AI is a game-changer for fund managers
18 November 2020
- Gillmore Centre for Financial Technology hears latest research on AI
- Technology offers a way to reverse declining returns from active funds
- Webinar hears new cutting-edge methods on how industry is using AI
- New AI model produces vastly better returns than usual factor methods
AI offers a chance for fund managers to step away from the crowd and end years of declining returns, a webinar for investors and researchers heard.
At the latest webinar held by the Gillmore Centre for Financial Technology, Dan Philps, Head of Rothko Investment Strategies, revealed how AI is changing the face of investment for fund managers.
Warwick Business School’s Gillmore Centre for Financial Technology has been established to research emerging technologies in the financial sector such as AI, blockchain, mobile payments, cryptocurrencies and crowdfunding platforms.
Those tuning in for its webinar heard how over the last five years traditional quantitative managers have seen returns dwindle, being outperformed by passive funds that have lower fees. But Dr Philps believes AI can reverse this trend and help separate fund managers from the crowded space that traditional factor investing has created.
Dr Philps said: “The explosion in data alongside rapid advancements in AI are a potential game-changer. It has created a fantastic opportunity across capital and investment markets.
“Factor-based investing assumes linearities and research has shown that returns have been exaggerated plus diversification has been illusory. It is also now a crowded area and in our view dangerous.
“Traditional quant investing is monopolised by factors, so funds are all using very similar models. This is producing a crowding effect, with capital moving into and out of similar trades at similar times. Factor strategies are hitting capacity limits in many equity asset classes, which might be leading to the lower returns from factor investing we have seen in the last five years. AI offers an alternative away from the crowd.”
Hedge funds and high-frequency traders have been using machine learning to spot arbitrage opportunities across exchanges for some time, but long-term investment funds have been slow to adopt the technology.
However, Dr Philps revealed his research into the many branches of AI, detailing the perfect blend of cutting edge technologies that produce accuracy and interpretability while aiming to avoid over-fit. This will allow fund managers to understand the stock-picking decisions made by an AI approach, rather than being required to trust ‘black box’ decisions from AI techniques like random forests, deep neural networks or similar.
“The solution we favour is an ensemble of simple models,” said Dr Philps. “This makes the outcome interpretable while being driven by many, many input series, analogous to many expert perspectives producing one, objectively justifiable decision.
“The decision boundaries we use emulate human thinking with many perspectives assessed to give a high dimensional picture based on earnings, the value of stocks and downside protection. It creates a landscape with a high dimensional region of stable defensive stocks. It is complex and non-linear.”
In the example shown in the webinar Dr Philps stated turnover was between 40 and 60 per cent, and when tested against factor modelling the AI system produced significantly better alpha – ie excess return above the market.
“AI can reveal complex interactions and predictors, which factor modelling can’t, and it can find them in a non-linear way,” said Dr Philps. “AI can also be designed so it doesn’t forget significant events like the 2008 financial crash, unlike factor models, which have to be retrained at each step.”
The webinar was part of a series of events being organised by the Gillmore Centre for Financial Technology, which was launched in October 2019 thanks to a £3 million donation from Clive Gillmore, founding member and CEO of Mondrian Investment Partners that holds Rothko as a subsidiary.
Ram Gopal, Pro Dean for Research and Director of the Gillmore Centre for Financial Technology, said: “The centre has three research streams; the first looks into consumers and their financial behaviour; the second will be focusing on firms in the finance industry and how they are using Fintech for their decision-making and portfolio construction; and the third will be on micro issues like cryptocurrencies, cybersecurity, regulation and societal impacts.
“Technology like AI is rapidly transforming the finance industry, creating questions, uncertainties and challenges which the Gillmore Centre for Financial Technology will seek to investigate and bring fresh perspectives to this fast-changing sector.”
To watch the Gillmore Centre for Financial Technology webinar in full click here.