One doesn’t read about financial markets long before encountering the efficient market hypothesis (EMH). A great deal has been written about the EMH and it has been contemplated by some very smart people. It is one of the hallmark concepts associated with the rise of financial engineering, the age of the quants (for a well-done summary of this history see Bookblogging: The Rise of the Efficient Markets Hypothesis by John Quiggin). Therefore, it is highly unlikely that anything I think or write about the EMH is novel. Nevertheless, I find most descriptions and discussions of it lacking some (to me) obvious points. (A recent exception is the 16 July 2009 article in the Economist titled Efficiency and Beyond.)
First of all, what is the efficient market hypothesis? Actually there are three versions of the EMH, each one more controversial and unlikely than the previous. The generally accepted version is the weak form: that the current price of an asset reflects all past prices. If true then it is not possible to gain any systematic trading advantage by analyzing past prices (sorry chartists!).
The semi-strong form is that asset prices incorporate all known public information and adjust instantly to any new public information. If true it means you can’t systematically gain an advantage by trading on the news no matter how closely you follow it. Finally, the strong form is that asset prices incorporate and adjust instantly to all information, whether public or private.
All forms of the EMH are argued and debated but particularly the semi-strong and strong forms. For examples cited as evidence against the EMH see the Wikipedia entry on the topic.Some participants in an argument about the EMH seem to overlook the fact that it is possible for some markets to be efficient while others not, or for a market to be efficient at one point in time but not another. Thus, one can take a position that a market is currently efficient even if there is past evidence that it was not. This perspective makes the notion of efficiency a handy descriptor though less useful as a lynch pin of financial theory.
It is instructive to consider how or why the EMH would or could hold in a market at a point in time. There is a one word answer: arbitrage. If the EMH doesn’t hold then there is the possibility of financial gain through exploitation of the information not incorporated in an asset’s price. Skilled market participants will quickly bid the price up or down in attempt to profit from the imbalance.
In fact, some will profit. The early birds. No, not you. I mean the really early birds, the arbitrageurs: the institutional investors with teams of smart analysts poised to take advantage of the slimmest of margins. What I have just argued is that if the EMH holds (in any form) it does so precisely because it did not hold for some brief moment during which somebody made a profit and the margin was arbitraged away (an argument made recently in the Economist). Therefore, the EMH cannot hold all the time everywhere, particularly not the stronger forms. (This line of thinking is similar to that expressed in Grossman’s and Stiglitz’s 1980 paper On the Impossibility of Informationally Efficient Markets, The American Economic Review 70(3)).
When markets present arbitrage opportunities they’re usually minuscule. The only way an investor can exploit one at a meaningful level is to trade an enormous sum and/or with high leverage. Very few individual investors can trade in high enough volume even if they could identify an arbitrage opportunity. And, if they could, they’re probably insufficiently diversified when they do so. Throwing all of one’s money into one asset or asset class to exploit what one believes to be a tiny imbalance is insanely risky. Therefore, while EMH cannot hold all the time everywhere, practically speaking, for the individual investor (you and me), it does.
The words “rational” and “rationality” are tossed about frequently in discussions of the EMH. Is rationality different from efficiency? The notion of rationality seems to be connected to the EMH in the following way. For the EMH to hold then market participants must collectively exhibit rational expectations, on average their assessment as to the effect of information on price is an optimal forecast. For agents to exhibit rational expectations in aggregate it is not necessary for each individual agent to be rational in the sense of maximizing his or her own benefit.
Finally, even if the EMH holds it does not imply that asset prices are “correct.” In fact, as expressed recently by James Kwak on Baseline Scenario, it isn’t even clear what “correct” means. It is possible to rigorously define an optimal price in the sense of incorporating information under some criterion of optimality. But “correct” (in a hindsight sort of way) is a different idea. It may be easy to make a subjective judgement about the correctness of a past price. But if one concludes it was not correct that does not imply the market was either inefficient or participants irrational. And even if it (they) had been it would not have been wise to have tried exploit either.
