A comparison of the forecasting accuracy of the downside beta and beta on the JSE top 40 for the period 2001-2011

Abstract

The purpose of this research report is to determine whether the use of a Downside risk variable – the D-Beta – is more appropriate in the emerging market of South Africa than the regular Beta used in the CAPM model. The prior research upon which this report expands, performed by Estrada (1999; 2002; 2005), focuses on using Downside risk models mainly at an overall country (market) level. This report focuses exclusively on South Africa, but could be applicable to various other emerging markets. The reason for researching this topic is simple: Investors – not just in South Africa, but all across the world – think of risk differently to the way that it is defined in terms of modern portfolio theory. Beta measures risk by giving equal weight to both Upside and Downside volatility, while in reality, investors are a lot more sensitive to Downside fluctuations. The Downside Beta takes into account only returns which are below a certain benchmark, thereby allowing investors to determine a share’s Downside volatility. When the Downside Beta is included as the primary measure of systematic risk in an asset pricing model (such as the D-CAPM), the result is a model which can be used to determine cost of equity, and make forecasts about share returns. The results of this research indicate that using the D-CAPM to forecast returns results in improved accuracy when compared to using the CAPM. However, when comparing goodness of fit, the CAPM and the D-CAPM are not significantly different. Even with this conflicting result, this research shows that there is indeed value in using the D-Beta in South Africa, especially during times of economic downturn.