by David Dreman
We see them consistently in far more normal market settings. Take the behavior of earnings surprises described in chapter 9. The “best” stocks, those with the most promising prospects by various yardsticks from high P/E to high price to book value to high price to cash flow, always underperform significantly as a group when they have earnings surprises. Similarly, the “worst” stocks, in terms of the same benchmarks, almost always outperform. The amount of both under- and overperformance is large, about 7 percent a year, which is 70 percent of the average 9.9 percent annual return on stocks since the mid-1920s.
An important cause of the consistent underperformance of the “best” stocks comes from investors’ and analysts’ overoptimism about their ability to pinpoint earnings, even though the empirical evidence shows emphatically that it can’t be done. This overoptimism and the focus on predictable growth, both manifestations of Affect, are the most important components of security analysis in our times and are growing more important each year. Forget Graham and Dodd, who concentrated on a wide range of fundamentals and financial ratios and would completely reject the narrow focus of analysts today. Most of us either directly or indirectly still follow these newer analytical guidelines, unsuccessful as they have proved to be in recent decades.
In the preceding chapter we saw how out-of-favor stocks have sharply outperformed the favorites for many decades. Statistics by Kenneth French, often Eugene Fama’s coauthor, on his Web site indicate that this outperformance has taken place in every decade since the 1940s.*60 As Figures 10-3 and 10-4 showed, an investor who purchased $10,000 in low-P/E stocks in 1970 would have thirteen times the money of someone who purchased $10,000 of the high-P/E index at the same time. Yet a study by Mark Peterson, then of Deutsche Bank, indicates that only about 3 percent of stocks held by equity mutual fund investors are contrarian holdings.1
Again, the answer is Affect. We inherently like the “best” stocks or industries and stay well away from the “worst.” It doesn’t matter if we’ve made the same error dozens of times; the influence of Affect is too strong not to influence our decision-making processes. Affect is very easy to detect in the case of a bubble or a mania—after the fact, of course. But it can also act very subtly when most people select “best” over “worst” stocks or when they ignore the evidence that earnings surprises favor “worst,” not “best,” stocks. If we look at mental disorders, we find that it’s easier for a psychiatrist to diagnose severe paranoia or schizophrenia than to identify neurosis, in which people’s behavior often seems normal. The situation is similar with Affect; it’s far easier to look at the effects of manias and bubbles than to see the influence of Affect in more normal markets, where price fluctuations are far more moderate. But the data indicate that Affect has a robust effect in these cases, too.
The evidence of consistent and predictable overreactions to events through this book, some new, some discovered over recent decades, led me to the investor overreaction hypothesis, which I introduced in the original edition of Contrarian Investment Strategy in 1979. Although the term “overreaction” is almost as old as markets themselves, I developed a testable hypothesis. At that time there were only a handful of anomalies to test, notably the superior performance of out-of-favor stocks and the consistently better performance of lower-rated bonds. Both anomalies had shown superior results for many decades. Sanjoy Basu,2 in 1977, and behavioral finance pioneers Werner De Bondt and Richard Thaler refer to investor overreaction in their 1985 paper,3 but limit it primarily to the performance of “best” and “worst” stocks. A stronger version of this hypothesis, reinforced by new psychological discoveries in both Affect and neuroeconomics, is presented here. The number of anomalies in which overreaction takes place has jumped manyfold in the intervening decades.
THE INVESTOR OVERREACTION HYPOTHESIS DEFINED
The investor overreaction hypothesis (IOH) states that investors overreact to certain events in a consistent and predictable manner.
This is based on the psychological forces we examined in detail in the research findings in Part I and the new biologically demonstrated discoveries of neuroeconomics, as discussed in chapter 9.
Three central predictions of IOH are:
1. Investors will consistently overvalue favored stocks and undervalue out-of-favor stocks.
2. Investors will find themselves overoptimistic on the forecasts of “best” stocks and too pessimistic on those of “worst” stocks over time. As a result, earnings surprises tend to favor “worst” stocks as a group in a predictable and consistent manner.
3. Over time both favored and out-of-favor stocks will regress to the mean because of earnings surprises and other fundamental factors, which will result in the underperformance of the “best” stocks while those considered “worst” will outperform.
These predictions are borne out in the market history that we have examined in detail earlier. Unfortunately, we won’t run out of good illustrations anytime soon. Let’s look at a few more that are in line with IOH-predicted investor behavior.
Investors extrapolate positive or negative outlooks well into the future, pushing the prices of favored stocks to excessive premiums and those of out-of-favor stocks to deep discounts. (The performance of “best” and “worst” stocks can be directly compared, of course, but “best” and “worst” investments can also be instruments other than stocks, and “best” might be in a different market from “worst.”)4 In 1980 and again in 2009 through August 2011, for example, “best” investments included gold—which peaked at $850 an ounce in 1980 before regaining this high again in 2008 and moving up to $1,892 in August 2011—and “worst” included tax-exempt municipal bonds, yielding as high as 15 percent as bond prices plummeted. Premiums or discounts on favored or out-of-favor investments can be substantial and last for long periods of time.
Similar reactions have been noted since the early 1900s with lower-rated bonds, which provide higher returns after adjustments for default levels over time. Such consistent above- or below-average returns are also likely to exist in other financial markets.
The investor overreaction hypothesis states that it is far safer to project a continuation of investor overreaction based on what we know psychologically, backed by the robust findings we have reviewed, than to attempt to project the visibility of the stocks or other investments themselves.
TWELVE KEY PREDICTIONS OF THE INVESTOR OVERREACTION HYPOTHESIS (IOH)
Here’s a full list of the predictions based on the IOH:
1a. Absolute contrarian strategies provide superior returns over time.
1b. Out-of-favor stocks, as valued by at least three different major fundamental measurements—low price-to-earnings, low price-to-cash-flow, and low price-to-book-value ratios—will outperform the market as a group over longer periods of time, normally five to ten years or more.
1c. Favored stocks, as measured by high price-to-earnings, high price-to-cash-flow, and high price-to-book-value ratios, will underperform the market over the same time periods.
1d. Out-of-favor stocks should significantly outperform favorites over the same time periods.
2a. Relative contrarian strategies provide superior returns over time. (For details of this strategy, please refer to chapter 12, “Contrarian Strategies Within Industries,” page 310.)
2b. The out-of-favor stocks in each industry will outperform the market over longer periods of time, normally from four to six years.
2c. The favored stocks in each industry will underperform the market over the same time periods.
2d. The out-of-favor stocks in each industry should outperform the favorites in each industry over the same time periods.
3. Favored stocks as a group are overvalued and out-of-favor stocks undervalued. Over time both groups will regress to the mean because of earnings surprises and other fundamental factors, which will result in the underperformance of the “best” stocks, while those considered “worst” will outperform, as both move toward a more average val
uation.
4a. IOH posits that the outperformance of out-of-favor stocks and the underperformance of favorites is caused primarily by behavioral influences (Affect, cognitive heuristics, neuroeconomics, and other psychological variables).
4b. The role of Affect applies on an industry-relative basis for most and least favored stocks, as it does to stocks on an absolute basis.
5. There are two distinct categories of earnings surprises:
a. Event triggers—significant positive surprises on the lowest-valued group of stocks and major negative surprises on the highest-valued group. Both result in a consequential impact on the categories’ price movement.
b. Reinforcing events—negative surprises in the lowest-valued group and positive surprises in the highest-valued group. Both have small relative impacts on stock movements.
6a. Surprise will normally have a much smaller impact on the 60 percent of stocks in the three middle quintiles, which are less over- or undervalued using the fundamental value criteria previously noted.
6b. Even without the occurrence of an event trigger, the “best” and “worst” investments regress toward the market average over time, with the “best” stocks underperforming the “worst” ones, whether high price-to-book-value, high price-to-earnings, or high price-to-cash-flow ratio is considered, as chapter 9 demonstrated.
7. A significant part of current investment theory is dependent on accurate forecasting. The IOH predicts that:
a. Analysts’ and economists’ consensus forecasts are overoptimistic over time.
b. Analysts’ consensus forecasts on individual stocks will show substantial errors over time, resulting in significant mispricing of stocks on an almost continual basis.
c. Analysts’ consensus forecasts on industries will also show substantial errors over time, resulting in significant mispricing of stocks on an almost continual basis.
8. Investor overoptimism occurs regularly in a number of important market activities, including:
a. IPOs.
b. Analysts’ and economists’ overoptimistic earnings estimates.
9. The overreaction occurs prior to the event trigger or other factors that lead to a reevaluation of the “best” and “worst” stocks. After the event trigger other reevaluation forces occur that continue the reversion to the mean. “Worst” stocks continue their rise in price and “best” stocks their fall for a period of years.
10. The overreaction-versus-underreaction debate in current financial theory is actually about the two parts of the same process, as findings demonstrate.5*61
11. Because mispricing of securities is constant, markets are continually readjusting securities values. Stock and other financial markets are thus never in equilibrium, contrary to the teachings of EMH and most economic theory.
12. The efficient-market risk hypothesis is questionable,6 and there is no robust evidence that greater volatility results in higher returns or lower volatility results in lower returns.
To recap, our five contrarian strategies, which are all in accordance with the assumptions of the investor overreaction hypothesis, 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
THE PSYCHOLOGICAL CHOICE: EMH OR IOH?
The efficient-market risk theory and a good part of economic theory are built on the unproved assumption of almost omniscient investor rationality dating from the eighteenth century. As we know, the assumption disregards recent major psychological findings as well as other behavioral work over at least the last century. The efficient-market hypothesis has not worked and cannot work for this reason. By contrast, the robust psychological finding that investors frequently overreact supports the investor overreaction hypothesis.
Because the investor overreaction hypothesis is based on psychological principles, it is also likely to apply in other fields where risk and uncertainty exist. The current empirical evidence of the findings supporting the IOH is significant. It is expected that with time the list of such evidence should grow considerably. A part of the evidence used to build the IOH was considered to be a series of aberrations by efficient-market believers and dismissed as anomalies.
CRITICAL IMPLICATIONS OF IOH
1. The fact that “best” and “worst” stocks continue to exist in all markets indicates that constant rational pricing, though a cherished EMH concept, has never existed. The investor overreaction hypothesis rejects EMH teaching and conclusions as they are currently presented almost en masse.
2. In many instances IOH is diametrically opposite in its assumptions and conclusions to EMH, and IOH reaches very different conclusions from EMH given the same facts. As we’ll soon see, a very different and hopefully improved method for measuring risk will be put forth. In the 1987 crash, for example, the investor overreaction hypothesis would have cautioned against the interaction of index arbitrage and portfolio insurance that led to a lethal market overreaction and the devastating panic that followed. One of IOH’s goals is to decrease the causes of major market overreactions and their sometimes devastating consequences. High levels of leverage and liquidity have a long-standing record of creating serious downturns, often accompanied by panic. IOH would thus caution against high leverage and massive illiquidity, because of the dangers of a major overreaction. EMH risk theory ignores them entirely, because, under its assumptions, the only risk is volatility, although, as we saw in chapter 5, this situation resulted in three major panics and crashes.
3. Unlike EMH and general economic theory, the IOH does not accept the belief that equilibrium in markets has ever existed or will exist. In a dynamic, continually changing global economy, thousands of new political, economic, investment, and company-related inputs come up almost instantaneously and are immediately integrated into the marketplace. Decision making, as a result, is always changing, and this makes equilibrium as elusive as finding the Fountain of Youth.
4. The IOH attempts to see the world as it is, not in the idealistic manner that all too many economists have adopted using the rationality assumption. By comparison, the IOH has built its assumptions on recent well-proved findings of human behavior, and its conclusions are supported by strong statistical findings that investors do behave in the manner that the IOH predicts. It may fall short of a ten commandments for contrarians, but remember, all twelve predictions of the IOH have been confirmed to a high level of statistical probability in the studies reviewed in chapters 9 and 10.
Section II. Four Contrarian Strategies Derived from the IOH
Next we are going to discuss how to use contrarian strategies to boost portfolio results using the findings we have just garnered from the investor overreaction hypothesis. It’s important to understand that with ideas, just as with a college course, it’s not the amount of information we cram into our brains that does any good but how we apply critical thinking to the subject matter.
MORE RECENT CONTRARIAN PERFORMANCE
You are certainly entitled to ask if these contrarian strategies or any others still work in this new, sometimes alien, investment world, where many of the rules seem to have been washed away. The answer is a definite yes, as Figures 11-1 and 11-2 demonstrate. They cover a very short but explosive period of time, including the bursting of the dot-com bubble in the 2000–2002 period, followed almost immediately by the housing bubble, the ensuing financial crisis, and the market rebound in 2009.
This time was anything but a cakewalk, but out-of-favor stocks appear to have weathered the storms much better than favored stocks. The chart shows that $10,000 invested in any of the four contrarian strategies from 2000 through 2010 would have outperformed the market through this period. Low-P/E stocks provided the highest return, 11.7 percent annually versus 5.6 percent for the market in what is becoming known as the “lost decade.” The three other metrics, low price-to-cash flow, low price-to-b
ook value, and high yield, also outperformed the average, again the Compustat 1500. All four metrics of favored stocks again underperformed the market and all but the low yield had negative returns. The best was low yield, at 2.4 percent annually. The worst were high price-to-cash flow (–2.9 percent annually) and high price-to-book value (–3.0 percent annually). The differences in performance between “best” and “worst” stocks were large. Both low P/E and low price-to-book value returned 11.7 percent more than their higher growth counterparts.
CONSTRUCTING YOUR PORTFOLIO
How, then, can you build a portfolio that should outdistance the market while providing better protection when the bear growls? And, since a good sell discipline is one of the hardest things to develop, what guidelines should you use? The guidelines I’ll add in this chapter and the next, including several new ones based on the 2007–2008 market collapse, though certainly not guaranteed to get you out at the top (if you know of any that are, please write), have a high probability of success.
But before figuring out when to sell to protect your profits, let’s figure out what to buy. Fortunately, there are four proven ways to do this.
CONTRARIAN STRATEGY 1: LOW-PRICE-TO-EARNINGS STRATEGY
The low-P/E strategy is the oldest and best documented of all the contrarian strategies and the one most used by market professionals today. Although there are many ways to calculate a P/E ratio, the most common is to take reported earnings for a company (before nonrecurring gains or losses) for the last twelve months and then divide them into the stock price. The strategy has outperformed in both up and down markets since the mid-1930s and probably will for a good deal longer.