There are stark differences in schooling and retirement decisions across time and space. In the U.S. in 1900, 60 percent of men aged 65 or older were in the labor force, and the average worker had 5.4 years of schooling. In 2000, only 20 percent of men aged 65 or older were in the labor force, while the average worker had 12.4 years of schooling. Also in 2000, the labor force participation rate for men aged 65 or older was only 5 percent for a group of European countries. For these countries, the average years of schooling was 9.4 years.
In this paper we study a model of effective lifetime labor supply that is rich enough to analyze the effects of changes in wages and taxes on the quantity dimension of lifetime labor supply (i.e. career length) as well as the quality dimension which includes choices of the quantity and quality of schooling as well as investments in on-the-job training. We use this framework to explain the patterns of schooling and retirement across time and space, and to explore the effect of changes in retirement tax/benefit schemes.
The model that we study blends elements of the standard life-cycle model which are essential to understand retirement decisions with a dynastic preference structure that guarantees that the long run behavior is easily characterized. We assume that individuals choose the amount and quality of schooling and that, when in the labor force, decide how to allocate their time between producing and increasing their human capital (on-the-job training). In this version, we study a model with indivisible labor and we assume that, once the worker retires, he can enjoy leisure. It is this extra utility associated with not working that drives retirement. We model retirement benefits as a non-linear function of income that includes a fixed payment.
We use the model to explore, both theoretically and quantitatively, the impact of some policy changes (e.g. changes in the nature of the retirement system), demographic shocks (e.g. changes in population growth rate or life expectancy) and total factor productivity, which allows us to model variation in space (e.g. productivity differences between Europe and the U.S.). In general, a given change, say, a change in retirement benefits financed by changes in taxes, affects several margins. Higher taxes –holding retirement decisions constant discourage individuals from acquiring human capital and, at the same time, can induce workers to retire earlier. Thus, it is possible that a policy that does not have a large impact on the length of a career (individuals join the labor force earlier and retire earlier as well) can have a significant impact on effective labor supply through their effect on human capital.
Theoretically, we show that in the long run, reducing the progressivity of the tax code and changing the composition of government expenditures (net of retirement benefits) in the direction of more goods and fewer transfers, increases the retirement age. We also explore the effects of unanticipated shocks. We show that wealth shocks, e.g. a drop in the stock market, increases the retirement age and that the effect is stronger for older workers. A permanent decrease in the wage rate (a TFP shock) has non-monotonic effects on the equilibrium ageearnings profile: on impact, effective labor supply increases, and over time it decreases to fall below the pre-shock level at retirement. We show that lower wage rates unambiguously induce older workers to retire earlier. Finally, we study the effect of an unanticipated change in the stock of human capital (e.g. a loss associated with reallocation in the presence of firm or sector specific human capital). We find that older workers respond by supplying less effective labor and retiring earlier.
Download
Lifetime Labor Supply and Human Capital Investments
