Robust Standard Errors for Panel Data with Serial Correlation

• Econometrics

Date and Location

Monday, October 11, 2021, 3:40 PM - 5:00 PM
ARE Library Conference Room, 4101 Social Sciences and Humanities

Abstract

We propose new robust standard error estimators for panel data. The new estimators account for the cluster dependence within i, the cluster dependence within t, and serial dependence in the common time effects, whereas each of the existing methods fails to account for one or more of these three features. Simulation studies show that the new standard errors produce robustly superior coverage performance than existing alternatives, including the heteroskedasticity robust estimator (Eicker-Huber-White; also known as HC0), the panel Newey-West estimator, the cluster robust estimator within i, the cluster robust estimator within t, and the two-way cluster robust estimator. We also argue that our proposed method is relevant to empirical analysis in economics and finance based on data used for market value equations.