9730155 Abrevaya This project investigates a specific branch of the semiparametric econometrics literature known as rank estimators. The project focuses on the development of new rank estimators for various econometric models and consideration of their asymptotic theory. This research should be of interest to both theoretical and applied researchers in statistics and economics. The estimators are simple to compute and avoid the well-known problem of bandwidth selection. More specifically, existing U-statistic asymptotic theory will be applied and extended, when necessary, in order to derive asymptotic formulas for the following: rank estimators for the censored and truncated regression models; new estimators of the transformation model based on a fourth-order U-process; a new estimator for duration models that estimates the coefficients of a proportional hazards model with unobserved heterogeneity; and an estimator for the semiparametric binary choice models with symmetric errors. Monte Carlo simulations will be used in order to examine the performance of the new estimators relative to existing estimators. An exhaustive library of routines for the rank estimators in the literature will be developed and made available over the Internet so that practitioners will be able to use the estimation techniques. ??
