This project should make important contributions to theoretical econometrics and to applied economics. This project addresses problems in two major subfields of econometric theory: latent variables and quantile regressions. The theoretical research is accompanied by significant empirical applications. The results of this project should provide applied economists with more general and more powerful statistical tools for studying the behavior of financial markets and the characteristics of labor markets. More specifically, this project develops new methods for using longitudinal data to allow for latent variables in nonlinear models. The latent variables are individual effects that vary across individuals but are constant over time. The project will focus on relaxing strict exogeneity assumptions, so that covariate values may be partly determined by previous values of the response variable. Empirical applications of the techniques will be developed. The project will also examine inference procedures in cross-section studies using quantile regression. Extensions of quantile regression for use with longitudinal data will be explores.
