Mises Daily Articles
Following the Efficient-Markets Hypothesis into Absurdity
Austrians in the tradition of Mises and Rothbard stress the resilience of a free-market economy, and they believe that the institution of private property — along with the profit-and-loss test — will steer resources into their most valuable niches.
However, most Austrians stop short of following Chicago School economists' advocacy of the "efficient-markets hypothesis" (EMH). In its most extreme form, the EMH becomes a caricature of itself in which asset bubbles are not just unlikely but logically impossible.
In the present article, I'll illustrate one of the most severe cases of EMH-itis, in a recent blog post by Scott Sumner.
Sumner Defends QE2
Sumner is a very sharp, Chicago-trained economist who teaches at Bentley University. I go back to his blog again and again because I find his economic views insane in the sense that G.K. Chesterton described: Even though Sumner reaches conclusions that strike me as patently absurd — for example, he blames our woes of the last two years on Ben Bernanke's tight-money policies — Sumner nonetheless reaches these conclusions through an exquisite deployment of internally consistent logic and a wide range of empirical data. In truth, I would much rather debate Paul Krugman than Scott Sumner.
With that background, let us turn to Sumner's recent defense of the second round of quantitative easing (QE2) against its conservative critics. Here is Sumner responding to one of the standard objections to Bernanke's $600 billion shopping spree:
[Conservative Objection #]6. If the policy does raise [nominal] GDP, interest rates will also rise, causing the Fed to suffer capital losses on its large bond portfolio.
[Sumner's Response:] Conservatives presumably believe in efficient markets, and thus the expected loss is approximately zero. The term structure of interest rates has already priced in the expected increase in rates that will occur as the economy recovers. Yes, there is some risk, but far less than people think. The Fed is mostly buying medium term bonds, for which the price risk is rather low. And if the recovery is much stronger than expected, the gains to the Treasury would far exceed the losses to the Fed. This is NOT an argument for leaving millions of workers unemployed. Especially given that the Fed took far greater risks to save the big banks.
I imagine Sumner did not spend much time reflecting on the above paragraph, and yet if we do so we will find that there are an awful lot of problems tucked into this tidy application of the EMH.
What about the Housing Bubble?
The first reaction an Austrian economist might have is to ask, "Well if Sumner thinks there can't be a bubble right now in US government bonds, how does he explain the housing bubble? What happened to 'efficient markets' back then?"
Ah, but Sumner has an answer: There was no housing bubble, at least if that term is to have any operational meaning. The fact that people like Mark Thornton quite clearly called it, back in 2004, means nothing. In fact, with so many economists running around, you would have expected a bunch of them to "call" every downward fall in asset prices.
Elsewhere I have dealt with Eugene Fama's rejection of the very notion of asset bubbles, so we can move on.
Simplified Modeling Assumptions
One of the fundamental problems with the EMH — or at least, the popular applications of it — is that it refers to the "market expectation" of some future event. For example, Sumner says that financial markets are presumably efficient, and so the "expected loss" on US government debt must be "approximately zero."
Of course, in reality there is no "market expectation" at all; the market is a collection of individuals, and they each have vastly different views of the future. What is true is that entrepreneurship tends to eliminate (nearly) pure arbitrage opportunities. But this is a far cry from saying that "the market" quickly zooms in on a unanimous assessment of the future path of all asset prices.
Another problem with extreme models of the EMH is that they don't take into account the constraints facing traders in the real world. For example, to defend his statement about bond prices, Sumner would probably argue like this: "It can't possibly be the case that investors believe the Fed is going to lose money on its bond portfolio over the next three years. If that were the case, then people who were holding similar portfolios would dump them immediately, before interest rates spike. But this dumping would cause the bonds to lose their price now, rather than over the next three years."
But this argument breaks down once we consider the actual complexity of the financial markets. Drawing on our earlier observation that "the market" is composed of individuals, we can say that currently, some individuals are expecting price inflation and interest rates to climb very rapidly over the next, say, five years. On the other hand, another large segment of the market is forecasting the opposite and thinks we are in for a Japanese-style "lost decade."
There is nothing illogical about someone believing that bond prices are too high, or that gold prices are too low. Furthermore, it doesn't follow that such a person must "put his money where his mouth is," by continuing to short bonds and buy gold. For one thing, people don't have single-point estimates of future prices, but rather have a general view of how likely certain price moves are over certain time frames. (For example, I am very confident that the average dollar price of gold will be higher in 2013 than it was in 2010. I am fairly confident that it will be at least 20 percent higher. I am not sure that it will be 100 percent higher. And I have no idea if gold on January 21, 2011 will be higher than on January 20, 2011.)
Another real-world limitation to the EMH logic is that people have limited wealth. Suppose Scott Sumner hits his head in the bathroom and invents the flux capacitor. He then uses it to go into the future and obtain a copy of the Wall Street Journal from the year 2015. Sumner compares the 2015 stock prices to today's levels and discovers the stocks that will appreciate the most over the next four years.
Undoubtedly, Sumner will be a very rich man by 2015. But what happens to the EMH logic in this scenario? Although Sumner could sell his house and everything else he owns, run up his credit cards, borrow money from family and friends, etc., etc., in order to pump as much as possible into his hand-picked portfolio, he won't completely draw the future price increases into the present. A stock that trades today at $1, but will reach $1,000 by 2015, might get pushed up to $50 because of Sumner's buying. Yet eventually he could push it up no more, even though he is dead certain it is still "undervalued by the market."
Yet another real-world consideration is the need for liquidity. A retired couple might expect severe inflation to kick in sometime during the next decade. Even so, it doesn't follow that "if they really believed that" they would dump all of their bonds. Rather, they would do what such people in the real world actually do: they would scale back the proportion of fixed-income assets in their portfolio and increase their holdings of gold, real estate, stocks, or some other price-inflation hedge.
In light of all the above considerations, I personally interpret the current situation not as one where "the market" has modest expectations of price inflation, but rather one in which the inflationist camp has flocked to gold and other commodities, while the deflationists have flocked to bonds. In the short run, market prices adjust until everybody is happy with his or her holdings. And yes, one of these groups is going to be proven right, at which point they will have every right to say, "We told you so."
Sumner could ably rebut these remarks, but that would just reinforce my view that the efficient market hypothesis is a perspective, a way of viewing the world. It is a tautology. There are no facts that could cause the true believer in the EMH to say, "Oh, I guess financial markets aren't efficient after all."
What If Richard Pryor Ran the Fed?
Before closing, let me address one last interesting implication of Sumner's view. He is effectively arguing that the Fed couldn't lose money if it tried. Anyone who enjoyed the Richard Pryor film Brewster's Millions can be thankful that the script was not vetted by a Chicago School economist.
Consider this: Before Bernanke announced his intention to buy $600 billion worth of government bonds, their price (and yield) was a certain amount. At that point, had we asked Scott Sumner if government bonds were undervalued, he would have said, "We can have no reason to expect so, otherwise the market would have bid them up."
Now when Bernanke announces QE2, the prices of the targeted bonds change. Maybe they go up (pushing down yield), just as the conventional Keynesian rationale suggests. On the other hand, they could conceivably go down (pushing up yields), as investors expect higher price inflation and/or stronger economic growth because of the Fed's announcement.
But either way, in the new equilibrium Sumner would once again say, "Now bond prices are correct, in the sense that we have no reason to expect them to go up or down."
This is a very odd result. There is nothing paradoxical per se about both the pre- and post-announcement bond prices being "correct" in Sumner's book, even though they are different prices. After all, things change when the Fed injects $600 billion in new reserves into the financial system, so what was the "correct" price before need not be the "correct" price afterward.
Yet what is odd is that Sumner's argument doesn't rely on the Fed investing wisely. In other words, the reason that financial markets are supposed to be "efficient" is that investors are presumably interested in increasing their wealth. Yet if there is one major investor with trillions of dollars at his disposal, and who buys assets in order to achieve nonpecuniary goals, then shouldn't this affect the argument?
Let me put it differently: If a hedge fund begins buying a stock with a small amount of outstanding shares, this increase in demand would push up the price. An EMH proponent would presumably say that the hedge-fund manager had inside information, i.e., had reason to believe the stock had originally been undervalued, and so now the higher price reflects this information.
But what if it's Ben Bernanke who is snapping up shares of the stock? Surely the same buying pressure would push up the stock price, at least in the short run, even though Bernanke has no special reason to assume the stock is undervalued.
As I have stressed before, Sumner is an incredibly clever and slippery fellow. He could argue that investors would take into account the Fed's purchase of the stock, and thus completely offset the Fed's increased demand, such that the price didn't budge. Or Sumner could acknowledge that the Fed's buying would raise the price, and that this higher price was now correct precisely because of the buying.
If the reader really wants to push this logic to the extreme, he can consider a situation in which the Fed begins a massive campaign to add US quarters to its balance sheet at the price of 26 cents each. From one point of view, it would be obvious that the quarters were being held above their "fundamental" value, and that the Fed is bound to eventually lose money on its portfolio. Yet, on the other hand, Sumner could argue that people could freely exchange their quarters for 26 cents, not only to the Fed but also among each other, so long as everyone expects the Fed's price-support program to last at least another week.
Would quarters be in a bubble? That would depend on one's philosophy. And if the Fed decided to start buying quarters at a price of 30 cents each, I would say that the bubble got larger, setting the holders of quarters up for an even bigger crash.
As the example of quarters demonstrates, economists do not all agree on the notion of equilibrium in the setting of financial markets. Even the critics of the EMH concede that there can be a self-sustaining equilibrium in (what they would call) an asset bubble. After all, the reason home prices didn't collapse in 2004 is that people believed — correctly! — that they would continue to rise in 2005.
Yet the further a bubble proceeds, the more fragile it becomes. It is analogous to a marble perched atop a steep hill, as opposed to a marble at the bottom of a basin. As the Fed continues to pump billions more into the financial sector, the various marbles are being pushed farther up the hill.