Testing Stock Market Efficiency and the Asset Pricing Model: Some Evidence from Sri Lanka

Abstract

This paper examined the Price-Earnings anomaly and the Capital Asset Pricing Model (CAPM) for the Colombo Stock Exchange (CSE) in Sri Lanka over the period 2004 to 2013. Stock market anomalies inevitably challenge the validity of the Efficient Market Hypothesis and the accuracy of the asset pricing model used in measuring stock returns. Therefore, the accuracy of the asset pricing model used to measure stock returns is in the centre of the market efficiency debate, and the CAPM is an extensively applied asset pricing model in the Sri Lankan capital market. A stochastic version of the CAPM time-series regression was therefore adopted in this study to estimate portfolio abnormal returns as measured by Jensen’s Alpha. The analysis found systematically high abnormal returns from portfolios of stocks with a low Price-Earnings ratio relative to the portfolios consisting of stocks with a high Price-Earnings ratio. This suggests anomalous pricing behaviour in capital market operations in Sri Lanka. Price-Earnings anomaly hypothesises either the CAPM is mis-specified and/or capital markets are inefficient. If the CAPM appears to be a correctly specified asset pricing model that accurately explains stock returns in response to their true level of systematic risk, Price-Earnings anomaly evidently undermines the semi-strong form stock market efficiency. To shed light on this joint hypothesis problem, the study investigated the accuracy of the CAPM. Results of a two-pass regression approach similar to Black, Jensen and Scholes (1972) and Fama and Macbeth (1973) empirical work indicated that the CAPM-beta does not appropriately explain stock returns in Sri Lanka. As such, it appears that the CAPM suffers from omitted variable bias problem and hence mis-specified. In the absence of statistically significant evidence to constitute accuracy of the CAPM, Price-Earnings anomaly observed in this study appears to have been produced either by measurement inconsistencies of the mis-specified CAPM or by a combined effect of both CAPM-misspecifications and capital market inefficiency. From the policy front, market efficiency is an important attribute and the accuracy of the asset pricing model is a vital strand in market efficiency tests. An investigation of market anomalies using alternative asset pricing models such as Fama-French (2015) five-factor model might be a potential area for future research. Asset pricing models are being constantly investigated and these studies may find further specifications in the contemporary asset pricing model.