Contrarian Investment Strategies

by David Dreman

  In this part we will present a new paradigm (or model) of investing that will both utilize the state-of-the-art psychology we viewed in Part I and combine it with investment methods based on it that have proved highly successful over many decades. The methods also rely on us to disregard some of the conventional teachings we have received about how we should evaluate stocks, bonds, or other investments. But they are anything but bugle blasts ordering you to follow the new knowledge blindly. Each has been proved over time either by psychology or by superior investment performance.

  Sounds a little heady, doesn’t it? Now let’s start at the beginning and see exactly what’s behind these assertions. We’ll begin with a bit of discussion to reveal the elemental structure, function, and statistical justifications of five major variations on the contrarian strategies, each highly successful to date. Using this key, you’ll see how the pieces of the puzzle rapidly fall into place—although I’m afraid it won’t get you a discount on one of those Paul Stuart suits or Ferragamo ties. The five important strategies are:

  1. Low-price-to-earnings strategy

  2. Low-price-to-cash-flow strategy

  3. Low-price-to-book-value strategy

  4. High-yield strategy

  5. Low-price-to-industry strategy

  Each of these strategies has a number of subcategories to allow you to custom-tailor it to your own requirements. Strategy 5, a very new and different type of contrarian strategy, will be introduced in chapter 12.

  The Investment World Turned Upside Down

  What seemed apparent from our review of experts’ forecasts is that the companies they liked the best tended to be the wrong ones to buy. Therefore, let’s ask another question: should you avoid the stocks the experts and the crowd are pursuing and pursue the ones they are avoiding? The answer, as we shall see, is an unqualified yes.

  And we can document the consistent success of this investment strategy going back almost eighty years, a strategy that dramatically opposes conventional wisdom—a prime reason it works.

  For the findings show that companies the market expects the best future for, as measured by the price-to-earnings, price-to-cash-flow, price-to-book-value, high-yield, and low-price-to-industry strategies, have consistently done the worst, while the stocks believed to have the most dismal future have always done the best. The strategy is not without an element of black humor. In fact, to the true believers at the shrines of efficient markets or other contemporary investment methods, the approach may appear to be a form of Satan worship: the “best” investments turn out bad, and the “worst” ones turn out good.

  However, the findings are not in the least magical and don’t contain even a trace of voodoo. Most investors do not recognize the immense difficulty of predicting earnings and economic events, and when forecasting methods fail, a predictable reaction occurs. Here we confront the main irony: one of the most obvious and consistent variables that can be harnessed into a workable investment strategy is the continual overreaction of people to companies they consider to have excellent or mundane prospects. This works just as surely with investors today as it has worked with investors in any market in the past.

  The Boot Hills of Concept Stocks

  In order to make evaluations of “best” stocks, forecasts extending growth well into the future must be made with extreme accuracy. As we saw in chapter 8, the reliability of these forecasts is ridiculously low. We also know that when the earnings of a favored company fall below the forecast, the earnings surprise, even if it is minute, has a devastating effect on its stock prices. Not surprisingly, investment strategies based on precise estimates have performed erratically, to say the least. There is a Boot Hill full of Internet names from the 1996–2000 dot-com bubble, as well as a Boot Hill containing similar stocks from earlier bubbles and another one of “must-own” concept stocks to be bought at any price.

  The term “best,” used by so many professional and individual investors, also filters out the inherent risk of the situation. Conversely, the lowest-visibility stocks have been shown to be significantly underpriced and, when they have positive surprises, are subject to sharp upward reappraisals. As we saw in chapter 2, inherent risk is understated for high-expectation situations; it is often exaggerated, sometimes dramatically, for low-expectation stocks. This puts the master key of the book into our hands—and explains why the investment strategies that we’ll examine work remarkably well over time.

  The Success of Low-P/E Strategies: The Early Evidence

  Beginning in the 1960s, researchers began to wonder if visibility—the crucial pillar of modern security analysis—was actually as solid as generally believed. The original studies were done with price-earnings ratios because of their ready availability in the early databases. One of the first of those researchers asked, “How accurate is the P/E ratio as a measure of subsequent market performance?”

  Francis Nicholson, then with the Provident National Bank, was the interrogator. In one comprehensive study done in 1968 that measured the relative performance of high- versus low-P/E stocks, he analyzed 189 companies of trust-company quality in eighteen industries over the twenty-six years 1937–1962.1 The results are given in Figure 10-1.

  Nicholson divided the stocks into five equal-sized groups according to their P/E rankings. These quintiles were rearranged by their P/E rankings for periods of one to seven years. When the quintiles were recast annually on the basis of new P/E information, the stocks most out of favor showed a 16 percent annual rate of appreciation over the total time span. Conversely, switching in the highest P/Es on the same basis resulted in only a 3 percent annual appreciation over the period. Although the performance discrepancies were reduced with longer holding periods, even after the original portfolios were held for seven years, the lowest 20 percent did almost twice as well as the highest.

  With remarkable consistency, investors misjudged subsequent performance. The results are completely uniform. The most favored stocks (quintile 1) sharply underperformed the other groups, while the least popular (quintile 5) showed the best results. The second most popular quintile had the second worst performance, while the second most unpopular quintile had the second best results.

  Benjamin Graham’s The Intelligent Investor cites a second study, this one involving the thirty stocks in the Dow Jones Industrial Average (Table 10-1). The performance of the ten lowest- and ten highest-P/E stocks in the group and of the combined thirty stocks in the industrial average was measured over set periods between 1937 and 1969. In each time span, the low P/Es did better than the market and the high P/Es did worse.

  The study also calculated the results of investing $10,000 in either the high- or low-multiple groups in the Dow Jones Industrial Average in 1937 and switching every five years into the highest P/Es (in the first case) and the lowest P/Es (in the latter). Ten thousand dollars invested in the lowest P/Es this way in 1937 would have increased to $66,866 by the end of 1962. Invested in the highest P/Es, the $10,000 would have appreciated to only $25,437. Finally, $10,000 invested in the stocks of the Dow Jones Industrial Average would have grown to $35,600 by 1962. (Some thirty-five years later, Graham’s findings were reintroduced as a hot new investment strategy, “The Dogs of the Dow.”)

  A number of other studies in the 1960s came up with similar findings. The conclusion of these studies, of course, is that low-P/E stocks were distinctly superior investments over an almost thirty-year period. But theories, like sacred cows, die hard, so these findings created little stir at the time.2

  The Days of Disbelief

  If the low-P/E results were analyzed at all, they were criticized. For one thing, the growth school has always had a major following among investors. Many institutional investors could not bring themselves to believe the efficacy of the findings. After all, like the studies that discredited earnings forecasts, these findings did seem, cavalierly, to toss aside our years of practice (or perhaps brainwashing) to the contrary. When I published a paper in early 1976 summarizing s
ome of the previous research, a number of professionals told me that such information was history: “Markets of the 1970s are very different.”

  The low-P/E researchers were onto something, but the efficient-market hypothesis was rapidly ascending, and the work was mainly ignored. After all, in the late 1960s, our financial friends in academia were tightening the final nuts and bolts of the formidable efficient-market hypothesis. According to this academic-launched dreadnought, such results simply could not exist—whether they did or not. And that was that!

  Further EMH criticism asserted that low-P/E stocks were systematically riskier (in the parlance, had higher betas) and therefore ought to provide higher returns. We have already looked closely at the failure of risk measures that the EMH academics used in chapters 5 and 6.

  However, even in science fact has little chance against entrenched belief. For the coup de grâce, methodological criticisms of the studies were wheeled into action. They were mostly hairsplitting and not convincing—at least to me. But recall Einstein’s dictum “It is the theory that describes what we can observe.” Those findings could not exist if the dreadnought was to proceed merrily along annihilating traditional investment practice.

  Buying low-P/E stocks appeared successful in studies of past performance. As a practical matter, it had worked for me—no small inducement to belief. Consequently, I updated several studies that I had used first in Psychology and the Stock Market (1977) and second in Contrarian Investment Strategy (1979). The low-P/E findings were still robust.3

  My findings once again showed the clear-cut advantage of a low-P/E strategy. The study covered the two-tier market of 1971–1972, the bear market of 1973–1974 (the worst in the postwar period up to then), and the subsequent recovery. It did not matter whether investors started near a market top or a market bottom; superior returns were provided in any phase of the market cycle.

  Through the 1980s to the mid-1990s, I completed (with various collaborators) half a dozen separate studies on low-P/E strategies, most of which were published in Forbes.*57 One study measured the 1,800 largest companies on the Compustat tapes between 1963 and 1985. The lowest-P/E quintile returned 20.7 percent annually against 10.4 percent for the highest.4 Another divided the 6,000 stocks on the Compustat tapes into five equal groups according to their market size, for twenty-one years ending in 1989. Each group then was divided again into five subgroups according to P/E rankings. The low-P/E groups handily outperformed the high-P/E groups for all market sizes, ranging from the smallest, at around $50 million in market value, to the largest, at nearly $6 billion.5

  Another study measured how low price to cash flow performed for the twenty-two years ending March 31, 1985, using 750 large companies. The results are shown in Table 10-2. The stocks were again separated into five equal groups and ranked each year according to the ratio of price to cash flow.6 As the chart shows, the most out-of-favor stocks—lowest price to cash flow—almost doubled the annual performance of the favorites through this extensive period.7

  But what of some of the past criticisms? Were any of them still valid?8 Our experimental design adjusted for those criticisms and still provided the results shown above.

  All this has been reconfirmed by other research in the late 1970s and early 1980s. Three carefully prepared studies by the late Sanjoy Basu* came up with similar results.9 In his study, published in the Journal of Finance in June 1977, Basu used a database of 1,400 firms from the New York Stock Exchange between August 1956 and August 1971. He took 750 companies that had year-ends of December 31 and turned over the portfolios annually, using prices on April 1 of the following year. As in most of the previous studies, he divided the stocks into quintiles according to P/E rankings. The results (again using total return) are shown in Table 10-3.

  Basu found, as we did, that the low-P/E stocks provided superior returns and were also somewhat less risky. In his words, “However, contrary to capital market theory, the higher returns on the low P/E portfolios were not associated with higher levels of systematic risk; the systematic risks of portfolios D and E [the lowest] were lower than those for A, A*58, and B [the highest].”10 This and subsequent work, updated and adjusted for previous criticisms through the mid-1980s, thus added many links to a chain extending over seventy years, documenting the superior performance of low-P/E issues.

  Whatever Took So Long?

  Given the weight of the evidence, you’d think that contrarian value approaches would have captured the imagination (and wallets) of investors long ago. By the 1990s, perhaps we should have been looking back at the golden age, when only a handful of pioneers reaped the rewards and only the favored few knew what a powerful tool these strategies were. But that’s not the case. Even today, in 2011, contrarians are a distinct minority—and, remarkably, are likely to remain so.

  One reason for the situation is historical. You’ll remember how EMH swept the land and banished the heathens from Wall Street. As the new orthodoxy, EMH had everything to gain from a long and prosperous reign—not to mention a good deal to lose if competing ideas shouldered their way onto the Street. That, of course, is as much innate human psychology as any of the other crowd reactions we will examine.

  In the late 1970s and through most of the 1980s, our work did not stand tall with efficient-market advocates, particularly those residing at the high shrine of market fundamentalism, the University of Chicago. As a leading heretic, I often experienced the wrath of the true believers. My work came under sharp attack, and my Forbes columns even had the distinction of being assigned to classes of students to hammer apart. On several occasions my unfortunate editors had sets of 10 or more letters attacking the work, letters which they on occasion good-naturedly published.

  With the publication of my books and additional articles, the barrage intensified. Letters attacked every part of the findings, saving some choice remarks for the author. None found any statistical fault with the work, but they questioned the Neanderthal beliefs of a writer who couldn’t understand the overwhelming sweep and beauty of efficient-market thinking.

  When I submitted a paper reporting the findings to the Financial Analysts Journal in 1977, it was not accepted or rejected but left in some purgatory reserved for ideas that don’t fit the prevailing paradigm.

  The low-P/E anomaly was large enough to continue to attract attention from academia and Wall Street. Academics also dismissed the work using their old standby, risk measurement. Low-P/E stocks might provide higher returns, but they were far more risky, said the critics: rational investors accepted the higher risk only by demanding higher returns. Unfortunately, that, too, turned out to be not quite accurate.

  The Great Discovery

  Through the 1980s, Wall Street became increasingly interested in contrarian investment strategies. With continuing improvements in databases, confirmation that these strategies worked grew stronger and stronger.

  The tide continued to gain strength. Dennis Stattman and Barr Rosenberg, Kenneth Reid, and Ronald Lanstein, for example, found that low price to book value outperformed high price to book value and the market.11 At the same time, the evidence mounted that beta had no value as a predictor of stock prices. Though the economic fundamentalists attempted to rationalize away their existence for almost three decades, those exasperating black swans just did not have the good manners to fly away. That low-P/E strategies worked and volatility didn’t work was a death knell for efficient markets.

  What was a poor apostle of this new religion to do? Since neither of these results could be denied, the answer, of course, was to be the first to discover them. That is exactly what Eugene Fama, the apostle of efficient markets, did. In the 1990s, academics firmly planted their own flag on the newfound world of contrarian strategies.

  In a revolutionary paper that thrust boldly thirty years into the past, Professor Fama discovered precisely what Nicholson, other researchers of the 1960s, Sanjoy Basu, Stattman, Rosenberg and colleagues, and I had found in the 1970s and ’80s: that contrarian
strategies worked. Worse yet, as we saw in chapter 6, beta didn’t work. The three decades of stating that contrarian strategies provided better returns because they were riskier were swept away, as we saw, by the discovery that beta, the risk measure the apostle and his disciples used, was valueless.

  To be fair, Fama and French12 do reference Basu, Stattman, and Rosenberg and colleagues in passing, as well as Ray Ball,13 who argued that low- P/E is a catchall for all risk that cannot be explicitly separated or even found. Still, Ball’s explanation was widely accepted by efficient marketers for years.

  The reasoning in Ball’s paper is not unlike the phlogiston theory of heat popular in the eighteenth century. According to the theory, some elements are more combustible than others because they contain more phlogiston, while others are less so because they contain smaller amounts. Phlogiston was supposedly weightless and odorless, and could not be detected; nevertheless, it was there. How else could combustion occur? (Or how could low-P/E strategies beat the market if they were not risky, regardless of the fact that said risk could not be detected?) Circular and fallacious reasoning, yes, but Ball, Fama, et al. used precisely this logic to defend EMH. In both cases, the theorists defended themselves against phenomena they cannot explain through the creation of ingenious fudge factors. (The efficient-market hypothesis, as we have seen, has a grab bag full of these defenses.)