mortality improvement and projection

People are living longer than they used to. Thanks to better medicine, sanitation, and lifestyles, the death rate at almost every age has been falling decade after decade — a 65-year-old today faces lower mortality than a 65-year-old did thirty years ago. This downward drift over time is called mortality improvement, and any table built only from past data will quietly become too pessimistic about death (and too cheap on annuities) as the years pass.

To handle this, actuaries project mortality forward rather than freezing it. The starting (base) rate at each age is reduced year by year using an improvement scale — a set of assumed annual percentage reductions, often larger at younger ages and smaller at the oldest. A simple version multiplies each base rate q-x by (1 minus the annual improvement) raised to the number of future years. More sophisticated approaches fit statistical models to decades of national data (the well-known Lee-Carter model is one) to extrapolate the trend, with uncertainty bands around it.

Getting this right is enormous in money terms, especially for things that pay while you are alive. Underestimating improvement is exactly the longevity risk that has strained pension schemes and annuity providers worldwide: liabilities that look fine on frozen mortality balloon when people stubbornly keep living longer. The honest difficulties are real — improvement is not steady (it has stalled and even reversed in some recent years and populations), and projections decades out are genuinely uncertain, so prudent valuations carry margins and test alternative scenarios.