A Period TFR with Covariates for Short-Panel Data. In: Stockholm Research Reports in Demography 2013: 5

This paper extends a recently proposed approach aimed at reconciling the most widely used macro-level indicator of fertility, the Total Fertility Rate (TFR), with estimates that derive from applications of Event History Analysis (EHA) to micro-data. Using cross-sectional or, as in this case, short panel data, group-specific fertility estimates can be obtained that are consistent with the period TFR for the entire population. Short panels are now relatively frequent in socioeconomic research, and extending their use to demographic analysis allows researchers to encompass exogenous variables that are only rarely available or very imperfectly measured in retrospective surveys, among which income or, more generally, individual-level economic data. An additional merit of the proposed approach is that is avoids a few of the selection problems that frequently emerge with EHA (e.g. when applications are specific by birth order, or by marital status). An application to Italian data reveals that the fertility of all the subgroups that can be formed is extremely low, and, therefore, that no structural modification, no matter how large, would suffice to bring fertility back to replacement level. Behavioural changes are required to that end.