An Estimable Dynamic General Equilibrium Model of Work, Schooling and Occupational Choice

This paper develops and estimates a dynamic general equilibrium overlapping generations model of career decisions. The model is fit to data on life cycle employment, schooling and occupation decisions, and on life cycle labor earnings, within and between cohorts observed in the US between 1968 and 1993. Based on the estimates of the model, the impact of cohort size on skill prices and, consequently, on career decisions is assessed. Compared to a baseline case in which cohort size increased at a steady (average) rate and the fertility process was stationary, it is found that the male baby bust generations born in the 1930s and 1940s faced higher skill prices (by as much as 2.0%), completed college at a higher rate (by as much as 0.3 percentage points) and worked more over the lifetime (by as much as 0.1 years). In contrast, the males from the baby boom generations born in the 1950s and 1960s faced lower skill prices (by as much as 1.5%), completed college at a lower rate (by as much as 1.0 percentage points) and worked less over the lifetime (by as much as 0.1 years). The impact of cohort size is found to differ across gender due to the differential impact of fertility on the value of home production: unlike males, the female baby bust generations completed college at a lower rate (by as much as 1.5 percentage points) and worked less over the lifetime (by as much as 0.5 years) while the female baby boom generations completed college at a lower rate (by as much as 1.5 percentage points) and worked more over the lifetime (by as much as 0.2 years). A tuition subsidy experiment is also performed both in a general equilibrium setting in which skill prices are endogenous and in a partial equilibrium setting with fixed skill prices. Based on the estimated model, a 1% increase in tuition is predicted to reduce college enrollment rates by 1.27% in partial equilibrium, consistent with previous estimates, and by 1.05% in the general equilibrium case. Thus, for this policy experiment, the partial equilibrium analysis of schooling choice would not differ significantly from the general equilibrium analysis.