The capital asset pricing model (CAPM) is the default model for risk in equity valuation and corporate finance. It has retained this position since it’s introduction by Jack Treynor in 1961. Unusually this is a position it has maintained despite it having been proven it is not the most accurate model by a long shot.
This is reflected in the graph below plotting CAPM predicted returns against actual excess returns.
In Graham and Harvey’s 2000 survey we see that around 74% of CFOs still use CAPM in their work, and in ‘The Consensus Estimate for the Equity Premius’ by Academic Financial Economists in 2007, 75% of finance professors recommend CAPM for corporate capital budgeting purposes.
What is it that has allowed CAPM to maintain its position despite a long list of shortcomings and detractors?
It is not some sort of sentimental value that has allowed CAPM to hold it’s position of importance in the financial world, but the inherent failings associated with other models developed to this point.
Through exploring the prominent alternatives that have emerged as opponents to the CAPM we can see that while offering statistically significant results in comparison to CAPM, the extra work associated with finding these inputs do not yield proportionally substantial results. Simply put, the work associated with CAPM’s alternatives outweighs any improvements in results from these models.
The simplicity of the inputs associated with the CAPM is a double-edged sword. The risk free rate, beta and expected return on the market portfolio are the foundations of the CAPM and are relatively easy to calculate. Their simplicity to calculate is marred by their overlooking of several areas that have proved significant in calculating risk in equity valuation and corporate finance.
This is where alternatives such as the Fama-French and Carhart Four-Factor model have made inroads.
Fama and French’s model has made the greatest inroads into displacing CAPM yet it is still a long way from being the ‘go to’ model for analysts and academics. They started with the observation that some stocks tended to outperform the market. Small caps and stocks with a high book-to-market ratio were seen as these regular outperformers. Based upon these observations Fama and French constructed a 3 factor pricing model.
This model was based upon the idea that variations in market returns can be explained by exposure to three elements – the marketplace as a whole, small stocks and value stocks.
In their 1993 paper Fama and French claim that their 3-factor model can explain over 90% of the variability in returns in comparison to the rough figure of 70% accounted for by CAPM.
For an example of the 3 factor model’s effectiveness in comparison with CAPM we can look at the equal weight Centre for Research in Security Prices plugged into the models and compared by a fellow blogger and advocate for the Fama-French model.
We can see here the alpha’s vary greatly in the two outcomes. The alpha of the EW CRSP index when put through the Fama-French model being roughly 10% of the CAPM alpha.
It’s clear to see from this example that the Fama-French model offers a much fuller view of what is driving returns when compared to the CAPM.
A rational thinker, unfamiliar with the world of finance, would surely see the Fama-French model as superior to CAPM and would wonder as to why it is not favoured by those in the field of finance. This is where the worlds of academia and application collide. Fama and French’s model proves significant on paper but the real world application of models to the world of finance is more than just results based. The time taken to calculate all the essential inputs for the model is often deemed relatively more costly than the results achieved.
This flaw in the application of models to real world analysis extends to other models such as the 4-factor pricing model.
The fuller views, compared to CAPM, explaining what is driving returns have been seen to get even better when exposed to the Carhart 4-factor pricing model in some cases. This model takes into account the system of momentum investing or ‘fair weather investing’, which has been reported to yield average returns of 1% per month by some academics.
Unfortunately there is no consensus as to the importance of the results shown from this model, which further cements CAPM’s position as the consistent yet flawed model of choice.
The Arbitrage Pricing Model has been called the ‘supply-side’ response to the CAPM’s ‘demand-side’ basis. It is a less restrictive model that assumes an investor will hold a unique portfolio instead of the market portfolio assumed in CAPM. It has been said that the CAPM is a special form of APM.
CAPM has maintained its advantage over the Arbitage Pricing Theory due to the simplicity of its inputs compared to the more complicated nature of the APM.
Having looked at the effectiveness of CAPM in comparison to it’s competing models whether they be APM or a multifactor model it’s clear that CAPM is not the model that offers the clearest picture in equity valuation and corporate finance. This does not change the fact that it remains the default model for risk in these areas.
The CAPM has survived due to its simple and effective nature. The alternative models have a clear advantage in explaining past return but their use in forecasting future returns are about as effective as the CAPM.
It is well and good to herald another pricing model as more effective in the academic world, but in the application of financial know-how to real world markets the expected returns from alternate models are not substantially different enough to warrant the complicated job of calculating up to four additional betas.
Analysts are under no illusions as to the effectiveness of CAPM, but concede that its benefits outweigh the negatives for now.
This has been the saving grace for the CAPM and until the day where a model who’s extra work yields proportionally better results CAPM will survive as the default model for risk in equity valuation and corporate finance.
“Returns to buying winners and selling losers: Implications for stock market efficiency,” Jegadeesh, Narasimhan and Titman, Sheridan. Journal of Finance 48 (1993).
“Common risk factors in the returns on stock and bonds,” Fama and French. Journal of Financial Economics, 33 (1993).
“The Consensus Estimate for the Equity Premium,” Welch, I. Academic Financial Economists, (2007).
“How to use the Fama French Model,” Empirical Finance Blog.
“Toward a Theory of Market Value of Risky Assets” Treynor, J.L. Asset Pricing and Portfolio Performance: Models, Strategy and Performance Metrics, (Unpublished until 1992), (1962).
“Grading the performance of market timing newsletters,” Graham and Harvey,
Association for investment management and research, (1997).
Prof. Brian Lucey, TCD. JS Applied Finance lecture slides.