Does the Colombo Stock Exchange Overreact?

Abstract

The efficient markets hypothesis has been the central theme in finance for nearly 35 years. Fama (1970) defined an efficient financial market as one in which security prices fully reflect the available information. However behavioralists such as Daniel, Hirshleifer and Subrahmayam (1998) Barberies, shleifer and Visny (1998), argue that financial markets are not efficient due to two investor biases These are over reaction and under reaction. De Bondt and Thaler (1985) have written several papers in which they argue that investors over react New York Stock Exchange. Particularly they find that stocks that are the most extreme losers in the past 3 – 5 year period have abnormally high returns in the subsequent 3 – 5 year period and vice versa. Hence they argue that investors over react for the extremely good news and extremely bad news. Objective of this study is to test whether the long term overreaction hypothesis is predictive in the Colombo Stock Exchange. This study uses monthly return data of common stocks for Colombo Stock Exchange for the period between January 1993 to December 2005. De Bondt and Thaler define the over reaction hypothesis as (~ | ) 0 1 < Wt t− E u F and (~ | ) 0 1 > Lt t− E u F . The first term explains that residual returns of high returns portfolio formed conditional upon the information set at t-1 are negative and the second term explains that residual returns of low returns portfolio formed conditional upon the information set at t-1 are positive. In this study t period has been taken as 24 months and 36 months. The residuals are estimated as it jt mt U~ = R − R . There is no risk adjustment except for movement of the market as a whole and the adjustment is identical for all the stocks. For every stock j cumulative excess returns (CUj) are computed for the prior 36 months and formed in to 5 portfolios. The highest CU portfolio is named as the Winner and the lowest CU portfolio as Loser. Then, for both Winner and Loser next 36 months Cumulative Abnormal Returns (CAR) are computed. If a security’s return is missing more than 80% of the months with in a period that security is removed from the portfolio and the CARs in order to handle the thin trading problem. Findings of the study shows that past period losers are mostly generating positive returns in the subsequent 36 months and only 10 months show significant returns. However prior period winners reflect on average positive returns also in the next period and all of these returns are statistically significant after the 25th month. Therefore it can be concluded that there is no long term over reaction to new information at the Colombo Stock Exchange.