Period and cohort life expectancy: what the number really measures

Period life expectancy: one year, frozen

To calculate life expectancy at birth for a year, demographers take the death rates of that year at every age, from babies to people over 100. They then imagine a group of newborns who live their whole lives with exactly those rates and ask how long they would live on average.

By the death rates of 2023, that average for the world is 73.2 years, by the UN estimate. In 1950 it was 46.4 years.

Nobody lives a whole life under the death rates of a single year. The number is a summary of how deadly that year was at each age, which makes it useful for comparing places and years.

Cohort life expectancy: real lives, counted afterwards

A birth cohort is everyone born in the same year. Its life expectancy is the average age at death of those people, measured as their lives actually unfold. Each of them lives through decades of changing health care, living conditions, wars and epidemics.

That number can only be known once almost everyone in the cohort has died, so it exists for people born a century ago, not for anyone born recently. For younger cohorts it can only be projected.

Why the two differ

When death rates keep falling, every year of a cohort’s life is safer than the same age was in its birth year. The cohort then lives longer on average than the period life expectancy of its birth year suggested. When a crisis such as an epidemic raises death rates for a short time, the period number for those years drops sharply, while each cohort lives through the crisis for only a few years of its life.

How Anthrotempo uses it

The longevity tool and the longevity topic show period measures from the UN life tables, always with their year. They describe a year’s mortality, never a forecast of anyone’s lifespan, and the site never subtracts your age from a life expectancy to give a “time left”.

Last updated 26 October 2026