Mercer aims to explore predictive power of AI
11 April 2019, by Anna Devine.
Mercer, one of the world’s largest investment consultants, has set its sights on tapping artificial intelligence to give its ratings on asset managers more predictive power.
The firm is at the early stage of exploring whether computer systems mimicking human intelligence could help in forecasting investment strategy performance.
The move could make the asset management industry more competitive, says Jo Holden, UK chief investment officer at Mercer.
She says the firm is considering whether there is the potential for AI to give “a bit more of a predictive power to our ratings so that our clients can move more quickly”.
Ms Holden says: “We do a huge amount of research to come up with manager ratings.”
Higher-rated managers on average outperform, she says, but Mercer is now also considering whether it should look at more data to predict trends and give its ratings more predictive power.
“It’s fair to say we ought to explore it,” she says.
However, Ms Holden adds that the firm has not yet identified what issues it needs to address in order to use AI. The firm may speak to asset managers to get their input on what could be considered.
She says Mercer is looking at AI now as it considers how it needs to adapt its business to help clients further.
She adds that the more data Mercer and its clients have, the more asset managers could have access to its clients in terms of sales opportunities.
“Better informed clients surely gives more asset managers an opportunity,” says Ms Holden.
However, the development could negatively impact some fund firms too.
“It could hurt the [asset managers] that aren’t on their game,” says Ms Holden.
“If we’ve got asset managers who perhaps weren’t on the radar of institutional investors […] then it challenges the rest of the market to up their game.”
In October last year, Mercer entered an alliance with Morningstar.
The consultant is connecting its data, which is largely institutional, with Morningstar’s mostly retail data to do a “deep dive on a much more extensive data set” and look at the asset management industry through one lens, says Ms Holden.
The firm is initially looking at how it can enhance governance of clients’ assets by analysing asset flows, fees and costs.
Mercer is interested in answering questions such as ‘Do our clients want to know how many people have looked at this strategy?’
Clare Flynn Levy, founder and chief executive officer of Essentia Analytics, a behavioural analytics service, says Mercer appears to be “dipping [its] toe” into AI.
“AI has the power to, if not transform the industry, certainly up the ante and level of professionalism that [firms] can serve,” she says.
Governance is one of the “easiest” places to start because data on asset flows and pricing will probably be quite granular, she adds.
However, to give its ratings more predictive firepower, Mercer will need much more granular data, she says.
“Really the asset managers themselves are the ones with access to that data,” says Ms Flynn Levy.
She says that AI could be used to predict which fund managers will be the best performers and that if investment management strategies can be assessed in a more predictable manner, then it will be “less scary” for investors when things go wrong.
However, the participation of asset managers in the process is essential, she adds.
If asset managers have access to the same information as Mercer then there is an opportunity for a “massive improvement” in investment performance, and the way investors and managers communicate with each other, she says.
“But if analytics are kept for the judges and not shown to managers […] and they can’t see or understand it, then it can’t be as useful.”
Amin Rajan, CEO of Create-Research, a consultancy, says: “With the rise of AI, this move by Mercer has long been overdue.”
He adds that the partnership between Mercer and Morningstar is “encouraging”.
“[These] two big brands […] have the capability and resources to exploit the power of predictive analytics that is set to change investing,” he says.
However, the firms face challenges. Mr Rajan says: “The rise of big data is welcome, but many investors worry about their noise-signal ratio.”
He adds: “The second challenge is to construct predictive models that actually deliver actionable insights.
“In investing, driven as it is by greed and fear, that has proved difficult in the past.”