Predictability of Turn-of-the-Month Effect at Stock Markets in Malaysia, South Korea and Japan

Abstract

Financial market efficiency has been constantly challenged with new evidence of calendar anomalies in the stock market. This paper examines the existence and predictability of the turn-of-the-month effect at Kuala Lumpur Composite Index (KLCI), Seoul Composite Stock Price Index (KOSPI) and Tokyo Nikkei 225 Stock Average (JNK) by employing Kalman filter technique. The advantage of using the Kalman filter technique is that it allows the mean returns to change over time and from the analyses, it can be said that the turn-of-the-month effect exists in more than fifty percent of the months during the 16-year period from 1995 until 2010 for the three indices. The findings from the Kalman filter model contradict the results from the regression model and that it also shows that there is inconsistency in the absence or existence of the turn-of-the-month effect. Further analyses show only little evidence for predictability for the occurrence of the turn-of-the-month across the sample period for the three countries.

Keywords: Calendar anomalies, Efficient Market Hypothesis, State-space representation

JEL Classification: C22, G14, O16