On Structural Changes in the Holiday Effect

Abstract

From 1926 to 2016, the average stock return on the day before holiday market closings is up to 15 times the average return on all the other days of the year. We study whether this holiday effect is contingent on the subperiod over which it is estimated and locate the critical break dates that delineate each subperiod. We find the holiday effect is critically dependent on the sample period over which it is estimated and that there is no statistically consistent set of results in each subperiod. Nevertheless, the holiday effect is statistically significant in the CRSP value-weight stock market portfolio and in the low-size stock portfolio in every subperiod from 1926 to 2016.