HindsightThe first thing to realize is that many of the so-called "mispricings" that people observe are just hindsight at work. The dot-com bubble, for example, is widely cited as an obvious failure of EMH. The logic usually employed looks something like this:
- Semi-strong EMH predicts that prices reflect all publicly-available information
- Prices during the bubble were too high, given what we now know
- Therefore semi-strong EMH didn't apply during the bubble
Capital limitationsBut what if we replace #2 with this: "Prices during the bubble were too high, given what we knew at the time". While this assertion is far from proven, I concede that it's possible.
To explain this, I employ a modified version of EMH that I refer to as the Capital-Limited Efficient-Market Hypothesis:
While fair prices reflect all known information, actual market prices reflect only the information known to those with the capital and willingness to drive prices toward their fair value in search of profit.Thus, one way to explain an asset mispricing would be to demonstrate how investors are deficient in at least one of three areas:
- The dot-com bubble was characterized by a dramatic increase in capital invested according to market momentum and wildly optimistic guesses of future performance.
- The relatively few investors who knew stocks were overpriced could drive stock prices down only by short-selling them.
- Short-selling carries the risk that the investor will lose a great deal of money (more than invested) should prices continue to climb. This reduces investors' willingness to risk a lot of capital.
- Therefore, the inefficiency was able to persist because market was unable to eliminate it.
CorollariesThis rule is a lot more "fuzzy" than the original EMH rule. I wish I could evaluate this hypothesis more rigorously, but I lack the expertise to do so. I think it would stand up well. I also worry that it is so vague as to be meaningless, but I don't believe it is.
Does this capital-limited EMH make any actual concrete predictions? I think so.
The first prediction would be that bubbles crash more quickly than they rise. This would be because investors are always less willing to short a stock than to invest in it. Is there evidence of this? Not being a statistician, I'm not even sure how to check, but it does "seem" right.
A second prediction would be that Warren Buffet, an investor who excels in identifying mispriced assets, must necessarily become less effective as his available capital becomes so much that he becomes more effective at canceling the very inefficiencies he notices. This is borne out by the long-term price trends of Berkshire-Hathaway stock, and I don't think this would be a surprise to anyone.
A third prediction would be the existence of the risk premium. Higher risk reduces the amount of willing capital available to drive a price up, thereby leaving it lower than the fair price. Again, I'm no statistician, but I think it's fair to take "implied volatility" as a proxy for risk, and if you look at this chart comparing NASDAQ "implied volatility" versus price, to me it's clear there's an inverse relationship.
As a software developer, I see the market as a kind of debugging exercise in which efficiencies can continue to exist only until they are noticed by someone with enough capital to cancel them. Benjamin Graham gives several examples of these inefficiencies that got "debugged" out of existence:
In 1949 we could present a study of stock-market fluctuations over the preceding 75 years, which supported a formula—based on earnings and current interest rates—for determining a level to buy the DJIA below its “central” or “intrinsic” value, and to sell out above such value. It was an application of the governing maxim of the Rothschilds: “Buy cheap and sell dear.” And it had the advantage of running directly counter to the ingrained and pernicious maxim of Wall Street that stocks should be bought because they have gone up and sold because they have gone down. Alas, after 1949 this formula no longer worked. A second illustration is provided by the famous “Dow Theory” of stock-market movements, in a comparison of its indicated splendid results for 1897–1933 and its much more questionable performance since 1934.
ConclusionsI started writing this post in November, hoping to support the hypothesis with additional evidence and reasoning, but since I seem to have neither the time nor expertise to achieve this, I've resigned myself to posting it in its current half-baked form.
By and large, I no longer try to pick stocks. I don't have the advantage over the general market in any of the three required qualities: knowledge, capital, or willingness. Following Graham's advice, and being unwilling to become an "enterprising investor", I have no option but to become a "defensive investor", and my life is much less stressful for it!