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

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.