效率测度与S型生产函数
In the literature of efficiency analysis, the most frequently used empiricalfrontier is the piecewise linear frontier exhibiting variable returns to scale(VRS). It has a convex production set and cannot model a conventional propertyin economics textbooks: the production function has S-shape curve along any rayin the input space. Petersen (1990) introduced empirical production frontiersto model nonconvex production set with convex input and output sets. However,these frontiers do not model variable return to scale. Li (2019) has introduceda new empirical production function that is able to capture variable returns toscale without convexity. The focus of Li (2019) was on nonparametric test. Thediscussion of this production function was limited. This paper provide detailsframework of Li’s empirical production frontier and its relationship withefficiency measurement. It turns out that the conventional VRS frontieroverestimates the technical inefficiency and underestimates the scaleinefficiency of firms.
This new empiricalfrontier is applied to the tourism industry in China. It is found that theconventional VRS frontier fails to capture most of the scale inefficiency ofsmall firms in this industry. Instead, the poor performance of small firmsusing conventional empirical production frontier is attributed to technicalinefficiency. The new frontier indicated that a high proportion of the overallinefficiency of small firms is due to their small scale. This leads toimportant policy implications to the development of the tourism sector inChina.