Search Results

Chapter One: The literature on occupational matching has traditionally assumed that an individual's match in one occupation is completely uncorrelated with their match in every other occupation in the economy. In this paper I develop and estimate a model that relaxes this assumption. The key feature of the model is that as an individual learns about their occupation specific ability in one occupation, this experience will be broadly informative about their ability in other occupations. Workers continually process their entire history of information which they use to determine when to change careers, as well as which new career to go to. Endogenizing information in this manner has been computationally prohibitive in the past. I extend new methods of solving dynamic discrete choice models with unobserved heterogeneity in a unique way that facilitates the estimation of an occupational choice model where individuals have correlated unobserved heterogenous ability in every occupation. The model is estimated on the National Longitudinal Survey of Youth 1997. The results suggest that occupational ability shares a rich correlation structure. While the gains from search are large, a counterfactual shows that search frictions, including: risk aversion, destruction of occupation specific human capital, entry costs, and switching costs, limit worker's ability to find their comparative advantage.

Chapter Two: Economists typically take one of two approaches when estimating models containing omitted variables and measurement error: reduced form methods using instrumental variable, or structural methods that formally model the selection process of individuals. This paper demonstrates a new estimation strategy that conjoins these two methods in a useful way which takes advantage of each approaches individual strengths. The motivating example will focus on the literatures interpretation of the returns to education.