High Frequency, Fat Target

Michael Lewis misses the competitive benefits of computerized Wall Street trading.


Flash Boys: A Wall Street Revolt, by Michael Lewis, W.W. Norton and Company, $27.95

For the last five years, the press has been sounding alarms about high-frequency trading (HFT), a practice in which investors use fast computers driven by secret algorithms to rapidly trade securities. Time wondered in a 2012 headline whether the practice is "Wall Street's Doomsday Machine." Mother Jones in 2013 worried it could "set off a financial meltdown." In March of this year, 60 Minutes aired an infomercial-toned segment promoting the new Investor's Exchange (IEX) trading venue, which, according to IEX's website, is "dedicated to institutionalizing fairness in the markets" by slowing down trades.

Now we have Flash Boys, Michael Lewis' highly lauded attempt to explain the dark ways of Wall Street to the masses. The book seems aimed at convincing the public that high-frequency trading is bad and that firms should be trading on IEX, because the other "markets are rigged." This emotion-provoking tagline has prompted cries for regulation, launched legislative committee hearings, and forced trading firms to defend the integrity of the financial markets.

Lewis' book is primarily about the individuals who developed the IEX. These characters are compelling, and the book provides interesting insights into the types of people lurking around Wall Street's back rooms. The hero is Brad Katsuyama, a former Royal Bank of Canada trader who became frustrated by his inability to beat high-frequency traders to a particular stock price.

Katsuyama found that when he wanted to purchase 100 shares of a stock at $50 each, he might only be able to purchase 10 shares at that price even when it appeared that 100 were available when he executed his order. The remaining 90 shares might have to be purchased for a penny or two more. Through trial and error, Katsuyama discovered that HFTs were beating him to the stocks, because getting 100 shares at $50 required him to trade across multiple trading venues. The computers used by HFTs were so fast at processing data, they were able to detect the demand for the stock at $50 and rush to the other venues to buy it before he could. The HFTs then sold it back to Katsuyama and others like him whose orders were executed more slowly.

Trading venues exist at specific physical places. When Katsuyama, or any trader, placed an order on his computer, it would arrive at the physically nearest venue first. At that point HFTs picked up the signal and raced to the more distant trading venues to beat traders to the $50 price. To resolve this problem, Katsuyama developed an algorithm that slowed down the speed of his trades but changed the order in which they arrived at different exchanges. By sending his orders to the most distant exchanges first, he was able to beat HFT algorithms to the stock.

It's a story well told, but it doesn't support Lewis' thesis about markets being dangerously rigged. The fact that this new strategy was indeed able to neutralize the HFT strategy suggests that markets are excellent at producing competitive corrections to any distortion. If the financial markets were truly rigged, Lewis' band of merry men and women wouldn't be able to take from high-frequency traders and give to IEX.

Logical gaps like this appear throughout the book. Lewis raises several important issues, but time and again, he either misses the solution or fails to see the problems in a broader context, preferring a simplistic if entertaining heroes-and-villains narrative.

The book does provide some value to the lay reader. Lewis is particularly on point in describing the unintended consequences of the Securities and Exchange Commission's Regulation National Market System (Reg NMS), implemented in 2007. Reg NMS was designed to encourage competition in the marketplace and to ensure that investors receive the best price execution for their orders. To this end, traders are required to buy or sell equities for their clients on whichever exchange will give them the best price at any given moment.

This has created at least two important unintended consequences. The first is what Lewis describes as market "fragmentation." In the past, most trades were handled by either the New York Stock Exchange (NYSE) or the NASDAQ. Now, trading is divided between 13 public exchanges and about 50 alternative trading systems within the United States alone. Reg NMS was successful in making markets more price-competitive by breaking the domination of the NYSE and NASDAQ. This very fragmented environment, Lewis claims, laid the groundwork for high-frequency trading.

Fragmentation did indeed create the need for computerized trading systems to quickly process price and market information across multiple venues for the best price. But this alone needn't have been a problem. Faster and more efficient markets with increased price competition have driven trading costs down for everyone, and that's a good thing.

The problem arises with the Securities Information Processor (SIP), which consolidates and disseminates prices across trading venues. The SIP is managed by a 15-member committee of exchanges, market participants, and regulators, and it is operated, via contract, by the company that owns the NASDAQ stock market. To facilitate discovery of the best available price across so many different venues, Reg NMS required that the nationwide best bid and offer price be compiled in one place. The venue that this information was presumed, but not required, to be compiled, was in the SIP.

The problem is that in order to provide the national best bid and offer, the information has to be collected, calculated, and disseminated by the SIP. The technology used to do so is old and much slower than that used by high-speed traders.

To overcome this gap, HFTs began creating their own internal versions of the SIP, but ones that process the information much faster. Because these homegrown clearinghouses process information so much faster than firms still relying on SIP-distributed data, they can detect demand for equities well before the competition. Thus HFTs are able to buy at better prices and sell for a penny or so more to those using price data direct from the SIP.

Lewis suggests that the SIP's sluggishness means we should eliminate high-frequency trading technology. But high-speed traders jumping ahead of slower competitors is just a symptom. Since the slowness of the SIP caused the perceived problem, why not fix the SIP, or require centralization of national best-bid and-offer information in a more modern system? Or maybe we should repeal Reg NMS instead of creating a Rube Goldbergesque tangle of regulations designed to regulate the regulations. Blaming high-speed trading technology for informational inequities is like blaming the train for a derailment caused by a stalled vehicle on the rail.

Lewis also fails to mention just who is most interested in banning HFT: the exchanges and legacy trading firms that have not competed as well in this new environment. The first firm to trade on the new IEX venue was Goldman Sachs-and after Flash Boys was published, Goldman's chief operating officer declared HFT bad for U.S. financial markets. Now the New York Stock Exchange is lobbying against HFT.

Why? Fragmentation creates competition, and both Goldman Sachs and the NYSE are losing market share in this new environment. They want their customers back.

Lewis' markets-are-rigged mantra is gaining purchase. After the book came out, the FBI launched a probe of HFT and New York Attorney General Eric Schneiderman made his year-old investigation of the practice public. Some brokerages have received requests from their clients to have their securities traded on the new IEX venue, and IEX's trading volume has risen dramatically.

Lewis and his compatriots have rallied the villagers to take up torches and pitchforks to fight market monsters that they know little about. They are arriving at their brokers' doors demanding they trade with IEX, and they are calling their legislators to insist on new investigations and regulations.

Yet the jury is still out on whether the high-frequency traders' overall market impact has been negative or positive. HFT is not taking over markets; both its share of trading volume and its profits have been declining steadily since 2009, according to The New York Times. There is also evidence that HFT has dramatically decreased the costs of trading for all market participants, increasing liquidity and decreasing spreads.

It is attractive to believe that markets are rigged, since someone always seems to be better at investing than we are. If Lewis' book and the emotions it provokes manage to have more influence on public policy than reasoned analysis does, then the crusader against rigged markets may have pulled off one of the greatest market manipulations of all time.