The impact of incorporating a bond index into the proxy for the market portfolio.
The Capital Asset Pricing model (CAPM) is the most widely used equity valuation model in both the United States of America (U.S.) and South Africa, thus its importance in corporate finance cannot be underestimated. The largest criticism of the CAPM lies in the difficulties with estimating its parameters and in particular the return on the market parameter. Roll (1977) believed that it is impossible to estimate the market portfolio let alone find a good proxy for it. The common trend amongst practitioners is to use a broad based stock index such as the S&P 500 or in South Africa‟s case the All Share Index (ALSI) as a proxy for the market portfolio. However these methods are questionable, as the market portfolio theoretically contains all risky assets held in proportion to their market value, and stock indices ignore large asset classes such as bonds. Furthermore, using a broad based stock index in the South African context ignores South African specific problems such as the supposed segregation of the market to the Resource and Financial and Industrial sectors. Therefore the purpose of this study was to determine whether simply using the broad based stock index, the ALSI, as a proxy for the market portfolio would suffice or whether the inclusion of debt instruments and the acknowledgement of the segregation on the JSE would enhance the proxy‟s performances. First a set of theoretical requirements that a proxy must satisfy to be considered a suitable proxy for the market portfolio were derived. Then a review of literature on the matter was undertaken, which showed that studies in both the U.S. and South Africa had had mixed results. Next, the various proxies were formed, and tested using the two-pass regression method. The two-pass regressions that were run with the model comprising solely of the ALSI as a proxy, produced a negative sloping SML. This result suggested an inverse relationship between risk and return, which is contradictory to the theory set out in chapters two and three. Thus robustness tests were performed on the model, but none solved the problem. Next the proposed multifactor models were tested to see if they would enhance the results of the first model. Although the results improved slightly, they too did not solve the problem. Thus, in conclusion it was found that incorporating a bond index into the proxy for the market portfolio did not significantly enhance the use of the CAPM in South Africa.