MODELLING EMPIRICAL REGULARITIES FOR BANKING STOCKS IN PAKISTAN

Abstract

This paper examines empirical regularities in terms of dayof- the-week anomalies for the banking stocks at the Karachi Stock Exchange for the period 1996 to 2008. Non normality of the data with excess kurtosis suggested application of ARCH and GARCH models which proved uncertainty of returns from banking stocks. Nonparametric analysis of the data reveals evidence of day-of-theweek effect in Pakistan as contrary to the findings of some of the previous studies on Asia-Pacific markets. The evidence of the dayof- the-week anomaly might be attributed to the strength of the nonparametric estimation methods which are more robust when the data does not meet assumptions of normal distribution. The study also finds relatively higher risk associated with returns on the last day of the working week in general. Negative skewness for most of the return series indicates probability of yielding loss for investors in the banking stocks. Volatility test proves asymmetric results for various banking stocks. Results of the nonparametric tests also reveal significantly different median returns on various days of the week for these stocks. The study finds out relatively greater risk associated with Faysal Bank, Jahangeer Siddiqui Bank, Meezan Bank, National Bank and Prime Commercial Bank on Mondays. For the days in the middle of the week the risks associated with the banking stocks are not asymmetric. For Friday, the closing day of the week, risk in respect of Bank of Punjab, Faysal Bank, Muslim Commercial Bank and National Bank is significantly large.