MPIDR Working Paper
A behavioral Gompertz model for cohort fertility schedules in low and moderate fertility populations
MPIDR Working Paper WP-2010-021, 20 pages.
Rostock, Max-Planck-Institut für demografische Forschung (Februar 2010)
In this paper, I re-introduce the Gompertz model of age-specific fertility. This model has been rejected by past authors as fitting poorly to cross-sectional, or period rates. However, I find that the model fits very well to recent medium and low fertility cohort schedules in France, Italy, and Japan. A distinct advantage of the Gompertz model is that it has a simple behavioral interpretation as the result of social diffusion of fertility behavior in a cohort competing with the fertility-depressing effects of older age.
The Gompertz model, and further refinements that include a better specification of the biological limits of childbearing, offer a means for forecasting future fertility, describing temporal change, and assessing the fertility-limiting effects of older entry into motherhood. In addition, the Gompertz function allows for traditional uses of model age-schedules such as smoothing and correction of data.
The model estimates the completed cohort fertility of French, Japanese, and Italian cohorts born in 1965 to be 2.0 , 1.6, and 1.6, respectively. For France, this represents only a minor decline from earlier cohorts, but for Japan the decline in cohort fertility is marked. The model with infertility suggests that recent decline in Italy is largely due to shifts to older ages of childbearing.