Microsimulation of demographic behaviours using 2 alternative data sources

A microsimulation of demographic behaviours using the 1997 Insee Survey about Young People and Job Histories (Robert-Bobée 2001) leads to the following results: if demographic behaviours remain the same as the ones observed in 1995-1996, completed fertility may decrease to less than 2 children per woman born around 1970 and remain constant about 1.9 children per woman born after 1975. These simulations rely on several assumptions which may prove inadequate in the future. In particular, we assume that future behaviours will remain identical to the ones observed in 1995-1996, when fertility was quite low. Some other uncertainties stem from the size of the sample. For instance, estimated probabilities for second unions after a long time spent without a partner are difficult to estimate: only people from 19 to 45 were interviewed, and those who have been living alone for a long time are not numerous. In this paper, we investigate the robustness of the simulation using another data source for estimating behaviours. The new data base is the 1999 Insee Survey on Family Histories. Behaviours are estimated in 1996-1998, years of higher fertility. Results are found to be very consistent with the previous ones.