Three and a half years have passed since the afternoon when the stock markets went into a trillion-dollar free fall and just as suddenly reversed course, recovering 80 percent of that loss. It all happened in less than 45 minutes.
The "Flash Crash" of May 6, 2010, was the unintended result of high-frequency trading (or HFT), in which heavy-duty computers execute sophisticated trading strategies in millionths-of-a-second time frames. HFT has been the source of many smaller crashes and other mishaps, occasionally shutting down trading venues and even putting firms out of business. Yet the practice continues to grow, while the transaction times get shorter and shorter.
A few weeks ago, the Commodity Futures Trading Commission, which has jurisdiction over almost all derivatives markets under the Dodd-Frank financial reform law, published a paper evaluating various "kill switches" that might contain the damage when the machines go haywire again.
When, not if, because, as the CFTC's Concept Release points out, HFT trader bots now control more than 90 percent of all exchange-traded derivatives – the "financial instruments of mass destruction" that Warren Buffett warned about, five years before the 2008 crash proved him right. In other words, we have reached the point in the sci-fi movies when the machines truly take command of the nuclear arsenal.
So it is time to ask a few basic questions about HFT. Starting with: What good is it? How does society benefit when we reduce the average trade-completion time from, say, 125 milliseconds to 5 milliseconds?
By society, I do not mean a single financial institution or even the entire financial sector. The increasing speed of transactions is driven by competition among trading firms. Speed is crucial to any trader's ability to act before other traders do. But no one has identified an advantage to the overall functioning of the markets or the wider economy when traders get sucked into an "arms race" of escalating speed and automation. In fact, even the most passionate HFT advocates would largely agree that the public ceased to benefit several developmental stages ago.
The case for this type of trading rests on the premise that it lubricates the markets by adding more "liquidity." A liquid market, however, is characterized by plentiful and reliable offers to both buy and sell; it must be balanced, since the market crashes if everyone wants to sell at once.
No doubt, HFT increases volume by allowing transactions to be made and reversed with blinding speed. But volume and liquidity are very different. In markets dominated by HFT, new information stimulates the trading algorithms of multiple firms, which behave like herds of impala stampeding across the Serengeti, this way and that. Prices jump up and down until the pointless activity peters out. The Flash Crash was no more than an extreme example of what goes on day in and day out in today's markets.
HFT traders often do supply executable price quotes, which superficially increase liquidity. True liquidity, however, comes when offers can be relied upon, allowing investors to predict whether the transactions they seek can be completed within their preferred price range. Because HFT traders can morph from providers to consumers of liquidity whenever the herd abruptly shifts from buy to sell, they create uncertainty rather than predictability.
Ultimately, the diminished reliability of the financial markets is a burden on businesses and governments that seek to raise capital. By inducing short-swing market volatility, HFT traders make money from each other, but also from investors. Investors pass the cost of uncertainty on to businesses and governments, burdening their productive activities. Thus, the public pays for the non-productive churning of securities and derivatives markets by HFT traders.
The goal of each trading firm is to profit from the ability to perceive and act on information more quickly and accurately than others do. But there is no rule of nature or humanity which dictates that a spiral of such advantages must translate into improved intermediation of capital between investors and those seeking investment. Indeed, HFT has made the markets less efficient, by making them more volatile. Since investors must price this added unreliability into their decisions, raising capital for productive purposes becomes costlier, and the extra cost is passed on to the rest of us as consumers and taxpayers.
In fairness to the CFTC, it was not asked to decide whether HFT actually serves any valid purpose. Yet the whole premise of its Concept Release is that such trading creates new and serious risks. Surely there should be some compensating benefit to justify those risks. The Concept Release does not attempt to identify any such benefit.
Regulators have been hesitant to slow traders down, in part because doing so might be seen as stifling innovation. But HFT is not innovation in any real sense; it is just the accumulation of computing capacity and efficient wiring into the exchanges, with software that simply encodes trading strategies that have been around for years.
Unfortunately, businesses that can afford to compete by trading at nanosecond speed also have the wherewithal to exert tremendous influence in Washington. The big financial firms will use this influence to protect HFT because it gives them advantages in the markets and the ability to extract ever more value from the real economy. There has been plenty of conversation about HFT by lawmakers and regulators. Identifying ways to manage the risks is laudable, though it is far from clear how effective such techniques will really be. In any case, Washington needs to go further, and muster the independence to put some meaningful speed bumps in place to slow the bots down to a really safe speed.
Wallace Turbeville is a former vice president of Goldman Sachs, a fellow at Demos and a member of the derivatives task force of Americans for Financial Reform.