The Faltering Rise of Robo-Advice
In September 2013, Carl Benedikt Frey and Michael Osborne, two researchers at the Oxford Martin School, estimated that by 2033 some 47% of jobs in the US could be replaced by computers.
For many working in financial services, this will not have come as a surprise. Machines are already carrying out a significant percentage of financial market trades and there has been a widespread assumption for some time now that certain types of financial advice will also be automated in the foreseeable future, particularly where customers are looking to address common personal finance priorities such as setting up and running a basic pension or investment portfolio.
In other areas, non-advised sales of financial products, for example standard insurance policies sold through price comparison websites, have become a familiar feature of the market based on fairly limited information gathering exercises to establish basic underwriting information. Surely it is a relatively small step from automating sales of insurance policies to robo-advised investment decisions?
That may be true but for all its attractions robo-advice has yet to enter the mainstream. Services such as Wealthfront and Betterment in the US and Nutmeg in the UK, which automate the gathering of customer data and use it to recommend portfolios of low-cost index funds, have attracted much interest and comment but remain some way short of critical mass.
As The Economist pointed out last autumn, this is partly because their low-fee model means they must gather substantial assets under management before profitability becomes possible while spending heavily on marketing to establish their challenger brands.
Even Nutmeg in the UK, an online discretionary fund manager with a slightly higher fee, has never published figures for assets under management or the number of accounts it manages. In May, founder Nick Hungerford stepped down as chief executive to take on a non-executive role. His successor was Martin Stead, previously chief revenue officer.¹
Does robo-platforms’ limited success so far indicate that belief in the potential of automated investment advice is misplaced? Maybe. But there are other plausible interpretations.
It seems clear, first of all, that advances in computer science over the past few years in areas such as machine learning and data analysis have brought closer the arrival of machines with the level of cognitive function required to automate complex processes that today depend mainly on humans. It does not therefore look as though technology is likely to be the major stumbling block, and in any case it is likely to improve significantly in the medium term.
Instead, it is possible that we are simply witnessing a phenomenon that has played out in numerous other businesses: early movers blaze the trail but frequently give way to the companies that follow them and ultimately win.
Examples abound but consider the fate of GoTo.com, the pioneer of paid internet search. The company, which later changed its name to Overture and sold itself to Yahoo in 2003, was founded by US entrepreneur Bill Gross and created the first paid search service to take off.
But Google soon caught up and overtook it, leaving GoTo.com as a footnote in internet history and Google with advertising revenues for 2015 of $67.4 billion, the vast majority from paid search. Similar things could be said of Boo.com in online fashion retailing, Napster in music and Sir Clive Sinclair’s early home computers and his battery-powered C5 vehicle.
Indeed, although public appetite for robo-advice seems so far to be confined to the early-adopter market, the idea of using it does show some signs of moving more into the mainstream: Accenture’s 2016 North American Digital Banking Survey found that around 80% of respondents could see themselves using fully-automated investment advice in future.
But if it is an investment truism that “to be early is to be wrong”, why might those following on a little later turn out to be right? There are several credible reasons to think so, the most important being that regulators, politicians and the industry all have an interest in seeing robo-advice take off, although to date they have tended to work at crossed purposes for much of the time.
Since the financial crisis, politicians and regulators have focused on promoting higher standards of professionalism and transparency among financial advisers and wealth managers, and enhanced consumer protection.
This has led to more stringent rules on investment suitability including Finra’s Rule 2111 in the US, Mifid II in Europe and similar moves from the Monetary Authority of Singapore and the Hong Kong Monetary Authority, along with a wholesale redrawing of the rules on financial advice in many parts of the world, starting with the UK’s Retail Distribution Review.
The RDR’s insistence on higher-level professional qualifications and the banning of payment for advice through commissions on product sales has dealt with certain perceived problems in the advice market. But in doing so it has pushed up the cost of providing compliant advice to the point where many financial advisers can no longer make money by serving clients with relatively modest sums to invest.
With an ‘advice gap’ opening, partly caused by problems of affordability, it is hardly surprising that non-advised sales have climbed rapidly in the UK. “The FCA’s product sales data suggests that the proportion of retail investment products (which includes pensions, retirement income products, and investments) sold without advice has increased from around 40% in 2011/12 to around two thirds in 2014/15,” according to the Financial Advice Market Review’s call for input 2015.
At the same time, research consistently suggests that only a small minority of retail investors is willing to pay the typical £1,000 cost of providing face-to-face financial advice, although 47% would be willing to pay some fee. It is therefore hardly surprising that both policymakers and regulators believe that automated services could offer a low-cost way to serve consumers who have modest sums to invest and little appetite to pay for full-service financial advice.
The obstacles to realising this vision so far have largely centred on concerns among wealth managers over the weight of regulation that applies to robo-services. Harriet Baldwin, the UK’s Economic Secretary to the Treasury, told a Financial Conduct Authority conference in September 2015 that a would-be entrant to the automated advice market had been told it would need to ask users 247 questions to comply with existing regulation.
Firms have also called for a better definition of what constitutes a ‘personal recommendation’ delivered by an online service, and remain worried by the possibility of systemic problems if flaws are discovered in the algorithms used to generate automated investment recommendations, which could result in large numbers of customers having the right to redress. This concern also figures prominently in the list of risks cited by the European Supervisory Authorities in their discussion paper on automation in financial advice, published in December 2015.
These are genuine obstacles, but in time they are very likely to be resolved given the desire among policymakers to ensure more people have access to affordable financial advice. Initiatives such as the UK FCA’s Project Innovate, which provides a regulatory ‘safe space’ to help firms develop new types of service that might benefit consumers, illustrate how some regulators at least are trying to smooth the way for innovators including those building automated advice services.
On the other side of the table, despite their reservations about the regulatory risks surrounding automation, advisers and wealth managers can also see the potential for big productivity benefits if these services can be made to work.
There are plenty of experiments already under way. Bloomberg reports that Vanguard’s Personal Advisor Service in the US attracted $12 billion between May and December 2015 to an offering that uses an online front-end to gather customer information followed by a conversation with an adviser before the plan is implemented.
Others are using less-expensive para-planners to collect customer information and draw up a financial plan that is then checked and signed off by a fully qualified adviser. In truth, these are just two approaches to the same central issue: how to rethink and reconstruct the information gathering and financial planning processes to reduce costs and increase productivity.
There are still a lot of wrinkles to be ironed out before robo-advice will really take off, and the likelihood is that we will see robo/human hybrid models as well as the ‘pure robo’ services that first movers such as Betterment, Wealthfront and Nutmeg have attempted.
But the modest progress so far does not indicate that the ultimate opportunity is small. It simply means that automating financial advice at the same time as regulators are pushing for stronger assurances in areas such as risk disclosures and investment suitability remains a work in progress that will require policymakers, regulators, advisers and asset managers to adapt their approach.
So far, the robo-advice field has been left to the start-ups with no legacy business to lose. But once the major incumbents can get comfortable with the regulatory framework and the technology risks, they are very likely to want to get more closely involved. And if history is any guide that is the point at which we will start to see the longer-term winners emerge – even if they do not ultimately include the names we read about today.
¹ See: Nutmeg chief steps down in shake-up at online wealth manager (3.5.2016) The Financial Times
Photo: © Niki Natarajan 2013