Search This Blog

Sunday, December 04, 2011

A nagging question about the rationality of stock markets

For some time, I have been puzzling over media reports and economic analyses of the movement of stock markets.  Almost invariably, the goal is to uncover/understand the thinking of investors, on the assumption that this is what is moving markets.  So it makes a lot of sense that the Nobel Prize for Economics was awarded to a pair of economists for their work on "rational expectations" in economic theory.

But the question that has been nagging at me is this:  does it really make sense anymore to assume that the stock markets' behavior reflects actual (human) thinking, rational or otherwise?  Depending on the market and the source of the estimate, forty to sixty percent of stock-market trading is driven by the algorithms of computer programs.  It goes by a variety of names - high-frequency trading, flash trading, automated trading, program trading, robotic or robo-trading - but its essence is that trading is based on pre-defined parameters written into a program.  "Thought" obviously goes into the construction of algorithms (though they are proprietary and not open to public inspection), but once a program has been written and put into effect, there's no thinking behind the trade that it executes, no individual's assessment of market conditions as the basis for specific trades.

So the ink, digital or otherwise, that is spilled daily in an effort to discern the thinking of market participants as the key to the markets' behavior seems increasingly futile.  Program trading needs to be pulled from the margins of our attention to the center, which, according to a recent report in the New York Times, may actually be happening.  As the Financial Times' Jeremy Grant wrote recently of the UK, "there is growing concern that exchanges have moved too far from their traditional role in facilitating capital-raising as they chose other business streams to satisfy shareholder returns."  Understanding stock markets - their behavior and their practical function - now requires as much, if not more, attention to the thinking of algorithm writers than to the thoughts of sentient traders.