Calendar anomalies; evidence from Pakistan stock market

Abstract

This paper analyzes daily returns of PSX-1 00 index for the period of 17 years started from April 2001 and end to March 2017. The Efficient Market Hypothesis discourages calendar anomalies and recommends that all investors have equal access to information. In an efficient market, no one can outperform the market using any analytical tools. Also, no chances exist to earn excess yields at a particular time. We have tested this concept in the PSX-1 00 index, searching for January and June effect for months, Monday and Friday effect for days. The primary null hypothesis stated that no significant differences exist for daily returns from a specific day of the week or month of the year during the period of study (200 1-20 17). The initial analysis relied on descriptive statistics, and it indicates that Mondays has higher mean returns as compare to others. The skewness and kurtosis values suggested that the data is not normally distributed. We coded dummy variables for use in the ordinary linear regression (OLS), but only Monday returns were significant at 5% level in the PSX- 1 00 index. Moreover, January has significant impact at 10% level of significance but Friday and June has no significant impact on returns. However, additional analysis demonstrated that daily returns have auto correlated residuals. Thus, the linear regression model is not a suitable choice for the data, and so we introduced a GARCH model. After running regression model we used correlogram to confirm the autocorrelation in residuals and also we conducted several analyses to confirm the outcomes of the correlogram, such as normality tests (JB statistic) and a test for ARCH effects. Again, the results confirmed that the daily returns are not normally distributed. But as per regression GARCH model results only Monday have negative impact on daily returns of PSX-1 00 index whereas Friday, January and June has no impact on PSX-100 index return as per OLS and GARCH model. Therefore, the market was not entirely efficient and it can presume that fund managers and market participants can adjust their investment decisions accordingly.