Nonparametric estimation of an incomplete model of English auctions: An application to online judicial auctions of used cars
报告人： Nianqing Liu （Shanghai University of Finance and Economics）
地点：Room 217, Guanghua Building 2
This paper proposes to estimate the bounds of valuation distribution by data driven smooth estimators in an incomplete model of English auctions. Bound crossing happens when the upper bound estimate is below the lower bound estimate at some value quantiles. It is a serious issue in a robust inference. To address such an issue, Haile and Tamer (2003) estimated the tight upper and lower bounds by smooth weighted averages of upper and lower bound candidates, respectively, with a prior tuning parameter. We propose a data driven method to choose the tuning parameter minimizing the mean squared error. We also establish the validity of our estimators in large sample and demonstrate their finite sample performance in a simulation study. We then apply our approach to estimate the bounds of value distribution and optimal reserve price in the online judicial auctions of used cars in China. Among our major findings, we first obtain very tight bound estimates of the value distribution. Our nonparametric estimates are shown to be significantly different from a maximum likelihood estimate of value distribution under a lognormal specification. Second, the observed reserve price of an average auction is well below the optimal one when the seller has a high cost, but is above the optimal one otherwise. Third, if the optimal reserve price were implemented, the change of profit as well as probability of no trade vary much across the cost of seller.
About the Speaker:
上海财经大学经济学院数量经济系副教授。2013年获得宾夕法尼亚州立大学经济学博士学位。2013年8月至今在上海财经大学经济学院工作。主要研究领域为应用计量经济学，产业组织和劳动经济学。学术论文发表在Journal of Econometrics，International Economic Review，The Japanese Economic Review，《应用概率统计》，《预测》等中英文学术期刊上。
Your participation is warmly welcomed!