Calendar anomalies in the stock markets: conventional vs Islamic stock indices

Purpose In this study, the authors evaluate seven calendar anomalies'-the day of the week, weekend, the month of the year, January, the turn of the month (TOM), Ramadan and Eid festivals-effects in both the conventional and Islamic stock indices of Bangladesh. Also, the authors examine whether these anomalies differ between the two indices. Design/methodology/approach The authors select the Dhaka Stock Exchange (DSE) Broad Index (DSEX) and the DSEX Shariah Index (DSES) of the DSE as representatives of the conventional and Islamic stock indices respectively. To carry out the investigation, the authors employ the generalized autoregressive conditional heteroskedasticity (GARCH) typed models from January 25, 2011, to March 25, 2020. Findings The study's results indicate the presence of all these calendar anomalies in either conventional or Islamic indices or both, except for the Ramadan effect. Some significant differences in the anomalies between the two indices (excluding the Ramadan effect) are detected in both return and volatility, with the differences being somewhat more pronounced in volatility. The existence of these calendar anomalies argues against the efficient market hypothesis of the stock markets of Bangladesh. Practical implications The study's results can benefit investors and portfolio managers to comprehend different market anomalies and make investment strategies to beat the market for abnormal gains. Foreign investors can also be benefited from cross-border diversifications with DSE. Originality/value To the authors' knowledge, first the calendar anomalies in the context of both conventional and Islamic stock indices for comparison purposes are evaluated, which is the novel contribution of this study. Unlike previous studies, the authors have explored seven calendar anomalies in the Bangladesh stock market's context with different indices and data sets. Importantly, no study in Bangladesh has analyzed calendar anomalies as comprehensively as the authors'.