The Capital Asset Pricing Model as a tool for investment decisions in South Africa

Abstract

ABSTRACT The Capital Asset Pricing Model (CAPM) is widely used to estimate required equity returns to inform investment decisions. Significant empirical research has shown that the CAPM does not effectively predict the risk-return behaviour of portfolios of assets or individual assets on a number of different stock exchanges. The problem addressed here is the impact of errors in the CAPM on effective investment decisions, especially for single companies unable to diversify risk. Daily share price data for twenty companies on the JSE, as well as data for various JSE indices and government bonds, were sourced from Bloomberg for the period 1 January 2000 through 31 December 2012. The results show that the CAPM does not effectively predict the risk-return behaviour of the twenty companies considered here, and often leads to significant errors in the prediction of Net Present Value of an investment (by as much as 100 percent). A number of different approaches to application of the CAPM were considered, with the single firm conclusion that the longest possible data period should be used to estimate the parameter (the factor that represents the responsiveness of the company’s return relative to that of the market). The estimation of the equity premium to be used is problematic and makes application of the CAPM difficult. A multiple linear regression using a Fama and French 3-factor model showed that equity returns have a statistical correlation with equity premium, but limited ability to predict future behaviour. Investment decisions underpinned by the CAPM should be taken with consideration of the often large inherent errors.