LAUSR

Abstract In this paper, based on the cumulative distribution function (CDF) method (Jradi et al., 2021) for finding the optimal quantile when estimating stochastic frontier models (SFM) with normal–exponential composite error term, we derive an expression to find the optimal quantile for the SFM with normal–half normal composite error term. We then use Monte-Carlo simulations and the same data set as Jradi et al. (2019) to compare the difference of iteration method (Jradi et al., 2019) and CDF method for the SFM with normal–half normal specification. The simulations and empirical application illustrate that both methods work well.