Fintech: Friend or Foe? A "ViewPoint" Webcast - December 18, 2018
In this "ViewPoint" webcast, representatives from the Atlanta Fed's Supervision, Regulation, and Credit division discuss the significant impact that fintech has had on banking.
Lali Shaffer: Hello, and welcome to the Federal Reserve Bank of Atlanta's "ViewPoint" webcast. I'm Lali Shaffer, your host, and today I'm joined by Cynthia Goodwin, interim head of our Supervision, Regulation, and Credit division, and by Scott Royster, senior examiner. Our topic today is an innovation called financial technology, also known as fintech. As a reminder, the views expressed here are of the presenters and may not reflect the views of the Federal Reserve Bank of Atlanta or System.
I titled today's segment, "Fintech: Friend or Foe?" and we use the definition of fintech from the Financial Stability Board, which is "technologically enabled financial innovation that could result in new business models, processes, or applications, with an associated material effect on financial markets and institutions and the provision of financial services." I will add to this that almost all of us are touched by financial innovation—that is, directly and indirectly. How? Because we are all consumers, in a digital age—more notably in the retail sector, but more importantly in the banking sector, which is why we are here today.
So welcome, Cynthia. I believe this is your first ViewPoint webcast as interim head of Supervision. Would you like to share any current issues that you're seeing in the banking sector?
Cynthia Goodwin: Yes, thank you, Lali. I'd like to highlight credit card growth at small commercial banks. While the exposure isn't great, it has been increasing. My concern, basically, is that the amount of delinquencies and charge-offs has been growing fairly significantly, and when you take that in totality, recognizing that household debt is also increasing, that gives you a bit of pause. You might know that prior to the crisis, household debt had been contracting for the few years prior to 2012. Well, now it's back over precrisis levels. So I think banks should be very attentive to their underwriting and ensure that it's prudent.
Shaffer: Great. So have fintechs made this easier, and where do they get their funding?
Goodwin: I think fintechs do have the opportunity to expand access to credit, so that is a positive—although it brings risks as well. Their funding comes from a variety of sources: private equity funds, venture capital, angel investors, family and friends, your own personal wealth. But there are also incubators, accelerators, and innovation hubs that provide funding and support.
Shaffer: Thank you. Cynthia. How has financial technology evolved, both as a technology and how we view it?
Goodwin: It's fairly interesting—a lot of these fintech firms took the opportunity during the crisis to really brand themselves and expand their footprint. A lot of that began in the peer-to-peer lending space. It was fairly simple at that point—basically matching a borrower with an investor, using the technology. But that has expanded because of the benefits that accrue to the user—not only is it convenient, it's also faster, and you can do it on your telephone. So I think that has caused it to proliferate—that and the fact that we have a lot more data and a lot more computing power, so not only are you seeing it in the lending space, you're seeing new technologies, such as machine learning and artificial intelligence, that have been employed in a number of use cases that really have brought a lot of benefit, not only in the credit space but also in the anti-money laundering space.
Shaffer: Okay. Scott, any thoughts to add on that?
Scott Royster: Sure. There are a number of technologies that we're keeping an eye on, as Cynthia mentioned. I wouldn't necessarily characterize it as a technology, but the proliferation of data is where we start, right? And there's a statistic that IBM Cloud felt compelled to mention at an investor conference: that in 2016, at the end of 2016, of all the data that was available digitally, 90 percent of it was generated in the previous two years. I can't even imagine what that number looks like today. So how do we interact with that data? How can we use that data and make sense of it?
Well, as Cynthia alluded to, we're using techniques, statistical techniques, that were known to us for decades—linear and nonlinear functions—but were not accessible to us until the computing power that has made such a leap in the past decade. But with that power—with great power comes great responsibility. Understanding that if the data contains biases, it will certainly have an impact on your outcomes, and so you need to be aware of that.
Interpretability—some of these functions are very, very complex, and the process is very complex. To understand what you're looking at and being able to explain it in certain situations, especially if there's a situation where you're decisioning credit—by law, you are to explain that. So interpretability is a big thing that we're looking at in the machine learning space.
Then there are some other drawbacks to machine learning, like not being able to predict black swan events. It's very good to know the weaknesses of machine learning, and if you're using data—and it's obviously historical, backward-looking data—we aren't at the place yet in machine learning where it can make the leap and actually predict those low-probability but very high-impact, novel events that haven't happened before.
Other technologies that we're looking at: distributed ledger technology. While that isn't widely used yet, it does have the potential to be a very disruptive technology, and it's something that we definitely are paying attention to. Other things could be like open banking, through application programming interfaces. Some of that is coming from market forces, where people want access to their data that their banks hold, and some of it is also being imported from Europe through their passing of the General Data Protection Regulation, or GDPR, that gives consumers power over their data.
Lastly, I would say that we are keeping an eye on something called quantum computing. While that is still in the near future, or distant future—five to ten years—I predict that you will hear about it with greater frequency in the future.
Shaffer: All right, thank you. So a basis of this is power—computing power, consumer power. Scott, since the Federal Reserve is not a chartering agency, what options currently exist for fintechs and banks to level that playing field regarding regulation, access to financial services, and also to experiment?
Royster: Well, whenever you embark on this discussion it goes to a matter of strategy. You can put fintechs into largely two buckets, if you're looking at it through that purview. One is a standalone entity competing with banks, versus a partnership or ownership or some sort of relationship with banks. Now, there are some exceptions that straddle that or may even be outside of that classification. But by and large, that's where the discussion goes.
And some arguments on both sides are that fintechs are largely viewed as regulatory arbitrage, versus, "Well, fintechs shouldn't have access to the national payment system, or shouldn't be included in the fractional banking system and have access to the discount window." We aren't going to settle that debate here today, but perhaps we can put a finer point on it, and what I'm referring to is, actually, if a fintech wanted to go on its own, what is available to it today?
I would first look at the industrial loan company, or industrial loan bank charter, offered by the FDIC [Federal Deposit Insurance Corporation]. It has been offered in the past, and they haven't recently granted a charter, but it would provide some advantages to a fintech—but it also comes at a cost. Those advantages are a multistate platform, a near-national platform, from which to build your business, access to the national payment system. However, it would open up and, as Cynthia spoke about earlier, a fintech that is owned by a venture capitalist firm would now, as the parent, become open to certain aspects of the Volcker rule—and that may not be in the long-term strategy of the venture cap, so they may steer away from that option.
The other two are more strategic, and disadvantages. The one is, who has applied for an industrial loan bank charter in the past, and who was interested—of course, I'm speaking about Walmart and Amazon. For a fintech to push for that, to be offered to everyone, could be inviting some rather large participants into the marketplace that they may not want to deal with.
Next, you can look at the OCC [Office of the Comptroller of the Currency] charter, fintech charter. There is some uncertainty with the fintech charter, as with the industrial loan bank, in that the fintech charter has not been given yet. So when that first one will come out, we don't know. But there is some uncertainty for a fintech firm there. What it does offer is for a fintech that is owned by a venture capital firm, they may avoid some of the Volcker implications. It is also a national platform from which to launch. Some of the other uncertainty in that charter is that they will not necessarily—from the Fed—be granted access to the national payment system, and if that is their business model, well, then that isn't a viable option for them.
The next I would look at would be the Conference for State Banking Supervisors, the CSBS, who has a target for 2020 to standardize their licensing and examination supervision for fintechs. That is in 2020, so it's not available to fintechs now. Whether or not that will be hit—right? We're talking about 50 different entities standardizing something that could be rather complicated, so there's some uncertainty there as well.
Lastly, there are sandboxes, which Cynthia mentioned, and even a no-action letter from the CFPB [Consumer Financial Protection Bureau], which has been recently issued to one fintech in 2017, that basically those solutions allow the fintech to operate under a little bit greater scrutiny and provide complete transparency to the regulatory body sponsoring the sandbox—but there are some limitations there.
Then we move to the other bucket, right? Partnering with a bank. There are many different positives why a fintech would go that route instead of the standalone, especially today, immediately. The first is access to the national payment system, a national platform, potentially, access to cheap credit, access to a customer base, access to data that they had never had previously.
And then there are two sort of less obvious. We talked about machine learning being used to really interact with the data in a precision that creates many different credit profiles when you're applying it to consumer lending or small business lending. That can create a portfolio that is a little heterogeneous, not homogeneous, and that lends itself to a balance sheet lender and less so to capital markets—at least for the time being. And while we're on the subject of capital markets—the second sort of not-as-obvious concern is stable access to capital. After all, in the 2008 crisis it was the capital markets that really froze up, and if that's your only source of funding—well, you probably want to team up with something that has been tested through a recession.
So regardless of those two buckets, there is something that is impacting all of the banking industry right now, and that is what is approaching a self-regulatory sort of approach—and there are many fintech associations that are pushing for certain regulations that are beneficial to their cause, right? And so I'm talking about a bill that was passed in California that shares characteristics of the general data protection regulation that I mentioned earlier, as well as things like the Small Business Borrowers' Bill of Rights. So there are a lot of things working, a lot of moving parts to find that level playing field that you talked about, and we're keeping an eye on it. It's a full-time job.
Shaffer: Yes, definitely. So we talked about many different options that the fintechs have with chartering and regulation. Cynthia, what types of lending are we actually seeing from these fintech firms?
Goodwin: It runs the gamut. As mentioned, it mostly started out in consumer lending, but a lot of these firms do small business lending, mortgage lending, student loans—so just about anything you could think about. But I want to mention just one thing about GDPR. Within Dodd-Frank—I believe it's Section 1033—we also assert that consumers own their data. The CFPB is accountable for that. They have issued some principles, so our guidance around it isn't as prescriptive as GDPR, but we at least have something on the books to leverage. So that's also something that could also be in play.
And besides just lending—and Scott talked about some of the other technologies—you also have the data aggregators, who might help to advance DFA [Dodd-Frank Act section] 1033 because by virtue of the fact that consumers will give them access to their information, they can tap into their accounts and see exactly how they manage their money and what their flow of funds are and underwrite them as well for credit using those kinds of sources.
Shaffer: Okay. So we spoke about fintechs, we spoke about the consumers. Cynthia, how is the typical bank engaging these firms or utilizing the technology, and what are some of those risks?
Goodwin: For the most part a buy, build, or partner strategy, and that differs based on the size of the firm. And this is not exclusive. The larger firms tend to buy or build their own proprietary systems or partner with these firms. For the smaller commercial banks, they tend to buy their services from a significant service provider, some of the larger ones—like First Data and FIS, Amazon Web Services—who will deliver that type of technology and capability to them without them having to have their own expertise in-house.
That service delivery model, though, does not…the firm still owns the risk, so even if you go that route you need to fully appreciate what it is you're getting into, have very effective due diligence, very rigorous due diligence, and put in place all of the right risk management and controls in order to execute on that, because there is no relief from regulatory compliance.
Shaffer: That's right. And speaking of regulatory: Cynthia, what is the Federal Reserve Bank of Atlanta, and Supervision, doing to stay ahead of this rapidly changing topic?
Goodwin: We are certainly trying to stay abreast of not only what our firms are doing, but what kinds of technologies are being utilized and the risks and benefits inherent therein. On December 3, FinCEN and the federal banking agencies had a press release encouraging innovation, particularly in the AML [anti-money laundering] space. So again, we're starting to move forward and trying to encourage this type of innovation because it does bring benefits. Not only can it increase the ability to comply, it also has some cost-effectiveness capabilities. So we do want to encourage it but to go into it with your eyes wide open and with the appropriate guard rails in place.
Shaffer: Thank you. Scott, what are you seeing that the Federal Reserve is doing locally or on a system level?
Royster: As Cynthia alluded to, we are actively engaging our supervised institutions. We want to have an open discussion with them on what they see, how they're evolving in this ever-changing landscape that fintechs are bringing. We want to know what they're developing internally, what they're not, what they've chosen not to develop and embark on, and to understand why.
Additionally, we are engaging with local fintech firms and national fintech firms through our partners in the System, as well as conducting meetings here. It's been very beneficial, and that's something that we definitely want to keep: an open dialogue with our supervised institutions, to share with them what we've learned and what we see.
Shaffer: Great, thanks.
Goodwin: Two other areas in the Reserve Bank, too, that have their eye on this: the Retail Payments Risk Forum—their Take On Payments is a blog that would be very insightful. And also in our research department, the Center for Financial Innovation and Stability has done work in this area as well.
Shaffer: That's right—lots of work being done on this. Cynthia, do you want to speak to the future about any enhancements that might be coming through?
Goodwin: Well, we are looking at our supervisory processes, looking at what type of guidance and regulations we have, whether or not that's impeding innovation—and whether or not we should maybe re-specify some of that. I don't want to go into the specifics of those regs, but yes, we're very attentive to that. We're also trying to isolate which risks a lot of this innovation presents, not only in the consumer protection and fair lending space, but model risk management, data protection, privacy. So yes, just trying to make sure that we understand the technologies, the use cases, and ensure that our guidance and regulations don't impede but draw out the benefits.
Shaffer: Great. For either of you, and you can both answer if you want to: are there any other points that you want our banks to know?
Royster: Any sort of contracts that our banks enter into, or agreements that our banks enter into, with fintech firms, we want to make sure that our banks fully understand those contracts—and not only those contracts, but any relationships that the fintech firms may have with other firms. And if the fintech firm is given any, I guess, purview over bringing on risk to the bank, whether it is entering into credit agreements with consumer lending or small businesses, or anything to expand the footprint of the bank, that our banks fully understand those contracts as well and those implications.
Schaffer: Okay. Well, I want to say "thank you" to Cynthia and Scott for joining me today, and thank you to the audience. As usual, you can find more information than what was discussed here today on the Federal Reserve Bank of Atlanta's website, such as articles and other ViewPoint webcasts.