Close Close

Technology > Artificial Intelligence

MIT's Lo Predicts the Future of Finance

Your article was successfully shared with the contacts you provided.

Financial regulators are “three or four steps behind” technological innovation — and that gets dangerous,” argues Andrew W. Lo, finance professor at MIT Sloan School of Management and director of the MIT Laboratory for Financial Engineering, in an interview with ThinkAdvisor.

The “biggest challenge” for regulators is trying to keep pace with the years-long pick-up in tech innovation, says Lo, who is on FINRA’s Economic Advisory Committee.

Financial regulation is just one area of his current research.

For another, he is a principal investigator at the MIT Computer Science and Artificial Intelligence Laboratory.

Time magazine named Lo one of “the 100 Most Influential People in the world” in 2012.

Co-founder of the tech-driven QLS Advisors, he is on the board of biopharmaceutical company Roivant Sciences.

His newest book is In Pursuit of the Perfect Portfolio, co-written by Stephen R. Foerster and published in August.

In the interview, Lo cites the necessity for “additional oversight” of robo-advisors because of their inherent risk, which the typical retail investor isn’t prepared to manage.

He also discusses new technologies, like cryptocurrencies, which are difficult to monitor and regulate, he says.

Lo’s AI research focuses on “artificial humanity,” which he calls “a second-generation AI.” 

“Until we have an algorithmic understanding of how humans would likely behave [in market downturns], you’re never going to have truly intelligent software,” he contends.

This involves incorporating into algorithms investors’ “freak-out factor,” as he terms it, and their “pain point.”

ThinkAdvisor recently interviewed Lo, speaking from his office at MIT. The conversation embraced what regulators are doing to try to avoid another devastating financial crisis and his research on systemic risk, which is about how the financial system’s subcomponents are interconnected and how that affects risk.

The professor is also studying the financial impact of impact investing and is at work on a framework for “discharging fiduciary duties” when shareholders, plan sponsors or plan participants are interested in making such investments.

Here are highlights of our interview.

THINKADVISOR: The SEC has found that many financial advisors haven’t been supporting the best interest regulation, Reg BI. Do advisors need more regulation?

ANDREW LO: Regulation has to be done very carefully because when you over-regulate markets, you create unintended consequences that can take years or decades to undo; in some cases, you can never undo them. So you have to proceed cautiously.

However, in the area of robo-advisors, we need to understand what the systemic implications are, particularly because you’re dealing with retail investors, and not all of them are prepared to fully understand and manage the kinds of risks that come with robo-advisors.

So I do think that additional oversight is necessary there. But that will require a fair amount of study before we can implement the proper regulations.

The SEC has also found broker-dealers who were seemingly ignoring Reg BI and even offering risky and more costly products. Your thoughts?

It’s a growing problem because financial markets are becoming much more complex, and financial services are now being offered by a number of new entities that may not look like your traditional financial institution. 

In fact, that’s the whole point of fintech — to be disruptive in providing services to individuals at lower cost in unconventional ways. That’s creating challenges for regulators to keep pace with.

Regulators need more resources to address these issues and deal with them. The budgeting process [needs] to reflect the new realities of market complexity.

Such as?

Cryptocurrencies are a case in point. Many of the players are not your traditional broker-dealer. These exist both onshore and offshore in ways that are very, very difficult to monitor and regulate.

The SEC and CFTC [Commodity Futures Trading Commission] are both looking at this very carefully, as are a number of the Feds [Federal Reserve banks]. 

This is a good example of where the technologies move so quickly that it’s not even clear what the right answer is — never mind that we don’t have the answers yet. 

What’s the biggest challenge for regulators?

The main issue is the pick-up in the pace of technological innovation.

Over the last 20 years, the speed of innovation in technology has greatly increased. It’s been going great guns. 

Historically, regulation has been one or two steps behind. Now they’re maybe three or four steps behind, and that gets to be dangerous.

Regulators are definitely making progress and have hired a number of people at the SEC and the CFTC that have the kind of experience to be able to address some of the technological advances.

But [keeping pace] is the biggest challenge right now.

What are regulators doing to try to prevent another financial crisis like the one in 2008?

They’re collecting a lot more data than they used to and are trying to keep track of it in a more systematic fashion. 

The Office of Financial Research is a new agency that was [established] precisely to focus on the data-collection problem and monitoring emerging risks to financial stability.

The Financial Stability Oversight Council is another [monitoring body] that’s meant to try to keep pace with the changes going on in financial markets.

So a combination of collecting more data, engaging in more analysis and being able to apply these ideas to oversight is where I think regulators have been going over the last 10 years.

Does that include FINRA?

Yes, all of the regulatory agencies including the Federal Reserve Bank and the regional Feds. They have upped their game because they had to, given how many changes were going on, particularly with respect to fintech.

You’re on FINRA’s Economic Advisory Committee. What are some of the main concerns?

They’re very much concerned with the exact same thing as the SEC and the CFTC, which is the impact of financial technology on fiduciary responsibility but also on financial stability, and then the emergence of new asset classes and what that means for the same issues.

The third topic that has been of great interest is cybersecurity. All three of those are related because they’re all consequences of the innovations in financial technology. 

It’s about trying to keep pace with all of the new ways of generating value for investors but also their not being taken advantage of by unscrupulous individuals.

What are your thoughts about a fiduciary standard for all financial advisors? 

We definitely need to have clearer guidelines on who is or isn’t a fiduciary and exactly what fiduciary standards really mean — the “prudent man rule” [making prudent decisions] needs to be updated to reflect the realities of current financial technology.

The notion of making prudent decisions has been somewhat vague, so providing more guidance on what that means would probably be helpful.

This is particularly important with new asset classes that are emerging because NFTs [nonfungible tokens] and digital assets really push the boundary of what a prudent investment might look like. 

Generally, there’s a belief that Republicans want less regulation of the financial markets and Democrats want more. Is that so?

That used to be how you’d characterize the differences. But it’s changed, and it’s changed in some strange ways. 

For example, there are situations now where the Democrats are the ones that are shying away from regulating certain markets, and it’s the Republicans that are pushing for it. 

So it’s hard to know exactly what the politics will bring given how complex the landscape is.

What are you working on now in the area of artificial intelligence?

To have truly intelligent algorithms, we need to understand human foibles and frailties. Until we have an algorithmic understanding of how humans would likely behave, you’re never going to have truly intelligent software. 

Artificial humanity is what we really need to focus on. We have to have an algorithmic understanding of how people actually behave rather than how we think they ought to behave.

We’re 80 to 90% there. What’s missing is the ability to take that research and encode it in algorithms. So we need to do more engineering — taking the science and putting it into practical use.

The hope is that we’ll get there in another 10 or 15 years — artificial humanity, not so much artificial intelligence. 

In the case of robo-advisors, if you’d like to have software that will help you navigate through difficult challenges like the financial crisis, you really need to understand, first of all, what the human tendencies are.

When we lose more than 10% or 15% [in the stock market] in a matter of days, we all freak out and tend to withdraw money from the market. 

It’s this “freak-out factor” that’s part of what we need to incorporate into our algorithms to anticipate when individuals are likely to freak out and to be able to apply AI to address those moments and take some risk off the table and put it back on again when the market turmoil subsides.

I’m proposing a second-generation AI that would say, “I know you’re nervous. I know you’re going to want to take money out. So I’m going to do that for you now.” 

But — and this is the most important thing — our statistical analysis shows that people lose more money not by pulling out of the stock market during these downturns.

When they really suffer losses is by not putting it back in after the market turmoil is over. They wait too long. They wait till the stock market goes up to such a point where the fear of missing out forces them back in. 

But by that time, they’ve already missed the recovery and then some.

So you need an algorithm that understands what your pain point is — how much you can withstand — to pull the money out before you freak out and to put it back in as the market starts to recover and as you’re more comfortable with that. 

We’re still probably five to 10 years away from a robo-advisor that can really engage with us on a personalized level. 

Since 2008, you’ve been conducting research on systemic risk. Please explain.

Systemic risk is a new way of thinking about old issues, like the kinds of risk that financial regulators try to guard against.

It focuses on how to think about the financial system, which is interconnected with many different subcomponents, and how the interconnections are related to certain kinds of market events. 

We’ve had a number of findings over several years and produced software that allows people to download and calculate these statistics and see how they perform. It’s a systemic-risk toolbox, available on the Office of Financial Research website as well as on my website: the Laboratory for Financial Engineering.

(Photo credit: Erica Ferrone)