It turns out that the robots are human after all.
There has been an explosion of interest in computer-powered investing of all kinds in recent years. Barclays estimates that the assets managed by so-called quantitative hedge funds have doubled over the past decade and hit a record $500bn last year.
Unfortunately, the performance has atrophied in tandem with their trendiness. The average equity hedge fund has gained 7.7 per cent this year, according to HFR, while quant equity funds have gained only 4.9 per cent. Quant “macro” funds, which invest across markets, have lost 1.4 per cent.
“There have been some unison screams this year,” says Wesley Chan, a quantitative fund manager at Acadian Asset Management.
There have been no major blow-ups, but for many funds until recently basking in the ravenous appetite for nearly all algorithmic strategies, it makes for an uncomfortable period and raises questions over whether quants are merely suffering a sour spell, or if something more fundamental is occurring.
Neal Berger, chief investment officer of Eagle’s View Capital Management, a fund-of-funds, thinks it may be the latter. In a letter to investors he said that the “fantastic returns” of many quants has attracted too much money. This is in turn eroding the opportunities for everyone, and turning once-profitable strategies into duds, a phenomenon known as “crowding”.
“With all the geniuses in quant, high-powered computers and enormous data, where are the ‘suckers’ who are providing the juice for all of these absolute return quantitative strategies?” Mr Berger asked. “We have a condition among the traditional quantitative strategies whereby we have robots trading against robots.”
Quants scoff at this. Philippe Jordan, the president of Capital Fund Management, points out that eight months of performance is far too short a timeframe to make sweeping judgments, and highlights how many funds and strategies are still doing well. “These numbers are completely within the normal distribution,” he says.
Indeed, quantitative investing is a broad church that can include everything from the relatively simple to achingly complex algorithmic strategies that mine vast seas of digital data for faint but profitable signals across financial markets. That makes it difficult to draw general conclusions.
“The quant term is kind of useless,” says Anthony Morris, head of quantitative strategies at Nomura. “It’s a sloppy description, which like ‘hedge fund’ can mean almost anything.”
Indeed, HFR’s average performance numbers obscure a wide divergence in performances among different funds and strategies, which offer clues on what has worked and what has fizzled this year — and why.
Some traditional powerhouses of the quant investing world continue to do well. Renaissance Technologies’ main equity fund is up 10 per cent in the year to August 4, and two of its other funds have returned 7.6 per cent and 11.3 per cent, according to a person familiar with the matter. But these are primarily equity funds with a bias towards betting on stocks gaining — which they have for most of the year.
Mr Jordan says other quant equity strategies, such as “market neutral” or “statistical arbitrage”, have had a harder time. But many of the poorer-performing funds are so-called “trend followers”. Again, the details vary greatly, but they primarily surf market momentum, going short when an asset class is falling, or piling in when the trend is positive.
Several big names have struggled. AHL’s Alpha and Dimension funds have gained only 1-2 per cent this year; BlueTrend, a fund run by Leda Braga’s Systematica, has lost 6.4 per cent, while David Harding’s $9.9bn Winton Futures Fund has trod water.
But some have done well. For example, AHL’s Evolution fund is up 9.6 per cent in the year to the end of July, and Systematica’s Alternative Markets Fund has returned 11.4 per cent. Crucially, however, these two funds take advantage of momentum in less liquid, less efficient markets, not the major ones that tend to be popular with trend-followers.
Mr Morris points out that the divergence is probably caused by their asset mix. Some trend-followers lean towards equities — which have enjoyed strong, positive momentum this year — while others tilt towards fixed income or commodities, which have been jittery. “Asset allocation matters in quantitative investing, as it does in everything else,” Mr Morris says.
That’s not to say that crowding isn’t part of the answer in some cases. A 2016 paper by the academics Jeffrey Pontiff and David McLean kicked the tyres of 96 separate investment “factors” and discovered that their market-beating returns on average halved after they became known. In other words, once a signal is discovered, it rapidly loses its value, a process known as “alpha decay”.
Nonetheless, quants are well aware of this, and are constantly scouring markets for new signals to mine. And investors are likely to keep faith. Emma Bewley, head of fund investment at Connection Capital, points out that all strategies suffer periodic downturns, and expects quant allocations to continue.
“They can’t all perform brilliantly all the time, and they don’t,” she says. “I’m not convinced that crowding is an issue. It might become one eventually but I don’t think that is the issue today.”
Copyright The Financial Times Limited . All rights reserved. Please don't copy articles from FT.com and redistribute by email or post to the web.