Trading Day Effect and Volatility Clustering in NEPSE Returns: An Empirical Analysis of Market Anomalies

Abstract

This study examines the seasonality of stock returns across trading days in Nepal’s stock market. The analysis is based on 4,504 trading days of the NEPSE composite index from 2005 to 2024. The study utilized descriptive statistics, OLS regression, and EGARCH (1,1) estimation incorporating weekday dummy variables. A nonparametric Kruskal–Wallis H Test is also employed to measure the robustness of the results. The study found negative average returns on Sundays and Mondays, and positive returns on Tuesdays, Wednesdays, and Thursdays. NEPSE operates five days a week, with Friday and Saturday being closed. During the study period, mean returns on Wednesday were higher than on any other day of the week. The ANOVA (one-way) results report a significant difference in mean returns on Wednesday and Sunday. The regression results reveal that the returns vary systematically across weekdays. The study found significantly higher returns on Wednesday and Thursday compared to Sunday, indicating seasonality in returns across operating days in a week. The results of the EGARCH (1,1) model, significant ARCH and GARCH coefficients indicate that shocks immediately increase volatility and that volatility persists over time, confirming both short-term clustering and long-term dependence in NEPSE returns. However, the insignificant volatility term (γ) suggests there is no asymmetric volatility response to negative shocks. The findings of this study are beneficial for stock market investors in Nepal, suggesting that they should buy stocks at the start of the week and sell them by the end.