A period TFR with covariates

Gustavo De Santis, University of Florence
Sven Drefahl, Stockholm University
Daniele Vignoli, University of Florence

We propose a method for the study of fertility on cross-sectional survey data that explicitly aims at reconciling micro- and macro analysis: the group-specific fertility estimates that it produces are consistent with the period TFR for the entire population. Incidentally, this approach also avoids a few of the selection problems that frequently affect other applications (e.g. specific by birth order). First we estimates the probability of birth (of any order) in each subgroup (specific by age x and other characteristics c), and then correct this estimates with a proper parameter k, so that the weighted and corrected sum of group-specific birth probabilities gives precisely the TFR. An application to the four waves of the Italian section of the EU-SILC (Community Statistics on Income and Living Condition), 2004-2007 is proposed.