How can humans work with machines in the age of digital investing?

In recent years, a technological revolution has begun within the fund management industry, with advances in fintech making information more accessible than ever before. New algorithms are being developed all the time to parse the colossal amounts of digital data generated every day.

For those who have witnessed the evolution of the industry over the past couple of decades, the changes have been vast. ‘Back when I started in 2002, my access to data when it came to analyzing funds and looking at which ones to recommend was a combination of what was printed in the Sunday papers and CDs with the latest fund data on them, which were updated monthly,’ says Martin Bamford, managing director at financial planning firm Informed Choice.

Now, nearly 20 years on, we live in a world where managers and fund selectors across the globe are expected to spend more than $1 billion a year on digital information and skilled analysts by 2020. It means that almost every manager has their own USP when it comes to how they utilize technology and digital resources.

At Myria Asset Management, investors use vast data warehouses and various mining tools to gather information that helps to influence decisions on which stocks to buy and sell. ‘For now, one of the biggest ways that new technology has helped us is in terms of organizing and being able to browse lots of data efficiently,’ says Pierre Bismuth, director general at Myria. ‘We can then use this to create models for every company that we invest in.’

For other managers, one of the main advantages of digitization is the increased transparency that it provides for clients, particularly around the investment process. This is especially relevant for the tech-conscious millennial generation. ‘We are a small investment boutique, so we have quite a range of clients, and with that, a range of attitudes about how much the clients care about all this new technology,’ says Madalena Teixeira, a senior portfolio manager at ASK Wealth Management. ‘For example, some of our clients still want to receive their monthly statements in the post rather than by email, but we have digital solutions which enable younger clients to monitor how the portfolio is doing on a daily basis, how well orders have been executed and how the portfolio is performing compared with other strategies.’

So far, though, the biggest beneficiaries of technological advances across the industry have been passive funds, through the rise of ETFs. ‘A fund manager being able to replicate an index’s return cheaply is probably the biggest benefit from technology so far,’ Bamford says. ‘That’s along with ease of communication, ease of access to data and the ability to see valuations in your portfolio whenever you want, rather than having to get them manually valued.’

While there has been considerable hype regarding whether advancements in artificial intelligence (AI) will make it possible for machines to take over the fund selection or stock selection process, few managers see this taking place in the very near future. Bismuth believes that within the next five to 10 years, funds will emerge that specialize in the use of AI, in the same way that hedge funds and quantitative funds have become mainstream. However, most of this technology will be put to work carrying out natural language processing for conducting comparisons between potential investments, he argues.

‘AI is very efficient for reading text,’ Bismuth says. ‘When used in conjunction with databases and data management, it can be used to sweep all the information regarding a company, create some assumption of where that company is going and conduct comparisons across that sector and indices. This is the next step for us.’

So don’t expect to see robots replacing advisors anytime soon. While digital applications are available that can quickly simulate an optimised portfolio for any investor, there are no fallback mechanisms in place should anything untoward occur. ‘If you have a small portfolio, you don’t want to expend too much. You don’t have time to look into orders, so you just follow what the computer says instead,’ Teixeira says. ‘It’s cheap, but it won’t advise you if there’s anything wrong with the fund or the fund manager or if the liquidity is big enough, and it won’t do anything differently if the markets go down.’

What’s more, successful investing is typically the product of decisions made based on a diverse range of human opinions within a firm – something that machines cannot yet replicate.

‘What computers are very good at is collating a consensus, but to be a successful investor, you often need to be a contrarian,’ Bamford says. ‘If a computer is simply looking at what everyone else is doing and trying the same thing, that’s not going to be a good outcome. But if you use all that data and do the opposite, then maybe you stand a slightly better chance.’

Given the current state of AI, the lack of genuine cognitive understanding means that in a world of entirely machine-driven investing with no humans making the final calls, disaster could easily strike. Bamford cites the example of the 2016 ‘flash crash’ in the Asia-Pacific market following comments on Brexit from then French president Francois Hollande.

‘We have seen these examples in the past of computer-driven algorithms misinterpreting headlines and triggering a big sale,’ he says. ‘There are dangers if machines have too much autonomy. Investing should be a long-term thing. It should be based on long-term visions, not knee-jerk reactions. So while tech can be used to probe company accounts and identify opportunities, the key is this blend of human input and tech input. One without the other is fairly ineffective.’

But Bismuth says that while much of the drive towards the introduction of AI-driven strategies is coming from within the industry, so far there is little client demand for newer and more sophisticated technologies. Instead, he says, the majority of clients are more focused on ESG screening than machine learning.

‘Clients are keener on the environment than the use of AI in creating or managing portfolios,’ he says. ‘Things such as Big Data or AI are more essential to businesses for efficiency and saving costs than to the clients. The consumer doesn’t really care about you using this kind of technology.’