Scott Sumner is a terrific economist, a creative thinker and a formidable debater. He not only knows a lot about economic theory and history, he has an uncanny knack for knowing how to draw interesting and useful empirical inferences from the theory. Plus, he’s funny and writes really well. No wonder he is just about the best and most prolifiic blogger there is. He is also a very nice guy, and has the added virtue of agreeing with me about 97 percent of the time. So why am I about to start an argument with him?
Well, for a couple of years, Scott has been using his blog periodically (here is the latest) to take on critics of the efficient markets hypothesis (EMH), and doing an excellent job, pointing out many of the lapses in the reasoning of EMH critics, for example that supposed anomalies in pricing that prove that pricing is inefficient may simply be statistical flukes incapable of providing a basis for profitable trading, which is what a true exception to EMH would provide. And Scott very effectively asks why all those people who were so convinced that there was a real estate bubble before 2007 weren’t out there shorting the market.
So let me give Scott his due and say that I don’t know of a more effective defender of EMH than he is, but I still can’t accept EMH. What’s the problem? Well, the most important empirical claim of EMH, the one that Scott uses relentlessly to bludgeon EMH critics, is that future prices cannot be predicted from past prices. The best predictor of the price tomorrow is the price today. That powerful empirical regularity seems to imply that today’s market price has already processed all the available information about the price tomorrow, so that, given the information today, the price tomorrow is already where it should be. The price tomorrow will be different from today’s if and only if some new information not now available will cause it to change. But new information, by its nature, can’t be anticipated, so even though new information causing us to revise the estimate of tomorrow’s price incorporated in today’s price might arrive, that possibility provides no basis for revising today’s price before the new information reaches us. If you could predict when new information would arrive and how it would affect tomorrow’s price, that information, insofar as it really was predictable, would not be new.
This view of how asset markets operate is related to the idea that market prices are ultimately determined by fundamentals, demand, supply, cost, taxes, etc., objective magnitudes that can be ascertained, or at least estimated, by doing enough research into the asset that one is trying to evaluate. Markets reflect the central tendency of all the various judgments about an asset being made at any time. That explains why the current market price is more likely to predict tomorrow’s price than is any single person, no matter how knowledgeable or astute, and why it so hard for any individual to outguess the market consistently.
In a famous discussion in Chapter 12 of the General Theory, Keynes offered a different view of how the stock market operates, comparing the stock market to a newspaper competition
in which competitors have to pick out the six prettiest faces from a hundred photographs, the prize being awarded to the competitor whose choice corresponds to the average preferences of the competitors as a whole; so that each competitor has to pick, not those faces which he himself finds prettiest, but those which he thinks likeliest to catch the fancy of the other competitors, all of whom are looking at the problem from the same point of view. It is not a case of choosing those which, to the best of one’s own judgment, are really the prettiest, nor even those which average opinion genuinely thinks the prettiest. We have reached the third degree where we devote our intelligences to anticipating what average opinion expects the average opinion to be. (p. 156)
When I first read this chapter as a graduate student already trained to think of markets as efficient processors of information, I thought it quite remarkable that an economist as great as Keynes could have entertained such a primitive notion of how the stock market operates. Sadly, I can no longer indulge myself with feelings of superiority when reading that chapter, because I am no longer convinced that Keynes was totally misguided in his characterization, indeed, caricature – but caricatures are effective by identifying and emphasizing some salient feature of reality — of how the stock market works.
I would note parenthetically that on Keynes’s view, asset markets are no less unpredictable than they are under EMH. So, despite the identification in the EMH literature of efficiency with unpredictability, the two concepts are not necessarily equivalent.
So why do I think that Keynes was on to something in describing the stock market as a kind of beauty contest? The key point, it seems to me, is that it is misleading to believe that there is a clear distinction between fundamentals and opinions. In Keynes’s beauty contest, the fundamental question is which are the prettiest pictures. Working with the fundamentals, you would compete in the contest by doing fundamental research on the pictures. The contest would be efficient if the winner selected the prettiest pictures. Keynes said that whether the prettiest pictures are chosen is irrelevant, because to win the contest what you have to do is to guess who the other competitors will think are (or will select as) the prettiest pictures.
The problem that I see with Keynes’s analogy is that he implicitly admits a dichotomy between fundamentals and opinions, between the prettiest pictures and pictures selected by those trying to guess who will be selected. I take Keynes a step further; beauty is in the eye of the beholder, so the pictures selected as prettiest have as much claim as any others to be the prettiest, and there are no fundamentals by which to determine which pictures are the prettiest apart from the process that has chosen them.
Let’s try to bring the argument back from beauty contests to markets. We all know that expectations sometimes can be self-fulfilling. That’s what network effects teach us. The market goes where it is expected to go. Sure there can be exceptions; expectations can be disappointed. But when expectations point toward a particular outcome, it can be very difficult to avoid that outcome, especially when network effects are strong. When expectations determine the outcome, the distinction between the expectations of traders and fundamentals starts to disappear. If depositors expect a bank to go insolvent, it goes insolvent. If traders expect the price of oil to go up, it goes up. If they are pessimistic about the economy, the stock market goes down and the economy may follow. So Keynes was right, traders in the stock market are trying to figure out where everyone else thinks that stocks are headed. That’s not the whole story, but it is part of the story, and a part of the story that is left out of EMH.
Let me cite a specific example. An important milestone on the way to the development of EMH was Milton Friedman’s paper “The Case for Flexible Exchange Rates” in which he argued that currency speculation would be stabilizing rather than destabilizing. Friedman’s reason was that in order for speculators to earn profits, they would have to buy low and sell high, but buying at low prices tends to make the prices higher than they would have been and selling at high prices tends to make them lower than they would have been. So speculators are smoothing out the fluctuations in exchange rates, raising the lows and bringing down the highs. The argument presumes that speculators on average are earning profits, which may or may not be true, but leave that aside. I want to address another unstated presumption of Friedman’s argument: that there is in fact an exchange rate consistent with the fundamentals and that sooner or later the exchange rate always comes back to the equilibrium level determined by fundamentals. The fundamentals consist of the real exchange rate determined by real factors and the monetary policy of the government and the monetary authority. Suppose speculators believe that the monetary policy will become more lax and drive the value of the currency down. The government and the monetary authority are now confronted with a choice: accept the depreciation or tighten money, raising interest rates, perhaps risking a recession to restore the old exchange rate. The government might decide that it is just not worth it to take the actions required to restore the old exchange rate even though it had no intention of loosening monetary policy in the first place. Thus, it was the expectations of speculators that created the change in fundamentals. And their expectations were both destabilizing and profitable.
So when Scott defends EMH against its critics and points out that it is very hard to find a way to make profits systematically by exploiting inefficiencies in the asset prices, I agree with him, sort of. But I can’t figure out what this has to do with efficiency.