30 Sep 2010

Ten Things I Hate About the Efficient Markets Hypothesis

Financial Economics 14 Comments

Well maybe not ten, but Scott Sumner hits a bunch of my pet peeves in this post:

Commenters often present me with market anomalies, which supposedly “prove” the efficient markets hypothesis is wrong.    I always respond that they’re just engaging in data mining.  They retort that no theory that can’t be disproved is worth anything.  But the EMH can be disproved.  The tests have been done, and it passes.  Sort of.

OK, that “Sort of” is a little odd, but clearly at this point, Scott is saying the EMH is a falsifiable theory. Let’s continue:

Finance professors have done many different tests of the EMH, and I’d guess 90% of the published tests (but only 5% of the actual tests) show that the EMH is wrong.  (Yes, I’m pulling these numbers out of thin air, but you get the point.)  They’ve found January effects, small stock effects and value stock effects.  They’ve found the market does better when P/E ratios are low.  They found the market does worse on rainy days (a study published in the AER!).

Boom, that’s it, right? These are all things that, prima facie, shouldn’t be true if the EMH is right. In fact, Scott himself agrees that these tests “show that the EMH is wrong”; don’t take my word for it. So case closed, right? Nope:

Of course you’d expect to find 5 anomalies for every 100 tests you do, and for the most part you only get published if you find an anomaly, and finance professors have a lot of computer power, so . . .

Here’s my analogy.  Suppose Stephen Wynn was concerned that some mysterious gamblers were getting away with cheating at one of his casinos.  He’d heard rumors, but had no proof, or even suspects.  You are statistician brought in to investigate.  You study 600 slot machines, and find 30 of them produced three cherries more often than you’d expect from mere chance.  You suggest that the anomalous slot machines be destroyed.  How should the casino owner react to that “investigation?”  I’m guessing Wynn wouldn’t be impressed.

I’m not sure that’s a good analogy. But doesn’t matter, suppose it is. Let’s move on to what really bugs me:

To find out whether cheating is occurring you need to look at whether the winnings of gamblers are serially correlated.  Are those who win once, more likely to win next time.  That’s the proper test, indeed the only practical test, of whether people are cheating the casino.  And it’s also the only test of the EMH.  Don’t look for “the system,” the secret way to beat the stock market.  Look for whether other people have found it.  Look to see whether people who did better than average one year, tended to do better than average the next year.  Don’t look for market anomalies—look for evidence that other people have found market anomalies.

Of course the study has been done.  I recall that Fama and French found that mutual returns were approximately serially uncorrelated, but not exactly.  There appears to be a slight serial correlation among the very best funds (top 3%), but not enough to give the average investor any advantage.

And that’s what I would have expected.  The EMH is approximately true; indeed it’s almost impossible for me to imagine any other model of financial markets.  But it’s not precisely true, again, just as you’d expect.  After all, if the EMH were perfectly true then no one would have any incentive to estimate fundamental values.

Huh? This is the same trick that the evolutionary biologists pull–”Our theory is so right, that even when the data don’t support it, it just proves how right it is.”

Note that I’m not saying, “The EMH is wrong,” or even that, “Evolutionary biology doesn’t fit the data.” What I have said over and over is that the EMH is a way of viewing the world. No matter what happens, Scott Sumner would say, “See? Just what I would have expected.”

Don’t believe me? Look at this:

A smart person like Eugene Fama should have been able to come up with both the EMH, and its limits, by just sitting in a room and thinking.  Much as David Hume got the QTM by imagining what would happen if everyone in England woke up one morning with twice as much gold in their purses.  Or Fisher’s theory of inflation and nominal interest rates.  Or Cassel’s purchasing power parity.  Or Friedman/Phelps’ natural rate hypothesis.  Or Muth and rational expectations.  Certain ideas are simply logical, and that’s why I have no doubt that despite all those economists on the left arguing the EMH has been discredited, it will still be taught in every top econ/finance grad program 100 years from now, whereas fiscal stimulus will be long gone from macro textbooks.

See? If you can derive a theory by sitting in a room and thinking, then it is not an empirical theory. So let’s drop the charade and stop pointing to all of Fama et al.’s “tests” of it. Just admit it is a very useful way of viewing the world, like supply and demand. Have the courage, as Ludwig von Mises did, to say that it is an a priori approach that could not possibly be falsified.

But instead, just about every EMH supporter I have read thinks it is an empirical claim, open to falsification. They just don’t realize that they can explain everything. If sophisticated hedge funds are making a bunch of money, that just proves how hard it is to “beat the market”; you need to spend a bunch of quants and computers.

And if the hedge funds all blow up, while guys like Mark Thornton called the housing bubble in 2004? Nope, just shows Thornton got lucky, and how hard it is to beat the market. (I’m not exaggerating; that’s exactly what Jeremy Siegel said in the WSJ, defending the EMH, as I explain at the bottom of this article.)

So whether hedge funds make a killing or blow up, it just shows how rational markets are. The EMH–a scientific theory, subject to falsification–passes its test with flying colors.

14 Responses to “Ten Things I Hate About the Efficient Markets Hypothesis”

  1. Jon O. says:

    Anyone who actually works in financial markets knows EMH is bunk. Unfortunately (most) finance professors have no clue about the real world, that’s why they do what they do.

    All you have to do is take a look at a firms like Renaissance Tech or SAC or Goldman.

    I thought the LTCM blow-up proved this point 10 years ago but I guess not…

    • bobmurphy says:

      The LTCM blow-up is just what we would expect to see if the EMH were true. In fact, if we didn’t see a once-in-a-hundred-year event now and then, we would know something was wrong with our theory…

    • Peter St. Onge says:

      I’d disagree that ‘everyone’ in finance knows EMH is bunk. The consensus is, I think, closer to Graham’s ‘short-run voting machine, long-run weighing machine’. Finance, as a whole, pays attention to fundamentals as well as technicals.

  2. Martin OB says:

    So … are you a creationist?

    • bobmurphy says:

      Probably not in the sense that you mean by the term. But I definitely think the popular discussions of how well evolutionary theory passes its empirical “tests” are fooling themselves.

      I wrote a lot about this stuff a few years ago, but unfortunately those essays are no longer online. Some day I will see about getting them re-posted.

      • John B. says:

        I find this very interesting, especially since many Austrians seem to be so religious (correct me if I’m wrong). In fact, I personally find the market process and evolution to be almost the same in principle. In both you have chaotic, random systems with innumerable, independant interacting variables. In both, these chaotic systems are self-organizing through a relatively simple success/fail mechanism; in one it’s profit and loss, the other, natural selection. I don’t even think it too far of a stretch to say that the market processes as explained by the Austrians are simply peaceful, human-only continuations of the evolutionary processes themselves (assuming animals as property either directly through ownership or usury rights in wild animals).

  3. Peter St. Onge says:

    I think what Scott’s getting at with “After all, if the EMH were perfectly true then no one would have any incentive to estimate fundamental values.” is Andrei Shleifer and Robert W. Vishny, 1997, ‘The Limits of Arbitrage’, The Journal of Finance, American Finance Association. Basic idea is the market will be ineffient (i.e. EMH violated) to the extent it remains profitable to ferret out anomalies.

    Now, Shleifer, Vishny may work for small anomalies (calendar effects, rainy days, etc), but doesn’t quite capture the dot-com boom. I wonder if part of your differences with Scott here is that he takes the (many) small-effect papers and extrapolates to the (fewer) bouts of exuberance.

  4. Brian Shelley says:

    Didn’t we just prove EMH this past decade when housing prices never exceeded the present value of rents?

    • bobmurphy says:

      Once you take into account asset price appreciation, it was all perfectly rational. After all, if people knew that houses were overvalued, then they would sell. Bubbles are impossible. QED.

  5. Maurizio says:

    It seems to me you did not understand what Scott was saying. I think he is saying that it is a _logical_impossibility_ that the EMH be exactly true (because, if it were true, then it would not be true; more precisely: if it were exactly true, then there would be no profit opportunities, so there would be nothing to keep the market efficient, and therefore the EMH would not be true).

    So there is nothing to be tested here. We can say a priori that the EMH is not true. It is contradictory to even state that it is true.

    What remains possible, however, is that EMH is approximately true, i.e. that a _weak_ form of the EHM is true; and this theory is indeed falsifiable, and can be tested empirically. And empirical tests suggest it is true. (At least do not disprove it).

    Does this make sense?

    • bobmurphy says:

      Yeah it makes sense. But I’m not sure that it is correct that empirical tests suggest it is true. What else would have to happen, besides the dot-com and housing bubbles, for people to think markets are not “efficient” in the way that the EMH says?

      Believe me, I know what the responses are–I read Scott’s blog. My point is, this isn’t an empirical test; the EMH gives you a framework within which to evaluate financial activity.

      Look at what Scott said in his post: That he couldn’t imagine any other model of financial behavior. That’s not an empirical theory, open to verification. The world HAS to fit into Scott’s understanding of the EMH framework; he can’t imagine it otherwise. And in fact, a guy locked in a closet could have come up with it.

  6. david (not henderson) says:

    EMH depends, for its approximate accuracy, on the possibility of human error in the face of uncertainty, incomplete information and incomplete understanding. It’s that possibility presumably that motivates investors to believe that they might be able to consistently occupy the upper tail of the distribution or at least to avoid being in the bottom half. Along those lines, I read of a study not to long ago that found a positive correlation between fund managers’ IQ and supra-market returns (although I dismissed those results based on their obviously fallacious implication for my own IQ).

    I’m inclined to think that the error (committed on all sides of the debate) is to assume that EMH necessarily implies that bubbles cannot develop or exist. It seems to me that, from a trader’s perspective, money supply in excess of demand IS a fundamental valuation factor, at least in the short/medium term.. Also, I’m not sure what happens to EMH when non-profit-seeking “big players” (a la Koppl) like the central bank are inserted into the process.

  7. Silas Barta says:

    Great post, Bob. (I know, not very much content, but I can assure you, as an expert information theorist-by-night, that the surprisingness of me agreeing with you *makes* this comment informative. Indeed, only surprising data can be informative.)

    • bobmurphy says:

      True story: Once John T. Kennedy agreed with me on something, and I said, “JTK, the fact that you agree with me–which almost never happens–makes it all the more likely that what I am saying is right.” And he said, “No it doesn’t.”

      (Not exact quotes, but that’s the gist of the conversation. Ah, good times.)