Could your next financial advisor be a robot?
John M. Mulvey, professor of financial engineering and operations research at Princeton University, thinks that’s a definite possibility, and so do many of the other finest minds in the world of finance. For many years hedge funds and big investors have used increasingly sophisticated algorithms to pick stocks and to buy and sell securities of all kinds.
Unlike a human advisor, an algorithm can analyze vast amounts of data: everything from weather reports to balance sheets, public statements, and news items can be taken into account. (As an experiment, reporters from the Planet Money podcast made a bot that monitors Donald Trump’s Twitter feed and automatically buys stock when he praises a company, and sells short the stock of any company he criticizes. The bot had not made any trades as of press time.)
Financial advisors have long used these sophisticated algorithms to inform their decisions. In some cases, programs have automatically bought and sold securities on their own, most famously in the realm of high-speed trading.
The next step is to bring the power of computer decision-making to the average consumer, not just to gain returns, but to help average people build investment portfolios for retirement planning or to achieve other financial goals.
“Robo-advisors” are now available to the average consumer, not just wealthy investment firms. Companies like Charles Schwab and Vanguard have launched automated advisors for customers who want to give active investing a try, but who can’t afford a human advisor, or who don’t want one.
Even companies that aren’t brokerages are delving into robo-advisors. Alibaba, the Chinese e-commerce company, is planning to offer investment bots to the 400 million users of its Alipay service.
Business leaders and academics are sure that the automated investment wave will have far-reaching effects. Human investment advisors could be put out of work.
“Potentially it’s disruptive technology,” Mulvey says, “sort of like cars that drive themselves. Now we have investment systems that drive themselves. Will it be disruptive? It depends on how widespread it becomes, and how much pressure it puts on investors to lower their fees.”
Experts will discuss these and other questions currently facing the world of financial technology at an upcoming conference hosted at Princeton University. Financial technology, or fintech, is what experts are calling a new industry that uses computing technology to create new business models and gain advantage in the world of financial transactions. Princeton University’s Operations Reesearch and Financial Engineering department, where Mulvey works, was founded in 1999 and is deeply involved in this new arena. The department conducts research in game theory, statistics, and quantitative analysis that is central to FinTech.
The upcoming FinTech conference hosted at Princeton is meant to delve into this business frontier. Three days of related events begin Wednesday, April 26, and continue through Friday, April 28, at Princeton. For more information, visit www.edhec-risk.com/events or fintech.princeton.edu. Tickets for the April 26 wealth management systems for individual investors conference are $150. The next day’s tutorial classes and demonstrations are $100.
The first day will focus on wealth management systems for individual investors. The event, the first of an annual series of FinTech conferences rotating between four institutions, will feature talks by Mulvey, Lionel Martellini, professor of finance at the EDHEC-Risk Institute, Woo Chang Kim, associate professor at KAIST Center for Wealth Management Technologies in Korea, John Mashey, consultant at Techviser, Andrew Yao, professor at Tsinghua University, and a report by John C. Bogle, founder of the Vanguard Group. An afternoon panel will be moderated by the Wall Street Journal, and include leaders from high profile consulting companies and investment firms such as E*Trade, Merrill Lynch, and Betterment.
On Thursday, April 27, a more technical event will focus on demonstrations and tutorials on financial technology products such as the WealthBalancer system, how to preserve privacy in big data, goal-based investing, and artificial intelligence.
On Friday, Princeton will host a conference on the impact of robo-advisors. Speakers include Larry Summers, former U.S. treasury secretary, and Michael Evans, president of Alibaba. Panelists at the event include John Schindler of the Federal Reserve, Chen Long of Alipay, Mike Angus of Mastercard, Robert Sams of Clearmatics, Lamar Wilson of Fluent, James Evans of Bain Capital, Dan Ciporin of Canaan Partners, and Ulrike Hoffman-Burchardi of Tudor.
Throughout his career, Mulvey has gone back and forth between developing financial technology techniques as a university researcher and implementing them in industry. He grew up in Chicago, where his father was a police officer and his mother was a homemaker. He earned a bachelor’s and a master’s at the University of Illinois, then completed a doctorate in management at UCLA. His academic work has focused on large-scale optimization models, and he has created risk management programs for American Express, Towers Perrin, Tillinghast, Pacific Mutual, and other companies.
But Mulvey’s vision of financial engineering doesn’t come only from the world of engineering. He is married to a doctor, and his daughter is also a physician, and he has seen many ideas in medicine that are also useful in finance.
“In medicine, you have to have evidence. You have experiments to see what types of drugs are effective, and what the downsides of those drugs are. On ads, you have a long list of problems that could occur. The side effects are understood, or at least attempted to be understood, and you track them for a long time even after a drug has been approved,” Mulvey says.
Mulvey says new tools in finance should similarly be studied and monitored, and their risks understood, before being widely used.
Another idea from the medical world is alignment of interest: Ideally, the doctor always acts in the best interests of the patient. However in finance, not every kind of financial advisor has a “fiduciary duty” to act in the best interest of a client. For example, the advisor might be getting a commission to sell certain financial products whether or not they are good for the client.
Mulvey believes robo-advisors should be more like doctors, and should serve the interests of the investor. “Alignment of interest doesn’t always work but it’s a goal to have,” Mulvey says.
“With robo-advisors as with any new technology, when it is first brought out there is often a great enthusiasm for it,” Mulvey says. He points out that subprime mortgages, junk bonds, and portfolio insurance were also enthusiastically embraced but proved to be problematic. In the case of subprime mortgages, they contributed to a catastrophic economic downturn with effects that are still being felt a decade later.
Mulvey says it would be wise to understand and appreciate the risks of new financial tools, just as new pharmaceutical treatments are tested for problems.
For the individual investor, robo-advisors could actually help with risk management and contingency planning, Mulvey says. Currently, many individual investors use investment vehicles like index funds, which are large cross-sections of stocks, so that no one company’s stock tanking will bring down the entire portfolio.
But Mulvey observes there are periods of “contagion” in the stock market, where entire classes of assets will fall at one time. “Many things that seem to be diversified have losses together,” he says. “Whether it’s interest rate risk, or currency risk, or stock market risk,” he says, “on average you’ll make money, but you’re periodically going to lose money, and more than you thought. Robo-advisors could potentially help you get more diversification than you would otherwise.” With robo-advisors, a small investor without much money to spread around between different funds could “diversify a portfolio against factors that are diverse themselves.”
Lionel Martellini, professor and director of the EDHEC Business School’s Risk Institute, has also studied robo-advisors, and also likes to draw medical analogies.
Martellini compares the financial market to a pharmacy stocked with medicines, and a financial advisor as an expert doctor who knows a patient’s needs and can advise them on which medicine to take, and how much. Currently, online trading companies like E-Trade are like pharmacies, with different investments being equivalent to drugs. To Martellini, robo-advisors won’t be much good if they are ultimately just mechanisms to buy into funds or buy particular financial product — they would just be “pharmacies.” A good robo-advisor would be like a doctor, who could create a strategy and then explain it to the client.
Will human wealth managers be put out of business by computers? Martelllini doesn’t think that will happen any time soon. So far there is only about $200 billion under management by robo-advisors, he says, although some estimates say the figure will rise to $1 trillion by 2020. (The size of the entire global financial market is estimated at $156 trillion.)
There is however a definite shift towards robo-advisors and away from traditional actively managed funds. The world’s largest fund company, BlackRock, recently announced it was consolidating many of its actively managed funds into algorithm and model-driven funds.
Martellini believes that millennials are likely to use robo-advisors rather than human financial advisors, but that as a group they do not yet have enough money to make a real impact.
“But eventually millennials will have the capacity to invest,” he says. “Pretty much everything else they are doing is digital, so there is a case to be made that this is definitely the future of investing.”
Currently, Martellini says, there are three types of companies engaged in building robo-advisors. First, there are companies like Charles Schwab that have done business as online brokers since their inception. “They don’t have to start from scratch. They have an existing client base,” Martellini says.
The second group is composed of digital wealth management companies like Betterment, which are built entirely around robo-advisors.
The third category is technology companies that have nothing to do with finance, but which have a large user base. The trend is taking off in Asia, as exemplified by Alibaba and Samsung, which are both offering wealth management programs. “It may sound a little sci-fi-ish to think about those guys deciding to up and enter the wealth management space,” Martellini says. But there is a real possibility of a company like Amazon or Facebook getting into the business.
The main advantage of robo-advisors is their low cost. “Even if they don’t do better than a human advisor, they will do it cheaper,” Martellini says. “In the investment management industry, a lot of people complain about a lack of transparency and fees that are extremely high. Even if you start with a good looking financial product, most of the performance is eaten up by fees.”
But the low fees are a double-edged sword. For example, if a robo-advisor only charges $50 a year in fees on a $10,000 account, it will be difficult to turn a profit. Savvy investors might also create a small account to get the financial advice, then duplicate its decisions using a low-cost online trading platform, essentially getting the advice for next to nothing.
To Martellini, robo-advisors make the most sense for companies like Vanguard or BlackRock, which have their own financial products that robo-advisors could recommend.
Robo-advisors aren’t just a hypothetical idea — they already exist. So, how are they doing so far? Martellini says that’s a harder question to answer than it appears. Most financial products can be assessed by performance, but the goal of a robo-advisor is not to maximize performance, it’s to meet the goals of individual clients. For example, a 50-year-old preparing for retirement in 10 years may have a particular income level in mind. A younger investor might want to make greater returns, accepting greater risk, over a longer period of time.
Martellini says currently, only Wealthfront offers data on how its robo-advisor performs on goal-based investing.
“We are merely looking at robo-advisor 1.0,” Martellini says. “We expect them to improve on this early stage, and what they should be doing is reporting in terms of probabilities of reaching goals.”
Another potential risk, although a remote one according to Martellini, is that tech companies that don’t have expertise in finance will create flawed robo-advisors because they don’t understand investment. “They need to have teams of experts in place to make sure whatever algorithm is put in place is not just something that looks good, but something that is fully consistent with state-of-the-art money management techniques.”
Martellini has also considered the issue of alignment of interest between the robo-advisor and advisee, but isn’t worried about it. Will robo-advisors serve the interests of their human clients, or will they follow some hidden agenda? He doesn’t see any pitfalls with robo-advisors that are not shared by their human counterparts. After all, another person is always a “black box” whose motivations are impossible to discern. “How do you know a financial advisor is not acting in the interest of maximizing fees from their clients?” he says.
Similarly, Martellini doesn’t see much to worry about when it comes to the possibility that robo-advisors will set off a cascading market crash such as the “flash crash” in 2010 that was caused by runaway trading algorithms. Because robo-advisors are geared towards buying and holding assets over long periods of time, they won’t create the kind of feedback loops that caused the flash crash.