Reflections from Europe | SEPTEMBER 7, 2009
Thank Heaven for an Inefficient Market: A Tale of Zombies and Speculators
For a recent podcast on the efficient market hypothesis and risk, see Justin Fox on the Rationality of Markets with host Russ Roberts on EconTalk.
It is now argued that it was being drilled at university and business school in this theory that misled a whole army of young bankers into trusting the "efficiency" of asset markets and entering into vast commitments that have ended in astronomical losses. Their mathematics were sophisticated and worthy of the schools they graduated from, but their underlying idea was only moderately clever: if asset prices were broadly right, there was a probability distribution that favoured the hypothesis of their staying at their present level, i.e. the risk was moderate and seemed to lend itself to calculation. Moreover, future bad news impacting one type of asset would not impact other types; risk was particular and not general.
Misplaced confidence in this reasoning was far from being the only cause of the shambles of 2007-2008, but it was an important cause.
Why is the "efficient market" theory a poorly conceived one? When it speaks of "all available information that is relevant", it must mean all information that actually reaches the attention of investors, rather than information they did not bother to notice or have put in the in-tray to be discussed at next week's investment committee meeting. On this definition, there is a fraction of the body of investors who think the news they just received makes the stock they hold worth more than the top of the price range at which they were prepared to just hold it. They should therefore try and buy more of what they have, (or buy the stock they previously thought too dear to hold at all). Another fraction of investors may react to the news by concluding that the stock in question is not worth what they have previously believed, and will try to sell it. The balance between the two groups of investors would generate a net buying or net selling interest. The price would instantaneously adjust to it, perhaps on minute additional volume or no additional volume at all, since market-makers would just mark the price up or down to choke off the buying or selling interest. They should do this if they believed in the "efficient market" theory they have heard spoken of. After all, it was the price at which net buying or selling in the wake of the new information was choked off that fully "reflected" that information.
Retreat from the logic of instantaneous adjustment to the new "right" price would allow investors to react to the new information, not when it becomes "available", but next week, next month or next year. This would render the scenario more realistic, but render the hypothesis of the "right" price and the efficient allocation of capital useless because it would make it unfalsifiable. The actual price will reflect the news one day, but that day may always lie in the future—next week, next month, next year or whenever—especially as the impact of one piece of news will with the passage of time be overlaid by additional news. Under these circumstances, one could say both that the price is right because it (ultimately) reflects the news, and that it is "wrong" because full adjustment to the news is indefinitely delayed. It is "jam, jam tomorrow, but never have jam today".
For the idea to be worth discussing, we must revert to the version that supposes instantaneous adjustment. Under this version, the immense majority of investors—in the limiting case, all investors—would remain passive. They would buy only index funds for their retirement, never favouring one stock or one industry, for the relative prices of different assets would be just right and nothing could be gained by investing in one rather than another. Nor would they try to "time" the market, buying the index fund when they thought the market was low and postponing the purchase when they thought it was too high. They would sell their index fund to pay the tuition fees of their daughters at an expensive college, but never because they expected it to slump. In short, they would be the model investors populist politicians and leader-writers dream of when they call for "moral capitalism" and call the stock market a "casino" for greedy gamblers.
Where does all this leave the "efficient" asset market? On a rigorous look, it leaves it nowhere at all, because if investors always just blindly accepted any asset price as the right price, they would never influence the allocation of capital between companies, market sectors, asset types, regions and continents.
This is not to say that as zombies hypnotised by a false or at least misinterpreted theory, they would be acting irrationally when they just passively accepted a market price as the right price. It is perfectly rational for the average investor to stay passively invested in the index, since if he is average, he can by definition not do better than the average.
However it must be clear that if he acts in this manner, he is doing nothing whatever to allocate capital, let alone to allocate it "efficiently". The theory tells investors that if they are rational, they must understand that future prices are intrinsically unpredictable, hence it is a mistake to try and anticipate them. But it is only by anticipating them that investors are adjusting the allocation of capital to what they think is going to happen in the future and that will be an efficient allocation if they guessed the future right. If they have guessed wrong, they will sway the flow of funds in the wrong way, allowing ultimately less productive uses to absorb too much capital and more productive ones too little. But in doing so, they will also lose their ammunition that would permit them to make a similar mistake next time.
Those who get it right get rich and with their more ample ammunition can have a greater impact on the market next time, pushing it closer to the price pattern that would best adjust the flow of capital to what future events and developments demanded. This supposes that successful speculators remain mostly successful and unsuccessful ones wither away; a supposition that is more plausible than its opposite would be. Hoping for the best on this score, let us conclude that the efficient market theory promotes zombies, but speculators promote an efficient market.
Available information testifies about the past efficiency of capital allocation. Adjusting asset prices to available information may or may not do anything for efficiency now. Only the information that becomes available as the future unfolds will tell whether capital allocation was what it should have been. It was what it should have been if and only if present asset prices leave no room for successful speculation—if semiconductors are never overproduced, if power generating capacity is always just adequate, if shipping freight rates are broadly stable, if no industry is made suddenly obsolescent by the rise of another, if there is no recovery from a slump and no bubble is ever pricked. The inefficient asset market allows things to happen that should not, and also permits successful speculation to mitigate them by changing the allocation of capital in anticipation of them, and mitigating their ill effects. Perfect foresight would do an even better job, but let the best not be an enemy of the good.