Anomaly within the calendar anomalies: a comparative study of the GFC and COVID-19

Purpose In this paper, we examine the January effect (April effect in India), the turn-of-the month (TOM) effect and the day-of-the-week (DOW) effect in small-, medium- and large-stock indices in the Indian stock market using daily data from January 1, 2004, to December 29, 2023. We further attempt to compare and contrast the presence of these calendar anomalies during both the COVID-19 and the global financial crises. Design/methodology/approach To address the issue of volatility in returns clustering over time which may spike sporadically, we use the Generalized Autoregressive Conditional Heteroskedastic regressions to control for conditional heteroskedasticity and the possible effect of outliers. Findings The January effect is evident only in the small-cap stocks, while the TOM effect is significant among the small- and medium-cap stocks. The returns on Wednesday are positive and significantly different from zero; however, the returns on other days are not different from the returns on Wednesday. Finally, we create a trading strategy based on these anomalies which generally outperforms the traditional buy-and-hold strategy. Overall, our results indicate that the calendar anomalies are more pronounced in the small-cap stocks and behave differently during the two crises periods, akin to an anomaly within the reported calendar anomalies. Originality/value First, we provide a pioneer study on the presence of calendar anomalies in the Indian equity market using 20 years of data from January 2004 to December 2023. Second, we intend to capture the size effect in our considered calendar anomalies by analyzing three important stock indices. Third, we attempt to compare and contrast the presence of these calendar anomalies during both the COVID-19 and the global financial crises. Finally, we develop an implied trading strategy based on which the investors could time the market and earn excess returns over the passive buy-and-hold strategy for the considered time period.