Topics in Microeconometrics: Estimation of a Dynamic Model of Occupational Transitions, Wage and Non-Wage Benefits Cross Validation Bandwidth Selection for Derivatives of Various Dimensional Densities Testing the Additive Separability of the Teacher Value Added Effect Semiparametrically
- Author(s): Baird, Matthew David
- Advisor(s): Buchinsky, Moshe
- et al.
I study three separate questions in this dissertation. In Chapter 1, I develop and estimate a structural dynamic model of occupation and job choice to test hypotheses of the importance of wages and non-wages and learning in occupational transitions, and find that wages are approximately 3 times as important as non-wage benefits in decisions and that workers will pay 70 cents of their hourly wage to avoid the uncertainty surrounding occupational choice. Chapter 2 develops and tests criteria for cross-validation bandwidth selection for derivatives of multidimensional densities; for conditional density, joint estimation of the numerator and denominator bandwidths performs best. Chapter 3 tests the additive separability of the teacher effect assumption common in the teacher value added models using data from the Los Angeles Unified School District, and finds that interacting the teacher indicator variables with a function of the students' lagged test scores captures most of the nonlinearities, preserves the heterogeneity of teacher effects, and provides more accurate estimates.