Cohort Parity Analysis and Fertility Transition Dynamics: Reconstructing Historical Trends in Fertility Control from a Single Census

The Cohort Parity Analysis (CPA) method for the indirect measurement of the extent of marital fertility control makes use of parity distribution data of the kind reported, for specified marriage duration and marriage-age (or current age) groups of women, by a fertility census. Corresponding to every such distribution, CPA yields efficient upper and lower bounds on the proportion of controllers in the specified cohort, as well as bounds for the parity distribution of the controllers. As the women who belong to different marriage cohorts are observed at different durations of marriage in a census, it is possible to infer inter-cohort trends in the extent of fertility control (at some specified marriage duration(s)) if one knows the time pattern of the adoption of control among the members of each cohort. It is shown in this paper that robust estimates of the intracohort diffusion pattern can be obtained from the census cross-section itself, by supplementing the basic assumptions of the CPA framework with a plausible assumption about the behaviour of the ‘representative’ controller in a given cohort. The estimation technique is illustrated by applications to data for married women in urban Eire from the Census of Ireland in 1911, and to native white married women in the U.S. South from the U.S. Census of 1910.