基于第一价格拍卖的广义分段伪最大似然估计

Generalized Piecewise Pseudo-maximum Likelihood Estimation:Based on the First-price Sealed-bid Auction

越来越多的研究表明投标人是趋于风险规避的,而作为一种经典第一价格密封拍卖的结构估计方法,Donald和Paarsch的分段伪最大似然估计(PPMLE)只适用于风险中性的情形,为此本文把PPMLE方法推广到投标者为风险规避的情形。为便于估计风险规避参数和分布函数的特征参数,本文总结了一个简单的估计过程;为验证估计过程的可行性,本文考虑了四个不同的风险规避参数(包括风险中性情形),分别用蒙特卡罗模拟实验进行拟合效果分析。实验结果表明,无论是对风险中性还是风险规避的情形,本文所提出的估计过程对第一价格密封拍卖模型的结构参数均能做出很好的估计。

A growing number of studies have shown that bidders tend to be risk-aversive, but as the classical structure estimation method of the first-price sealed-bid auction, the improved maximum likelihood estimation （PPMLE） proposed by Donald and Paarsch only applies to the case of risk neutrality. Thus, this paper generalizes the method to the risk aversion case, and gives a simple estimation procedure for risk averse parameters and characteristic parameters of distribution function. In order to verify the feasibility of the estimation procedure, four different risk averse parameters including the risk neutral case are considered. And Monte Carlo simulation is utilized to analyze the goodness-of-fit of parameters and private cost. The experimental results show that the estimation procedure can work well in estimating the structural parameters of the first-price sealed-bid auction for risk aversion as well as risk neutrality cases.