The extended skew-normal-based stochastic frontier model with a solution to 'wrong skewness' problem. Issue 6 (2nd November 2021)

• Record Type: Journal Article
• Authors: Wei, Zheng; Zhu, Xiaonan; Wang, Tonghui
• Is Part Of: Statistics. Volume 55:Issue 6(2021)
• Page Start: 1387
• Page End: 1406
• Publication Date: 2021-11-02
• Subjects: Extended skew-normal distribution -- stochastic frontier model -- technical efficiency -- wrong skewness problem; Mathematical statistics -- Periodicals; 519.505
• DOI: 10.1080/02331888.2021.2004142
• Languages: English
• ISSNs: 0233-1888

Abstract

Abstract : In this paper, a new extended skew-normal-based stochastic frontier model (SFM) is proposed to investigate the relationship between the disturbance U (representing technical inefficiency) and V (representing noise). Two methods for deriving the distribution of the error, E = V − U, in the model are studied, together with the discussion of the model with wrong skewness problems in detail. The conditional distribution of U given E = ϵ and technical efficiency of the model is obtained. For the illustration of our results, simulation studies on the performance of the proposed two new models are given with comparisons of the classical normal-half normal SFM. Two real data examples, including the one with the wrong skewness under the classic SFM, show that our skew normal based SFMs perform better than the classic SFM. More importantly, the extended skew-normal based SFMs provide a non-zero inefficiency estimates among firms with the correction of the wrong skewness problem.