数据包络分析与随机前沿分析的比较研究--基于 Monte Carlo 模拟研究

目的：比较研究数据包络分析（ DEA）和随机前沿分析（ SFA）方法。方法通过蒙特卡罗模拟方法，比较两种方法效率值估计的准确性和效率值排名的一致性（ Spearman相关系数）以及评价指标选择对两者的影响。结果DEA与SFA结果一致性较高；删减必要指标对两种方法评估的准确性和一致性的影响较大，而增加无关指标对两者的影响较少，增加无关指标前后效率值一致性较高，Spearman相关系数大部分在0．9以上。结论选择指标时要慎重，对效率评估结果影响较大的指标应保留。

Comparative study of DEA and SFA-A Monte Carlo simulation study

Objective To compare the data envelopment analysis and stochastic frontier analysis.Methods The study compared the efficiency evaluation accuracy and rank consistency （ Spearman correlation coefficient） and the impact of indicator selection on the efficiency evaluation of two methods by the Monte Carlo simulation.Results The results showed that the consis-tency of two efficiency analysis results was high, and the omission of relevant indicators had great impact on the accuracy and consistency of both two methods, while adding irrelevant indicators had less impact, and most of the spearman correlation coeffi-cient between, before and after adding irrelevant indicators were above 0.9.Conclusion Be careful when selecting indicators and the indicators which have greater impact on the efficiency assessment result should be retained.