Contrarian Investment Strategies

by David Dreman

  The computer proved fickle. Though helping the chartist when in his hands, it was also turned against him. In one such test, a computer program analyzed 548 stocks trading on the New York Stock Exchange over a five-year period, scanning the information to identify any one of thirty-two of the most commonly followed patterns, including “head and shoulders” and “triple tops and bottoms.” It was programmed to act on its findings as a chartist would. It would, for example, buy on an upside breakout after a triple top, a strong technical indicator that the stock would go higher; or it would sell after the market had plunged through the support level of a triple bottom, indicating that the stock would drop lower. The computer measured its results, based on these signals, against the performance of the general market. No correlation was found between the buy and sell signals and subsequent price movements. Once again, our old friend the buy-and-hold strategy would have worked just as well.

  Price-volume systems met with the same fate. Although this is an important technical tool, the size of neither price nor volume changes appears to have a bearing on the magnitude or direction of the future price; stocks going down on heavy trading may reverse themselves and go up in the next period, as may stocks currently going up on large volume.13

  All the tests indicated that mechanical rules do not result in returns any better than the simple buy-and-hold strategy.14 The evidence accumulated is voluminous and strongly supports the random walk hypothesis. With a very minor caveat, some tests have shown a degree of dependence (nonrandom price movements), indicating that a number of marginally profitable trading rules and small filters appear to work consistently. The problem is that the numerous transactions involved in such systems generate substantial commissions, which absorb the expected profit.15

  In numerous tests, then, no evidence to date has been able to refute the random walk hypothesis. Technicians nonetheless claim that their methods work, and if you look at their examples they certainly appear to. But, as we have seen, their success is actually only chance in accordance with the laws of averages. Also, of course, their methods work much better with hindsight. Technicians, being human, forget their “misses” and remember their “hits.” If they were wrong, the cause wasn’t the basic technique but its misapplication or the fact that another application or supplementary information was required. Technicians have also claimed that some technical systems work, supported by computer evidence of a correlation over certain periods of time. Undoubtedly this is true, but when a portion of these results was tested more thoroughly, using different time periods and more extensive price information, the correlation disappeared, again showing that the results of the systems were based simply on chance.

  When all is said and done, it is impossible to absolutely prove that the random walk hypothesis never works, for that would mean testing not only all the hundreds of systems but also the hundreds of thousands of possible combinations, with the final decision depending on the technician’s own interpretation. An enormous number of tests would be required to do this. Technicians can, quite rightly, say that not all systems have been examined and that in any case their decisions were not based on any one method but were the outcome of judgment and experience. Still, from the substantial evidence accumulated, no system of technical analysis has as yet been found that can put a dent in the random walk hypothesis.

  Even though the academic findings have been strongly refuted, chartists and other technicians continue to flourish. They disregard the findings—if they are aware of them—because “their” system is different and hope their clients also ignore the research. Occasionally they let off steam at their antagonists, but usually their protests are without factual support.

  When they are not otherwise warning investors to beware, the academics appear to regard technicians with a detached amusement that some may reserve for witch doctors or primitive soothsayers in bygone cultures (I won’t include astrologers today). This tough and dedicated cult of financial forecasters has been taking its lumps for many years, not only from academics but also from the proponents of fundamental analysis. Some unkind fundamental analysts I have known would go so far as to propose a new experiment to the academics: a survey would be taken of shiny suits, frayed collars, and sundry holes in the attire of a sample of technicians, to be measured against a control sample of other Wall Streeters. They believe the findings would show technicians to be far the worse for wear and tear, since most tend to follow their own pronouncements.

  Even so, many fundamental analysts are part-time dabblers in the technical mystique. Although fundamental analysis is dominant on Wall Street, most members of this group at one time or another take a peek “to see what the charts tell them,” probably more often in periods of crisis but also as a final affirmation of a decision to buy. We find that in spite of the accumulation of evidence for more than five decades on the unproductiveness of technical analysis, it continues to be widely practiced by investors.

  The Academic Blitzkrieg Rolls On

  Unfortunately for most money managers and analysts, the academics did not rest on their laurels after this one rather clear-cut victory. Beginning in the mid-1960s, a much more ambitious operation was launched when the researchers asked whether fundamental analysis, the gospel of the large majority of Wall Street professionals, is of any use in obtaining above-average returns in the market.

  The fundamentalist school believes that the value of a company can be determined through a rigorous analysis of its sales, earnings, dividend prospects, financial strength, and competitive position, and other related measures. Fundamental analysts have been trained in its many complex nuances and applications, in both undergraduate and graduate schools,*17 and have expanded their knowledge through daily application in their work. Such analysis is used by the great preponderance of mutual funds, bank trust departments, pension funds, and investment advisers, as well as most brokers.

  Yet despite their impeccable credentials, the money managers’ record has been anything but awe-inspiring over the years. One of the first studies of money managers’ results was undertaken by the SEC, which measured the performance of investment companies from the late 1920s to the mid-1930s. The report stated, “It can then be concluded with considerable assurance, that the entire group of management investment companies proper (closed end funds) failed to perform better than an index of leading common stocks, and probably performed worse than the index over the 1927–1935 period.”16 Alfred Cowles, an American economist and businessman who founded the Cowles Commission for Research in Economics Research and its journal, Econometrica, analyzed the performance of investment professionals in 1933 and concluded that stock market money managers do not beat the market. The study was updated in 1944, with the same results.17

  Numerous scholarly studies of the lackluster performance of money managers were published in the 1960s and 1970s. One of the most exhaustive and devastating studies was by Irwin Friend, Marshall Blume, and Jean Crockett of the Wharton School in 1970.18 The report was widely read and discussed by both academics and professional investors. One hundred thirty-six funds produced an average return of 10.6 percent annually between January 1960 and June 30, 1968. During the same period the shares on the New York Stock Exchange produced an average return of 12.4 percent annually.

  If fundamentalists were perplexed by why their results weren’t better, the academics certainly were not. Starting in the mid-1960s, they shifted their research firepower from technical to fundamental analysis. Extensive studies have been made of mutual funds and other professional performance results, the great majority of which subscribe to security analysis. The professors demonstrated once again that the funds and, for that matter, other large accounts of money managers did not outperform the market.19

  A powerful cast of financial academics led the charge of EMH against the then-conventional wisdom, including the Nobel laureates Marshall Blume, Merton Miller, William Sharpe, and Myron Scholes, as well as distinguished academics such as Professors Euge
ne Fama and Richard Roll.

  Academic analysis proved as unsparing of the fundamental practitioner’s sensibilities as of the technician’s. Other prevailing beliefs were treated as harshly. No link was found between the widely held belief that higher portfolio turnover led to better performance. Rapid turnover does not improve results but seems to damage them slightly. No relationship was found between performance and sales charges, although mutual funds with higher sales charges often claimed they provided better results.20 In sum, the reports firmly concluded, mutual funds do not outperform the market.

  The results hardly comforted fund managers, who like to represent to clients that they provide superior performance. Mutual fund managers not only underperformed the market but, if we adjust for risk as the academics defined it, often fared worse. At the time, the money managers appeared to be routed as badly as the Inca armies who fled from Francisco Pizarro and his terrifying cannon and cavalry.*18 But, as we’ll see, it’s one thing to overthrow the ideology of the native rulers with computer “horses,” and another to rule over the theoretical world.

  The Efficient-Market Legions Surge Further

  As we’ve seen, the academic investigators proposed a revolutionary new hypothesis that we briefly examined earlier, the efficient-market hypothesis (EMH), which holds that competition between sophisticated, knowledgeable investors keeps stock prices where they should be. This happens because all facts that determine stock prices are analyzed by large numbers of intelligent and rational investors. New information, such as a change in a company’s earnings outlook or a dividend cut, is quickly digested and immediately reflected in the stock price. Like it or not, competition by so many market participants, all seeking hidden values, makes stock prices reflect the best estimates of their real worth. Prices may not always be right, but they are unbiased, so if they are wrong, they are just as likely to be too high as too low.

  Since meaningful information enters the marketplace unpredictably, prices react in a random manner. This is the real reason that charting and technical analysis do not work. Nobody knows what new data will enter the market, whether they will be positive or negative, or whether they will affect the market as a whole or only a single company.

  A key premise of the efficient-market hypothesis is that the market reacts almost instantaneously (and correctly) to new information, so investors cannot benefit. To prove this contention, researchers conducted a number of studies that they claimed validated the thesis.

  One important study explored the market’s understanding of stock splits. In effect, when a stock is split, there is no free lunch—the shareholder still has the same proportionate ownership as before. If naive traders run up the price, said the academics, knowledgeable investors will sell until it is back in line, and market efficiency will be proved. And, said the researchers, this is indeed the case. Tests have confirmed that stock prices after a split was distributed maintained about the same long-term relationship to market movements as before the split.21

  Another study measuring the earnings of 261 large corporations between 1946 and 1966 concluded that all but 10 to 15 percent of the data in the earnings reports were anticipated by the reporting month, indicating the market’s awareness of information.22 Other tests came up with similar results, demonstrating, the professors said, that the market quickly adjusts to inputs.

  Did these tests actually prove what they claim and cinch the case that markets react quickly to new information? Remember these and other EMH cornerstones. We’ll see shortly whether these tests, along with many others, will boomerang back on the researchers who tossed them so confidently at investors.

  The Capital Asset Pricing Model: EMH Victory or Trojan Horse?

  With the stock market declared efficient, the theorists could tell investors they should expect only a fair return—that is, a return commensurate with the exact risk of purchasing a particular stock. The capital asset pricing model (CAPM), a younger brother of EMH, comes into play here. Risk (as CAPM defines it) is volatility. The greater the volatility of the security or portfolio, measured against the market, the greater the risk.

  The most common quantitative measure of volatility is designated as beta.*19 It can be the volatility of a mutual fund, a portfolio, or a common stock. To calculate it, a mutual fund, a portfolio, or a stock must be measured against a benchmark, normally a stock index. Stock portfolios of larger-capitalization mutual funds or other similar portfolios are often measured against the S&P 500, which is assigned a beta of 1. If the mutual fund has a higher beta, the academics say, it is more risky than the index. Conversely, a lower beta is considered less risky. Over time, risk and return must always be in line, say the theorists. Securities or portfolios with greater risk should provide larger returns; those with less risk, lower returns. Thus the money manager whose portfolio outperforms the market by 3 percent a year might have a much higher beta for his portfolio. Since the academics assume that there is a direct correlation between risk (volatility) and return, after adjusting for volatility, the manager who outperforms the market by 3 percent might actually be doing worse than a manager who outperforms by 1 percent. The academics call this measurement risk-adjusted return.

  The efficient-market hypothesis, elegant in its simplicity, is intuitively appealing because it explains the single most obvious mystery about investing: how can tens of thousands of intelligent, hardworking professional stock pickers be endlessly outwitted by the market and embarrassed by their selections?

  EMH has much wider implications than the random walk hypothesis,*20 which said only that investors would not benefit from technical analysis. If correct, the new argument tears the heart out of fundamental research. No amount of fundamental analysis, including the exhaustive high-priced studies done by major Wall Street brokerage houses, will give investors an edge. If enough buyers and sellers correctly evaluate new information, under- and overvalued stocks will be rare indeed.23

  The implications are sweeping. If you’re in the market, the theory tells you to buy and hold rather than trade a lot. Trading increases the commissions you pay without increasing your return. The theory also tells you to assume that investors who have outperformed the market in the past were just plain lucky and that you have no reason to believe they will continue to do so.24

  The semistrong form of EMH contends that no mutual fund, million-dollar-plus money manager, or individual investor, no matter how sophisticated, can beat the market using public information. This is the most widely accepted form of EMH today.

  If the “weak form” of EMH, the random walk hypothesis, made some dramatic claims, the “strong form” doubled down. The stronger form claims that no information, including that known by corporate insiders or by specialists trading the company’s stock (who have confidential material about unexecuted orders on their books), can help you outperform the market. In the few studies done to date, some evidence has surfaced that both insiders25 and specialists26 display an ability to beat the market. However, the strong form of EMH is generally considered too extreme and is not widely accepted.

  Further Academic Backing of EMH

  Professor Eugene Fama,*21 the leading advocate of EMH (Fortune magazine once referred to him as the Solomon of stocks), reviewed the literature and development of EMH in December 199127 and again in 1998.28 Fama’s reports were thorough, covering hundreds of papers published since his last major review twenty years earlier. The papers strongly supported the semistrong form of EMH.

  Despite the tens of thousands of academic papers written about EMH in the last forty-five years, relatively little new research supporting the efficient-market hypothesis has been produced, with two exceptions. Some new studies show daily and weekly predictability in price movements from past movements. But after transaction costs, there is nothing much left to put in your wallet.

  A second area of new support for efficient markets, according to Fama, comes from event studies, the study of specific events and how they affect a stock or the
market. In the past twenty years, hundreds of such studies have been undertaken. Fama concludes that “on average stock prices adjust quickly to information about investment decisions, dividend changes, changes in capital structure, and corporate-control transactions.”29 He also refers to another large body of research from event studies showing the opposite conclusion: rather than adjusting rapidly to new information, prices adjust slowly and thus inefficiently. Nevertheless, he concludes his review article with this statement: “The cleanest evidence of market-efficiency comes from event studies, especially from event studies on daily returns.”30