Definition
The Survivor Selection Effect is the compositional mechanism by which a selective mortality event — one that kills disproportionately from already-high-mortality subgroups — produces a surviving population with lower expected future mortality than the pre-event population. It is a second-order effect: the first-order impact of a mass mortality event is to increase deaths; the survivor selection effect is a partial, delayed offset arising from the fact that the newly dead were disproportionately drawn from the frailest members of the pre-event cohort.
Key Ideas
- Selection vs. randomness: A random mortality shock leaves the survivor population's health composition unchanged. A selective shock (targeting the frail) leaves survivors disproportionately healthy — shifting the distribution of health risk downward.
- Scale constraint: Even perfectly selective mortality produces only modest survivor selection effects if total deaths are small relative to population size. Wettstein et al. (2022): 300,000 COVID deaths / 78M Americans 60+ ≈ 0.4% of population → ~1 pp decline in 10-year mortality for the oldest groups.
- Age gradient of selection efficiency: Selection efficiency (actual decline / maximum possible decline) decreases with age. For 60-somethings, ~80% of the maximum is realized; for 90-somethings, only ~50%. At very old ages, nearly everyone has high mortality risk, so COVID cannot further discriminate between frail and robust individuals.
- Temporary compositional shift: Survivor selection produces a one-time shift in population composition. Unlike true improvements in biological mortality rates, it does not generate sustained mortality decline; as cohorts age, new entrants with average mortality risk dilute the effect.
How It Works
- A selective mortality event kills disproportionately from the high-mortality tail of the health distribution.
- Survivors are a biased sample: health-conditioned survival probabilities shift the surviving population toward lower-risk individuals.
- New life tables computed from the post-event population show lower mortality rates — not because anyone's underlying health improved, but because the composition of the population changed.
- The effect attenuates over subsequent years as new entrants to each age group have more average mortality risk.
Why It Matters
- Social Security (SS) finance: COVID's first-order effect on SS finances is positive (fewer surviving beneficiaries, lower payouts). The survivor selection effect adds a modest second-order improvement (survivors are slightly healthier than expected). Wettstein et al. conclude the selection effect is likely swamped by ongoing acute and long-COVID mortality increases.
- Actuarial pricing: If COVID was sufficiently selective, life insurers and annuity providers may need to revise survivor life tables. The magnitude suggests revisions are warranted for the oldest age groups.
- General epidemiological pattern: Any mass mortality event concentrated in frail subpopulations — pandemics, heat waves, famines — may produce short-run survivor selection effects. COVID-19 is an unusually well-documented natural experiment in this mechanism.
Open Questions
- How does long-COVID mortality risk offset the selection effect? If survivors of COVID infection face elevated long-run mortality, the compositional improvement is partially counteracted.
- How persistent is the effect? Does the one-time compositional shift dissipate within a few years as cohorts age?
- Do selection effects compound across pandemic waves, or does each wave select from an increasingly non-frail pool?
Related
Sources