Abstract
Estimation of technical efficiency lies at the core of stochastic frontier analysis. However, it is common that only the conditional expectation of technical efficiency for each observation is calculated. If one were interested in alternative features of the conditional distribution, such as quantiles or the mode, these are commonly unavailable in closed form and require simulation methods. Here we propose a simple nonparametric approach that can provide an array of features of the distribution of technical efficiency, for any set of distributional assumptions.
•Nonparametric estimation of technical efficiency.•Estimation mean, quantile, mode.•Portable across a range of distributional assumptions.•Easy to deploy strategy.