Erasmus university rotterdam

Difficulties with the CAPM

Although the CAPM has dominated finance theory over thirty years, evidence shows that the cross-section of stock returns cannot be described solely by the classical CAPM. The CAPM had a lot of support before the discovery of anomalies in the 70’s. Further research on the CAPM showed that the CAPM performed poorly in explaining stock market returns. Beta alone does not seem sufficient to reflect the risk in the market. As a result researchers have suggested adding other factors to the model to complete the beta in explaining stock price movements. Evidence was found that stock market returns are related to some fundamental factors such as size, book-to-market equity and momentum (Banz, 1981;Rosenberg et al., 1985; Jegadeesh and Titman, 1993). Other studies have shown that average stock returns are related to earnings/price ratios (Basu, 1983), to past sales growth (Lakonishok et al, 1994) and display long term reversals (De Bondt and Thaler, 1985).

To address the anomalies that have been found, alternative models were examined to explain the cross-section of stock returns. It has to be mentioned that some of these anomalies could also be a result of the survivorship bias. This bias is created because data of bankrupt companies are no longer available in databases.

These models take three different directions (Ben Naceur and Chaibi, 2007). The first directions are the multifactor models that add a factor to the market return, such as the Fama and French (1993) model, in which two other variables are added, the return of high book to market value minus low book to market value stocks (HML) and the return on small minus big stocks (SMB). Second, a different asset pricing theory, the Arbitrage Pricing Theory (APT). Thirdly, the non-parametric models that criticize the linearity of the CAPM. These separate directions will be briefly explained hereafter.

Basu (1977) proposes that other factors should be considered besides beta. He states that the price earnings ratio has a great influence on the market return. Fama and French (1993) found that size and book-to-market provide a better explanation of stock market returns than the CAPM. This resulted in the extension of the one factor model to a three factor model, including average stock market sensitivities to size and book-to-market ratio. The three factor pricing model captures most of the market anomalies except the momentum anomaly (Fama and French, 1996). Carhart (1997) proposes a four factor model; he adds momentum to the Fama and French (1993) model as a result of the findings of Jegadeesh and Titman (1993, 2001). Their findings suggest that momentum trading strategies can exploit the phenomenon that stocks that perform the best (worst) over a 3-12 month period tend to continue to perform well (poorly) over the subsequent 3-12 month period. These multifactor models tend to address some criticisms of the classical CAPM, however these additional factors are not motivated by theory (Ben Naceur and Chaibi, 2007)

The APT allows for the individual modeling of macro-economic factors, such as inflation, sovereign, political and exchange risk. This way, typical components of country risk can be modeled as explanatory factors. Evidence shows that in a quasi-efficient market, APT shows a better predictive power than the CAPM (Copeland et al, 1994). The main problem with using APT in emerging markets is the availability of data. Most of the analysts dealing with emerging economies are confronted with data series that are usually incomplete, extremely short, very volatile, and highly unreliable (pereiro, 2002). Only 8% of the corporations and financial advisors use APT when calculating the cost of equity capital (Pereiro 2001)

Non-parametrical models include additional moments into the CAPM (Harvey and Siddique, 2000; Dittmar, 2002). Dittmar (2002) added kurtosis to the three moment CAPM and provide evidence that the inclusion of this factor leads to superior performance. Ben Naceur and Chaibi (2007) argue that investors incorporate skewness and kurtosis in the portfolio decisions in addition to the first two moments.