Will tech-savvy amateurs be able to consistently beat the market and take on the hedge funds?
By day Dan Houghton helps run Chilango, a Mexican restaurant chain in London, that he co-founded. By night, he is an algorithmic stock trader, coding complex, automated investment strategies once his children are in bed.
In 2013 Mr Houghton sold an online texting company that he founded while studying mathematics at Cambridge university. The proceeds were ploughed into his restaurant and a new family home, but rather than deposit the remainder in a staid mutual fund he decided to use his maths skills to try to beat the market.
After reading books on finance and scouring the internet for articles on “quantitative” investing, he stumbled over a company tailor-made for people like him: Quantopian, an online trading platform and “crowdsourced” hedge fund. The company is designed for a new breed of do-it-yourself algorithmic traders who develop automated investment strategies. Each month, the best “algos” win trading money from Quantopian and a select few are included in its embryonic hedge fund.
“My wife thought I was crazy,” Mr Houghton says. He had to learn Python — a programming language popular among quantitative analysts, or “quants” — from scratch, and after experimenting with the platform for a few months is trying to get one of his algorithms included in Quantopian’s hedge fund.
So far he has made $ 7,971 after trading commissions, but making money is not the main driver. “It’s primarily intellectual curiosity but I’m fairly competitive and the chance of getting included [in Quantopian’s fund] makes me stay up for the extra hour working on a strategy,” he says.
“Burrito-Dan”, as he is known at Quantopian, is one of a swelling number of mathematicians, programmers, data engineers and physicists who are taking advantage of cheap, powerful computers and the availability of financial data to code their own trading strategies. These DIY algo traders are an updated version of the army of day-traders that emerged during the 1990s stock market boom.
Some analysts believe this new wave of amateur algo traders can be harnessed using crowdsourcing techniques to disrupt one of Wall Street’s elite professions — the hedge fund manager.
“If our model is successful there will be no need for hedge funds any more,” says Martin Froehler, an Austrian mathematician who created Quantiacs, one of several online platforms for DIY algo traders. “A smart guy with a laptop will be able to start his own hedge fund. It will be very challenging to the big incumbents. A very simple idea can prove very powerful.”
Most day traders are ultimately unsuccessful — many were wiped out when the dotcom bubble burst — and are often derided by professional money managers as “dumb money”. The new DIY algo traders may be smarter and more computer-savvy than the average day trader but many experts warn they will find it just as hard to consistently beat the market. They scoff at the idea that their industry will face any threat from nerds coding in their basement.
But Quantiacs and its rivals are emerging as the classic hedge fund model is showing signs of strain. Investors have been pulling billions of dollars from hedge funds after a long period of underwhelming performance. And the fact that some technologists scent an opening underscores the sense of a looming industry shift, as more scientific, computer-driven approaches to investing gain favour.
Monetising pure data
Khaled Sharif, a young Jordanian computer engineer, arrived in the US in June to work at Nasa’s Ames Research Center, the space agency’s Silicon Valley outpost. Mr Sharif is a research assistant who specialises in deploying artificial intelligence techniques such as “machine learning” to crunch data on satellite images, weather patterns and earthquakes. But on the side he has an avid interest in another data-intensive area: the stock market.
He started using Quantopian a few years ago, attracted by a platform that allowed him to deepen his interest in using computer-based forecasting methods on the stock market. More recently, he found something that suited his expertise even more: Numerai, a small San Francisco-based hedge fund founded by Richard Craib, a South African mathematician.
Numerai provides users with encrypted, raw financial data to develop models using approaches like machine learning — Mr Sharif’s field — to help them make stock market predictions. They submit their predictions to Numerai, which awards prize money for the best models and incorporates them into automated trades for its hedge fund.
The prize money idea is similar to Quantopian’s but Numerai differs in that it offers a “pure” data challenge, leaving how to solve it up to users.
In this respect it is similar to Kaggle, a data science website popular among computer scientists, where people post challenges and offer rewards for solutions.
The concept is esoteric and ambitious but in April Numerai raised $ 1.5m from investors led by Howard Morgan, one of the founders of Renaissance Technologies, a computer-powered hedge fund group. Numerai represents an intriguing return to his past and a pointer to a possible hedge fund model of the future, Mr Morgan says.
“Numerai could be disruptive because you can get a lot of people doing this part-time at home,” he told the Financial Times. “We’re very excited about the democratisation of problem-solving.”
The hedge fund industry could be ripe for disruption, as investor dissatisfaction is rife. While the money trickling out is modest compared with the industry’s $ 3tn size, even insiders admit that the old band of iconoclasts, including people such as George Soros, Michael Steinhardt and Julian Robertson, has been replaced by a far larger, institutional industry where the sheer number and size of hedge funds hit returns.
Tony James, president of Blackstone, the US asset management group, has predicted “a day of reckoning” that will contract the hedge fund industry by as much as a quarter. “There will be a shrinkage in the industry and it will be painful,” he told a conference in May.
However, quantitative hedge funds such as Two Sigma, Renaissance, DE Shaw and Winton Capital, which eschew the gut feelings of human traders in favour of the logic of computer programs and mathematical models, have generally performed much better. As a whole, the quant hedge fund industry is on its seventh straight year of client inflows, with total assets going from $ 407.7bn in 2009 to $ 878.7bn this year, according to data from Hedge Fund Research, accounting for almost a third of the overall industry.
Talent has emerged as the leading challenge, with an intense battle among funds to recruit from rivals and Silicon Valley. But many people with the requisite skills have little interest in working for a big, established hedge fund in New York or London. Some merely want to dabble on the side of a full-time job.
This is where the new platforms — like Numerai, Quantopian, Quantiacs and QuantConnect — come in. They want to enable a new generation of iconoclasts who will help upturn the traditional hedge fund model.
Quantiacs is a good example of this vision. Like Quantopian and Numerai, it attracts computer scientists to its platform by awarding trading funds to strategies that perform best in its tests. The Silicon Valley company invests some of its own money in the competition winners, but Quantiacs is primarily focused on becoming a “marketplace” for automated investment strategies, and plans to open up the platform to individual and institutional investors to invest in algos directly.
While hedge funds traditionally charge “2 and 20” — a 2 per cent management fee and 20 per cent of profits — Quantiacs does not charge management fees and the 20 per cent of profits are split equally between the company and the quant. It hopes to funnel millions of dollars directly from investors to quants in their basements, dorms or research departments.
“There are a lot of people that don’t want to work at a big hedge fund,” Mr Froehler argues. He set up Quantiacs after working for a hedge fund in Switzerland and becoming disillusioned with the industry. “It is elitist, closed to outsiders and wastes a lot of talent. I would have loved to do this myself when I was studying mathematics.”
These attempts at harnessing the talents of these DIY traders are nascent, and the new platforms may never attain the scale of the quant industry, let alone the broader hedge fund world. Most money is invested by pension funds, endowments and insurers that want a more traditional, institutional set-up.
Many hedge fund managers are also sceptical, arguing that DIY algos often underestimate the differences between financial theory and practice. An algorithm that might do well in a simulated test environment can swiftly unravel when exposed to the chaos of real markets. Even strategies that work with a modest amount of money can turn into duds when they are scaled up.
“We’ve seen very little evidence that lone geniuses sitting in their living room will be successful. There are some examples but very few,” argues one quant hedge fund manager. Even superficially successful ones are vulnerable to sudden reversals. “These guys might not understand the risks that they’re taking. They might be doing well but they might be exposed to a huge tail risk they’re just not seeing.”
However, some of the industry’s leading names think the idea has potential. This year Steven Cohen, the head of Point72 — a family office set up after SAC Capital, his former hedge fund, pleaded guilty to insider trading, and the Securities and Exchange Commission banned Mr Cohen from managing client money until 2018 — invested in Quantopian and promised up to $ 250m for its embryonic crowdsourced hedge fund.
“The tools can now be put into anyone’s hands around the world to do very sophisticated quantitative investing. That is very different, and very powerful,” says Matthew Granade, chief market intelligence officer at Point72 and a board member of Quantopian.
Numerai’s Mr Craib admits that the concept is “hard to get your head around but I think this is the future”. He is convinced that the crowdsourced alternative hedge fund models will supplant the dinosaurs that have lumbered around markets.
“Will the hedge fund of the future really be lots of people sitting in a big office? I think that in the post-artificial intelligence, post-internet world it will be far more democratised and diverse.”
Recruitment: asset managers search for tech talent
As banks, asset managers and hedge funds become dependent on technology, the demand for data scientists and programmers has become intense. Many hedge funds have been scouring Silicon Valley to attract some of the best minds in tech.
This year, Citadel, a Chicago hedge fund run by Ken Griffin, poached Microsoft’s chief operating officer Kevin Turner to be the chief executive of its securities arm, while Bridgewater Associates, the world’s biggest hedge fund manager, hired Apple’s Jon Rubinstein to be co-chief executive.
Bridgewater had already hired IBM’s David Ferrucci — the man behind the Jeapardy-beating Watson supercomputer — to lead its new artificial intelligence unit, and Two Sigma, a quantitative hedge fund, last year hired Alfred Spector, formerly Google’s vice-president of research and special initiatives, as its chief scientist.
But the flurry of high-profile hires is just the tip of the iceberg, with a war for technology talent building across and at every level of the finance industry.
“What is an absolute certainty is that the asset management sector is set to become even more of a technology industry than it already is,” Rick Lacaille, chief investment officer at State Street Global Advisors, wrote this year. “Those firms that have made the necessary investments in data and technology will have the edge.”
This is a prime driver behind the establishment of online platforms for amateur or part-time quantitative traders. While a top quant hedge fund might employ several hundred mathematicians, physicists and programmers, these start-ups hope they can gain an edge by harnessing the skills of potentially tens of thousands of tech-savvy people from the US, Europe or Asia.
Copyright The Financial Times Limited 2016. You may share using our article tools.
Please don’t cut articles from FT.com and redistribute by email or post to the web.