It is communly understood that markets are subject to "bubbles". Bubbles are consequential. When markets are irrational, people who behave irrationally, squander resources, and do foolish things may be handsomely remunerated, while people who act sensibly are punished. During bubbles, mountains are moved to create what should not be created, people devote years of their lives to developing skills for which there will be no good use, opportunities are lost and scarce resources are squandered. Estimating the costs of any particular bubble is impossible, because we can never guess what an economy would have done instead of the unproductive things that it did do. But it's obvious that the scale of malinvestments have often been vast.

There is a fuzzy line between legitimate enthusiasm and risk-taking and a destructive speculative bubble. Reasonable people can and do disagree. Nevertheless, I think any fair reading of economic history shows that speculative bubbles are frequent phenomena, which are often identified contemporaneously by market participants who find they have no way to profit from their correct, but disputed, judgement. So, these insights fail to be reflected in market prices, permitting bubbles to grow without restraint. This is the essence of Keynes' quote, and it is a remediable institutional failure that decades of much-hyped financial innovation has simply failed to remedy.

The structure of financial markets is systematically biased towards long positions, which reflect an expectation of price appreciation, and are biased against "shorts" who try to profit from expected depreciation. Why? Because a "long", someone who expects a company or composite to do well, can pay cash for a position, and not worry about the path by which the securities have purchased arrive at a rational evaluation. A "long" can ignore transient irrationalities in the path by which a security arrives at an expected price, and merely bet (correctly or incorrectly) that the expected price will be attained within a reasonable time-frame. This is the strategy that has made Warren Buffet rich. He scours the market for "50-cent dollars", buys them, and waits. If a stock "worth" a dollar drops from $0.50 to $0.25, or even to $0.01, Buffet, if he is confident of the security's intrinsic value, can just hold the stock, until eventually what is worth a dollar sells for a dollar. He can ride out transient market irrationality, and if he is correct in his valuation, he profits.

But what about the "5-dollar dollars" in a market? It is just as useful, from a resource allocation perspective, for investors to be able to profit from pointing out misallocations of capital as it is for them to remedy underallocation. But an anti-Buffet, who is as confident and correct as Buffet in his valuations and tries to short expensive stocks, cannot be blithely indifferent to the path a security takes in finding its correct price. Under existing market institutions, ordinary short investors have to maintain a collateral for stock they have borrowed and sold that varies with the path of the security price. If a stock "worth" 1 dollar is absurdly valued at $5, there is no reason to think it might not transiently reach the absurd value of $10. As David Merkel notes, "twice absurd is still absurd." But while a buy-and-hold investor on the long side can just hold an absurdly undervalued stock, a buy and hold investor on the short side has to "cover" the increased debt implicit in her short position by putting up hard cash. Any short investor can be put out of business by a sufficiently large spike in the price, however transient, low-volume, or downright absurd.

There are lots of differences in costs and risks between "selling short" and "buying long". There are even some asymmetries — only applicable to very large, very creditworthy, investors — that work in favor of shorts. But for the vast majority of buy-and-hold value investors, the possibility of being 100% correct on the fundamentals but wiped out by a spike makes shorting simply not worth the risk. It's value investors who force stocks to reasonable prices. So called "technical" investors and traders don't care, they'll follow momentum from one greater fool to the next without batting an eyelash. The only value investors who do have the capacity to sell, investors with an previously-purchased inventory, are an unrepresentative sample of the value investment community (bullish enough earlier on to buy), sticky (buy and hold investors usually hold), and subject to behavioral-finance effects like finding it hard to sell a position that's earning money. Under these circumstances, there is a systematic bias for prices to generally rise, but to fall suddenly when some news or trend calls attention to the mispricing way so vividly that it juices the traders and gets the portfolio investors nervous about holding.

In other words, with the short-side so under-represented among value investors, it shouldn't be surprising that bubbles tend to gently grow, then suddenly pop. And so they do. And the costs are enormous.

This is not rocket science or genius. This is obvious stuff. Again, that it hasn't been fixed, that this institutional bias in favor of one kind of opinion has never been controlled, is a blot and a shame on the financial community.

p.s. For some kinds of investments, the bias against shorting is very strong. It is very hard, for example, to short housing. Maybe new derivative instruments will partially change that, but the risk characteristics of futures and options make them also unpalatable to the value investor, and the arbitrage strategies that usually link derivative securities to fundamental investments won't work well with housing. Bubbles have lots of apologists, and it is easy to confuse the present institutional forms of markets for "nature", so the idea that markets as presently consituted are badly broken and should be intentionally reformed to track value with price more precisely comes off to many as utopian science-fiction. My opinion is that when the US housing and liquidity bubbles burst, this kind of talk will seem very moderate.