Contrarian Investment Strategies

by David Dreman

  I’m afraid we’re overconfident that we can overcome our tendency to be overconfident by simply recognizing that we tend to be overconfident. It’s just not that simple, alas.

  Analysts’ Consistent Overoptimism

  Given what we just saw, how optimistic do you think analysts’ earnings estimates are? Jennifer Francis and Donna Philbrick studied analysts’ estimates for some 918 stocks from the Value Line Investment Survey for the 1987–1989 period.27 Value Line is well known on the Street for near-consensus forecasts. The researchers found that analysts were optimistic in their forecasts by 9 percent annually, on average. Again, remembering the devastating effect of even a small miss on high-octane stocks, these are very large odds against investors looking for ultraprecise earnings estimates.

  The overoptimism of analysts is brought out even more clearly by I/B/E/S, the largest earnings-forecasting service, which monitors quarterly consensus forecasts on more than seven thousand domestic companies. Despite these allowable quarterly estimate changes, analysts tend to be optimistic, according to I/B/E/S. What seems apparent is that analysts do not sufficiently revise their optimistically biased forecasts in the first half of the year and then almost triple the size of their revisions, usually downward, in the second half. Even so, their forecasts of earnings are still too high at year-end.

  A study that Eric Lufkin28 of Morgan Stanley and I collaborated on for my Forbes column some years back (January 26, 1998) provided further evidence of analysts’ overoptimism. I updated the study to the end of 2006 in Table 7-1. The study measures analysts’ and economists’ estimates against the actual S&P reported earnings from 1988 to 2006, a nineteen-year period, which saw more than its share of bubbles and crashes, as well as economic booms and recessions. Analysts make “bottom-up” estimates; that is, they look at all the important fundamentals of a company and then make their estimate for a stock. They forecast company by company and then the companies are added up, with the proper weighting in the S&P 500 index for each, to arrive at a forecast. Economists, on the other hand, make “top-down forecasts”; that is, they look at the economy and then decide how their overall forecasts will trickle down to individual company estimates. A glance across each year shows the percentage increase or decrease in earnings the analysts forecast in column 2, those the economists forecast in column 3, and the actual increase or decrease in earnings that the S&P 500 showed for the year in column 4.

  What is striking is how overly optimistic the estimates by both analysts and economists actually are. To be fair, let’s look at analysts’ and economists’ estimates over the entire nineteen years of the study and then compare them with the actual earnings of the S&P 500. As Table 7-1 shows, the average estimate for analysts was a 21 percent increase in earnings, the average for economists was 18 percent, and the actual earnings increase for the S&P was 12 percent. As column 3 demonstrates, economists—supposedly charter members of the “dismal science”—made earnings forecasts that were overoptimistic by an astounding 53 percent (over S&P reported earnings) annually on average over the nineteen-year period. Could anything be worse? Why, yes: analysts were overoptimistic by 81 percent annually on average over that time, an enormously high miss. The study evidence: earnings forecasting is neither an art nor a science but a lethal financial virus.

  What makes analysts so optimistic? The subject is anything but academic because it is precisely this undue optimism that induces many people, including large numbers of pros, to buy stocks they recommend. As we have seen in the recent examples and will see more thoroughly in the chapters ahead, unwarranted optimism exacts a fearful price. A Psychological Guideline is again in order:

  PSYCHOLOGICAL GUIDELINE 9: Analysts’ forecasts are usually overly optimistic. Make the appropriate downward adjustment to your earnings estimate.

  Now, there are people with outstanding gifts for abstract reasoning who can cut through enormously complex situations. Every field has its William Miller or Bill Gross. But such people are rare. It seems, then, that the information-processing capabilities and the standards of abstract reasoning required by current investment methods are too complicated for the majority of us, professional or amateur, to use accurately.

  A Surefire Way to Lose Money

  At this point, you might be wondering if I’m exaggerating the problems of decision making and forecasting in the stock market. The answer, I think, can be found by considering the favorite investments of market professionals over time.

  Consider a large international conference of institutional investors held at the New York Hilton in February 1968. Hundreds of delegates were polled about the stock that would show outstanding appreciation that year. The favorite was University Computing, the highest-octane performer of the day. From a price of $443, it dropped 88 percent in less than twelve months. At an Institutional Investor conference in the winter of 1972, the airlines were expected to perform the best for the balance of the year. Then, within 1 percent of their highs, their stocks fell 50 percent that year in the face of a sharply rising market. The conference the following year voted them a group to avoid. In another conference, in 1999, a large group of professionals polled picked Enron as the outstanding performer for the next twelve months. We all know what happened to that hot stock.

  Are those results simply chance? In an earlier book, The New Contrarian Investment Strategy (1982), I included fifty-two surveys of how the favorite stocks of large numbers of professional investors had fared over the fifty-one-year period between 1929 and 1980. The number of professionals participating ranged from twenty-five to several thousand. The median was well over a hundred. Wherever possible, the professionals’ choices were measured against the S&P 500 for the next twelve months.*43

  Eighteen of the studies measured the performance of five or more stocks the experts picked as their favorites.29 Diversifying into a number of stocks, instead of one or two, will reduce the element of chance. Yet the eighteen portfolios chosen underperformed the market on sixteen occasions! This meant, in effect, that when a client received professional advice about those stocks, they would underperform the market almost nine times out of ten. Flipping a coin would give you a fifty-fifty chance of beating the market.

  The other thirty-four samples did little better. Overall, the favorite stocks and industries of large groups of money managers and analysts did worse than the market on forty of fifty-two occasions, or 77 percent of the time.

  But those surveys, although extending over fifty years, ended in 1980. Has expert stock picking improved since then? The Wall Street Journal conducted a poll on whether the choices of four well-known professionals could outperform the market each year between 1986 and 1993. At the end of the year, four pros gave their five favorite picks for the next year to John Dorfman, the editor of the financial section, who reviewed them twelve months later, eliminating the two lowest performers and adding two fresh experts. In sixteen of thirty-two cases, the portfolios underperformed the market—a somewhat better result than in the past, but no better than the toss of a coin.30*44

  Table 7-2 gives the results of all such surveys that I found through 1993. As the table shows, only 25 percent of the surveys of the experts’ “best” stocks outperformed the market. The findings startled me. Though I knew that experts make mistakes, I didn’t know that the magnitude of the errors was as striking or as consistent as the results make evident.

  But some of the studies go back to the 1920s, and the last one ends in 1993. Has stock picking improved since then? After all, in the past fifteen years, the performance of professional investors has been more carefully scrutinized than ever before. As we saw in chapters 1 and 2, money managers as a group have not outperformed the market. Another study done by Advisor Perspectives for the ten years ending December 31, 2007, analyzed how stocks in the S&P domestic indices performed against the S&P benchmark index. As Table 7-3 shows, using one of the S&P indices as a benchmark, the S&P indices outperformed the stocks in six of the nine sectors and tied on
ce.

  In only one case, the S&P 500/Citigroup Growth, did the mutual funds significantly outperform the S&P index. In brief, the averages beat the money managers two-thirds of the time over the ten-year period.

  Finally, for the five years ending in 2009, the S&P 500 Index outperformed 60.8 percent of actively managed large-cap U.S. equity funds, the S&P MidCap 400 index outperformed 77.2 percent of midcap funds, and the S&P Small-Cap 600 index outperformed 66.6 percent of small-cap funds.

  The studies now in place as a group for more than seventy-five years clearly demonstrate just how badly money managers and analysts underperformed the market. They also clearly demonstrate that professional investors, in the large majority of cases, were tugged toward the popular stocks of the day, usually near their peaks, and, like most investors, steered away from unpopular, underpriced issues, as the subsequent year’s market action indicated. Also interesting is that one industry—technology—was favored over the years, although there were dozens to choose from. And it was favored so unsuccessfully! Experts’ advice, in these surveys at least, clearly led investors to overpriced issues and away from the better values.

  What can we make of these results? The number of samples seems far too large for the outcome to be simply chance. The evidence indicates a surprisingly high level of error among professionals in choosing stocks and portfolios over six and a half decades.

  The failure rate among financial professionals, at times approaching 90 percent, indicates not only that errors are made but that under uncertain conditions there must be systematic and predictable forces working against unwary investors.

  Yet again, such evidence is obviously incompatible with the central assumption of the efficient-market hypothesis.31 Far more important is the practical implication of what we have just seen, another plausible explanation of why fundamental methods often don’t work. Investment theory demands too much from people as configural and information processors. Under conditions of information overload, both within and outside markets, our mental tachometers surge far above the red line. When this happens, we no longer process information reliably. Our confidence rises as our input of information increases, but our decisions do not improve. This leads to another Psychological Guideline.

  PSYCHOLOGICAL GUIDELINE 10: Tread carefully with current investment methods. Our limitations in processing complex information prevent their successful use by most of us.

  Though it is probably true that experts do as poorly under other complex circumstances, market professionals unfortunately work in a goldfish bowl. In no other profession I am aware of is the outcome of decisions so easily measurable.

  In examining the stock-picking record of money managers and other market pros, a critical question is: how accurate are analysts’ earnings estimates? Those are the key elements underlying stock selections and the heart of investing as it is practiced today. That’s the question to be examined next. The results of some very thorough studies on the accuracy of the top security analysts, the cream of the crop, will surprise you.

  Chapter 8

  How Big a Long Shot Will You Play?

  IN THE EARLY 1970s, when I was an analyst, nothing was online. Nada. Zilch. Today, the analysts at major brokerage houses have immediate access to competitors’ reports, estimate changes, and volumes of other information. There is exponentially more information available now than back then. It’s like moving from a hand-cranked telephone to the latest iPhone. Yet in spite of the information revolution, there is every reason to believe, as we’ll soon see, that earnings estimate errors remain enormously high, much too high to be of any use in determining the intrinsic value of most stocks. It isn’t a matter of my colleagues in the industry not giving their all, either. Rather, as you undoubtedly know, iPhone-equipped or not, forecasting is far from an exact science. The weather forecaster who sends you out for a sunny day only to leave you drenched in an afternoon shower has more in common with a Wall Street security analyst than you might think.

  If you’re like me, you’ll occasionally take a chance at winning a reasonable poker pot. But more generally, how big a long shot will people play? When analysts like an idea, it can be off the chart. As we saw in chapter 2, if players think they are going to win, they’ll readily pay the same price for a lottery ticket whether the odds are 10,000 to 1 or 10 million to 1 against winning. As we also saw, the possibility of winning, rather than the probability of doing so, can cause very low odds to carry great weight, even when the probabilities against the players are increased by 1,000 times.

  Are those people daft? Maybe a little, but, as we have seen, many investors have played with the same odds in every bubble and mania that we’ve looked at. They also play with odds that are against them—certainly not as spectacular but still high—in far more tranquil markets. Why do they do it so consistently?

  If you watch an early-morning market show on CNBC or Bloomberg or read The Wall Street Journal to start the day, you have undoubtedly noticed that the anchors or reporters attentively listen to or write about the advice of a collection of well-dressed men and women who seem to know everything about the markets. In this chapter, we’ll examine the reliability of the advice of this chic group who happen to be security analysts. Breaking news in the financial media is often an analyst’s raising or lowering his or her earnings estimates of a stock or industry. Upgrades or downgrades are listened to carefully, even when we don’t know a company’s name or industry. Changes in analysts’ estimates on major companies are showstoppers that freeze people in their tracks. If the downgrade or upgrade is major, it can move the stock and its industry noticeably.

  The people whom the media cover so attentively consider themselves to be hardheaded realists, not pointy-headed theoreticians. They’re immersed in the reality of daily market action, not writing math equations and running computer simulations, and they are confident that they are offering analytically tried-and-true expert advice to investors. Fair enough. Let’s put them to the test and track their actual performance. Their estimates and recommendations are the crucial factor the majority of investors look at when deciding what stocks to buy, hold, or sell. It’s time for us to analyze if CNBC, Bloomberg, or other media are giving us any more consistently accurate information than the Weather Channel.

  Forecasting Follies 1: Predicting Company Earnings

  Although they don’t agree on many issues, Wall Streeters and financial academics concur that company earnings are the major determinants of stock prices. Modern security analysis centers on predicting stock movements from precise earnings estimates. As a result, major brokerage houses still have eight-figure research budgets and hire top analysts to provide accurate estimates. The largest bank trust departments, mutual funds, hedge funds, and money managers demand the “best” because of the hundreds of millions of dollars in commissions they command.

  Several decades ago, Institutional Investor magazine formalized the process of determining the “best” analysts. Each year the magazine selects an “All-Star” team made up of the “top” analysts in all the important industries—biotech, computers, telecommunications, pharmaceuticals, and chemicals—after polling hundreds of financial institutions. There are first, second, and third teams, as well as runners-up for each industry. The magazine portrays the team on its cover each year, often dressed in football uniforms with the brokerage firm’s name on each star’s jersey. The competition to make the teams is fierce, as we’ll see later in this chapter.

  If a brokerage firm can boast a number of All-Stars, its profitability ratchets up accordingly. Some years back, the managing partners and the director of research of a large brokerage house decided to let one of their analysts go. The office executioner was on his way to inform the analyst when the research director came running down the corridor, grabbed his arm, and, gasping for breath, said, “Wait . . . we can’t do it . . . he just made the second team.”

  Salary scales, as you may guess, are in the stratosphere. Experienced analysts
make between $700,000 and $800,000 a year; standouts receive more. Then there is the million-dollar-a-year club, which includes several dozen of the Street’s outstanding oracles. Incomewise, they are in a class with popular entertainers and professional athletes.

  Some analysts earn in the eight figures, exceeding the pay of a number of CEOs of Fortune 500 companies. Jack Grubman, the once highly regarded telecommunications analyst, jumped ship from PaineWebber to Salomon Brothers in the mid-1990s. The price was a two-year contract with annual pay of $2.5 million. The salary was so high that his research colleagues jokingly referred to underwritings of the firm being offered to its clients in “Grubman units” of $2.5 million each. Needless to say, Wall Streeters often accord top analysts the same hero worship that teenagers reserve for rock stars and film heroes.

  The hero worship diminished sometime after the dot-com collapse in 2002, when SEC and state investigations of major analysts produced disturbing results. In late April 2003, the SEC and New York State regulators released thousands of documents that showed that the traditional rules of the Street had been violated, and the “Chinese Walls” supposedly keeping bankers from influencing the work of security analysts came tumbling down. Although the investigations focused on two of Wall Street’s most highly paid and powerful analysts, Jack Grubman of Salomon Smith Barney*45 and Henry Blodget of Merrill Lynch, it rapidly expanded to many dozens of other analysts. E-mails and other documents showed that many analysts had been pressured into giving favorable ratings to weak companies, a good number of which were wobbly Internet firms with almost no business plans, revenues, or viable platforms. Because underwriting IPOs had become so profitable during the dot-com bubble in the late 1990s, the pressure to serve major corporate clients rather than retail clients was overwhelming. In the mad scramble to bring in billions of dollars of investment banking fees from IPOs, analysts spoke in double-talk. Publicly they heaped praise on shaky companies they followed, rating them as strong or even “screaming” buys. In e-mails to investment banking clients they mocked the same firms, calling them “pigs,” “junk,” “crap,” and much worse. A distinct pecking order existed. Retail clients were encouraged by the analyst to buy these poor-quality issues, often at a substantial premium from where they were issued, while institutional investors were sometimes tipped off to stay well clear. Henry Blodget had a buy recommendation out on GoTo, a dot-com stock. When an institutional heavy hitter asked him what he liked about the company, Blodget flippantly replied that there was “nuthin” interesting about the issue except for the large investment banking fees Merrill was getting.1 Bad as it was, this game was only penny ante for some of the more serious analysts.