Valentin Zelenyuk, University of Queensland, Australia

Inference for Non-Parametric Stochastic Frontier Model

Abstract

In this paper we focus on statistical testing in the framework of nonparametric
and semi-parametric stochastic frontier via local least squares approach (due to Simar, Van Keilegom and Zelenyuk (2016)). In particular, we are concerned with inference about existence of production inefficiency in general and at particular points of interest as well as inference about significance of factors potentially influencing the conditional mean of efficiency. To reach our goals, we consider and adapt the bootstrap based approach proposed by Delgado and Gonzalez-Manteiga (2001).

This is joint work with Léopold Simar and Ingrid Van Keilegom.