortality assumption to the valuation date and to project it across the lifetime of a policy in the principle-based reserve calculation.

What makes the annual refresh consequential rather than routine is that each new vintage absorbs another year of insured mortality experience, and the years now entering the data are the ones distorted by the pandemic and its aftermath. An improvement scale built to describe a steady, gradual decline in mortality has to digest a period in which mortality first spiked and then partially receded, and the factor a company carries determines whether it assumes the recent improvement continues, stalls, or reverses. Two reserving teams using the same base table and different defensible improvement assumptions can produce materially different reserves and prices, which is why the scale, and the judgment layered on top of it, sits at the center of this year's assumption setting.

How Mortality Flows Into VM-20

VM-20, the principle-based reserving framework for individual life, does not take a single prescribed mortality rate. It builds the company's prudent estimate from its own experience where that experience is credible, grades toward an industry table where it is not, and then applies the mortality improvement scale to project the assumption forward. The base table is typically the 2015 Valuation Basic Table or the 2017 Commissioners Standard Ordinary table blended with company experience, and the improvement scale is the dynamic overlay that ages it. The deterministic reserve and the stochastic reserve both depend on this assumption, so a change in the improvement factor moves the reserve directly, and through the reserve it moves the capital strain and the price the product has to carry to hit a return target.

The pricing chain runs the same way. A term or universal life price is, at its core, the present value of projected death benefits net of premiums and expenses, discounted and loaded for profit and risk. Mortality is the largest single driver of that projection, and the improvement assumption compounds over the full policy term, so a small annual difference in assumed improvement becomes a large difference in projected claims twenty or thirty years out. For a competitive level term product priced on thin margins, the improvement assumption is not a rounding detail; it can be the difference between a price that wins business profitably and one that wins business at a loss. The same assumption that sets the reserve sets the floor under the price, which is why valuation and pricing actuaries have to reconcile to one mortality view rather than carry two.

Separating Durable Change From Pandemic Noise

The technical difficulty this year is a credibility problem dressed as a data problem. The post-pandemic experience is both short and unusual, so the actuary has very few years of clean signal and those years contain a structural shock whose permanence is unknown. Some of the elevated mortality of 2020 through 2022 was a one-time cull of frail lives, which would imply lower mortality afterward; some reflected deferred care, undiagnosed conditions, and long-term effects that could persist; and the data through June 2025 showing a small residual of excess mortality in the 65 and older band is consistent with more than one of these stories. The actuary cannot resolve which dominates from the experience alone.

That is why the improvement assumption cannot be set mechanically by blending recent experience into the base. A blend treats the pandemic years as random fluctuation around a stable mean, but the question is precisely whether the mean shifted. The defensible approach is to decide, as an explicit judgment, how much of the recent deviation to treat as durable and how much as transient, to document that decision, and to build margins that acknowledge the uncertainty rather than pretend it away. Insured lives are not the general population, and the selection effects of underwriting mean the insured improvement pattern can diverge from the population trend, which adds another layer the actuary has to reason through rather than read off a chart.

Two Improvement Scales, Two Purposes

It is easy to conflate the SOA's individual life improvement scale with the better-known general-population improvement scales, and the distinction matters for which actuary uses which. The Retirement Plans Experience Committee produces the MP series of mortality improvement scales used to value pension liabilities, built from national population data, and in October 2025 it again decided not to release a new MP scale, citing insufficient post-pandemic data to support an update, leaving the prior scale in place. The individual life scale is a separate product, built from insured experience for the specific purpose of AG-38 and VM-20, and it did update for 2025.

The divergence is instructive. A pension actuary valuing a retirement obligation and a life actuary pricing a term policy are both projecting future mortality improvement, but they draw on different data, serve different regulatory frameworks, and reached different conclusions about whether the post-pandemic record yet supports a new scale. A life actuary who reaches for the pension MP scale, or assumes the two move together, imports a general-population trend into an insured-lives calculation where selection and the purpose of the measurement both argue against it. The two scales exist because the two problems are not the same, and 2026 is a year when treating them as interchangeable would produce a measurably wrong assumption.

Pricing Sensitivity in a Competitive Term and UL Market

The assumption reset lands in a market that is selling well. LIMRA reported 2025 US individual life new annualized premium up 10 percent to a record $17.5 billion, with policy count up 7 percent and records set in four of the past five years. For 2026 LIMRA forecasts more measured growth: continued double-digit gains in indexed universal life, roughly flat term sales, and moderating variable universal life as equity volatility weighs on market-linked lines. A growing book magnifies the consequence of the mortality assumption, because every new cohort is priced on the current view and locks that view in for decades of projected claims.

In the most price-sensitive segments the effect is sharpest. Level term and accumulation-focused universal life compete on small margins where the mortality and improvement assumptions, the lapse assumption, and the discount rate together determine whether a given price clears the return hurdle. An aggressive improvement assumption lowers projected claims and supports a lower price, which wins business, but if the post-pandemic improvement does not materialize the cohort is underpriced for its entire duration with no opportunity to reprice in force. Reinsurance and underwriting class add further leverage, because a reinsurer's own mortality and improvement view sets the cost of ceding risk, and a mismatch between the direct writer's assumption and the reinsurer's shows up immediately in the economics of the treaty.

Setting the Assumption You Can Defend

The mortality reset of 2026 is less about a single new number than about the discipline of setting an assumption under genuine uncertainty and documenting why. The SOA's 2025 scale gives life actuaries an updated, regulator-adopted projection basis, but the scale does not decide how much of the lingering 65-and-older excess mortality to treat as permanent, how insured selection modifies the population trend, or how much margin the uncertainty warrants. Those are the judgments that separate a reserve and a price that hold up from ones that look competitive until the experience arrives. The actuary who states the assumption explicitly, reconciles pricing and valuation to one mortality view, keeps the insured scale distinct from the pension scale, and carries margin sized to what the data cannot yet confirm will be the one not revising the assumption under pressure two valuation cycles from now.