Long-Run Performance Evaluation: Correlation and Heteroskedasticity-Consistent Tests

Abstract

Although much work has been done on evaluating long-run equity abnormal returns, the statistical tests used in the literature are misspecified when event firms come from nonrandom samples. Specifically, industry clustering or overlapping returns in the sample contribute to test misspecification. We propose a new test of long-run performance that uses the average long-run abnormal return for each monthly cohort of event firms, but weights these average abnormal returns in a way that allows for heteroskedasticity and autocorrelation. Our tests work well in random samples and in samples with industry clustering and with overlapping returns, without a reduction in power compared to the methodologies of Lyon, Barber and Tsai (1999).