Risks and benefits
Using machines to give financial advice can make investing and other financial decisions cheaper, turn swathes of left-out consumers into new savers and investors and remove human bias in deciding where and when assets should be bought or sold.
But so-called robo-advisors also bring new risks, ranging from technological glitches to making mistakes by misinterpreting what a consumer or investor really want and, perhaps more importantly, need.
This was the message from US and European experts brought together to discuss the rise of machines using algorithms rather than humans to give advice on investment decisions and other relevant choices related to retirement, pensions, mortgages and debt management. It is a fast-growing trend that is bound to increase in size and scope as automation swallows up a broad range of human activity.
The discussion – on October 4 in Syracuse, Italy -- was part of ING’s Think Forward Initiative, which together with the London-based Centre for Economic Policy Research seeks to bring together global experts to find out how and why people make financial decisions. The initiative’s goal is to inspire solutions to everyday money problems and future financial challenges. The Think Forward Initiative's goal is to find out how people make certain choices, and help them make better ones. It was initiated by ING, Dell EMC, Deloitte, Dimension Data and the Centre for Economic Policy Research. The wider TFI network consists of more than 200 different organisations (including universities, NGOs, companies, etc.) and more than 1,200 individuals, of which approximately half are researchers.
The case of robo-advising goes straight to the heart of the matter – essentially, can ordinary investors now and in the future rely on technological systems that cut way back on direct human interaction?
As chief executive of Betterment, a New York-based online investment house that uses tech to decide portfolios, Dan Egan clearly thinks so. He told the group that investing properly is time-consuming and complex. Robo-advising is a way of pushing that to one side. “We…try to make (clients) spend as little of their life as possible worrying about money,” Egan said.
Betterment has also shown the growth potential of robo-advising. Egan said that when he joined the company five years ago it had around 15,000 customers and just less than $100 million in assets under management. Today’s equivalent figures are 400,000 and $15 billion.
Karen Croxson, head of research and deputy chief economist at Britain’s Financial Conduct Authority (FCA), also saw great potential in robo-advising as a way of ensuring that financial advice is suitable, consumers are protected and more people are brought into the investment world. Indeed, the FCA supports the concept of robo-advising to help financial markets work for consumers, and particularly to safeguard the vulnerable.
Croxson said robo-advisors can close what she called the “advice gap” which exists between sophisticated, often wealthy, investors and those who are poorer with less knowledge of financial systems. We see evidence of poor decisions in many areas of personal finance, she said, so robo-advising can help to lower costs, provide 24/7 tailored advice, make faster and less human-biased decisions, and provide regulators such as the FCA with an easily auditable trail for accountability.
Is widespread robo-advice really a panacea?
But there are pitfalls, particularly in mismatching individuals’ risk preferences. Algorithms, for example, may be fine for people fitting a specific profile, but not good at handling outliers. Similarly, Croxson said studies have shown that the surveys or questionnaires that robo-advisors require from clients in order to make decisions are not always as good as humans at picking up biases, such as over-confidence. "There have been some suggestions that human advisors can consider a broader range of variables, soft variables that are harder to capture with automated tools,” she said.
Similar risks were highlighted by Francesco D'Acunto, assistant professor of finance at Boston College, who presented some academic research into robo-advising. Like the others, D’Acunto said machines could broaden the investor base by cutting costs, particularly given that affordability is often a reason given by people for not investing.
But he noted that using robo-advising tended to lead to the investor’s disengagement – that is, they leave it to the machine. This may be fine when it comes to accumulating wealth, he said, but not at, say, retirement when a person needs to decumulate it.
In the meantime, D’Acunto also cited evidence that people were more likely to use automation if they are asked to approve the machine’s decision beforehand. In other words, they wanted to maintain some control.
This gelled with Croxson’s view that a human may need to be involved to ensure that automation does not create too much risk for people.
So robo-advising is clearly on its way to becoming a big player in the world’s financial system, but it is perhaps not yet time that the machines take over completely.