Mortality Improvement Assumptions After Structural Shocks

Why Mortality Experience Keeps Missing the Table After a Major Shock

Every life and health reinsurance pricing model rests on an assumption that mortality keeps getting better each year at a predictable rate. That assumption was built during a run of relatively stable health and social conditions, and it works fine until something breaks the pattern it was calibrated on.

A pandemic, a public health crisis, or a sustained shift in chronic disease trends can all do that. When they do, actual experience starts drifting away from the mortality improvement assumption baked into the reserves.

The gap does not announce itself with a single bad quarter. It shows up slowly, as a persistent, direction-consistent miss between expected and actual deaths that keeps recurring long after the acute event has faded from headlines.

What Is a Mortality Improvement Assumption, And Why Does It Break After a Shock?

A mortality improvement assumption is the projected annual rate of decline in mortality, and it breaks when the health and social conditions it was calibrated on change.

Reinsurers do not price mortality risk off a single static table. They apply an improvement scale on top of a base table to project how mortality will keep falling over the life of a treaty, often for decades.

That scale is a statistical summary of past behavior, not a law of nature. It assumes the medical, behavioral, and social drivers that produced past improvement will keep producing similar improvement going forward.

A structural shock breaks that assumption at its root, because it changes the underlying population, not just a single year's death count. Comorbidity patterns shift, healthcare utilization changes, and behavioral risk factors move in ways the original scale never saw.

The table keeps producing numbers, but the numbers describe a population that no longer exists in quite the same form.

How Did the Pandemic Change the Baseline Mortality Trend?

The pandemic did not just add a temporary spike in deaths, it appears to have altered the underlying trend line for several years afterward.

Research from the Society of Actuaries Research Institute's Retirement Plans Experience Committee (RPEC) shows this directly. As of its 2025 mortality improvement update, RPEC stated that "there is not yet sufficient post-pandemic data upon which to develop an updated MP scale."

It also noted that "excess mortality rates have continued to decline since the peak of the pandemic, and emerging data through June 2025 suggests that there is still a small amount of excess mortality for the 65+ population." Read carefully, that is a remarkable admission from the industry's own actuarial research body.

Years after the acute pandemic period, the data still was not clean enough to safely re-baseline the improvement scale. Reinsurers pricing new treaties during that window were, by definition, using a scale calibrated on a world that no longer fully applied.

Is Declining Excess Mortality the Same as Assumptions Being Safe Again?

No, declining excess mortality only signals the acute crisis has passed, not that the pre-shock improvement trend has resumed.

This distinction matters more than it sounds. A portfolio can show excess mortality trending toward zero while still running on an improvement trend that is fundamentally different from what the pricing assumed five years earlier.

Some cohorts recover to the old trend line, some settle onto a new, permanently different trajectory, and some overshoot in the other direction as deferred care catches up. Treating "the crisis is over" as equivalent to "the assumption is valid again" is exactly the mistake that keeps margins quietly eroding after headlines move on.

This is covered in more depth from the profitability angle in the margin cost of mortality improvement assumptions after structural shocks.

What Other Structural Shocks Beyond Pandemics Can Break the Assumption?

Chronic disease trend reversals, substance-related mortality waves, and major healthcare access shifts can each break the assumption in the same way a pandemic does.

Obesity-related metabolic disease, opioid and other overdose mortality, and periods of reduced preventive care access have all been documented as multi-year drags or boosts on mortality improvement independent of any single pandemic event. What makes these shocks structural rather than noise is duration and direction.

A single bad flu season reverts. A multi-year shift in chronic disease prevalence, in contrast, changes the population's baseline health profile in a way that persists across many future experience periods, which is exactly what an improvement scale is supposed to be measuring in the first place.

None of these shocks announce themselves with the same visibility a pandemic does, which is precisely why they are more dangerous for pricing. A pandemic triggers an industry-wide response, updated guidance, and heightened scrutiny almost immediately.

A slower-moving shift, like a multi-year rise in metabolic disease prevalence among a specific age band, can accumulate for several renewal cycles before anyone frames it as a structural change rather than routine year-to-year variation. Decision makers who wait for an obvious, headline-grabbing trigger before questioning their assumptions will systematically miss the slower shocks until the pricing gap is already large.

Why Does This Matter More for Reinsurers Than for Primary Insurers?

Reinsurers carry a concentrated, long-duration version of this exposure, since treaty terms often span decades and pool mortality risk across many cedants at once.

A primary insurer writing a single block of business is exposed to one company's underwriting mix, distribution channel, and geography. A reinsurer assuming risk across dozens of cedants inherits a blended version of whatever assumption drift is happening across each of those books simultaneously, which means a structural shock does not just create one pricing problem, it creates a portfolio-wide one that shows up unevenly across treaties depending on each cedant's specific population.

That concentration is also why reinsurers are frequently the first to see a structural shift clearly, ahead of any single primary insurer. Aggregating actual-to-expected data across multiple cedants gives a reinsurer's actuarial team a larger, faster-converging sample than any individual insurer has access to on its own book, which is a genuine analytical advantage decision makers should be using deliberately rather than treating as incidental.

For a reinsurance CUO or Chief Actuary, this means the mortality improvement conversation is not just about defending margin on existing treaties. It is also about deciding how quickly to feed portfolio-wide signal back into pricing guidance for new cedant relationships, since a reinsurer sitting on early evidence of a structural shift has a pricing advantage over competitors still relying on lagging, single-company data.

How Should a Reinsurer Detect That Its Assumptions Have Actually Drifted?

The clearest signal is a persistent, direction-consistent gap between actual and expected deaths across several consecutive experience periods, not a single deviation.

A single quarter's actual-to-expected ratio running above or below 100% is normal statistical noise. What signals real assumption drift is a run of periods pointing the same direction, especially when it holds across multiple cohorts, product lines, or geographies at once.

Granular mortality data and anti-selection monitoring give the resolution needed to separate real trend drift from routine volatility, a capability covered in individual life reinsurance's mortality data revolution.

How Does This Interact With Mortality-Linked Risk Transfer Structures?

Instruments like catastrophe mortality bonds and other index-linked risk transfer structures are especially sensitive to assumption drift, since their triggers and payouts are typically defined against a specific mortality index calibrated before the shock occurred.

Unlike a traditional treaty, where an actuarial team can revisit pricing and terms at each renewal, an outstanding mortality-linked bond or similar structure is locked into its original trigger definition for the full instrument term, which means a structural shock that moves the underlying mortality index can change the practical probability of a trigger event without any opportunity to renegotiate mid-term. This is discussed in more technical depth in the context of pandemic-linked structures in catastrophe mortality bonds and pandemic risk, but the core implication for a reinsurance decision maker is straightforward: any post-shock assumption review needs to extend beyond traditional treaty pricing to cover outstanding index-linked exposure as well, since that exposure cannot simply wait for the next renewal cycle to be reassessed.

A reinsurer holding both traditional treaty exposure and mortality-linked instrument exposure to the same underlying population is effectively running two different assumption-review clocks at once, one that resets at each renewal and one that does not reset until the instrument matures, and a mature assumption governance process needs to track both explicitly rather than defaulting to the renewal-driven cadence that governs the rest of the book.

What Belongs on a Reinsurer's Post-Shock Assumption Checklist?

A short, standing checklist keeps the response to a structural shock disciplined rather than reactive, and gives a reinsurer's leadership a shared reference point across pricing, reserving, and treaty negotiation teams.

At minimum, that checklist should confirm: whether cohort-level actual-to-expected data is being produced on a cadence tighter than the normal annual cycle; whether the deviation, if any, has been checked for correlation with a known or suspected structural driver rather than assumed to be noise; whether reserving assumptions have been checked against the same data used to flag a pricing concern; and whether the finding has been translated into a specific recommendation for new business terms, not left as an open research question.

Reinsurers that keep this checklist current across every treaty renewal cycle following a shock are the ones least likely to be surprised by a multi-year margin correction, because the checklist forces the same question to be asked and answered on a schedule the underlying data actually supports, rather than whenever it happens to surface informally in a pricing meeting.

The closing question every actuarial and reinsurance risk team should be asking after a structural shock is not whether the acute event has passed, but whether the improvement assumption underneath the pricing has actually caught up with the population it now needs to describe. Tables built on more stable conditions will keep producing confident-looking numbers long after the conditions themselves have changed.

The only way to catch the gap early is to keep measuring actual experience against assumption with enough granularity and enough patience to see the pattern before it compounds across years of in-force business.

Sources

• Society of Actuaries Research Institute: RPEC 2025 Mortality Improvement Update
• Munich Re: Group Life mortality trends, a post-pandemic perspective