Despite record-breaking bank earnings reports, it’s a bad time to be a stock trader on Wall Street. It may be hard to believe with the Dow Jones Industrial Average hovering around all-time highs, but declining profits and falling trading volume have dogged stock trading desks for the last five years. As profits and trading volumes continue to disappoint, critics have pointed the finger at computerized trading programs that are rapidly transforming the business into something completely different.
“The art of trading has died,” says Joseph Saluzzi, founding partner of Themis Trading, an independent brokerage firm based in Chatham, New Jersey. “I know tons of traders that have gone out of business.” Saluzzi and other critics argue that computerized trading has created a zero-sum game, in which ever-more complex high speed trading programs vie to buy and sell stocks milliseconds ahead of ordinary investors – and each other – for tiny profits that accumulate over millions of trades.
“It’s the hollowing out of the equity market – it’s been mechanized to an extent,” says Saluzzi. Mechanization has long been blamed for job losses on factory floors and behind checkout counters. Trading desks can now be added to that list. It seems like the once-mighty trader, upon whom so much of Wall Street’s mythology is based, is starting to look like an anachronism.
This may be surprising for anyone following some of the more exuberant headlines of the past few months, but it’s true. While overall business earnings at the major banks have reached the second highest levels ever recorded, profits from stock trading were at their lowest levels since 2008, having fallen 40 percent in the last five years. The daily tally of shares bought and sold has fallen as well, dropping 5.3 percent from this time last year. Both of these measures are registering their fifth straight year of declines. In response, the biggest banks fired 14 percent of their stock trading staff last year, citing falling revenues and excessive costs.
The proliferation of high speed trading even seems to be backfiring on the high-speed traders themselves. Just last month, Getco, the largest high frequency trading firm revealed that its profits in the first nine months of last year had fallen 82 percent from the previous year. The revelation highlighted an ongoing trend: As computerized algorithms continue to assert their dominant position in the market, the volume of trades paradoxically drops along with trading profits. As Reuters’s Felix Salmon explains, high frequency trading programs have become so evenly matched that they don’t even execute trades a lot of the time — they simply send out thousands of buy or sell orders, then retract them instantly when they see that no profit can be made from the potential transactions. Though the proportion of completed trades has diminished, the proportion of uncompleted orders has continued to rise.
High speed trading has also been blamed for some famous disasters. For market participants, the most harrowing was the “flash crash” of May 6, 2010, when the Dow Jones Industrial Average plunged over 600 points in five minutes, causing $1 trillion of value to temporarily vanish before the sell-off was recognized as a malfunction and reversed. In its report on the event, the Commodity Futures Trading Commission wrote, “High frequency traders aggressively trade in the direction of price changes . . . and may amplify price volatility.” In response to this, and other similar debacles, many large institutional investors pulled their money out of the stock market, leaving mutual funds, pension funds, banks, hedge funds and high frequency trading firms to trade with one another using this technology.
Some see this transformation as an obvious and inevitable progression driven by technology and efficiency. “In the past, if you wanted to make a trade you would call your sales trader and, based on your directions, they would place an order,” says Jeffrey Wallis, managing partner of Sungard Consulting Services, a financial services software company that advises large institutions on how to adapt to the ongoing transformation of the industry. “Now, you have a desktop tool into which you enter your order. All this is redefining what it means to be an investment bank. They are asking themselves whether they are an investment bank as defined in the early 2000s, or a technology company like Google that is driven by data and efficient transactional processing.”
Wallis believes the financial sector is in a state of flux, and is transitioning toward a future where “there will be less and less dependence upon trading skills and more focus on quantitative analysis capabilities that direct the algorithms. This will make [financial institutions] look more like technology companies than banks as we have traditionally understood them.”
Now it appears that this technology is moving beyond the stock market and into the highly profitable, less transparent markets for bonds and derivatives. The New York Times reported over the summer that United States Treasury bonds are starting to be traded electronically at some banks, while new financial regulations like Basel III and the Dodd-Frank financial reform are helping to move bonds and derivatives onto open exchanges, where high speed trading flourishes. “I’ve compared the high frequency trading guys to locusts,” says Joseph Saluzzi. “They’ll come through a field and they’ll strip it bare. That’s what they’ve done to the equity market. So what do the locusts do? They move on to the next field, whether it’s commodities or fixed income. Now the newest is going to be derivatives. And they’ll make a fortune, they’ll strip that market, and then they’ll move on to the next one.”
If investment banks are becoming technology companies, their indirect function appears to be the automation and streamlining of the financial markets that allocate capital in our economic system. Oddly enough, the biggest problem facing their operators seems to be that they are too good at it. In fact, computerized trading programs seem to be so good at seeking the best prices that when they trade among themselves, their algorithms command them to seek a profit that is sometimes impossible to find. In the decade or so since modern algorithm trading took hold, something approaching perfect competition seems to be have been created in the stock market.
This is what gives critics like Joseph Saluzzi nightmares. “The old specialists and market makers are out of a job, but if the high speed trading guys are saying ‘This is no longer profitable for me, I’m not making any money here’ who is going to fill the void? They could simply walk away. That’s what the flash crash was about.” Fears of this kind have led critics of high speed trading to petition Washington to crack down on the practice, and the Securities and Exchange Commission has recently announced its intention to re-regulate it.
If Jeffrey Wallis is right, however, and the automation of financial markets is inevitable, it seems worthwhile to contemplate what a more public-oriented application of these technologies could look like. Could high-speed trading algorithms be modified to seek different outcomes, based on a broader set of quantifiable variables? It’s not implausible to imagine public and non-profit institutions employing automated trading programs for capital allocation between themselves. In the era of the social impact bond and other experiments, the risk management tools of the financial sector may have applications that have only begun to be explored.
If financial markets are going to be mechanized, and drained of trading profits, shouldn’t they be put to some useful purpose beyond just private investment? As a thinking exercise, it’s worth imagining whether existing technology could be used as a jumping off point for an electronic network for public capital allocation. It’s too early to tell, but the tools that make our increasingly digital markets so efficient might constitute a vital part of a more modern, complex and egalitarian financial system to come.