The still unfolding economic and financial crisis has been hailed by some—too many—as the end of capitalism, justifying stronger, more pervasive and durable state intervention in the economy. A similar and related argument on an academic level is now gaining popularity, namely that the crisis has finally disproved the efficient markets hypothesis (EMH). The EMH, in a nutshell, asserts that markets are informationally efficient, meaning that relevant information is quickly reflected into market prices as rational investors try to exploit profit opportunities—and implying that it is impossible to systematically beat the market. Some critics now argue that the EMH had so far been universally accepted, thereby becoming a key cause of the crisis, but has now been once and for all disproved.
I find such assertions disingenuous, as well as internally inconsistent—disingenuous, because the EMH has been challenged for about thirty years, and internally inconsistent because the crisis has been brought about by behaviors that display a blatant lack of belief in the EMH. Recent popular criticism sets up the purest version of the EMH as a straw man to quickly burn down and argue that the idea that individuals are rational and markets efficient is stupid and should be discarded. This strikes me as misguided and troubling, as it is part of a broader trend that includes attempts to bury capitalism and claims that what we have witnessed is not just a financial and economic crisis but a moral crisis, brought about by the widespread abandonment of ethical or religious values. This is madness. Neither psychology, nor philosophy, let alone religion can or should provide the main guidance out of this crisis. Human beings are flawed and occasionally prone to mass delusions, but they do respond to incentives, and collective behavior is if anything more rational than we are ready to admit. And while the caricature version of the EMH sounds absurd, the underlying notion that if there are obvious profit opportunities people will flock to exploit them until they disappear is at least as reasonable and common-sense as the starting observations of behavioral finance. The EMH still provides an extremely useful working assumption. Academic research is in the process of augmenting and improving it with the contributions of behavioral finance, evolutionary psychology and bounded rationality to provide an even better framework—but simplistic attempts to throw the EMH out of the window will not improve our understanding of financial markets, nor help us strengthen institutions to limit the risk of future crises.
With celebrations of the end of anglo-saxon capitalism still in full swing, a more high-brow version has recently emerged in the form of sharp criticism of the EMH. The Financial Times’s Capital Markets Editor, Gillian Tett, citing surveys by the British Chartered Financial Analyst Institute, wrote recently that “A new realisation has dawned among the most fervent advocates of financial analysis and collective investor wisdom: markets are not always rational” (“Credit crunch causes analysts to rethink rational market theory”, Financial Times, June 16, 2009).
And always in the Financial Times, James Montier blames universal acceptance of the EMH for the benchmark performance management based on the capital asset pricing model, which has caused investors to become obsessed with trying to forecast the future better than their peers (“The efficient markets theory is as dead as Python's parrot”, Financial Times, June 25, 2009). He also blames the EMH for the prevailing approaches to risk management, shareholder value and market-based regulation that he says have been instrumental in bringing us to the current crisis. And all this when, he says, “The most damning evidence against the EMH scarcely merits discussion in academic circles. The elephant in the room for EMH is the existence of bubbles.” These arguments seem disingenuous, especially coming from someone who is a brilliant expert in behavioral finance.
It is internally inconsistent to argue that the EMH had been universally embraced by market participants, and thereby played a major role in triggering the crisis. I completely agree that asset bubbles flag a glaring contradiction with the EMH: investors who jointly pump more and more money into an asset in the common unshakable conviction that its price will continue to rise clearly do not believe that the current price reflects all available information. Similarly, investors willing to pay the “2 and 20” fees traditionally charged by hedge funds obviously did not believe that it was impossible to beat the market. But then I really do not see how one can argue that it was widespread blind faith in the EMH that brought us to the crisis. If anything, the opposite is true: it was the widespread conviction that systematic profit opportunities were available on a sustainable basis.
The EMH is also blamed for spawning the capital asset pricing model, which frames investment based on a trade-off between risk and return. Indeed, what could possibly be worse? Much better the soothing belief that high return can be obtained at zero extra risk, that is exactly the attitude that had led to unreasonably compressed spreads on risky assets on a global scale, right? Once again, I am puzzled at why the EMH would be the most obvious culprit here.
Moreover, the recent criticisms of the EMH are simplistic to a degree that borders on the intellectually dishonest: they set up the EMH as a straw man by outlining its most extreme version, then argue that (a) it has been universally and uncritically accepted until now, and (b) the recent crisis has finally exposed it as a glaring fraud. But this way of framing the issue conveniently ignores a few decades of intense academic debate. Some of the best academic economists have been criticizing the EMH well before the bubbles and crashes of 1987 and 2000 and others have been defending it after the same events, so the issue cannot be as simple, and the idea that the latest crisis will finally settle the dispute seems far-fetched.
In a nutshell, the EMH states that markets quickly reflect all publicly available information, so that it is not possible to earn above-average returns without running above-average risks. This also implies that asset prices behave as a random walk: tomorrow’s price equals today’s price plus an unpredictable shock, so that today’s price is the best predictor of tomorrow’s price. The hypothesis was developed independently by Eugene Fama and Paul Samuelson in the 1960s, and became dominant in the 1970s, in conjunction with the rational expectations theory (that economic agents form expectations using all relevant available information, and that expectations are not biased in a systematic way). To say that the EMH was “dominant”, however, does not mean that everyone believed it was a precise representation of real life markets. In fact, Fama (1970) himself reported some notable anomalies.
The 1980s (that is 20-30 years ago already…) saw a strong debate on the numerous anomalies identified in the markets, and on whether they meant the EMH should be rejected. These anomalies included: short-term momentum (prices react only slowly and gradually to new information); long-run return reversals (mean reversion over longer horizons of 6-10 years, seen as supporting contrarian strategies); size effects (small caps outperforming large caps over long periods); seasonal and day of the week effects.
Some of these anomalies are well-known to market participants, handed down from generation to generation in handy tidbits of popular wisdom such as “the trend is your friend”, “sell in May and go away”, etc. These are seen as anomalies in the sense that they appear to be systematic ways of earning excess risk-adjusted profits. Supporters of the EMH, however, counter that (1) most of these anomalies are not at all systematic, and in fact tend to disappear as soon as they have been discovered in the data; (2) some of them simply hide extra risk that justifies the extra return, as in the case of small caps; and (3) the excess returns are so small that they are wiped out as soon as you allow for any plausible level of transaction costs (see for example Malkiel 2003 and 2005).
A more serious challenge came from the observation that the S&P is more volatile than the discounted stream of dividends of the underlying companies. This seems to flatly contradict the EMH: under the EMH, stock prices are the optimal forecast of the present value of the future stream of dividends, and as such the former should be less volatile than the latter. This criticism, known as violation of the variance bound, was raised by Shiller (1981) and Le Roy and Porter (1981) (see also Shiller 2003). However, subsequent analyses showed that such violations can be consistent with the EMH if, for example (1) company managers smooth dividends (Marsh and Merton (1986)), or (2) individuals are risk averse (Michener (1982)).
In the 1990s, the EMH came under a second wave of attack from the new field of Behavioral Finance (BF). BF was rooted in the work of psychologists Tversky and Kahneman (1974, 1979, 1981), who showed that human assessments of probabilities display systematic biases: overconfidence, loss aversion, herding, regret, psychological accounting, all of which lead investors to make systematically sub-optimal and occasionally disastrous choices. BF proponents point out that the breath-taking speed at which bubbles inflate and the suddenness with which they pop cannot be explained by fundamentals, as these almost never change so much so fast. They point therefore to feedback models, where through excitement, word of mouth and bandwagon effects rising prices trigger expectations of further price rises in a typical bubble movement, until valuations reach such vertiginous levels that some investors suddenly take fright and sell, breaking the spell of the mass delusion.
Supporters of the EMH, however, note that markets can be efficient even if market participants behave irrationally, and that psychological factors can influence prices, although not forever (Malkiel 2003). This defense could be seen as sophistry, as it relies on the definition of what it means to be “efficient”: under the EMH, goes the defense, an efficient market is one that incorporates into the prices all available information, including presumably the behavior of other market participants, and only under some conditions will this ensure that prices reflect the optimal forecast of the discounted value of fundamentals. In other words, EMH allows for the fact that investors will heed Keynes’ paradoxical warning that “markets can remain irrational longer than you can remain solvent”: if everyone around you is swept away by collective madness, betting against the trend may not be the optimal/rational strategy.
Three decades of academic debate have left the controversy unresolved, and this is partly because the EMH is formulated in a way that requires additional assumptions to make it operational and therefore testable. For example, testing the assertion that prices fully reflect all available information requires making assumptions on what kind of information is reflected in the prices and according to what mechanism it is reflected. But this in turn implies that if your tests reject it, you do not really know whether you have invalidated the EMH or one of the additional assumptions (see Lo 2007).
But perhaps there is no need for empirical tests at all? Taken at face value, the strong form of the EMH seems not only implausible but internally inconsistent: information is instantaneously incorporated into the prices because market participants strive to collect it and attempt to profit from it; however, exactly as a result of their efforts, there is never any profit opportunity available; but that should then imply that market participants have no incentive to collect the information in the first place—the implication invalidates the initial assumption, and therefore the theory cannot hold (Grossman and Stiglitz 1980). But the inconsistency disappears if you take a more realistic stance: as long as there is an opportunity to reap excess risk-adjusted returns, investors will try to do so as long as the return is at least as high as the cost of collecting the information. And it is a fact that excess returns are harder to come by in markets that are highly transparent and liquid, such as US Treasuries, than in markets which are much less so, as in the case of some emerging markets equity and bond markets.
The true issue therefore is not whether the EMH is a faithful and precise representation of how markets work, but whether it is a good approximation that provides a useful working framework. Supporters of the EMH argue that this is the case, and point to the fact that professional fund managers do not systematically outperform the index, with those managers who outperform in one sample period unlikely to do so in the next. In other words, while the EMH might sound ludicrously implausible, its key implication that systematically beating the market is awfully hard seems to hold. When regulators insist on the caveat “Past performance is no guarantee of future returns”, they pay homage to the EMH.
While the pure or caricature version of the EMH may appear absurd, the underlying notion that if there are obvious profit opportunities people will flock to exploit them until they disappear is at least as reasonable and common-sense as the starting observations of behavioral finance. Everyone is familiar with the joke of the economist who walks down the street with a friend: as the friend spots a 100$ bill on the sidewalk and bends to collect it, the economist stops him, explaining that if there truly were a 100$ bill lying around, someone else would already have picked it up. The joke is beautifully effective at exposing the EMH’s paradox. And I will readily confess that, on the one occasion when I have happened to spot a $100 bill on the sidewalk (50 €, to be precise), I did pick it up. But I have not found another one since—and anecdotal observation will confirm that you cannot count on systematically finding 100$ bills along your daily commute.
The EMH is an imperfect but extremely useful guiding hypothesis, and academic research is constantly underway to augment it with other hypotheses to create an increasingly more precise framework: for example, the adaptive markets hypothesis (AMH) seeks to combine EMH with BF: based on evolutionary psychology and Simon’s (1955, 1982) “bounded rationality”, it argues that different trading strategies and different “species” of traders and investors coexist and adapt to a changing environment. Different investment strategies undergo cycles of profitability and losses as market conditions change. In this framework the EMH is the steady state to which the system converges under environmental conditions (Lo 2007).
To sum up: it is disingenuous to argue that universal and uncritical acceptance of the EMH was at the root of the crisis. The EMH has been challenged and criticized for the last thirty years, in a controversy that is still unresolved, and will not be resolved by the current crisis either. If anything, the crisis has been fuelled by behaviors that displayed a blatant disregard for the EMH. The latest bubble confirms that markets can be frighteningly efficient at amplifying periods of collective madness with disastrous consequence, and that the ideas of behavioral finance, bounded rationality and evolutionary psychology among others are extremely relevant to the analysis of financial markets. But the EMH’s basic underlying notion that if there are obvious opportunities to earn excess risk-adjusted returns people will flock to exploit them until they disappear is as reasonable and common-sense as anything put forward by the EMH critics. Systematically beating the market remains awfully hard, and the EMH remains an extremely useful working hypothesis. Augmenting it and improving it is extremely desirable, discarding it as hopelessly flawed and irrelevant would be just plain stupid.
