Interpreting Changes in Life Expectancy During Temporary Mortality Shocks

Abstract

Life expectancy is the most widely used measure of mortality, largely because of its intuitive interpretation as the expected age at death of an average individual. Changes in life expectancy are also used to assess mortality trends, but mortality change is incompatible with the assumptions underpinning that intuitive interpretation of life expectancy. To provide an alternative, I return to the two main interpretations of the period life table. The first interpretation, as a synthetic-cohort life table, allows for forward-looking probabilistic statements regarding the individual life course. The second interpretation, as stationary population characteristics, yields standardized demographic indicators of the (past) reference period. Interpreting life table indicators such as life expectancy in the first manner is often more appealing, but only the second one makes sense during temporary mortality shocks. The absolute change in life expectancy can then be interpreted as the standardized value of the average lifespan reduction in a past death cohort.