He looks really menacing, but he's a nice guy. Evangelista "Cyborg" Santos training at Chute Boxe. The video shows the high-octane MMA technical/conditioning circuit that is typical of the morning session (pro MMA team) at the Chute Boxe Academy.
Thursday, June 3, 2010
(Salvador Dali, The Persistence of Memory)
Today's post is the first of two that will focus on the strategy of "value investing", which typically entails a search for equities that have been mispriced by the market. The core claim of the value analyst is that he or she is in possession of superior valuation tools: the analyst can forecast future free cash flows for a given company with greater precision than the overall market can, and this creates an opportunity to "beat the market" by investing in target companies before the market realizes its mistake and corrects the mispricing.
As we will see down the road, there are strong, auto-reinforcing philosophical arguments underlying Keynesian economics, behavioral economics, and value-seeking investment strategies, and proponents of the three individual fields have tended to form alliances with one another. For brevity, I will simply term this alliance "the Trinity" here.
Today's post will comment on one of the theoretical foundations of value investing: the notion that market prices are not efficient (i.e., they fail to rationally account for existing information) and that, as a result, serial mispricings occur. Often the evidence that is presented to support this claim is the observation that market prices do not conform to a statistical straitjacket called the Gaussian random walk.
Behavioral Finance Takes on the TP Model and Market Efficiency
Let's begin with a theory of market behavior that is anathema to most members of the Trinity. To the economists of the Chicago school, free markets are "efficient"---all available information is embedded in prices. If one can equate efficiency with a sort of innocence, and equate inefficient pricing with a market that is guilty, the Chicago school advocates a particular default assumption that markets are efficient until we can establish otherwise---they are "innocent until proven guilty". This assumption has been specified as the "Tight Prior equilibrium" model: prior to any exogenous shocks, we should assume that the market clearing price was efficient and that equilibrium was achieved between supply and demand.
If supply and/or demand forces are being artificially constrained or subsidized by government interventions, then we of course have no guarantee of efficiency, since true costs and values are not being embedded in the prices that clear the market. The TP model assumes a free market pricing mechanism is being allowed to work; it is a normative model rather than a strictly descriptive one.
The Gaussian Battlefield
If one can imagine the existence of a Cold War strategic paradigm within modern economics, with Keynesians occupying the space formerly reserved for Communists and the Monetarists occupying the Western role, a logical extension would be the existence of skirmishes or proxy fights in which the major forces would clash over seemingly small, perhaps technical details. These details are actually very significant battlegrounds because of a version of the domino-effect theory that the West used to explain the risks of Communist expansion---for a dynastic antagonist, losing on one of these remote fronts can lead to a cascade of failures and, ultimately, to losing the whole war.
One such intellectual battlefield front specifically concerns the TP model. As noted previously, TP describes a default assumption about markets---markets are innocent (read that as "Pareto optimal") until proven guilty. The Pareto optimality provision means that there is no "free lunch" embedded in market prices---if prices are manipulated by an external force, someone will be made worse off in order to make someone else better off. The optimal distribution of resources in the economy will take place if markets are just left alone and allowed to freely find their own clearing price levels.
So, to give just one example, a government intervention policy that raises the costs of imported sugar in order to protect domestic sugar farmers does not really create jobs on a net basis---some jobs may be protected, but only at the cost of other jobs in the economy which can never exist because consumers and producers are paying higher prices for sugar than they otherwise would. Because these latter jobs may never get the chance to be "born", a particularly foolish or deceptive politician may claim that his protectionist policy has had a net positive effect on the unemployment rate.
TP is the sibling of another famous Chicago school intellectual tour de force, the Efficient Markets Hypothesis (EMH). Suffice to say that assuming prior equilibrium basically means that we should start any real-time analysis of a market with the anticipation that all currently available information has already been priced into its prices---any gross "irrationalities" have been removed by arbitrage traders who move in quickly to profit from such distortions (Chicago school economists frequently refer to the "no-arbitrage" condition---any free money will be quickly seized by traders).
Efficient=All (Publicly) Available Information Priced In
The equilibrium model allows for prices to move when new information becomes available and disturbs the market's current clearing price (what constitutes "new information" is quite slippery, since two people having access to the same report can leave with completely different interpretations). This has several important investment and policy implications: for instance, what some would term to be long-term, clearly irrational asset bubbles and crashes are not actually irrational. This interpretation would suggest that extreme values can occur for brief periods before arbitrageurs move in and correct the mispricings, but the distortions cannot persist unless aided by government policies that create moral hazard.
For instance, the dot.com bubble of the late 1990s and 2000 could be partially explained by the government-organized bailout of the giant hedge fund Long-Term Capital Management (LTCM) and Greenspan's surprise rate-cut in the bailout's immediate aftermath, which may have added fuel to the exuberant environment by creating the sense that the Federal Reserve would not allow a recession to take place. The Dutch Tulip Bubble---normally considered one of the top exemplars of financial folly---could be partially explained by the fact that the Dutch government had decided not to enforce futures contracts and mark-to-market margin requirements, causing speculative tulip-futures buyers to feel that they would be able to escape from obligations if the market was to collapse (and they were ultimately proven correct in this assumption).
If the market processes known information efficiently, the ability of an individual analyst working in a fair environment (i.e., without access to insider information) to spot great bargain stocks for buying or greatly overpriced stocks for short-selling is highly suspect; we should assume that bargain stocks are bargains for a good reason (risk), and that expensive stocks are expensive for a reason (strong growth prospects, niche dominance, etc.). Famous value investors like Warren Buffett have long felt insulted by the Efficient Markets Hypothesis, as it tends to explain their successes as being essentially the result of luck. In fact, it has been alleged that Buffett has been unwilling to give large donations to his alma mater, the Columbia Business School, in large part because CBS finance professors, like most financial economists, are disciples of the EMH.
The question of whether or not free markets are efficient is one of the most important in all of social science, with ramifications that can directly affect the livelihoods of billions of human beings. This question in part hinges on a particular application of statistical inference to market prices.
Efficient=Random: Enter the Gaussian Random Walk
In the natural sciences, a hypothesis must offer testable and observable predictions. A hypothesis which makes many successful predictions and which has coherent, tightly packed explanatory features may eventually be elevated to the status of a "theory." Given the Chicago school's general desire to grant the field of economics a mathematical rigor and elegance that was analogous to that achieved by physics, many have felt that the Efficient Markets Hypothesis and the TP Model should make testable predictions about market price behavior: unless we can test the claims of these arguments empirically, the EMH and TP should be considered merely ideological positions. To properly specify the EMH, the testable statement was made that prices in an optimizing equilibrium regime should be observed to follow a regular random walk.
This can seem counter-intuitive at first---we are conditioned to normally equate randomness with chaos or disorder; the idea that randomness is a produce of efficiency may seem alien at first encounter. The idea here is merely that efficient markets must have random price behavior because non-random behavior would be predictable, which would mean that participants would take advantage of the easy profits, and by doing so they would quickly push prices back to randomness again.
To illustrate this, let's pretend that the S&P 500 always rallied in the morning and then fell in the afternoon. Traders would observe this and would buy in the morning, which in turn would push the prices up and prevent the rally from happening. If they later sold in the afternoon, they would be doing so without capital gains from the morning rally and they would lose money. Soon you would have people buying low in the afternoon, after the selling took place, and then selling the next morning in the midst of the buying frenzy. Of course, everyone would try this, too. After a very short period of time, this "inefficiency" or pocket of predictability in the market would disappear because of the profit-seeking behavior of individual traders and firms.
Inside the Random Walk
This now gets very contentious, but at least a traditional, statistically-specified interpretation of the term "regular random walk" would indicate that price behavior will more or less comply with the normal distribution and a diffusion process called "Gaussian Brownian motion" (this is the "standard deviation ((sigma)) * t^.5 rule that was described in an earlier post on chaos, earthquakes, and market prices).
Using this form of random walk to model mark price behavior has the advantage of allowing for the importation of a number of useful mathematical techniques from the natural sciences, and for some well-known tools of statistical inference to be used as well. In fact, the desire to be able to use the financial analysis and portfolio management techniques that emerge if prices follow these rules is so great that even those who will cite Taleb and Mandelbrot and admit that price changes are not Gaussian will often schizophrenically want to retain the Chicago-developed Gaussian random walk armamentarium. The approach is widely taught in MBA and financial analyst training programs, and the toolkit includes the Black-Scholes Merton model for option-pricing, the Sharpe Ratio, mean-variance portfolio optimization, and many more. What you frequently find is that a few hundred pages of the popular textbooks are dedicated to tools and concepts that require a Gaussian world, and then a concluding chapter cautions the student that there is strong evidence that markets are not Gaussian animals. The student is left to deal with the dilemma that results from this on his or her own, and the author imitates Pontius Pilate and wipes his hands clean of the whole thing.
Conflict with Observed Market Price Behavior
(fat-tailed scorpions like Androctonus are among the deadliest in the world)
The EMH and TP are difficult to attack directly because they are largely conceptual, so the typical entry point is the regular random walk and the t^.5 rule. As mentioned in previous posts, market prices changes are not normally---or lognormally---distributed. Prices are far more "leptokurtic" ("fat-tailed") than they *should* be---a far greater frequency of large price moves is observed than would be even remotely conceivable if markets followed the t^.5 kind of random walk. Particularly after 2008, the question is not whether the strict random walk is actually obeyed in real-life; the real question is whether or not prices must conform to the random walk for market efficiency to be held as true. An even better question, to my mind, is whether markets can be efficient if governments continue to intervene in them in ways that create perverse incentives.
To date, there is no great consensus within the economics profession on this point. Behavioral economists have generally tried to fill the vacuum by using psychological biases in individual humans---mindless, lemming-like herding behavior---to explain these price divergences. Some see a multi-sigma one-day price drop as evidence that markets are flawed; others see it as evidence that markets work very well, and can violently adjust to new information---particularly after prices have been distorted by government policies for an extended period. There is also evidence that procyclical booms and busts are the result of homogenization within the investment products industry: the wide-scale availability of index-investing products, for example, may have increased the potential for tipping points and positive feedback loops to form that can contribute to large divergent moves in the market (more on this later).
Equity Risk Premium and Market Survivability
An obvious question that emerges from the discussion of random walks is why an investor would feel enticed to participate in the stock market if there is a 50/50 chance of an up or down move at any given point. The answer is that there is a long-term upward drift to equity market prices, and the effect of the drift is generally removed by economists before the price changes are analyzed. The random walk is meant to describe the behavior of prices net of the intrinsic return that the market must have to take into account the risk of the investment.
In one sense, then, the S&P 500 is not "random" at all: it is a Darwinian device, designed to attract and distribute capital to successful companies. Because the index is weighted by market capitalization, unsuccessful companies will have less and less of an impact on it over time, in some cases dropping out of it completely, while new companies that grow rapidly will be included.
The stock market can be thought of as an organic entity that wants to live and grow, just as you and I do. In order to live, it must pay participants---this payment is known as the "equity risk premium" and represents the competitive return that an investor must see in order to take on the risk of being the last guy in the capital stack to get paid if things go wrong. The return of the stock market obviously varies depending on the start and stop dates used as the window for calculation, but over the long term it has stabilized at approximately 7%. A concept called duration, which we will get into a bit more next time, can be useful for determining how long a simple buy-and-hold investor should anticipate having to maintain an equities index position in order to attain this fundamental return of about 7%.
Getting People to Play: Markets, Bookies, Lines, and Sports Betting
An equity market that has been around for a long time has, by definition, been successful at paying people enough to invest in it. A market that cannot attract players will die (and, historically, many markets have died). To illustrate the way that market prices move up and down in order to try to generate interest, entice trading activity, and encourage investor participation, we can consider the role played by a bookie in a sports betting operation.
The backbone of the sports betting industry and the standard NFL line bet is the "11-10 Pick 'Em". You may see that a betting line is listed as "Dophins -3 Bucs". The "11-10" descriptor indicates that you must put down $11 in order to win $10 (if you won, the bookie would have to pay you $21---your original stake of $11 plus the $10 in winnings); the "-3" means that the Dolphins must beat the Bucs by at least 3 points in order for the win to pay off.
If you wanted to pocket $100 on this game if you won, you would first need to take $100 (your payout goal) and divide by the $21 payout of a single successful bet; the answer is 4.762. So your necessary exposure to have a chance at winning $100 would be to take $11, or the amount you are required to put with the bookie in order to play 11-10 Pick 'Em, and then multiply by 4.762 (your answer being $52.38). To be able to walk away from a winning game of 11-10 Pick 'Em with $100 in your pocket, you would need to put $52.38 with the bookie. If you lost the bet, you would lose $52.38.
Many people think that the betting line that the bookie offers is designed so that he can win based on a particular outcome. That's not actually how the line is set: in an efficient gambling operation, the line is set by means of a market price discovery mechanism that is designed to find the game that would most closely represent a 50/50 chance of winning and losing.
In other words, a sports betting operation, like the stock market, will move prices up and own in order to reach the clearing level after which a random walk---50/50 odds---will be generated.
Why is this? Consider that you must be willing to put down $52.38 with the bookie in order to win $100 if the Dolphins beat the Bucs by at least 3 points. Now let's say that you want to simultaneously take the other side of the bet, so that you would win $100 if the Dolphins fail to beat the Bucs by 3 points; once again, this bet will cost you $52.38. In total, you must put down $104.76 in order to be guaranteed to win $100. You certainly won't do this, because it is clear to any rational individual player that the game would have a negative expectancy of $4.76 (this money represents the bookie's edge, which is the result of you having to risk $11 in order to try to win $10).
However, the bookie does not expect you to take both sides of the bet; the bookie merely wants to match you up with someone else who does not favor the Dolphins and who is willing to take the other side. Between the two of you, you will pay $104.76. One of you will win and the bookie will lose money on that person---the bookie will have to pay someone $100 after the game. But more than half of that money is actually coming from the $52.38 that the player put with him; the bookie is only losing $100 - $52.38 = $47.62 of his own take, and his own take is being financed by the losing player's complete write-off of $52.38.
The bookie doesn't care whether the Dolphins or the Bucs win because in this case he's making nearly 5% without risk. His fear is that he will not be able to match gambler against gambler and net out his risk exposure; if 50 people bet on the Dolphins and only 5 bet on the Bucs, the bookie could be annihilated if the Dolphins won (although he would make a killing if the Bucs did). Most bookies actually hate directional risk, so they make sure that the line---the betting odds---will move around to find the pool of greatest liquidity, or the point that will attract equal numbers on each side of the bet. If the Dolphins were more heavily favored by the majority of the sports gamblers, then the line might move to Dolphins -10 Bucs, or even higher, in order to entice people to take the side of the Bucs.
(by the way: online casinos will sometimes offer "21-20 Pick 'Em" odds, which are a bit better from the gambler's perspective. The way to make serious money in sports betting is to find two different bookies who are offering two different lines on the same game, and then to arbitrage them. This is called "middling the line").
In a free market, the "bookie" is basically economic growth under a capitalist system---when we are lured to play, we are providing risk capital to the Darwinian engine that drives growth. When a stock is currently priced at $100, it means that the last transaction that took place between buyer and seller took place at $100---$100 is what economists term the "price at the margin." The closing prices on a given day do not reflect the "average" of trades for that day; they simply reflect the final trades.
From that initial clearing price of $100, the stock's price may immediately go to $110, which would mean that the person who sold at $100 sold too low. Or perhaps the price will drift to $90, which will mean that the buyer at $100 bought too high. The point of market efficiency is that the marginal price reflects a bet condition in which a buyer and seller both felt that a price was fair enough for a voluntary transaction to take place. Barring some incredible information, the stock won't suddenly trade at $5 or $5,000 per share because no transactions could occur at such extreme levels.
When an extreme event occurs in the markets and prices move far more than they should in a Gaussian world, it is something like a sports betting line making a violent adjustment when it becomes public that a star quarterback is out because of an injury sustained in a closed practice. The U.S. equities markets are very sensitive to changes in the economic landscape because stock prices reflect anticipations of company-generated cash flows that extend literally decades into the future. Prices in the market will move to whatever clearing level is necessary to find the next marginal buyer and seller. If an analyst, behavioral economist, or pundit is bemoaning a given market's detachment from "fundamentals", he or she can always enter a trade at the margin.
Thus, in perhaps a counter-intuitive way, a healthy market is always trying to move the betting line---the price---so that a large number of eager buyers and sellers will participate and a random walk will ensue. When a major move occurs, it means that the betting line must shift in order to attract more players; where some consider this to be "market failure", I would submit that a more correct term might be "buyer failure" or "seller failure" depending on one's particular point of view (i.e., if you are a seller and upset because no one will buy your goods at the price you would like them to do, then you would call this "buyer failure"). The market only fails if prices are not able to adjust to find a level at which a deal can get done. Whether the adjustments happen gradually or violently is not really the market's fault.
Those who have already placed their bets may not approve of a shift in the betting line or a sudden move in market prices against their positions, but they would be well-advised to remember that the market, like the sports betting bookie, is not really interested in who is "right" so much as it is interested in making deals. If we don't like a particular price change, we should ask why there was no liquidity at the price we did like, and why our fellow investors did not want to participate at that time and serve as the marginal price-settlers. The fact of the matter is that, perhaps for a variety of reasons, no one wanted to do a deal, and now the market price discovery system is trying to find a price at which buyers and sellers can voluntarily come together.
The search for equilibrium does not mean that equilibrium is permanently achieved---it is a dynamic process. Uncertainty is what gives markets life, since a price structure that could not lure people to voluntary play (because buying or selling at those prices was obviously a bad deal) would cause a market to eventually die.
The Mysterious Levy Flight
I will briefly note here that the Gaussian is not the only type of random walk, and this may turn out to be an extremely important point. Another type, the so-called "Levy flight" (to appease the pedantic, I will note that there should technically be an accent over the "e" in Levy), can account for the market's tendency to occasionally generate spectacular price excursions (and to be able to do this for almost no reason at all; big moves can happen out of nowhere, on quiet news days).
When you graph the results of a Levy flight process, you do not get the familiar bell curve of the Gaussian random walk; instead, you produce a power law distribution. As a regular reader of this blog will recall, power law distributions are common in nature and are distinguished by having very long tails---extreme moves, while increasingly uncommon, are far, far less rare than they are in Gaussian bell curves.
(Gaussian bell curve next to power law distribution: note how the left and right tails of the bell curve drop off to nearly zero very quickly, while the right tail of the power law goes on and on. The tails are where the extreme events take place)
The problem with the Levy flight approach is that using it means abandoning many of the most beloved quantitative techniques and entering a far more uncertain and imprecise world. If shocks are unpredictable (i.e., random) and can cause monstrous moves away from prior "equilibrium" conditions, some currently popular investment and trading strategies would be seen as very dangerous.
Does Levy Flight Price Behavior Mean That Markets are Not Efficient?
Now things get more complex and subtle, almost to the point of representing a philosophical point more than an economic one. What we find in markets is that there are periods in which prices do fall well within the bounds of a Gaussian random walk, but they are punctuated by periods in which prices have a non-random trend component; markets can reveal persistence---positive feedback loops that reward strength with more strength and weakness with more weakness---as defined by analytical tools such as the rescaled range analysis (Hurst coefficient)that was described in an earlier post on earthquakes and markets.
The Levy flight is still a form of random walk---there is little evidence that forecasters can predict when and where markets will suddenly unleash terrifying, non-Gaussian price excursions. Many value investors have been caught on the wrong side of the downside moves, and have blamed them on "inefficient markets" or "animal spirits" (as if markets existed for their benefit and should be judged by how they reward any particular class of investment or trading strategy).
"Efficiency" in this context merely means that markets have a powerful way to react to the level of uncertainty in the environment, and of course the level of uncertainty is not stable. There is no requirement that markets cooperate by behaving in ways that are friendly to Gaussian statistical inference tools---markets do not "fail" when prices fall viciously to reflect an emerging reality, they would fail if they allowed investors to continue on the former course. Indeed, Nature itself is characterized by long periods of gradual evolutionary change punctuated by extreme events in which very aggressive selection pressures force massive changes. Natural history contains numerous mass extinction periods---"Great Dyings"---in which significant percentages of the dominant lifeforms on earth were wiped out, while former marginal players became the new big winners.
Crisis-Hunting Trading Stategies
In other words, markets are very efficient in that the transitions between the Gaussian world, which Taleb terms "Mediocristan" and which we could refer to as Regime A, and the Persistence-driven world, which Taleb terms "Extremistan" and we could refer to as Regime B, do not normally come with any warning. There is a deep unpredictability to their arrival, and attempts to find patterns or reliable indicators that would allow for the precise prediction of the timing of extreme moves have generally failed.
(billionaire hedge fund legend Julian Robertson shorted the NASDAQ bubble and, visibly shaken during a TV interview in 2000, was forced to shut down his previously successful Tiger Fund after sustaining large losses. As we all know, his investment thesis was later vindicated. Robertson is currently very bearish on the US economy and has put on a massive inflation bet by going long 2-year Treasury debt and short 10-year, thereby speculating that the yield curve will steepen sharply in response to a coming, possibly apocalyptic debasement of the US dollar. I am highly sympathetic to this macro position and will go over the mechanics of the problem in a future post)
This is why trading strategies based on "crisis hunting" or "persistence capture" are psychologically grueling: if the goal is to always make sure that the trading program is well-positioned within a breakout trend, the crisis-hunting program must treat every divergent move as if it was in fact the beginning of a breakout trend. Unfortunately, the majority of these divergent moves will be head fakes and the program will lose money (albeit small amounts of money) as Regime A dominates and prices quickly retreat back to their recent historical norms. The strings of small losers can go on for months until eventually Regime B arrives again and market behavior erupts to the upside or downside.
In an ideal world, the program would be able to tell that the markets were in Regime A and that any sudden price excursions were not going to turn into sustained trends. The program could ignore the head fakes and simply avoid trading during these quiescent periods (or even temporarily implement a naughty, pro-Gaussian counter-trend strategy), but then could jump back into the market just in time to get positioned in front of one of the monster waves of Regime B and surf it to great profits. I have never heard of a legit trading operation that has been able to pull this off repeatedly---the market's ability to switch between regimes is just too clever. In reality, the appearance of the monster waves appears to be unpredictable, and so crisis-hunting trading programs tend to have months of relatively flat, unexciting performance (characterized by small losers and winners), punctuated by dramatic periods of outsized gains. Possession of the ability to somehow systematically distinguish between the head fakes and the true breakout trends remains the Holy Grail of trading.
To summarize, the tricky part is this: the periods of divergence from historically quiescent periods come at random intervals, although both the severity and frequency of the extreme moves seems to be increasing as a result of the widespread use of some of the same innovations, strategies, and portfolio construction techniques. If a trader or strategic investor is equipped with an accurate analytical framework, then perhaps he or she can at least determine that certain environments are "primed" for big waves (relatively loaded up with latent potential for violence), but getting the timing right with any degree of precision is extremely difficult.
Efficient Markets Insult The "Elite"?
Many of us are very uncomfortable with the notion that our lives may be so governed by random processes. To someone who has aspirations of organizing and leading masses of people towards collective, political goals, the concepts of market efficiency and random walks go beyond discomfort and approach the level of true trauma, since market efficiency highlights distributed intelligence and individual decisions and sharply discounts the value of collective action, the acquisition of specialized training in forecasting techniques, and the prudence of commitment-based strategic planning (as we have discussed, a strategy based on heavy pre-commitment of resources towards a centrally-planned aim is only really rational if accurate predictions can be made regarding future states of the world).
"Progressives" who seek to harness the power of the masses for unified, centrally directed sociopolitical goals have long felt that the hands-off, essentially libertarian message of the EMH leads to a kind of anarchy and a celebration of self-interested behavior that subordinates the glories of collective action (and, of course, the attractive and intriguing "leadership" possibilities that such actions may present) to the insolent cult of the rugged individualist.
Here is Robert Shiller of Yale, one of the leading intellectuals of behavioral economics, writing on the notion of market efficiency in his widely-cited and enjoyable book Irrational Exuberance:
The theory that financial markets are very efficient, and the extensive research investigating this theory, form the leading intellectual basis for arguments against the idea that markets are vulnerable to excessive exuberance or bubbles. The efficient markets theory asserts that all financial prices accurately reflect all public information at all times. In other words, financial assets are always priced correctly, given what is publicly known, at all times. Price may appear to be too high or too low at times, but, according to the efficient markets theory, this appearance must be an illusion.
Stock prices, by this theory, approximately describe 'random walks' through time: the price changes are unpredictable since they occur only in response to genuinely new information, which by the very fact is new is unpredictable. The efficient markets theory and the random walk hypothesis have been subjected to many tests using data on stock markets, in studies published in scholarly journals of finance and economics. Although the theory has been statistically rejected many times in these publications, by some interpretations it may nevertheless be described as approximately true.
Note that Shiller's discussion of random walk rejection only applies to Gaussian random walks; as I mentioned before, non-Gaussian Levy flight random walks are in fact supported by the evidence, and they are even more vicious to the goals of elitist central planning than are their more benign and pleasant Gaussian brethren. If you tinker with a market that is prone to Levy flights and get something wrong, you can inadvertently unleash Hell.
("On my signal, unleash Hell.")
Perhaps attempting to garner support by revealing the great threat posed to central-planning enthusiast-type intellectuals by the EMH, Shiller goes on to neatly describe the social ramifications of the Chicago position on efficient markets (a ramification that creates a temporary alliance between two groups that would seem to normally be at odds with one another---government regulators and professional investors):
At its root, the efficient markets theory holds that differing abilities do not produce differing investment performance. The theory claims that the smartest people will not be able to do better than the least intelligent in terms of investment performance. They can do no better because their superior understanding is already completely incorporated into share prices.
In we accept the premise of efficient markets, not only is being smart no advantage, but it also follows immediately that being not so smart is not a disadvantage either. If not-so-smart people could lose money systematically in their trades, then this would suggest a profit opportunity for the smart money; just do the opposite of what the not-so-smart money does. Yet according to the efficient markets theory, there can be no such profit opportunity for the smart money.
Thus according to this theory, effort and intelligence mean nothing in investing. In terms of expected investment returns, one might as well pick stocks at random---the common metaphor of throwing darts at the stock market listings to choose investments.
Thus has the efficient markets hypothesis made enemies of two groups: active managers who claim that they are smarter than the market pricing mechanism (not all active managers make this claim, but those that say they have superior stock-picking skills certainly do), and governments who wish to intervene and change prices to reflect a social engineering mandate of some kind. The efficient market is above all a threat to the notion of an intellectual elite who are blessed with special insights into the correct pricing of assets.
In fact, the Efficient Markets Hypothesis (and Friedman's related concept of rational expectations) are incredibly democratizing, since they imply that the competence gulf between academic economists (and the politicians that they advise) and ordinary market participants decays to zero over time, and non-economists learn not to be tricked by inflationary policies and deficit spending programs. For those who peddle consulting and advisory gigs that, one way or another, seek to draw a sharp distinction between the illuminated individuals who can handle the truth and the dirty masses who cannot, Friedman became a devil figure, a heretic. Rational expectations was and still is seen by some as a serious blow to the prestige of the economics profession and a generator of problem children who "practice economics without a license," to use a phrase that I believe Paul Krugman has employed in the past.
Hopefully today's post has set the stage for the next, which will try to unpack the strategy of value investing and discuss its pros and cons.