面板数据随机前沿分析及其应用

Stochastic Frontier Analysis of Panel Data and Its Application

效率测度在经济学研究中占据着重要地位,是价值判断的重要切入点。面板数据较之于截面数据的信息优势,使得面板随机前沿分析逐步得到关注。但目前国内相关研究多停留于早期面板SFA,由此也导致了实践应用的局限甚至研究结论的冲突。文章通过对误差分布、时变特征、影响因素和模型形式的系统梳理,总结归纳了面板SFA模型的历史演变和最新进展,研究发现:(1)面板SFA模型关于误差假设分布和影响因素分析的研究框架基本定型,未来将更加趋向于"一步法"框架的创新应用;(2)统筹个体效应中的时不变效率与时变效率是未来面板SFA模型修正的方向;(3)对内生性和样本选择偏误问题的探讨、与非线性门限回归的结合及与空间计量理论的结合等将成为面板SFA模型的新热点。通过模型形式设定的改进以及与经典计量理论的多样化结合,基于面板SFA的效率测算将不断拓展出新的研究应用场景。

Efficiency measurement plays an important role in economic researches and it is an important entry point for value judgment. The information advantage of panel data over sectional data makes panel stochastic frontier analysis gain more and more attention. However, at present, most of the relevant domestic researches remain in the early panel SFA, which also leads to the limitation of practical application and even the conflict of research conclusions. This paper summarizes the historical evolution and latest progress of panel SFA model by systematically sorting out error distribution, time-varying characteristics, influencing factors and model forms. The research finding are as follows:(1) The research framework of panel SFA model on error hypothesis distribution and influencing factor analysis has been basically finalized, and will be more inclined to the innovative application of"one-step method"framework in the future.(2) Considering time-invariant efficiency and time-varying efficiency in individual effect is the revision direction of future panel SFA model.(3) The discussion on endogeneity and sample selection bias, the combination with nonlinear threshold regression and spatial econometric theory will become the new hot spots of panel SFA model. The efficiency measurement based on panel SFA will continue to expand new research application scenarios through the improvement of model formalization and the diversified combination with classical econometric theories.