The Theory of Stock Market Efficiency: Accomplishments and Limitations By Ray Ball
(From Donald Chew’s The New Corporate Finance Recap of article in Managerial Finance (1994 Volume 30 issue 2/3)
Markets are not perfect but pretty tough! Research and the idea of market efficiency have come a long ways in past 30 years. Many of the reported anomalies could be the result of mismeasurements and the failure to incorporate time-varying risks and returns as well as the cost of information.
The basic idea underlying market efficiency is that competition will drive all information into the price quickly. This idea got its start at least in part due to Ball and Brown’s 1968 paper looking at earnings announcements. The authors found that the market had forecasted 80% of the news BEFORE the announcement and the 3 and 6-month returns AFTER the announcement was approximately zero.
Of course following Fama, French, Jensen, and Roll’s 1969 dividend split paper (FFJR), which was the first true “event study,” researchers regularly found the market to be very efficient. These papers helped remove the generally prevailing view that market prices were noisy estimates that could not be trusted let alone used as a means of academic research. The more theoretical models of Modigliani and Miller, Sharpe and Lintner (CAPM), and Black and Scholes (1973) helped this idea that markets were efficient gain support. (Likewise the other models added credibility to the event-study findings.) Fama first defined the term “efficient market” in financial literature in 1965 as a market with a large number of profit-maximizers “with each trying to predict future market values” and “information is almost freely available.” Ball describes the 1969 FFJR paper’s introduction of “event time” as the event “which may well have been the single most important breakthrough in our understanding of how stock prices respond to new information. Ball, while still advocating the efficiency of the market, acknowledges several limitations that were found as researchers poured over the date looking for anomalies. Ball divides the research “into three overlapping categories “Empirical Anomalies: problems in fitting the Theory to the Data”: Price Overreactions (example DeBont and Thaler 1985), excessive volatility (Shiller 1981), Underreaction to good earning announcements, CAPM, Seasonal Patterns “Defects in ‘Efficiency’ as a Model of Stock Markets”. (This is possibly the strongest part of the paper)
The basic idea is that we assume information costs to be zero when in fact they are positive. Everyone recognizes that these costs should be incorporated into any efficiency finding, but as we do not know these costs, we assume them to be zero. Further, even if there are anomalies that researchers find, why publish them? If true anomalies, and not data quirks, then why not trade on the information? Publishing implies the benefit of this information is not great.
What we really need to know is the “expected gains from producing and trading on private information. Ball also makes the humorous comment that a trading-strategy (modeled on past data but using modern computing and statistical techniques) is similar to using a modern war technology to acquire “simulated gains” during the Middle Ages. (Note: for those of you uncertain of this, I suggest you read a Connecticut Yankee in King Arthur’s Court by Mark Twain). The fact that researchers can not detect abnormal gains is not surprising. Supposing marginal benefits = marginal costs, the marginal benefit of an analyst is quite small. Ball lays out an example where the cost of a given forecast is about $2000. This is only a “tenth of a percent” of a $2 billion firm. Ball also points out the obvious: certain investors (such as Warren Buffet) have access to better information and subsequently may have lower required returns.
Ball also points out that analysts may play an important role in reducing uncertainty by reporting on what other investors are doing and how they feel about the firm. “Problems in Testing Efficiency as a Model of Stock Markets” This section looks at problems in testing Market Efficiency Joint Hypothesis Problem: testing efficiency as well as the model of the market Changes in Risk free rate and Risk Premiums: serial correlation may be the result of time varying rates and premiums. Changes in Risk (even seasonal if firms typically announce major events on Monday or in December etc.). As a result, we can not say much definitively.
The last section (before the conclusion) asks “Is Behavioral Finance is the answer?” Ball answers his own question with a negative. He holds this position because (like Fama) he believes that the Behavorialist School has its own anomalies and “grossly inconsistent with competitive markets.” Ball concludes his paper by saying “the theory of efficient markets is, like all theories, an imperfect and limited way of viewing stock markets. The issue will be impossible to solve conclusively while there are so many binding limitations to the asset pricing models that underlie empirical tests of market efficiency.” Further he says that he has lived through the “transformation form the pre-EMH view of securities markets, and [he] is still impressed by how well prices respond to information” relative to what was expected thirty years ago.