More Accurate Inference with Small Bandwidths

• Activity Number: 600
• Type: Invited
• Date/Time: Thursday, August 5, 2010 : 8:30 AM to 10:20 AM
• Sponsor: IMS
• Author(s): Whitney Newey*+
• Companies: MIT
• Keywords: matching ; partially linear model ; inference

Abstract

Recently "adjusted" standard errors have been derived for a number of semiparametric estimators. They add an additional term to the variance that accounts for small bandwidths or many terms in series estimation. Examples include matching estimators, weighted average derivative estimators, partially linear model estimators, and instrumental variables. This paper shows that these methods have a common structure that is based on asymptotic normality of quadratic forms. The paper also shows that the adjusted standard errors can lead to more accurate inference for a range of bandwidths. Both theoretical and simulation results are reported.