Homogeneity Pursuit in the Functional-Coefficient Quantile Regression Model for Panel Data with Censored Data

Abstract

Abstract Homogeneity identification of panel data models has been popular in the literature in recent years. Most of the existing works only focus on the complete data case. This paper considers a functional-coefficient quantile regression model for panel data with homogeneity when its response variables are subject to censoring. In particular, we consider a more general censoring framework, i.e. different types of censoring are allowed to occur in the model simultaneously. For this, a “three-stage” method is proposed, which includes the preliminary estimation of subject-specific function coefficients based on data augmentation, the identification of group structure over subjects by clustering, and post-grouping estimation of function coefficients. Simulation studies considering the left-, right-, and double-censored data, are carried out to verify the finite-sample properties of the proposed method. Simulation results show that our method gives comparable performance to the complete data case. The application to the bank stock data further illustrates the practical advantages of this method.