Essays on the effect of employer-provided health insurance on job mobility and nonlinear measurement error models

Sequential decision making or choices made over time under uncertainty appears naturally in economic problems. Modeling the phenomenon often requires dynamic models, which accommodate persistent effects of individuals' decisions and states.;Chapter 1 presents a dynamic model of the joint employment and health insurance decision to investigate job mobility. The model is motivated by lacking work on modeling how health insurance decision affects health status. I estimate to the effect of employer-provided health insurance (EPHI) on job mobility via a dynamic model of joint employment and health insurance decision in the presence of uncertainty about wage rate and health status transitions. The model is based on a Markov decision process in which a hedonic wage approach provides an economic rationale for the different choices and health insurance serves as an input to the health production process. Including health transitions in the model helps us to understand how the availability of EPHI (positive job characteristic) and holding EPHI (the wage-health insurance tradeoff) enter into the individuals' decisions. The model is estimated using the 1999-2000 Medical Expenditure Panel Survey (MEPS) panel 4 and the results show that the “pure” effects of holding EPHI are negligible, the “full” effects of EPHI are significant, and the degrees of the inefficiency vary between 14% and 25% across different states.;Chapter 2 provides an alternative two-step sieve estimator for nonclassical measurement error models. Many structural econometric models can be stated in the form of operators. I provide a two-step sieve estimation method for a structural model identified by operators. The corresponding operator identifications are equivalent to infinite conditional moment conditions. My proposal is to extend the method of sieve minimum distance with finite conditional moment restrictions in Ai and Chen (2003) and Newey and Powell (2003) to estimate parameters identified by infinite conditional moment restrictions. I first show that the sieve estimator is consistent under the sup metric. Under the Fisher metric and additional assumptions, the sieve estimator is further consistent with a rate faster than n-1/4 and then the results n of consistency and asymptotically normality can be obtained. The application of the results are illustrated with nonclassical nonlinear errors-in-variables models with continuously distributed variables which are identified and estimated in Hu and Schennach (2008).;Keywords. Job Mobility, Employer-Provided Health Insurance, Joint Decision, Semi-/nonparametric Conditional Moment Restrictions, Sieve Minimum Distance, Nonclassical Measurement Error.;JEL Classification. J21, J31, J32, J63, C4, C5.