面板数据随机前沿分析的研究综述

近年来，面板数据随机前沿分析（SFA）越来越多地被用于测算各类决策单位的效率，取得了很多成果，但是国内外实证研究文献也存在过度依赖几种假设严格的模型和不注重模型局限性的问题。本文在统一的计量框架下，对面板SFA模型的发展研究进行了系统的梳理总结。本文将相关模型分为两大类：效率不随时间变化的模型和效率可随时间变化的模型，每一类又根据是否对效率项的分布做出假设分为：有分布假设的模型和无分布假设的模型。在明确和比较不同模型的假设、估计过程和局限性的基础上，我们对未来面板SFA模型的应用提出了建议。

A Literature Review on the Stochastic Frontier Analysis in Panel Data

In recent years, Stochastic Frontier Analysis (SFA) in panel data is being increasingly used to measure the efficiency of all forms of decision making units’ efficiency and obtaining many research achievements. However, most of these papers rely heavily on several models with strict assumptions and pay insufficient attention to the availability of these models. This paper summarizes the development of Stochastic Frontier Analysis in panel data under a unified econometric framework systematically. We classify these models into two parts based on whether efficiency term changing with time or not. In each part, we further divide the models into two parts according to whether the efficiency term and random error term having distributional assumptions. We emphasize and compare the assumptions, estimation procedures and availability of each model in order to provide potential and helpful suggestions for improving the methods employed in empirical research.