The battle of active versus passive funds has intensified in recent years – but at a smattering of funds, active managers work closely with computer algorithms to combine the two.

Gideon Smith, manager of the AXA Global Factors Sustainable Equity fund, has designed a computer algorithm which can pick out companies likely to deliver sustainable earnings growth and low volatility.

It scans company reports, the news and even social media to work out which stocks the fund should back and which it should avoid. “I used to do computer programming in the film industry but realised I could earn more using my skills in financial services,” Smith quips.

Launched in 2013, the fund looks to grow investors’ money by investing in a diverse range of companies across the globe. So far, so typical. The difference here is that a computer programme is making the decisions.

As well as looking at standard measurements such as profits and debt, the programme can interpret other factors to determine whether or not a stock deserves a place in the 400-strong portfolio. One data point it analyses is how many women there are on a company’s board.

“We know that companies with greater diversity on the board do better over the long term, so we look at that when we’re considering how sustainably a firm can grow its profits,” says Smith.

Similarly, the program will look at a business’s carbon footprint. “Because there’s always an associated cost with a high carbon footprint, whether its increased regulation or lost customers.”

Active vs Passive

In the ongoing – and likely never to be resolved – debate over whether active or passive funds are best for investors, the arguments are well-known: passive funds will deliver stock market performance at a low-cost where active managers bring the potential to beat the market – or to lag behind.

So where does Smith’s fund sit in that spectrum? Somewhere in the middle is the answer, in a pocket known as Smart Beta or Factor Investing. This is where a benchmark is constructed with specific rules for a computer algorithm to follow.

The computer is not just blindly following a stock market but taking into account any number of pre-determined factors to construct a portfolio.

Close-up of Facebook icon

Smith's computer predicted a spike in Facebook volatility after finding news stories only suggesting concerns about ethics and privacy 

The fund certainly seems to be bringing together the positives of both types of investing: it has delivered annualised returns of 9.3 per cent over three years and is up a hefty 19.3 per cent year to date. Meanwhile, the annual charge is just 0.3 per cent.

While the manager may create the rules, his role in running the fund is arguably reduced. “The way I see it is that the computer can do the automated stuff and that frees me up to do other things, where I can add more value,” says Smith.

“I’m not worried that I’m going to be out of a job any time soon. It’s the same as a plane – it’s mostly flown by autopilot, but passengers like the fact that the pilot is there and can step in if necessary.”

Certainly, it seems that there is an endless amount of tinkering and monitoring to do in overseeing the algorithm. There’s an ESG screen that Smith can use, whereby the computer looks for evidence that a company is “talking the talk, not just walking the walk” in its environmental, social and governance policies.

There is also a distress screen, which looks not just at how a company has coped with previous bouts of volatility but at whether they could be volatility on the horizon. “There is so much information out there that it’s easy to get information paralysis,” says Smith.

Recently, Smith reduced the fund’s holding in Facebook after a red flag was raised by the algorithm on this very screen. The company ticks many of the boxes that the manager is looking for with its sustainable earnings and steady growth, but the computer predicted a spike in volatility after finding news stories only suggesting concerns about ethics and privacy.

Information overload

It’s the stuff of sci-fi films, that a computer can read the news and scan the internet to make investment decisions. One algorithm is even trained to read social media and understand emojis. “How does it understand sarcasm and nuance? The same way you do,” says Smith. “By learning over time and getting more experience.”

The manager is regularly impressed by how effective the algorithm is. He remembers checking the portfolio after reading in the news about the Volkswagen emissions scandal, only to see the algorithm had already spotted volatility ahead and raised a red flag.

“Our computer even had a red flag on Enron 15 years ago before the scandal. It saw that the amount of tax the company was paying didn’t tally with its reported earnings and flagged that something must be wrong.”

Of course, Smith can’t entirely automate himself away. Fund management is a regulated industry and computer can’t hold the fiduciary responsibility to the fund’s investors. Finding the balance between how much of the fund’s management is automated and how much is overseen by the team is a constant fine-tuning process.

The computer used to rebalance the portfolio every three minutes, for example, but that was ruled to be inefficient and likely to add to costs so has been slowed down.

There are more than 400 holdings in the portfolio, which is a big list by anyone’s reckoning. While Smith points out that this is a small portion of the total global investable universe, he also concedes that having so many names in the fund limits the computer’s ability to damage performance by making a bad call.

“Fund managers need an edge and my computer programme increases my chances of getting an investment right from, say, 50/50 to 60/40 – so that’s my edge, but the buck still stops with me.”