The twin hypothesis of education and retirement

Abstract

Common models examining the effects of demographic change and the efficiency of policy reforms often suffer from two important shortcomings: (i) interaction between an individual-s schooling and retirement decision is disregarded and (ii) a realistic demographic population structure is absent. In order to overcome these limitations, we combine two papers of Ben J. Heijdra and Ward E. Romp who developed a continuous time overlapping generations model in the manner of Blanchard (1985) allowing for age-dependent mortality rates. In our modification of their work, agents choose both length of schooling and retirement age endogenously. Pension benefits are calculated from lifetime wage earnings which positively depend on the level of education. On the aggregate, dynamics are rather complicated due to generational turnover effects and a human capital externality. Nevertheless, we present a comparative static analysis of long-run effects as well as numerical simulations for the transition paths following demographic shocks and fiscal stimuli. We find that not controlling for individual adjustments in education and retirement at the same time will overestimate the negative impacts of aging on the macroeconomy. Similarly, the economic impact of education reforms is found to be much stronger if not only schooling but also retirement is treated as endogenous.