• Critical assumptions of CAPM
• Empirical tests of CAPM
• Readings on CAPM

Critical assumptions of CAPM

The CAPM is simple and elegant. Consider the many assumptions that underlie the model. Are they valid?

• Zero transaction costs. The CAPM assumes trading is costless so investments are priced to all fall on the capital market line. If not, some investments would hover below and above the line -- with transaction costs discouraging obvious swaps. But we know that many investments (such as acquiring a small business) involve significant transaction costs. Perhaps the capital market line is really a band whose width reflects trading costs.
• Zero taxes. The CAPM assumes investment trading is tax-free and returns are unaffected by taxes. Yet we know this to be false: (1) many investment transactions are subject to capital gains taxes, thus adding transaction costs; (2) taxes reduce expected returns for many investors, thus affecting their pricing of investments; (3) different returns (dividends versus capital gains, taxable versus tax-deferred) are taxed differently, thus inducing investors to choose portfolios with tax-favored assets; (4) different investors (individuals versus pension plans) are taxed differently, thus leading to different pricing of the same assets.
• Homogeneous investor expectations. The CAPM assumes invests have the same beliefs about expected returns and risks of available investments. But we know that there is massive trading of stocks and bonds by investors with different expectations. We also know that investors have different risk preferences. Again, it may be that the capital market line is a fuzzy amalgamation of many different investors' capital market lines.
• Available risk-free assets. The CAPM assumes the existence of zero-risk securities, of various maturities and sufficient quantities to allow for portfolio risk adjustments. But we know even Treasury bills have various risks: reinvestment risk -- investors may have investment horizons beyond the T-bill maturity date; inflation risk -- fixed returns may be devalued by future inflation; currency risk -- the purchasing power of fixed returns may diminish compared to that of other currencies. (Even if investors could sell assets short -- by selling an asset she does not own, and buying it back later, thus profiting from price declines -- this method of reducing portfolio risk has costs and assumes unlimited short-selling ability.)
• Borrowing at risk-free rates. The CAPM assumes investors can borrow money at risk-free rates to increase the proportion of risky assets in their portfolio. We know this is not true for smaller, non-institutional investors. In fact, we would predict that the capital market line should become kinked downward for riskier portfolios (ß > 1) to reflect the higher cost of risk-free borrowing compared to risk-free lending.
• Beta as full measure of risk. The CAPM assumes that risk is measured by the volatility (standard deviation) of an asset's systematic risk, relative to the volatility (standard deviation) of the market as a whole. But we know that investors face other risks: inflation risk -- returns may be devalued by future inflation; and liquidity risk -- investors in need of funds or wishing to change their portfolio's risk profile may be unable to readily sell at current market prices. Moreover, standard deviation does not measures risk when returns are not evenly distributed around the mean (non-bell curve). This uneven distribution describes our stock markets where winning companies, like Dell and Walmart, have positive returns (35,000% over ten years) that greatly exceed losing companies' negative returns (which are capped at a 100% loss).

Empirical tests of CAPM

How well does the CAPM hold up under the empirical microscope? Testing the CAPM is difficult because the model purports to compare expected risk and expected return. Our observations, however, can only be of actual results -- actual variability and actual returns. Nonetheless, if we assume over time that expected results match actual results, we can do some testing.

Annual Returns - Investment Types (1926-1992)

Source: Ibbotson Associates, Stocks, Bonds, Bills and Inflation (1992 Yearbook).

Readings on CAPM

To get a flavor for economic studies of CAPM, you might want to browse one of these important articles --

• Eugene Fama & Kenneth French, The Cross-Section of Expected Stock Returns, 47 J. Fin. 427 (1992) (collecting studies)
• more recent Fama study / most downloaded on SSRN