January Effect in the .Japanese Stock Market

Abstract

Extensive research finds that returns in January are substantially higher than returns in other months and this pattern of stock returns is known as the "January Effect". The January Effect has been studied in developed as well as developing stock markets. Gultekin and Gultekin ( 1983) studies January return patterns in 17 countries including the United States and finds that for all the countries January returns are higher than other months. Kato and Shallheim (1985) examine excess returns in January for the Tokyo Stock Exchange (TSE). They use both Tokyo Price Index and Nikkei 225 from 1952 to 1980 for the study. W.M Gunaratne Bandara (200 1) finds that January Effect is not evident in the Colombo Stock Exchange. Objective of this study is to test whether the January Effect documented by the Kato and Shallheim ( 1985) is alive in the TSE. This study uses monthly return data of Nikkei 225 for the period from January 1950 to August 2007. To test the January effect, the following regression equation is used. 17. R, =a,+ LfJ,D; + v; { .. /. Where, a" is January average returns. D, is a Dummy variable for month i (February to December) and v; is an identically and independently distributed error term. The estimate of each fJ; represent the average difference in January and other months. The null hypothesis tested is that a" = 0 and fJ; = 0, that is there is no January effect. Findings of the test reveal that January average returns are positive and statistically significant and average differences in January returns from the other months (each other month returns minus January returns) are also negative and statistically significant for the total sample period from 1950 to 2007. However sub sample analysis shows more important evidences than the total sample. Positive January returns are statistically significant and average differences in January returns form the other months are also negative and those are statistically significant for half of the months for the sub samples of 1950 to 1964 and 1965 to 1981. This pattern has changed in the sub sample 1982 to 1995 and 1996 to 2007. In the both sub samples January returns are not statistically significant. In the sub sample 1982 to 1995 December returns slightly higher than the January returns. In the last sub sample average returns of March, June and November are higher than the January returns. Therefore it can be concluded that January anomaly has vanished in the Japanese market after 1980s.