The following content is sourced from external partners. We cannot guarantee that it is suitable for the visually or hearing impaired.
Buildings stand on Wall Street near the New York Stock Exchange (NYSE) in New York, U.S.(bloomberg)
(Bloomberg) -- Millennium Management has shut down a quant hedge fund founded by two acclaimed physicists that suffered only one losing year since it began trading in the early 1990s.
The closing of Prediction Company, which Millennium bought in 2013, came as a surprise to employees because the firm was profitable, according to a person familiar with the matter. The hedge fund was started by Doyne Farmer and Norm Packard, who are known for their seminal work in developing chaos theory, and managed about $4 billion at its peak.
Izzy Englander’s Millennium shuttered Prediction, a statistical arbitrage fund, at a time when that strategy within the industry is struggling to make money. Returns for the approach are down less than 1 percent so far this year, according to a Scientific Investments index of 160 funds, marking its first year-to-date loss since 2010. While Prediction outperformed the average stat-arb fund, that wasn’t enough to keep it running, the person said.
“Prediction was certainly far ahead of its time in realizing that many aspects of trading could be made not just quantitative but algorithmic,” said Michael Kearns, a professor of computer science at the University of Pennsylvania and a senior adviser on machine learning and AI at Morgan Stanley. “What they did is far more difficult today because of increasing automation and competition.”
Millennium, which manages $35 billion, operates a multi-strategy fund that deploys capital to numerous teams. Prediction labored under the shadow of Millennium’s bigger systematic hedge fund WorldQuant, which also runs a stat-arb strategy. WorldQuant invests about $5 billion for Millennium and raised $2.3 billion in its first fund for outside clients earlier this year.
A spokesman for Millennium declined to comment.
Prediction traces its origins to a small adobe building in the artsy Southwestern town of Santa Fe, New Mexico. The location, far from the influence of Wall Street, offered the prospect that the founders would be less likely to follow the investing crowd.
Backed by a Chicago trading firm, Prediction jumped into systematic trading when few firms were doing it. The hedge fund started trading futures and eked out a small profit. By 1996 it began wagering with equities as well. That was immediately successful, allowing the firm to deploy significant amounts of capital, Farmer said in an interview. The fund maintained a Sharpe ratio of 3, a measure of risk-adjusted returns.
“The stock models were completely automated and we only overrode them occasionally, during times like 9/11,” said Farmer, who’s now a professor at Oxford University and no longer an owner of Prediction.
The hedge fund was acquired by UBS AG in stages beginning in 1999 and suffered its one down year in 2007 as the “quant quake” rocked systematic firms just before the financial crisis. The subsequent bailout of banks worldwide and restrictions on proprietary trading in the U.S. led UBS to sell the hedge fund to Millennium.
Prediction began life as a physics experiment of sorts. In the 1980s, Farmer, then at Los Alamos National Laboratory, and Packard, at the University of Illinois, were conducting research in chaos theory.
They developed algorithms that analyzed data and helped verify a hypothesis that turbulence in fluids is caused by chaos. The physicists also found that they could take advantage of chaos to make better short-term predictions about fluid flows despite limits to long-term forecasting.
From Flows to Markets
“That was the original motivation for starting Prediction,” said Farmer. “Since we could predict fluid flows, we thought maybe we could predict markets too.”
They did, posting robust returns in the late 1990s. But the novelty of their stat-arb strategy began to fade in the years leading up to the financial crisis and today there are hundreds of rivals vying for the same profits from pricing disparities in correlated securities, hurting returns.
When it closed, Prediction managed less than $350 million. The small number of employees who left are looking for a new partner.
“They are hoping to stay together as a group,” Farmer said.
(Updates with Prediction’s trading history starting in the 7th paragraph.)
To contact the reporters on this story: Vincent Bielski in New York at email@example.com;Saijel Kishan in New York at firstname.lastname@example.org
To contact the editors responsible for this story: Margaret Collins at email@example.com, Alan Mirabella, Josh Friedman
©2018 Bloomberg L.P.