The international capital asset pricing model : empirical evidence for South Africa.

Abstract

An integral component of all corporations‘ financial operations is the determination of the cost of equity of the firm. This input is required in many financial decision making processes, and the correct estimation of this value is therefore a very important issue. The Capital Asset Pricing Model (CAPM) of Sharpe (1964) and Lintner (1965) has filled this gap since its inception, and has been extensively used by both corporations and individuals in their estimation of expected return. Whilst the standard form of this model is intuitive and simple in its implementation, an additional issue faced when utilising it in the current day is that of global financial integration. Whilst the CAPM is suitable for use in a market which is completely segmented from the rest of the world, this is often not the case as the barriers across countries have gradually declined, with the result that much of the world is now internationally integrated. This therefore led to two extensions of the CAPM to the international environment by both Solnik (1974) and Grauer, Litzenberger and Stehle (1976). Whilst both are referred to as International CAPM (ICAPM) models, the difference lies in that Solnik‘s (1974) model incorporates the presence of exchange rate risk, whilst the Grauer, Litzenberger and Stehle (1976) one does not. This study therefore provides an analysis of the suitability of these two models to the South African environment, along with a comparison of the relative performances of each model against that of the standard CAPM model. The three different methods of analysis used are: the unconditional approach, a conditional GARCH approach, as well as the cost of equity approach. The analyses are applied to the data which consists of all listed firms on the JSE from 1990 up to 2010, with multiple methods of evaluation employed, such as information criteria and forecasting, in order to provide a robust analysis of all three models. The results of the analysis vary across the different methods used, however since a significant amount of evidence was found of the International CAPM models, it can be concluded that an international asset pricing model should be used instead of a domestic one. In the choice between the single-factor ICAPM model and the multifactor ICAPMEX, even though use of the Grauer et al (1976) model would not be inappropriate, it was concluded that use of Solnik‘s (1974) ICAPMEX model would be the best suited to the South African financial environment, as the presence of exchange rate risk factors in an asset pricing model is found to be an important inclusion which may lead to better cost of equity estimates.