Save, make, understand money

Experienced Investor

Are machines set to take over from your fund manager?

Written By:

Man versus machine has long been the stuff of dreams for science fiction authors but with the onset of robo-advice worrying some financial advisers, fund managers also fear the repercussions of artificial intelligence.

Only recently Seneca’s chief investment officer Peter Elston hit the headlines when he poised the question; is a computer going to steal my job as fund manager?

Elston pointed to three famous examples in modern history when a computer beat the world’s best player in three games; chess in 1996, the Chinese board game ‘Go’ in October 2015 and poker in January this year. His natural question to follow was how long will it be before computers are beating the world’s best fund managers?

While algorithm trading and quantitative strategies have been around for a numbers of years, the announcement of a completely new AI managed fund, will have done little to appease Elston’s fears.

Fund management group Sanlam Global Investment Solutions (SGIS) this week announced that its Sanlam Managed Risk Ucits fund is now managed solely by AI, which it claims makes it the first fund in the world that is solely driven by AI.

According to the marketing literature, the fund was previously managed by “systematic investment process”, whereas now it is run using an advanced AI and machine learning (ML) investment engine.

This would mark it out as different to other quantitative (quant) strategies – or black box strategies as they are often termed – which are more rules based. An example would be iFunds, which manage three multi asset absolute return funds run using a hard-coded rules based system that aims to take behavioural finance traits and emotions, such as fear and greed, out of the equation when it comes to investing.

To put this more basically, the computer is given a set of rules and it obeys them and aside from some checks and balances iFunds does not interfere with or argue with it.

However the key word here is “checks and balances”. While quant, or algorithm based processes as they are also called, rely mostly on a computer to make the investment decisions, there is an element of man working with machine. In the instance of iFunds, while the computer comes up with all the ideas, it is left to a team of humans to implement all the trades.

The new fund from SGIS however is absent of human input to make investment decisions, being run using a “totally automated investment process”. By doing this, SGIS says it is removing another problem with ‘human run’ funds, called key man risk.

A problem with successful actively managed funds is that when a manager does well they court the attention of other fund management houses. If they leave, investors in the fund face a tough decision of whether or not to stick with the fund they are in, or follow the manager to their new group.

The black box approach is often criticised for working until it stops working. This is because whereas the stockmarket is a very adaptive process, a problem with any quant approach is that by design, they are not very adaptive.

According to David Itzkovitis, the goal at SGIS was to find a solution that can adapt as quickly as the markets change.

“Our AI capability does this by applying the latest ML techniques which have multiple years of live operational experience and a real, impressive track record,” he says. “The AI engine derives its decision making in a different way to other investments and should be viewed within a client’s overall investment strategy as a diversifier of existing human manager risk.

“Because in today’s world, it’s not man vs. machine, it’s man with machine vs. man without.”

So, like Judgement day in the movie Terminator, are machines set to takeover the fund management world? Ben Yearsley, a director at Shore Financial Planning, says: “The answer is no. Just the same as passive investing isn’t the only option, and often active doesn’t work. This is just another way of managing money. Sometimes it will work, sometimes it won’t.”