截面数据的随机成本前沿面模型的半参数分析

Semiparametric Estimation of Cross Sectional Stochastic Cost Frontier Models

针对横截面数据的随机前沿面成本函数,给出了其半参数方法误差密度估计。因为随机前沿面模型中有2个随机误差因子,若用卷积进行密度估计将是非常复杂的。这里用特征函数构造概率密度估计从而避免了应用卷积进行密度估计的复杂操作。但是这种方法也有不足之处,就在于它对模型有一些弱的假定点估计依赖于误差因子与模型参数的假定,密度估计依赖于误差因子特征函数的假定。

This paper developed a semiparametric density estimation for cross sectional stochastic cost frontier models. As stochastic cost frontier models with two error components, convolution to get probability density will be very complex. This approach developed character function to gain density estimation avoided the complexity of using convolution. The disadvantages were that this method was based on assumptions on the model: point estimation based on parametric assumption and some properties of error components.