文章基本信息

标题：Edgeworth Expansions for Semiparametric Averaged Derivatives - (Now published in 'Econometrica', 68 (2000), pp.931-979.)
本地全文：下载
作者：Y Nishiyama ; Peter M Robinson
期刊名称：Japanese Studies Programme Papers
出版年度：1999
出版社：Suntory Toyota International Centres for Economics and Related Disciplines
摘要：A valid Edgeworth expansion is established for the limit distributionof density-weighted semiparametric averaged derivative estimatesof single index models. The leading term that corrects the normallimit varies in magnitude, depending on the choice of bandwidthand kernel order. In general this term has order larger than the n -½that prevails in standard parametric problems, but we findcircumstances in which it is O(n -½), thereby extending theachievement of an n -½ Berry-Essen bound in Robinson (1995). Avalid empirical Edgeworth expansion is also established. We alsoprovide theoretical and empirical Edgeworth expansions for astudentized statistic, where the correction terms are different fromthose for the unstudentized case. We report a Monte Carlo studyof finite sample performance
关键词：Edgeworth expansion; semiparametric estimates;averaged derivatives