Rank estimation of a generalized "xed-e!ects regression model

This paper considers estimation of a "xed-e!ects version of the generalized regression model of Han (1987, Journal of Econometrics 35, 303}316). The model allows for censoring, places no parametric assumptions on the error disturbances, and allows the "xed e!ects to be correlated with the covariates. We introduce a class of rank estimators that consistently estimate the coe$cients in the generalized "xed-e!ects regression model. The maximum score estimator for the binary choice "xed-e!ects model is part of this class. Like the maximum score estimator, the class of rank estimators converge at less than the Jn rate. Smoothed versions of these estimators, however, converge at rates approaching the Jn rate. In a version of the model that allows for truncated data, a su$cient condition for consistency of the estimators is that the error disturbances have an increasing hazard function. ( 2000 Elsevier Science S.A. All rights reserved. JEL classixcation: C23; C14