A0527
Title: The stochastic frontier model with ordered multiple choices
Authors: Yi-Wun Chen - Binghamton University, State University of New York (United States) [presenting]
Abstract: This paper develops the stochastic frontier model with multiple endogenous regimes to deal with the problem of biased and inconsistent estimates due to the sample selection bias. If we encompass all observations with different regression coefficients (heterogeneous observations) into one regression equation, statistically it also implies the sample selection bias. For the proposed model, I derive the closed form of the likelihood function and the estimator of technical efficiency index based on the sample selection information and estimate it by maximum likelihood estimation. In the empirical study, I studied the operating cost efficiency of doctoral granting universities with three levels of college acceptance rates in the United States and applied it to the proposed model with three regimes. All evidence suggests that ignoring selectivity and heterogeneity in the model may result in the wrong estimated parameters and thus the predicted TE indices are invalid. We can conclude that taking sample selection and multiple-regime model into account are necessary when observations are heterogeneous.
