As financial markets have rapidly digitised over the past decade, and the availability and pace of the information exchange has accelerated, the traditional barriers between human- and machine-based strategies are steadily being eroded.
Traditional discretionary hedge funds that utilise the portfolio-building skills and investment instincts of star traders and portfolio managers have often stood in sharp contrast to the tech-heavy, computer-based quant strategies built around algorithms, data and machine learning applications.
Now, though, a cursory glance across today’s industry landscape suggests more and more discretionary ‘human-first’ hedge funds are employing technology and data to both streamline and strengthen their front- and back-office functions.
“There is still a very valuable process of portfolio managers looking into the whites of the eyes of company operators and management,” says Finbourne co-founder and CEO Thomas McHugh. “But there is also a lot of data that now goes into that process.”
Scratch the surface of many discretionary hedge funds and long-only investment strategies today, notes McHugh, and you’ll quickly find managers using an assortment of P&L ratios, balance sheet numbers, ESG classifications, and “a whole raft of other ancillary data” to filter their respective investment universes down from tens of thousands of companies to a few dozen.
“That sounds an awful lot like a quantitative approach to narrow down that universe,” McHugh tells Hedgeweek. “What you tend to find is that the selection process for a lot of these funds is much more quantitative than people would lead you to believe. With data evolving so quickly, the question all funds need to ask themselves is – ‘is your technology set up in a way to quickly take advantage of new opportunities?’.”
Utilising technology to better handle the vast amounts of information strengthens managers’ credibility in the eyes of investors, market participants say. Specifically, discretionary managers are increasingly integrating systematic analyses into parts of their investment process to test trading ideas, more quickly, to better monitor real-time changes in trading signals, improve portfolio construction, and to run various risk, attribution and factor exposure analyses.
“Data is growing at an exponential rate. This explosion in the type, sources and quantity of data, when combined with increasingly sophisticated data science filters and techniques, and AI or machine learning processes, allows experienced managers to capture new alpha streams,” observes Dharmesh Maniyar, founder, CEO and CIO of Maniyar Capital, whose firm applies advanced quantitative and Bayesian inference techniques to global macro investing.
Maniyar, a trained machine learning scientist, was previously a senior portfolio manager and partner at Tudor from 2013, having previously spent five years as a macro portfolio manager at Brevan Howard. He describes the emergence of unstructured data as “one of the biggest advances of the 21st century”.
He explains: “We have seen this development in a number of industries and we are starting to see it increasingly in the investment world where computer science is being combined with the domain of macro-economics. We believe that our ‘mind + machine’ approach combines the best of macroeconomic analysis with computer science to generate attractive risk-adjusted returns for our investors.”
“Many of the large, successful global macro managers now have large divisions using quant models to assess risk and to create forecasts,” says Razvan Remsing, director of investment solutions at Aspect Capital.
“They might not be fully systematic in the sense that the ultimate construct of the portfolio choice is still a portfolio manager’s discretionary decision as to where they allocate. But they have big data budgets, and big data science teams – that tells you how important this is.”
However, it’s not entirely one-way-traffic. On the other side of the fence, some quantitative investment processes are increasingly pivoting towards aspects of human input, industry participants say.
Laurent Laloux, chief product officer at CFM, identifies certain spheres of the investment process where systematic strategies look to leverage some human expertise.
“You have certain sector specialists, and you want to find out how they value a company and then you look to try to systematise that,” he notes, adding that discretionary and quantitative managers are “coming from two extreme ends of the spectrum” and converging “in a kind of middle ground.”
“All of us are coming from different parts of the spectrum. But I think in a way we are somehow meeting in the middle. It is definitely something that’s very interesting happening in the industry,” he says of this ongoing cross-fertilisation of processes.
Alex Nixon, chief technology officer at Maniyar Capital, comments: “Having a human overlay and a human involvement in the right parts of the process is an important mitigator. I think any process, any automated pipeline, needs checks and balances of some kind.”
Paulo Remião, partner and portfolio manager at emerging markets-focused asset manager Broad Reach in London, says the interplay between humans and technology is a key component of his firm’s approach.
“Emerging markets were previously too illiquid, too idiosyncratic, too hard to access – but now you have an immense breadth of countries, asset classes and instruments, and that lends itself to a systematic process,” Remião tells Hedgeweek. “But in order to access and trade those markets, and be comfortable with the risks that exist in these markets, it still requires a level of understanding and oversight which, in a way, is completely foreign to the processes most systematic-only houses would be comfortable with. This is where the discretionary element helps the systematic process.”
Building on this point, he continues: “There’s a level of risk management which goes way beyond the traditional returns-based or data-based issues. Unlike in a typical systematic, mainstream, liquid environment, with emerging markets you can’t count on the liquidity, the transaction costs, or the market access to remain stable through time. In EM, there are capital control risks, political risks; the market structure can change completely from one day to the next; some instruments can become unsuitable for specific strategies.”
He adds: “While someone who trades US treasuries and the S&P 500 can count on the exchanges being open the next day to take profits, if you are trading emerging markets and frontier countries, some of which may occasionally become vulnerable to fundamental trading impact effects such as capital controls and sanctions – the most recent example being Russia – and there is significant risk of not being able to monetise those positions regardless of how profitable they may have been.
“This is where the discretionary oversight becomes essential, albeit in very occasional and exceptional circumstances, in order for you to be able to systematically deploy capital in emerging markets, and avoid the traps that a purely agnostic, fully systematic trading model may face.”
Read the full A Tech Revolution: How machines are reshaping hedge fund investment Insight Report here.