Introduction

Background of the study

It is a common issue to test whether the stock prices are efficient, in most of the researches. It is found that long term returns are more predictable than short term and thus mean reverting.

Studies mainly focus on autocorrelation of returns in different time lengths.

In finance theory it says when autocorrelation of stock prices is there, then prices are predictable. Therefore, it is evidence against the random walk hypothesis of stock prices.

Objective of the study

The objective of this research is to test the efficiency of stock prices in Sri Lanka studying autocorrelation of short and long term returns in the CSE.

Significance of the study

There was a previous research on the predictability of short term returns and mean reversion in long term return in Sri Lanka in two important ways.

1. Previous studies utilized only shorter time periods because lack of long time series, since Sri Lankan stock market commenced only in 1985 and some researchers only used limited data which is the shorter time series available. Whereas the present study uses all-time series of returns available as at the research date (18.5 years from 1985-2003).

2. All previous papers tested only the short term returns except one. But this study examines short term predictability as well as long term mean reversion. It is important to use a long time series when predicting the long term returns to obtain correct results. Therefore, this research contributes to predict both short and long term periods using returns data of a long time series with that data may be more valid and reduces the influences of temporary conditions.

Literature Review

In past many researches have been carried out to identify the predictability of stock returns. These can be divided in to two categories.

1. Short Term Return