Is the Singularity already happening at Goldman Sachs?
"Nearly everyone on Wall Street is wondering how hedge funds and large banks like Goldman Sachs are making so much money so soon after the financial system nearly collapsed," writes the New York Times’ Charles Duhigg in a front page piece that’s been the talk of the town in New York and Washington. "High-frequency trading is one answer." Duhigg writes, “High-frequency trading systems are so fast they can outsmart or outrun other investors, humans and computers alike."
The term “high-frequency” refers to fast entry and exit of trading positions, the process best executed by algorithms and dedicated computer programs employing artificial intelligence. However, this can turn into the intentional probing of the market with tiny orders that are immediately canceled at speeds that cannot be matched by individual human investors.
While this isn’t quite the "intelligence explosion" of machines foreseen by I.J. Good in 1965 and dubbed “the Singularity” by Vernor Vinge in 1993, the speed and sophistication of Goldman Sachs computer algorithms are indeed leaving humans –- at least, individual investors –- in the dust.
Investment Manager Irene Aldridge sees high frequency trading as having a bright future that is “bound to bring additional skill and capital” to markets. She points to Tabb Group estimates that high frequency trading (HFT) accounts for 73% of equities trading volume on U.S. exchanges. Aldridge: “HFT is characterized by fast turnover of capital. Instead of capturing large price changes over extended periods of time, HFT aims to book multiple small gains over short periods of time.”
Joe Saluzzi, Partner at Themis Trading, sees it differently. He bemoans the institutional use of co-located high-bandwidth supercomputers near the trading floor, “We are just mice dancing between the elephants of capital and their supercomputers. Just this past week, we found out that hedge funds have passed mutual funds in terms of volume of equity trading, despite controlling far less money.”
This video shows the intensity of the debate over HFT, with Joe Saluzzi expressing mock surprise that he isn’t debating a supercomputer (shades of Kubrick’s HAL 9000):
There are several possibilities about what’s going on here. One is that Goldman and others are literally using privileged information to make trades ahead of markets, in which case they are committing a felony. Specifically, this is known as "front-running," or trading ahead of customers, and it is an explicitly illegal form of market manipulation.
Another possibility –- mentioned by Robert Kuttner in a recent Huffington Post article –- is that “the Goldmans of the world” have found themselves a nice loophole. Tapping into the Stock Exchange’s own computers and other sources of trading activity is something that anyone in theory could do, but only a few privileged insiders have the sophistication to exploit what they find.
Duhigg’s New York Times article illustrates a case in point. Intel, the computer chip giant –- reporting robust earnings –-triggered some individual human investors to buy shares in the semiconductor company Broadcom, “The slower traders faced a quandary: If they sought to buy a large number of shares at once, they would tip their hand and risk driving up Broadcom’s price. So, as is often the case on Wall Street, they divided their orders into dozens of small batches, hoping to cover their tracks. One second after the market opened, shares of Broadcom started changing hands at $26.20.”
The slower traders began issuing buy orders. But rather than being shown to all potential sellers at the same time, some of those orders were most likely routed to “a collection of high-frequency traders for just 30 milliseconds — 0.03 seconds — in what are known as flash orders.” In just this fraction of a second, high-frequency traders determined that the market for Broadcom was growing. Their supercomputers began buying up Broadcom shares and then reselling them to the slower investors at higher prices. The overall price of Broadcom began to rise, “The slower-moving investors paid $1.4 million for about 56,000 shares, or $7,800 more than if they had been able to move as quickly as the high-frequency traders.”
Supercomputers began buying up Broadcom shares and then reselling them to the slower investors at higher prices. The overall price of Broadcom began to rise
Was the actual advantage the speed of the supercomputers? Nasdaq OMX Group Inc., Bats Exchange Inc. and Direct Edge Holdings Inc. hold these orders for milliseconds, giving their paying customers the opportunity to gauge demand before traders on other exchanges get the chance to bid. While the ethics and legality of this are debatable, Senator Charles Schumer has asked the U.S. Securities and Exchange Commission to ban such “flash orders,” saying the transactions give high-speed traders an unfair advantage over other investors.
“Flash orders allow certain members of these exchanges to obtain access to order flow information before that information is made available to the public,” writes Schumer. This allows “those members to use rapid trading programs to trade ahead of those orders and profit from advanced knowledge of buying and selling activity.”
Regardless of the debate over how HFT is used in market trading, it’s clear that supercomputers can already “outrun and outsmart” individual investors. This isn’t quite the Singularity envisioned by Good and Vinge, but it raises some perplexing questions about the use of artificial intelligence to gain market advantage when individual humans “with slide rules” cannot compete.
QUESTION FOR READERS: Are "flash orders" unethical and should they be banned?
(POST YOUR COMMENT BELOW!)