基于Copula方法处理随机前沿模型的正偏度问题  

Dealing with Positive Skewness Issue in Stochastic Frontier Model Based on Copula Method

作　　者：冯冬发 张涛[2] 李奥 FENG Dong-fa;ZHANG Tao;LI Ao(Graduate School of Chinese Academy of Social Sciences,Beijing 102488,China;Institue of Quantitative & Technological Economics,Chinese Academy of Social Sciences,Beijing 100732,China)

机构地区：[1]中国社会科学院大学(研究生院)数量经济与技术经济系,北京102488 [2]中国社会科学院数量经济与技术经济研究所,北京100732

出　　处：《统计与信息论坛》2021年第1期13-21,共9页Journal of Statistics and Information

基　　金：国家自然科学基金重大项目“宏观大数据建模和预测研究”(71991475);中国社会科学院大学(研究生院)研究生科研创新支持计划项目“大数据在宏观经济预测和政策评价中的应用”(2020-KY-048)。

摘　　要：随机前沿模型在小样本情形下容易出现正偏度问题,无法给出厂商技术效率的合理估计值。放松经典模型中关于复合扰动项的相互独立假设后,基于Copula方法刻画白噪声项与技术无效率项之间的相关性,采用数值模拟方法计算模型参数与厂商技术效率的估计值。蒙特卡洛实验表明,此估计策略具有良好的大样本性质,较之经典模型拥有更高的估计精度,能够在出现正偏度问题时给出合理的技术效率估计值,且估计结果在不同的Copula函数选择下保持着较好的稳健性。Stochastic frontier model is prone to positive skewness in the case of small samples,and cannot give a reasonable estimate of the firm’s technical efficiency.After relaxing the mutual independence assumption of the compound disturbance term in the classical model,the correlation between the white noise term and the technical inefficiency term is described based on Copula method,and the estimated value of the model parameters and the firm’s technical efficiency is calculated by the numerical simulation method.Monte Carlo experiments show that the estimating strategy are good for large sample,which has higher estimation accuracy than the classic.It can give a reasonable estimate of technical efficiency when there is a positive skewness problem,and the estimation result maintains good robustness under different Copula function circumstances.

关 键 词：随机前沿模型 正偏度问题 COPULA方法 极大模拟似然估计 

分 类 号：O212[理学—概率论与数理统计]