MPIDR Working Paper

Bayesian forecasting of cohort fertility

Abstract

There are already several documented examples of recent increases in cohort fertility in Scandinavia, but for most countries, cohorts are too young to see if cohort fertility has increased. We produce new estimates of completed cohort fertility for cohorts born in the 1970s. We combine the best of previous efforts, using cohort forecasting methods to preserve what demographers know about the age-pattern of fertility, and using trends in the age-period-cohort Lexis surface to tell us as much as possible about the way in which fertility appears to be changing over time. Our preliminary findings suggest that cohort fertility has stopped its long-term secular decline in the majority of low fertility countries around the world. In some cases, there is a clear suggestion of increase. As we further develop our models we expect to be able to make more precise statements about further trends and the certainty of our knowledge.