A model of female labor supply in Italy using cohort data.

Abstract

This paper develops a behavioral model of Italy's female labor supply using cohort data from 1959 to 1981. Both time-series trends and cross-section patterns by age groups of female participation rates in Italy show rather flat profiles. Using an approximate procedure to obtain an estimate of the relevant wage that is the mean of the wages distribution across the whole female population in a given cohort, results show that the flat time-series profile is presumbly produced by economic incentives effects that counterbalance each other. Interaction terms reveal that cross-section profiles by age group do not reflect steady-state dynamics; aging and the presence of children under 6 are less disincentive to younger cohorts. As is typical of most developed countries, this process implies for Italian female participation the emergence of the bimodal cross-section profile by age groups. In the last 20 years, participation rates for all women decreased slightly during the 1st 10 years and then increased, reaching in 1981 a higher level than in 1961. The dependent variable used in the estimation equation is the ratio between particpant (employed and unemployed) women belonging to a cohort; and the total number of women belonging to a cohort in a given year. Results show that as a cohort increases, the negative effects of age and of children are reduced. The results reveal some main points about rather flat participation rate profiles: 1) a very powerful female wage effect, 2) an equally powerful male wage negative effect, and 3) significant interactions,with a positive sign, between cohort and age and between cohort and children. Moreover, younger cohorts, other things being equal, participate more than older cohorts. This is attributable to the fact that the presence of children under 6 and the aging process are less of a disincentive for them than for older cohorts. Overall, the structure of lifetime participation is changing; younger cohorts participation rates, corresponding to child rearing age groups, are higher than older cohorts', and the decreasing effect of age suggests that a reentering phenomenon (after child rearing ages) could be emerging.