An underlying mortality assumption is an actuary’s estimate of current and future death rates for a defined population. It is typically expressed via a mortality table (life table) plus any adjustments and projection (improvement) scale. Insurers and pension plans use that assumption to price life-contingent benefits, set reserves and premiums, and measure long‑term liabilities.
Key takeaways
– The underlying mortality assumption translates observed mortality (from mortality tables) into the expected death rates used to value insurance contracts and pension obligations.
– Choosing and documenting the assumption is a core actuarial judgment that must follow regulatory guidance and professional standards.
– Small changes in the assumption can materially change liabilities: underestimating mortality (people live longer than assumed) increases pension liabilities and can reduce life-insurer payouts; overestimating mortality can cause underfunding or mispriced products.
– Ongoing monitoring, sensitivity testing, and governance are essential because mortality evolves (e.g., medical advances, pandemics, socioeconomic trends).
1. How actuaries think about mortality
– Base mortality table: A statistical table that gives death rates by age (often separately by sex and smoker status). Examples include industry tables published by the Society of Actuaries (SOA) and those adopted by regulators.
– Adjustments: Modifications to reflect the specific population (occupation, socioeconomic factors, product selection, underwriting, geography).
– Mortality improvement (projection): A scale that projects future changes in mortality rates (improvements or deteriorations) cohort‑ or period‑based.
– Deterministic vs. stochastic: Deterministic approaches use a single best‑estimate projection; stochastic models produce a distribution of possible outcomes to quantify longevity risk.
2. Why the underlying mortality assumption matters
– Life insurance: Mortality determines expected death claim costs and therefore premium adequacy and reserves. Underestimating mortality (expecting fewer deaths than occur) leads to losses; overestimating can make products less competitive.
– Pensions and annuities: Mortality assumptions determine the expected duration and amount of future payments. Underestimating life expectancy increases the present value of liabilities and funding requirements.
– Capital and solvency: Regulators and rating agencies scrutinize the reasonableness of mortality assumptions because they affect solvency metrics and required capital.
3. Special considerations that affect the assumption
– Age structure: Mortality patterns at birth differ from those at advanced ages. Use age‑appropriate tables (e.g., separate elderly mortality tables).
– Cohort effects: Generational differences (e.g., long-term smoking prevalence) can cause cohorts to experience different longevity trends.
– Pandemics and temporary shocks: Events like COVID‑19 can cause short‑term mortality spikes and may change long‑term improvement expectations; actuaries must decide whether effects are transient or persistent.
– Selection and anti‑selection: Underwriting, self‑selection into plans, and underwriting changes affect the experience of insureds vs. general population tables.
– Heterogeneity: Mortality varies by sex, smoking, health, occupation, income—one-size-fits-all assumptions can be misleading.
– Data quality and credibility: Small or skewed experience data require pooling with broader tables or credibility weighting.
4. Practical step‑by‑step guide to setting an underlying mortality assumption
Step 1 — Define the population and purpose
– Identify the cohort (age range, sex, smoker status, plan vs. insurer blocks, underwriting level).
– Define purpose (pricing, reserving, regulatory reporting, valuation).
Step 2 — Select a reasonable base table
– Choose an industry or regulatory table appropriate for the population (e.g., SOA tables, regulator‑prescribed tables).
– If small or specialized population, select a broader credible table and plan to calibrate it.
Step 3 — Adjust the base table for the plan’s experience
– Compare plan/portfolio experience to the base table (by age, duration, cause if available).
– Apply factors (scaling or age shifts) to align expected deaths with observed experience, using credibility theory where appropriate.
Step 4 — Decide on mortality improvement (projection)
– Select an improvement scale (e.g., SOA published improvement scales such as MP‑series) or develop a bespoke projection if justified.
– Decide whether to use period or cohort projection and the projection horizon. Document rationale.
Step 5 — Model stochastic/longevity risk (if applicable)
– For material long‑term liabilities, run stochastic mortality models (e.g., Lee–Carter, Cairns–Blake‑Dowd) or stress scenarios to capture uncertainty.
– Use this for solvency capital assessment or to price longevity hedges.
Step 6 — Sensitivity testing and scenario analysis
– Run sensitivity tests (e.g., ±10–20% mortality, slower/faster improvement) to quantify impacts on liabilities and pricing.
– Produce best‑estimate and conservative (prudent) scenarios for governance.
Step 7 — Governance, documentation, and review
– Document data sources, table choices, adjustments, projection methods, validation results, and rationale.
– Follow professional standards (e.g., ASOP guidance) and regulatory requirements.
– Set a review cycle (at least annually, and sooner after major events).
5. Testing and monitoring (practical checks)
– Back‑testing: Compare actual experience to expected deaths annually and update assumptions if persistent divergence appears.
– Experience analysis: Track mortality rates by key subgroups (age, sex, smoker, duration) and calculate credible mortality ratios.
– Validation: Use external datasets (national statistics, peer groups) to validate plan results.
– Triggers: Define quantitative triggers (e.g., experience deviation beyond band for X years) that prompt assumption review.
6. Common pitfalls and how to avoid them
– Relying solely on national tables without adjustment: adjust for selection, underwriting, and socioeconomic differences.
– Ignoring improvement: failing to project future improvements can understate pension liabilities.
– Overreacting to short‑term shocks: distinguish temporary blips from long‑term trend shifts; use scenario analysis to capture both.
– Poor documentation and governance: always document the rationale and have independent review.
7. Examples of practical application (brief)
– Pension plan: Use a credible base (e.g., SOA pension table), adjust downward if the plan’s retirees show lower mortality than the general population, and apply an improvement scale to project future lifespans. Test sensitivity by increasing life expectancy by 1–3 years and measure funding impact.
– Life insurer pricing: Start with an insured lives table, adjust for underwriting class (preferred/non‑preferred), apply a short‑term pandemic stress for pricing during volatile years, and maintain a reserve margin or reinsurance strategy for longevity risk.
8. Regulatory and professional guidance (high‑level)
– Follow applicable regulatory requirements and prescribed tables where required by insurance regulators or pension law.
– Apply actuarial professional standards (for U.S. practice, see Actuarial Standards of Practice such as ASOP No. 35 for selection of demographic assumptions for pension measures).
– Use industry research (e.g., SOA mortality studies) and national vital statistics (e.g., Centers for Disease Control and Prevention in the U.S.) for benchmarking.
9. Best practices checklist
– Use age‑ and subgroup‑appropriate base tables.
– Calibrate to your population when data are credible.
– Explicitly choose and justify a mortality improvement approach.
– Perform sensitivity and stochastic analysis for material exposures.
– Maintain clear documentation and independent review.
– Monitor experience and set formal triggers for assumption updates.
– Consider reinsurance or longevity hedges for significant longevity risk.
Conclusion
The underlying mortality assumption is a foundational input to pricing, reserving, and funding life‑contingent obligations. Making the assumption defensible requires good data, appropriate base tables, careful adjustments for the covered population, clear projection methods for future improvement, robust testing, and governance. Given the financial sensitivity and long time horizons involved, actuaries should use both deterministic best estimates and stochastic or stressed analyses to quantify uncertainty and inform business decisions.
References and further reading
1) Investopedia. “Underlying Mortality Assumption.”
2) Centers for Disease Control and Prevention. “Mortality in the United States, 2020.” (Accessed Jan. 24, 2022).
3) Society of Actuaries (SOA). Mortality & longevity research and published tables and improvement scales. /
4) Actuarial Standards Board. Actuarial Standards of Practice (e.g., ASOP No. 35). /
5) National Association of Insurance Commissioners (NAIC). Regulatory guidance for insurers.
Editor’s note: The following topics are reserved for upcoming updates and will be expanded with detailed examples and datasets.