Title: On asymmetry and quantile estimation of the stochastic frontier model
Authors: William Horrace - Syracuse University (United States)
Christopher Parmeter - University of Miami (United States)
Ian Wright - University of Miami (United States) [presenting]
Abstract: Quantile regression has become common in applied economic research. Recently, these methods have been adapted for use with the stochastic frontier model. However, the composed nature of the error term is ignored, drawing into question if a stochastic quantile frontier is actually estimated. Here we demonstrate that a particular distributional pair is consistent with the intent of these earlier proposals but is not, in fact, a quantile estimator. A unique feature of this distributional pairing is that both distributions can be asymmetric. We further discuss the identification and practical issues associated with this model.