Estimating the Labor Supply Dynamics of Older Workers Using Repeated Cross-sections

The empirical analysis in this paper adopts logit models to study the hazard rate of ceasing from work by the next year for Italian older employees. The specifications are estimated resorting to the framework proposed by Guell and Hu (2006), which extracts information from repeated independent cross-sections to recover the hazard rate of interest at the individual level. The sample is drawn from the ISTAT survey Aspects of Everyday Life and includes employees aged 50-65 in 1993-2002. Our results show that, even conditioning on a wide set of socioeconomic factors, the age profile of the hazard rate is increasing and confirms the low labor market attachment of older workers. Further, the time evolution of the risk of becoming not employed appears to be hump-shaped and achieves its maximum for the cohorts of employees at work in the period 1996-1998, which is characterized by the introduction of important changes in the Social Security system aimed at extending the working life of the elderly.