Semiparametric Estimation of a Censored Regression Model with an Unknown Transformation of the Dependent Variable

Preprint
Publisher Website
Google Scholar
    • preprint
      • Published in RePEc
Abstract
In this paper we develo psemiparametric estimators of L and y in the model L(Y) = min[b›X + U,C], where Y is a nonnegative dependent variable, X is a vector of explanatory variables, U is an unobserved random "error" term with unknown distribution function y, C is a random censoring variable, b is an unknown parameter vector, and L is an unknown strictly increasing function. This model includes as a special case the censored proportional hazards model with unobserved heterogeneity. Estimators of L and y already exist for the case where either L or y belongs to a known finite-dimensional parametric family, and methods for estimating b exist for the general case. In this paper we propose estimators of L and y which do not assume that L and y belong to known parametric families. We obtain their asymptotic distributions and investigate the small sample properties of the estimators by Monte Carlo simulation.
All Related Versions

This publication has 0 references indexed in Scilit: