The Day-of-the-Week Effect: Anomaly or Illusion? New Evidence from Sri Lanka’

In this paper the Day-of-the-Week effect is studied by comparing the empirical cumulative distribution functions (ECDFs) of the close-to-close log-returns of one day-of-the-week to the remaining weekday log-returns amongst the stocks comprising the S&P Sri Lanka 20 Index for the period January 2006 to December 2012. The sample period for the study is sub-divided into two periods: the civil war period (1st January 2006 to 18th May 2009) and the post civil war period (19th May 2009 to 31st December 2012). A modified Kolmogorov-Smirnov statistic, called the DoW-statistic, was then used to define the degree of DoW efficiency. The DoW-statistic for each stock is subsequently used to rank the S&P Sri Lanka 20 Index stocks according to their DoW efficiencies. The classical Kolmogorov-Smirnov test is then used to ascertain the statistical significance of these DoW statistics. With no adjustments for ARCH effects, the results indicate that DoW effects were weak during the war period but strong during the post-war period. However, when potential ARCH effects were taken into account, both periods showed evidence of strong DoW effects, indicating the Day-of-the-Week effect to be an anomaly rather than an illusion.