Chapter 8

Anomalies

I can calculate the motion of heavenly bodies but not the madness of people.

—Isaac Newton

If, as presented in Chapter 4, stock prices are basically a rational forecast of the present value of expected future dividends, then why do observed stock prices vary so much? In two influential papers, Robert Shiller (1981, 1984) argued that stock prices do not reflect values based solely on fundamentals. Had they, then the variability in stock prices would be roughly accounted for by the variability in the underlying fundamentals incorporated in earnings and dividends. Yet, stock prices often exhibit excessive volatility in the absence of news reflected in fundamentals (for example, the crash of October 1987). See Summers 1981 and Cutler et al. (1989) for more on this topic.

As we shall see further on, the documentation of significant departures in prices from fundamentals (so-called anomalies) and the attempt to attribute these departures to cognitive biases, limitations to arbitrage, and boundedly rational behavior has generated a large and diverse literature in behavioral finance. (For an excellent survey, see Barberis and Thaler 2003.) The seeds to this literature can be traced to tests of the CAPM, whose failure as a predictive model spawned the controversial literature on anomalies.

The CAPM that was developed in Chapter 5 posits a relationship between asset returns and the return to the market portfolio in an equilibrium setting in which supply and demand determine market capitalization weights that, in turn, define the market portfolio and where all investors share the same information set and preferences and are mean variance optimizers. These are a rigid set of assumptions. In fact, the observed market portfolio is, at best, a proxy; the S&P 500 for example, or the Russell 3000, or even the Wilshire 5000, are not the universe of all assets and therefore not the market portfolio on which the CAPM rests. In fact, no observed index can pass as the true market portfolio if for no other reason than it cannot include some assets, such as human capital. The observed market portfolio is therefore imperfect—we can hope that it is at least an efficient portfolio—but, in general, the CAPM in theory may be quite different from the CAPM in practice.