The Next Generation of Hedge Fund Stars: Data-Crunching Computers
The Next Generation of Hedge Fund Stars: Data-Crunching Computers
By ALEXANDRA STEVENSON NOV. 14, 2016
Every five minutes a satellite captures images of China’s biggest cities from space. Thousands of miles away in California, a computer looks at the shadows of the buildings in the images and draws a conclusion: China’s real estate boom is slowing.
Traders at BlackRock, the money management giant, then use the data to help choose whether to buy or sell the stocks of Chinese developers. “The machine is able to deal with some of the very complex decisions,” said Jeff Shen, co-chief investment officer at Scientific Active Equity, BlackRock’s quantitative trading, or quant, arm in San Francisco.
The future star of the hedge fund industry is not the next William A. Ackman, Carl C. Icahn or George Soros. Rather, it is a computer like the one at Scientific Active Equity, which sifts through data like satellite images from China every day.
Math whizzes have long dominated the hedge fund universe, but until recently, only a handful of well-known firms like Renaissance Technologies, the D. E. Shaw Group and AQR Capital Management used mathematical models and computers to plot out trading techniques. And other than the occasional blowup, as when Long-Term Capital Management went bust in spectacular fashion in 1998 after its models failed to factor in the possibility of a Russian government debt default, the world of quantitative trading has remained out of the limelight.
Now, as the financial world faces dismal returns and investor criticism over high fees, hedge fund managers are turning to computers to make decisions that used to be left to humans about which stocks to buy and sell, for example. Celebrity investors like Mr. Ackman are slowly being replaced by teams of Ph.D. holders who develop mathematical equations for trading and systems to scrape huge sets of data for patterns.
For instance, the billionaire investor Paul Tudor Jones, who runs the Tudor Investment Corporation, needed to make changes after investors pulled more than $2 billion from his firm, which now manages $10.6 billion. So he cut staff and brought in mathematicians and scientists to build up an analytical team. Other hedge funds have made similar moves.
“We’re seeing a kind of bifurcation among hedge funds, with some moving towards more quant-driven or automated style, while others are turning towards a more ‘long-only’ model, where they are judged on longer-term investment performance,” said Craig Coben, global head of equity capital markets at Bank of America Merrill Lynch.
Big institutional investors are also diverting more money to the hedge fund firms that use computer-driven hedge fund strategies.
While the hedge fund industry in recent months has suffered the biggest quarterly outflow since the financial crisis, investors continue to allocate money to hedge funds that use computer-driven strategies. Investors have put $7.9 billion into quantitative hedge funds this year, and the universe of hedge funds devoted to these strategies has more than doubled, to $900 billion from $408 billion seven years ago, according to Hedge Fund Research.
More broadly, money flowing out of the hedge fund industry as a whole comes at a time when performance has been disappointing. The Hedge Fund Research Composite Index, the broadest gauge of hedge fund performance, has lagged the Standard & Poor’s 500-stock index this year, gaining 3.56 percent through the end of October compared with the index’s 4 percent gain over the same period, accounting for reinvested dividends.
“Frankly, we expect to see assets move from human managers to machine managers,” Tony James, chief operating officer of Blackstone, told investors earlier this year. The Blackstone Alternative Asset Management arm, which manages $70 billion in hedge fund investments, is a big investor in quant-related hedge fund firms and has put billions of dollars toward these firms in recent years. The division now has $10 billion invested in quant-dedicated hedge fund firms, according to one person with direct knowledge of the firm; it has not publicly released the number.
Some industry observers warn that hedge funds building out new quant arms may simply be trying to capture investor money that is flowing into the strategy. But veterans in the quant world see the trend as an indication that the industry is finally catching up to other industries in which technology has disrupted businesses.
“The portfolio investment industry has been relatively late to adopting technology,” said Philippe Jordan, the president of Capital Fund Management, a 25-year-old quant hedge fund firm that manages $6.9 billion. “Finance is deeply conservative in nature,” he added.
Capital Fund Management has 160 employees, including 40 scientists, most of whom hold Ph.D.s in physics; 75 employees are focused on information technology, 20 of which are in data management. Like other types of hedge funds, the firm has a research department. The only difference is that at Capital Fund Management, the analysts who conduct research approach the work more like academics, and ideas are peer-reviewed.
With more investor money going toward firms that build models to trade on, there is some concern that these models will begin to look similar, potentially resulting in overcrowding. That could be a problem if there is a sudden event that drives everyone to start selling at the same time, something that happened during the “quant crunch” in the summer of 2007. Over one week in August, AQR Capital Management, D. E. Shaw and Renaissance Technologies were all hit with huge losses as the housing market began to show signs of collapse. With similar models and huge positions, the losses each firm suffered were amplified.
Mr. Shen at BlackRock thinks there are fewer risks this time around. “The diversity of data allows people to do a lot of different things,” he said.
Back in his San Francisco office, employees are using computers to create models for parsing the scripts from corporate quarterly financial earnings calls. At times, these computers are thinking faster than those who are using them.
“The machines are certainly doing more and more, so humans should worry there is a human replacement factor,” Mr. Shen said.
“But,” he added, “ultimately I do think it is the human who creates the machine and these techniques.”
A version of this article appears in print on November 15, 2016, on page F6 of the New York edition with the headline: Rise of Computers.