Quants – the use of mathematical models to drive investments – cover a broad spectrum of strategies from basic smart beta allocations to complex sci-fi-style algorithms. The term refers as much to the analysts behind models as well as the systems themselves.
“Quants are often portrayed as the computer making all the decisions but in reality, computers simply follow instructions,” said Matthew Beddall, chief executive at investment house Havelock London. “There is a lot of human input behind most quant strategies.”
“In quants, a person draws up the blueprint, and the computer builds it. But you still need someone to create the blueprint. You can’t just get a computer and expect it to make money.”
Unfortunately, quant strategies have not been making as much money as investors would like this year.
Global inflows into quant investment strategies almost halved to $4.6bn (€5.1bn) during H1 – the weakest first six months since 2009, according to data provider Hedge Fund Research.
Peter Sleep, senior investment manager at Seven Investment Management, said surging volatility and the gravity-defying performance of FAANG tech stocks have contributed to the slump.
Facebook, Amazon, Apple, Netflix and Google now represent 50% of the market cap of the Nasdaq 100 index.
The market’s overdependence on a handful of outperforming stocks has led many systematic quant-led funds to short the FAANGs, Sleep said, but tech growth stocks have continued to defy expectations this year – and as a result many quant strategies have suffered.
Moreover, the surge in volatility on global markets this year has made life more difficult for trend-following strategies.
Beddall, who spent 17 years at quant house Winton Capital, said a lot of money has poured into quant strategies over the over the last decade. “As quant strategies become more popular it becomes more and more difficult for people to make money,” he said.
Not enough funds have a clear understanding as to how a strategy can produce returns, said Philip Bagshaw, senior portfolio specialist at City Asset Management.
“A lot of funds point to their algorithm or base their expectations on past returns but that is a bad starting point,” Bagshaw said.
“It’s more about how well a strategy can deliver its on objectives. Sometimes that is via a person [an expert in quant analysis] sometimes it is more systematic. We’re fairly agnostic [about which one]”.
City Asset Management relies on two systematic or quant holdings: Old Mutual Global Equity Absolute Return (Gear) and ADG Systematic Macro.
“You have to start with understanding what are you exploiting,” he says. “If you don’t know what you are using your algorithms to exploit, you’re on a road to nowhere.”
The rise of data science
Despite the drop in H1 flows, advances in data science suggest that quants are poised to command an even greater share of the investment market in years to come.
Data science involves combining computing, statistics, and domain knowledge (knowledge of investment management) to identify where and how to use data more effectively.
Many asset management groups now have teams of dedicated computer scientists focused on how to better utilise the vast troves of data that are now available.
Havelock, for example, hired data scientist Kate Land to lead its data science effort, while Swiss manager Gam has employed Dr Camilla Schelpe as lead scientist across its portfolios.
Gam said that a rigorous scientific approach was necessary to get the best of out of ‘big data’ and avoid some of inherent risks that can be found in data-led quant investing such as relying on ‘unclean’ or mismarked data, succumbing to ‘overfitting bias’ with regards to machine learning.
The science behind quants is always developing, Beddall said. “Over the longer term, it is going to play an increasingly important role in investing.” But for the time being at least there remains a irreplaceable role for humans in the process.
For more insight on UK wealth management, please click on www.portfolio-adviser.com