A Defense of High-Frequency Trading
Wired magazine and the New York Times both recently published detailed stories on "flash trading" - the increasing use of high speed artificial intelligence algorithms in the financial markets. Both asked the same question: Will flash trading help the markets by improving efficiency - or will it destroy them?
But while both stories covered the same basic facts, they took strikingly different approaches. Wired discussed the technology in a generally balanced fashion, whereas the New York Times adopted a more alarmist attitude, including emphasizing the problems the technology would create for government regulators. However, the concerns raised by the Times against flash trading are variations of fallacies frequently raised against free markets. Hence, identifying and rebutting those fallacies will help one better appreciate and defend flash trading in particular, as well as market capitalism in general.
The Wired story discussed how traders are using increasingly sophisticated computer algorithms either to execute thousands of transactions per second (to exploit short-term fluctuations in stock prices) or to data-mine news and earnings statements to find attractive investment opportunities that others might have missed. Because these algorithms rely on fast computing power, they often adopt strategies human supervisors do not fully understand and could never devise on their own. Yet Wired notes that, "The result is a system that is more efficient, faster, and smarter than any human."
Of course, traders using such algorithms must necessarily cede a certain degree of decision-making to their software in exchange for higher profits. But this is no different in principle from a human motorist ceding some manual control of his automobile to his computer-controlled transmission and anti-lock brake systems in exchange for improved gas mileage and safety.
Naturally, the availability of such powerful technology elevates the financial stakes for traders. Traders who devise and utilize robust algorithms can quickly make a great deal of money, whereas those who use poor algorithms can quickly lose their shirts.
Furthermore, when multiple high-speed trading programs interact in the marketplace, they can sometimes create sudden sharp temporary swings in stock prices that human traders find difficult to understand or explain. Shrewd traders thus place a premium on high-quality algorithms able to correctly discern genuinely good buying (or selling) opportunities from spurious ones. Overall, the Wired story emphasized the opportunities - and challenges - flash trading creates for rational profit-seeking actors in the marketplace.
In contrast, the New York Times story covered many of the same basic facts, but with a noticeably different slant. In particular, it questioned whether it was "fairer" that trading firms investing in better technology should reap higher financial rewards, asked if "the technology [was] getting dangerously out of control," and fretted over the plight of government regulators "struggling to keep up with the pace of innovation." However, all three of these concerns are deeply misguided.
With respect to "fairness," any business willing to take risks and invest resources to become more productive rightly deserves any higher profits it earns as a result. Certainly, the New York Times executives have never complained about the fact that the Times is more profitable than, say, the Peoria Times-Observer because they've invested greater resources in their national and international news coverage, cultivated a larger readership, and are thus able to charge higher advertising rates.
With respect to "control," no trading firm is required to engage in flash trading. If a trading house chooses not to cede such control to software algorithms, it is free to rely on old-fashioned trading methods - and reap the financial consequences, for good or for ill. Given our still relatively-free markets, if a trading company believes that other companies relying on such algorithms are making bad trades (i.e., buying too high or selling too low), it can and should feel free to exploit those competitors' errors to its best advantage.
Finally, with respect to the difficulty of regulating flash trading, this is something desirable, not something to be feared. As Jonathan Hoenig pointed out in his superb critique of market regulations, "the purpose of a market is to bring together buyers and sellers to discover a true market price." In contrast, the regulators' basic premise is that they "know what's 'reasonable' more than those actual investors who have their own money on the line."
If the mutual interaction of trading programs creates inexplicable sharp swings in stock prices, the best way to "correct" them is not further regulation, but rather to let other profit-seeking traders (either computer or human) respond to those swings and drive prices back to their natural level. A free market is naturally self-regulating.
Government-imposed artificial limits such as trading halts and short-sale bans merely delay (or prevent) traders from finding mutually acceptable equilibrium prices. As Hoenig observes, government regulations only serve to increase the very market volatility, uncertainty, and fear they were supposedly intended to "correct."
Although I am not a trader, I benefit from efficient capital markets that allow investors to freely allocate their resources to maximize their profits. For me, efficient trading markets mean a better life in the form of cheaper food, exciting new consumer products (such as the iPad), and life-extending innovations created by medical device manufacturers (such as improved MRI scanners). If the widespread adoption of computer-assisted trading means that regulators are less able to hamper such markets, then I consider that a feature, not a bug.
Or to paraphrase Kent Brockman from The Simpsons, "I, for one, welcome our new robotic trading overlords!"