Are regulators right to focus on robo-investing risks?

The robo pioneers argue quant-based investing will be more complex, but also more efficient

BlackRock, Vanguard, Fidelity and other leading names in fund management have been buying robo-investment platforms, or starting their own. Industry watchers have been predicting a “race to the bottom” as fees are cut to persuade customers to sign up. Especially in Europe, where the robo platforms have been charging higher fees than their US counterparts.

But the costs of regulatory compliance could prove to be the real game-changer for the robos,  SEC commissioner Kara Stein recently expresed her concern that none of the laws regulating investment advice in the US were conceived for an age when that advice was communicated online, rather than in person. Investors, she said, may not realise that robo-advisors: “Will not be on the phone providing counsel if there is a market crash.”

And Europe’s regulators are raising a warning flag of the specific threat of IT risk facing robo-advisors.  The “super-supervisor” of European financial markets – the joint committee of the three European Supervisory Authorities (ESAs) – the EBA, EIOPA and ESMA – in December called for a discussion about the use of computer technology to provide automated investment advice to consumers. According to Steven Maijoor, chair of the joint committee: “Financial innovation is important and, at its best, contributes to economic growth. However, this can only be achieved and sustained where consumers have confidence in such innovations.”

While Maijoor’s committee says it sees the benefit of lower costs for more customers being delivered by robo advice, there is also potentially the risk of exposing those customers to IT problems they would not experience talking to a human advisors.

Robo-advisor managers argue that the vast majority of investors do not need to pay a premium for the kind of advice they would get from a human advisor. Are the interests of the moderately wealthy really best served by a human advisor who has maybe 500 clients to advise? How well is that advisor going to know their different needs?

And when it come to IT risk, the other big question is the extent to which automation of the investment process really does introduce a new level of complexity. The ESAs’ discussion paper defines automation in financial advice as where a financial institution provides advice or recommendations to consumers with little or no human intervention and relies instead on computer-based algorithms or decision trees.

It is true that some robo platforms are more complex than other in terms of their underlying methodologies. InvestYourWay, the UK-platform that has recently been acquired by IG Group, offers investors a decision tree that is no more complex than, say, Nutmeg’s. But while Nutmeg allocates the investor’s money to an ETF-based portfolio, InvestYourWay uses contracts-for-difference (CFDs). It had previously partnered with IG, which provided the CFD trading service.

Portfolio theory

In general, though, robo-investing could be described as a misnomer.  Most robo-advisor platforms don’t actually do “robot investing” in the sense that they are, for example, using complex algorithms to build bespoke portfolios for individual customers. Instead, they are allocating money to pre-determined portfolios on the basis of the answers investors provide to an online questionnaire.

The risk targets of those portfolios are based on modern portfolio theory as pioneered by Harry Markowitz, the 1990 Nobel prize-winning US economist.  And most robo platforms also simplify the underlying investments in portfolios, because most prefer to allocate assets via exchange traded funds (ETFs), because of the low costs involved compared to traditional managed fund structures.

Investor objectives

The world’s most successful robo-advisor, Welathfront, based in the Silicon Valley, asks potential investors just seven questions to score their risk profile on a scale of one to 20 and offer them a suggested investment portfolio, based on ETFs.  WealthFront charges nothing to manage portfolios of under $10,000 and just 0.25% annual above that, and it’s low cost approach has enabled it to gather $2.6 billions of assets under management. (By comparison, Nutmeg, the UK’s most successful robo-advisor, charges a 0.95% annual fee for management of up to £25,000 of investments.)

Robo platforms website are designed to establish three key investor objectives: how much does the investor want to invest, for how long, and at what level of risk tolerance? More detailed questions are designed to test how the investor would react if markets moved against them.

Investors are not typically asked questions about portfolio composition and, as a result, the risk view of the portfolio managers makes a big difference. A recent study by the Wall Street Journal found substantial differences in the allocations to US equities, for example,  recommended by the leading US robo-advisors to investors expressing an appetite for “moderate” risk. Wealthfront’s recommended portfolio allocation was 41%; Betterment’s was 34% and Schwab intelligent Portfolio’s was 30%.

Mourtaza-Ased-Syed-Yomoni

Mourtaza Ased Syed, Yomoni

In Europe, the problem is that the robo-advisors’ investment questionnaires are already dominated by the requirements of  regulatory compliance, argues Mourtaza Asad-Syed, founder of Yomoni, the French online wealth management platform. Compliance is a given, of course, but it can get in the way of understanding of the long-term needs of the customers.

The customers targeted by the robos are for the most part very risk-averse, says Asad-Syed. But he would like to see, over time, another next level of user experience introduced into the on-boarding process, one with more inter-action, or even gamification, so that investors learn more about investing and their own risk appetite.

Paolo Sironi, head of thought leadership for IBM’s wealth management analytics division, agrees that retail investment technology should be using even more complex – and engaging – tools. He believes asset managers should be moving on from modern portfolio theory to probabilistic scenario optimisation (PSO), which employs stochastic scenarios to derive multiple targets based around different time periods, for example. PSO should enable better goal-based investment, and help investors think harder about exactly how their retirement will work in practice.

This new phase of genuinely quant-based investing for the masses will potentially be more expensive to provide, says Sironi. He says robo platforms should plan on investing in software that will enable the graphical representation of quant-based investing for the masses.

To keep costs down, wealth management firms may spend less on their own platforms and more on delivering “robo-as-a-service” via APIs, says Sironi.  And artificial intelligence software such as IBM’s Watson will be used  during the on-boarding process to establish “cognitive dialogues” with clients that will seek out patterns of information that can flag up potential problems. “The challenge is that even with an expensive platform and analytics, some customers will not choose the right products,” says Sironi. “The solution is to funnel those customers based on compliance requirements as well as investment goals.”

Auomating admin

But compliance is expensive. Yomoni’s Asad-Syed argues that the real innovation of robo-advisors is that they can manage huge numbers of clients for very low costs. The advantage of the robos is they can automate the administrative process of on-boarding those clients, and the somewhat more complicated one of adjusting their portfolios over time to reflect changing circumstances, such as marriage, a new mortgage, college fees or divorce.

The asset allocation process is two-fold, and is not automated. First, long-term strategic allocations are decided by Asad-Syed, his quant manager and his portfolio manager. The decisions taken at this stage are all about how to optimise portfolios with key assets, and what the expected returns from those portfolios should be. The second, tactical, stage to the process is the regular, typically monthly, rebalancing of the portfolios. the aim here is to mitigate short-term risks such as country-risk, sector risk or inflation risk.

The portfolios are back-tested using standard industry tools such as Monte Carlo simulations, but tested more to ensure that risk is kept to a minimum than returns are maximised against benchmarks in volatile markets. “We know that clients who suffer in markets will quit,” says Asad-Syed. “So a lot of the back testing is not about expected returns or establishing how good we are as managers, but it is about calibrating risk.”

The Yomoni product is wrapped in a life assurance product for French tax reasons and, says Asad-Syed, the one day delay added to the client on-boarding process by the creation of that wrapper is frequently a more significant challenge to risk management than market volatility per se.

But where real problems with the process could arise for robo-advisors, says Asad-Syed, is in the assessment of  the client’s risk appetite at the outset; when that client fills out the online questionnaire. “That risk score defines the portfolio. If that score is wrong, that would be a big source of mistakes,” says Asad-Syed. There are a number of compliance-driven questions, notably KYC compliance, that have to be asked before the really interesting ones can be asked, and customers can only be expected to answer so many questions. “As an industry we should be looking to take a big step forward in risk profiling,” he says. “We have to ask a number of questions, say seven, that are compulsory, but really the limit on the total number of questions will always be 15 rather than 50.”

Robo advising sites in future will do more to engage potential customers at the emotional level so they can work out not just what their financial position is, but what their emotional attitude to money is. “A robot might say that a Formula One driver who is rich and takes risks for living might be very pro-risk in their investment choices. But that’s not the case,” says Asad-Syed. “That race driver would be more likely to say: ‘I take enough risks in my career, I want my investments to be super boring.”

Asad-Syed also believes that the robo-advisor business will be characterised by an “arms race” to improve the best customer experience. “UX will equal market share,” he predicts. “We have got  to stop annoying people and start entertaining them.” That might not be music to the ears of investment markets regulators. But then, when it comes to robos, those regulators may not be focused on exactly the right issues.

Additional information: Paolo Sironi’s blog is here. We also recommend this article by Professor Robert Merton’s on how goal-based investing should transform pensions saving.

Tweet about this on TwitterEmail this to someoneShare on LinkedIn

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

*