Politics & Policy

High-Frequency Trading: We Need More Data

The research we have is conflicting and often industry-funded.

Government shouldn’t rush to judgment on Wall Street’s high-speed traders. Granted, this whole business probably looks dodgy to Main Street’s 401(k) owners and stock pickers. Financial brainiacs are employing sophisticated computer programs called algorithmic robots to quickly scan financial data streams for subtle stock-market patterns. Trading, and lots of it, happens in a flash. The algobots see what big institutional investors are doing while they’re doing it. Every millisecond counts, which is why traders pay hundreds of millions of dollars a year to rent space for their computers in stock-exchange data centers. Half of daily U.S. trading volume is now of the superspeed variety, much of it passing through dedicated, specially built fiber-optic lines and microwave transmitters.

What in the name of Warren Buffett does any of this have to do with stock markets as places that enable public companies to raise money and let the little guy put his savings to work? Very little, according to critics who argue that high-frequency trading hurts small investors and poses systemic risk to the U.S. financial system as seen in the 2010 “Flash Crash.” The Justice Department, FBI, and New York attorney general are investigating whether markets are being manipulated and unfair advantage given to high-frequency traders. Author Michael Lewis is out with a much-hyped and highly critical book on the practice, Flash Boys, in which he claims the stock market is a rigged game. Congressional Democrats can almost certainly be expected to renew efforts to tax high-frequency trading out of existence.

Here’s one way to think about the problem: Are high-frequency algobots an innovation more like automatic-teller machines or collateralized debt obligations? In 2009, former Federal Reserve chairman Paul Volcker called ATMs the “most important financial innovation that I have seen the past 20 years” due to their convenience. Volcker was less enthusiastic about more recent financial advances such as CDOs, an innovation at the heart of the financial crisis. So which kind of innovation is high-frequency trading? Is it something that makes the system work better for everyone? Or is it something that profits a few while creating huge risks for the public?

Certainly, combining clever software with superfast computers and telecommunications to make massive trades at light speed is, borrowing a phrase from Robert Oppenheimer, a “technically sweet” innovation. Also, the billions in profits made by high-frequency trading firms would suggest, at least to some fervent free marketeers, that this innovation has passed the “market test.” And there is no doubt that high-frequency trading has dramatically lowered trading costs and enabled faster order completion, which is especially helpful for small investors.

But the final cost-benefit analysis of superfast, high-volume trading can’t be determined by competing claims of participants, nor in the compelling, character-centered narratives of bestselling authors. As much as anyone, Moneyball author Lewis should understand that data — and lots of it — is a surer guide for action than flashy anecdotes. Regulators, lawmakers, and other officials must carefully examine the research on high-frequency trading. Unfortunately, too much is either industry-funded or conflicting. For instance: A recently released European study found that high-frequency trading increases volatility and makes flash crashes more likely. Those findings directly contradict a 2013 U.S. academic paper that found “scant evidence” of high-speed trading increasing volatility. For the moment at least, there seems little reason for hastily assembled legislation or quick-trigger legal action. We don’t want a high-frequency-trading version of Dodd-Frank, a rushed regulatory fix that few observers think has ended the “too big to fail” government backstop.

So gobs more data and analysis are needed. High-frequency trading would be a perfect subject for the Office of Financial Research, located in the Treasury Department. Without a clear consensus, however, government’s default position should be inaction. Regulators must not be guided by the “precautionary principle,” where government permits innovation only if it’s proven harmless in advance; the result is too often crony-capitalist favoritism of whatever incumbent firms have the best lobbyists. If nothing else, high-frequency trading is a disruptive innovation that certainly annoys many institutional investors who can longer covertly trade big orders without the algobots taking notice — an intramural Wall Street battle that is so far benefiting Main Street.

Government should also keep an eye on market action. Whatever impact, good or ill, high-frequency trading is having, it is likely on the decline. According to a Rosenblatt Securities estimate, high-frequency trading used to account for some two-thirds of U.S. equity action. And industry profits from stock trading are down by 80 percent over the same span. Still, if the consensus research does eventually make a compelling case against high-frequency trading, even the most ardent libertarians or tea partiers should demand government action.

Thanks to automation and globalization, it is more important that workers today and tomorrow also be capitalists and earn capital income as well as labor income. Both the 0.1 percent and 99.9 percent need to participate in a stock market they can trust.

— James Pethokoukis, a columnist, blogs for the American Enterprise Institute.


The Latest