Improving the Statistical Power of Financial Event Studies: The Inverse Variance Weighted Average Based Test
Posted: 14 Oct 2008 Last revised: 22 Jan 2011
Date Written: June 9, 2008
Abstract
Some claim that the event-study methodology literature is "mature". But the overlooked IVWA-based test is more powerful than conventional tests. Its implementation requires the same inputs as the traditional test. Its functional form yields the power improvement. Using CRSP monthly data for 1944-1971 and 1980-2006, simulations indicate that IVWA-based test correctly rejects the null hypothesis substantially more frequently than the traditional test. Its superiority seems more pronounced over 1980-2006. This casts doubts over previous event-studies that failed to reject the null. They may have done so incorrectly due to lack of statistical power. Their reevaluation under IVWA-based test is advisable.
Keywords: event study, simulation
JEL Classification: G14
Suggested Citation: Suggested Citation