Search Results

Author: Houser, Daniel Edward
Resulting in 1 citation.
1. Houser, Daniel Edward
Bayesian Analysis of a Dynamic, Stochastic Model of Labor Supply and Saving
Ph.D. Dissertation, University of Minnesota, 1998
Cohort(s): NLSY79
Publisher: UMI - University Microfilms, Bell and Howell Information and Learning
Keyword(s): Bayesian; Capital Sector; Human Capital; Labor Supply; Life Cycle Research; Wage Dynamics; Wage Models

This thesis specifies and estimates a dynamic, stochastic model of life cycle labor supply. All of the past empirical work in this area has made one of the following assumptions: (a) there is no human capital accumulation, so that the wage stream is exogenous to the individual; or (b) capital markets are perfectly imperfect; or (c) that agents have rational expectations, or use some other rigid, expectations formation mechanism. While each of these restrictions may have justification within a particular application, it can be shown that imposing any of them may lead to biased and inconsistent parameter estimates. Accordingly, policy recommendations based on past empirical findings are open to question. The model that I specify and estimate allows for human and physical capital accumulation, and does not impose strong assumptions about the way individuals form expectations. I accomplish this level of generality by extending and employing an estimation methodology originally advanced by Geweke and Keane (1997). They were the first to point out that micro-level data on payoffs and choices could be used to estimate the parameters that determine preferences, as well as those that characterize expectations. All that one needs to assume is that expectations lie along some polynomial in the model's state variables. The coefficients of the expectations polynomial are estimated jointly with the model's structural parameters. That I do not need to impose the restrictions that are typically required in this literature allows me to take a first step towards assessing the effect they may have had on the estimates of policy-relevant parameters. I take my model to data drawn from the National Longitudinal Survey of Youth. I use a Bayesian approach to inference, and approximate the marginal posterior distributions of my model's parameters with a Gibbs sampling algorithm. I find that the uncompensated wage-elasticity of labor supply is very small, that wealth effects are very small, and that omitting savings decisions from life cycle models may have little effect on inference about labor supply decisions. Finally, I find that individuals are not myopic. Work experience and age play a significant role in expectations formation.
Bibliography Citation
Houser, Daniel Edward. Bayesian Analysis of a Dynamic, Stochastic Model of Labor Supply and Saving. Ph.D. Dissertation, University of Minnesota, 1998.