RegFi Episode 45: Challenges and Opportunities in Banking Today: An Economist’s Perspective

Orrick RegFi Podcast | Challenges and Opportunities in Banking Today: An Economist's Perspective

RegFi Episode 45: Challenges and Opportunities in Banking Today: An Economist’s Perspective
34 min listen

NYU economics professor Larry White joins RegFi co-hosts Jerry Buckley and Caroline Stapleton to discuss a range of current issues facing the financial services industry. Professor White provides data on the significant holdings of commercial real estate loans in the portfolios of community and mid-size banks and discusses the near- and mid-term challenges these are likely to present. He then turns to a consideration of monetary policy and the Fed’s delicate balancing act in timing a pivot from a restrictive to less restrictive stance as it tries to wrestle the inflation rate down to two percent. Finally, the conversation examines the increasing role that AI and large language models will play in the operations of financial services firms and the potential for fraudsters and others who weaponize AI for malicious purposes to erode trust in the financial system.

Links:

Episode Transcript

Jerry Buckley:	Hello, this is Jerry Buckley, and I am here with RegFi co-host Caroline Stapleton. Our guest today is Larry White. Larry holds the Robert Kavesh Professorship in Economics at NYU’s Stern School of Business, where he chaired the economics department from 1990 to 1995. He is a close observer and commentator on financial services industry and for three years served as a member of the Federal Home Loan Bank Board. He has also served as a senior advisor in the President’s Council of Economic Advisors and as director of Economic Policy Office at the Antitrust Division of the U.S. Justice Department. Larry, you and I regularly get together at the Smoke Tree Ranch events organized by our mutual friend, Chuck Muckenfuss. At those events, you often wear a T-shirt with the legend, “Mark to Market,” which would indicate that you are not a fan of lend-and-extend solutions when a bank encounters portfolio problems. In our current lending environment, we find that a number of banks have commercial real estate loan portfolios that might be a cause for concern given the high vacancy rates that have prevailed in many urban markets in the post-pandemic period. Could you give us some sense of the magnitude of this issue and maybe drill down a bit because not all commercial real estate loans are the same. Warehouses may be doing fine while office buildings may experience stress in certain markets. So, Larry, give us your insights.
Larry White:	Aww, shucks. Well, thank you, Jerry, and thank you for this opportunity. I’m pleased to talk about this. And you started that discussion with the T-shirt, the “Mark to Market,” and that is my-I don’t know, what’s the right metaphor-North Star here. I’m an economist and thinking about things in terms of market value is always the place that I try to start. And unfortunately, the history of bank — especially safety and soundness — prudential regulation, the history is that all too often we’ve had difficulties because there was delayed recognition of losses. That certainly goes back to the savings and loan debacle of the 1980s that I had firsthand experience with at the Federal Home Loan Bank Board. I worry that we are approaching — we may not be there yet — but approaching a similar set of problems because a lot of commercial banks make a good living by lending on commercial real estate, as you pointed out. That’s not a uniform category. It goes everywhere from shopping center to commercial offices to warehouses to even multifamily housing is often considered to be commercial real estate. And different parts of the country are experiencing different — especially office problems. But it’s there and, again, the tradition of the way the whole accounting framework works is it’s historically based, it looks backward, it doesn’t even look current, let alone looking forward. And so let me drill down a little more deeply. And at the moment, what we have at the aggregate, the end of the first quarter of 2024 data for all banks, the aggregate commercial real estate loans were $1.8 trillion. The net worth for all banks, all 4,500-and-some-change banks, the net worth was $2.3 trillion. So, commercial loans in aggregate, smaller than equity. And as you know, as a first approximation, when we talk about capital in the bank system, bank financial system, we’re talking net worth. That’s overall okay. But for the smaller community banks, and there are over 4,000 of them, they have $266 billion in net worth, and they’re reasonably well capitalized if you just compare their net worth to their total assets. It’s a shade under 10%. That’s good. But their commercial real estate lending is twice the size of their net worth position. So, again, that’s an aggregate; in a sense, that’s an average. There are banks with more than twice the level of commercial real estate as compared with their net worth. And there were 67 banks with $1.8 trillion in assets overall, that had a ratio of commercial real estate lending to equity that was greater than 300%. So, that to me is quite worrisome. We’re gonna have to see how it plays out, but man, I sure hope the examiners and supervisors in the banking system, whichever regulatory agency they are located, they’re looking hard at those commercial real estate-heavy banks, because those really could be near- to medium-term problems.
Jerry:	And that’s interesting in terms of your timing. You’re saying near- to medium-term. So, meaning in the next one to four years, we will have to address those issues one way or the other?
Larry:	That’s the way it feels to me. Now, if as expected, there is a decrease in interest rates announced by the Federal Reserve in a few weeks, that will help things, but it’s not going to cure the empty office building, the empty storefront building problem that a number of banks clearly face.
Jerry:	Larry, I’ll just say, you know, being in a law firm — and this is true of many businesses-it’s very clear that the culture has changed. And it changed so dramatically since COVID that the use of office space is just not the same as it was. It’s nothing that could have been expected. It was not imprudent activity by anyone. It was just a cultural change. It’s like a hurricane.
Larry:	Right. You know, I don't think of people as being culpable in this. Although not recognizing the likely losses, I understand that's the culture, but that doesn't make it any easier to deal with these problems in a timely fashion.
Jerry:	Sure. I mean, keep that T shirt.
Larry:	I don’t think it’s going to go out of style. Unfortunately, I don’t think it’s going to go out of style anytime soon.
Jerry:	Sure. Now, you mentioned the Fed. It appears that the Fed is about to lower interest rates, which seems to indicate that the Open Market Committee believes that inflation is under control or getting under control. The monetary policy game is really tricky. Though slowly headed downward, inflation is still running at a rate of nearly 3% a year, meaning that in three years, the price of goods and services goes up about 10% on a compounded basis. On the other hand, if the inflation rate goes down to 2%, the three-year price rise is closer to 6%. So, the percentage point on an annual basis seems to make a pretty big difference. It’s one-third difference. And, once prices go up, they stay up. So, while moderating inflation is better than accelerating inflation, the consumer’s significant loss of buying power over the last three years is not reversible. And they are well aware of that. That’s part of the malaise in the electorate. If the inflation rate stays at 3% or moves higher, who are the winners and losers in this economy, Larry? And second — wow, that’s a tough question. And the second one is, what will the Fed do if after it lowers rates, inflation starts to pick up again?
Larry:	Ahh. All right. The second one, man, you’re now asking for crystal balls and that’s not my strong suit. Let me just say the complete answer to the inflation question, gee, do we have two and a half hours or three and a half hours in which we can talk about this? As you know, the background on this is the last time that the United States had a serious bout of inflation was in the late 1970s, early 1980s, and, at that time, the Federal Reserve initially was too accommodative to the whole process and it took a substantial amount of time, a substantial amount of effort — primarily by Paul Volcker — to reverse that. The current leadership of the Federal Reserve is very aware of that. The whole issue of expectations and what expectations individuals have in their consuming behavior, in their investing behavior, markets as aggregations of individuals. The Federal Reserve has been slow to reduce interest rates because they really want to feel confident that inflation is slowing and they don’t want to have to address the second part of your question there, “What do we do if after we’ve cut interest rates, we find that inflation is quickening again?” My best guess on that second is they would grit their teeth and raise interest rates. They would say, “We were fooled.” But it’s so important that we not allow a further inflaming — to use another metaphor — of the inflation fires that we snuff it out.
Jerry:	But you know, Larry, I do want to have you address — and I don’t — I’m interrupting you and I know, but I do want you to try to address the winners and losers because — go ahead.
Larry:	And that’s where — exactly where I was going. Okay. And now, first, it’s important to remember, as you pointed out, the inflation rate is measured basically by a bundle of consumable goods. It’s a consumption-based measure. It does not include assets as a general matter, to the extent that housing is consumed and the price of housing it enters a little bit, but you know the stock market going up, going down, that has no impact on that rate of inflation. So, who’s going to benefit? Well, somebody whose income goes up faster than the price of a consumable bundle goes up, those are the people who will benefit. Those whose incomes go up more slowly, they’re not going to do so well. Those who have assets whose value rises faster than that consumable bundle, they’re going to do well. Those whose assets don’t rise as much, they’re not going to do as well. So that’s the general pattern. And then you’d have to drill down a little more deeply. That is an area of macroeconomics. I’m mostly a micro and financial guy. So, you need to get somebody else on the podcast to give you that more detailed, more drilled down macro story.
Jerry:	Well, I’ll just say this. I think you can listen to the people who are calling very strongly for reduction in rates and just imagine that they will probably do better even if there is inflation. And among those are some people who are running for president.
Larry:	Let me just say one other thing. Both you and I can remember the late 1970s, early 1980s.
Jerry:	Sadly.
Larry:	Yeah. And double-digit interest rates and et cetera. But one thing I do remember, and this is the economist in me coming out, is at the time, a lot of puzzlement by friends, non-economists, smart people who just didn’t quite understand what this inflation stuff was about. They couldn’t get their heads around it. They knew prices were going up, but they couldn’t quite sort of put it all together. And over the last two or three years, I’ve seen a replay of that. People, the prices have gone up, but they don’t fully understand the process. They understand what a recession is. And they did back in the 70s, because there had been recessions in the 60s and in the 70s. But inflation, it’s just a more puzzling phenomenon. People have harder times getting their heads around what’s going on.
Jerry:	Right. Unfortunately, it does it does affect their pocketbooks.
Larry:	No question, there are real consequences.
Jerry:	Well, we’ll move from this highly esoteric discussion of rates and monetary policy to Caroline’s questions.
Caroline Stapleton:	Yeah, sure. And Larry, thanks for being here. I think Jerry and I are going to try to see how many different topics we can get your thoughts on during this podcast. So, get ready. No, I’d like to pick your brain about a topic that’s a recurring theme on our RegFi podcast, and that’s the impact of technology on the financial sector. I know it’s a hot topic, especially now with recent developments in artificial intelligence that many industries, including banking and financial services, want to figure out how to take advantage of, hopefully in a safe and sound manner, of course. So in a recent speech to the 2024 Conference on AI, sponsored by the FSOC, the Financial Stability Oversight Council, acting controller of the currency, Michael Hsu, gave a talk on “AI Tools, Weapons, and Accountability: A Financial Stability Perspective.” And I reviewed the transcript of his speech. I found it to be very insightful and also a rather digestible discussion of the opportunities and challenges that the ongoing evolution of AI presents and will continue to present for the banking sector. And that’s, I guess, sort of a long wind-up to my question for you, Larry, which is this. You’ve served on several occasions in major economic policy positions in Washington, and you’ve seen the financial services sector evolve over time. So, from your perspective, is the rapid advance of AI, large language models, generative AI, are these advancements something different in magnitude from the other technology advances that have impacted the financial sector in the past? Or do you think it’s something that’s kind of on repeat?
Larry:	Again, gee, do we have two and a half hours to talk about all of this?
Caroline:	I wish we did.
Larry:	First, I had not known about Acting Comptroller Hsu’s speech, and you had referred me to it. I’ve read it, and I agree. It’s an interesting presentation and his metaphors of AI as a tool but it can also be a weapon is, I think, an interesting and useful set of metaphors. As a tool, sure, look if the technology that surrounds financial services has been getting better, I don’t know, ever since the telegraph, okay? Just ever since the telegraph, we’ve had improvements in financial services, financial services markets. And the other thing I think to remember is, the AI that we’re talking about now is, what should I say, deeper, more elaborate but it’s important to remember every time we use a spell check program in a word processing context, that’s AI. That’s basically AI. And that’s an algorithm that is trying to take the particular word that we’ve misspelled and given the context of the words around it, suggest a correct word. That’s a that’s a simple form of AI but that’s an AI. Similarly, something we’ve been comfortable with, I guess now it goes back 50 years, credit scores, particularly thinking the Fair Isaac, the FICO score, which was a major, major improvement, I think improvement. Now, clearly, we’ve been able to make improvements on the improvement and elaborate and refine, but being able to use technology use algorithms to provide predictions as to credit worthiness of potential borrowers that would go beyond just a loan officer reviewing a paper file and looking a potential borrower in the eyes. You know, that was a major major change. I think we’ll see AI improving on that, but my creativity, my imagination may not be broad enough. I don’t see it as revolutionizing financial services. It’s another tool. Again, as Acting Comptroller Hsu pointed out, it’s another tool, but it is just another tool. What he also points out, however, is it can be a weapon, and in that sense, a weapon for the bad guys, for the guys who are trying to commit fraud, commit embezzlement, and there, the ability of AI, say, to foil cryptology or to cause a recreation of somebody’s voice and thereby allow a bad guy to access a financial account. That’s scary. That is scary. My guess is there will be a, I don’t know what the right metaphor here is, an arms race between the good guys using AI to try to foil the bad guys’ use of AI. And I don’t know where that, I’m not a good enough technologist to know where that comes out. But it’s clearly, going to be an important development on both the good side and on the bad side. And we’ve got to pay attention to it. And of course, regulators need to pay attention to it as well.
Caroline:	You know, I think it’s so interesting. I agree with you that the idea of using complex algorithms in financial services really isn’t so new. It’s just that something about the recent leaps, I think, in how AI has developed, particularly chat GPT and this voice generation, kind of the fraudster toolbox that you were just talking about, has made the regulators focus a lot of attention in guidance and speeches and more informal communications to industry about putting the brakes on implementation until you’ve made sure it complies with all of the various regulations and statutes that apply to the industry. And I sense a lot of hesitance among financial services, industry participants who are thinking like, “Well can update our existing technological capabilities, but will we be criticized for it if we go first?” I think this gets back to the overarching fear that may, in my view, be driving what the regulators are doing, which is this concern about, as the acting controller said in his speech, the loss of trust. And that if the loss of trust happens because it may not be so new from an academic standpoint, but for some reason we all perceive it as new and we perceive it as not trustworthy. Therefore, do we lose trust in the banking system? And I guess I put it to you to the extent you have any thoughts on it. Is this something that the FSOC should be thinking about, or Congress should be thinking about? Do you have any advice for setting up appropriate guardrails at kind of the highest levels of government?
Larry:	Whoa, whoa, whoa. For sure, FSOC.
Caroline:	If I put you in charge.
Larry:	If I were the philosopher, I’d never be philosopher king, but maybe a philosopher prince. For sure, the FSOC. I’m not sure the Congress, I don’t know what useful things the Congress can do. One of the things that struck me again, that speech by the Acting Comptroller probably ought to be read by everybody in the financial services area. It’s not a long speech. Written out, it’s like 13 pages or something like that, double spaced. You can read it in 10 or 15 minutes. It’s really worthwhile. One of the things that I think comes out here is the issue of accountability, that if a financial services provider is contracting with a technology company to provide a particular piece of technology, first, you know, making sure this is a reliable provider and regulators get involved in that, but then the accountability. And if something goes wrong, not only the financial services provider itself, but the supplier has to be held accountable. I think that’s just terrifically important. The other thing about … Man , I knew I wanted to make one more point here, and now it’s just skipped out of my head. Let’s talk about other things. It’ll come back into my head.
Caroline:	No problem. I think Jerry has another question for you on AI, so it may even trigger that lost thought.
Jerry:	Actually, my question relates to something you’ve sort of already alluded to. Which is, so AI holds both opportunities and risks for banks. But it may also hold ability to significantly enhance bank supervision, providing early warnings or deeper insights into the risks that regulated entities are taking, including the risk of the way that they may implement AI. In a way, one might speculate that AI might be used by regulators to police the use of AI by regulated entities.
Larry:	Yeah, absolutely that. To the extent that it does improve, call it the oversight programming and functioning beyond just simply looking at a snapshot of a balance sheet which isn’t mark to market is important. Let me come back. It popped back into my head. Again, the Comptroller’s speech, the issue of frictions and trust. Trust, as I think we all know, in an important sense is a lubricant for the financial services business. The more trust there is, the better, the more easily, the faster things can happen. And that’s a good thing. The fact that we still take way too long to clear checks is something that drives many of us crazy. Jerry and I have been, as Jerry mentioned at the very beginning, at the Smoke Tree Ranch sessions where Aaron Klein from the Brookings Institution has reminded us of just what an outrage it is that we haven’t been able to speed up the clearing of checks, which includes payroll checks, which means people don’t get access to earnings as fast. But with the problems, potential problems of AI being used on the part by fraudsters and others, that erodes trust, that creates more friction and slows things down. And that really worries me. I just hope the good guys employing the good AI are able to get the better of this. But again, I’m not a technologist and so it’s hard for me if this may just be a continuing battle that is going to go on, that our great- great-grandchildren will still be talking about 80 years from now but this trust, this friction issue? It’s really important.
Jerry:	You know, Larry, a thought occurs to me as you’re talking, and that is in the cyber realm. And I don’t know that it’s been that successful. There is cyber insurance, cyber risk insurance. I’m wondering whether we will see the evolution of AI risk insurance, which would, if properly, how it’s administered, what conditions there are would be an interesting question, but it might help to reduce the friction because there would be a greater reliance on third party validation. Just a thought.
Larry:	No. All right. Look, in some sense, you know, you would hope vendors offer guarantees, money back, you know, if my AI that I’m providing to you, my chat bot, in the example that the Acting Comptroller mentioned in his presentation, if it doesn’t work or there are problems with it, I’ll fix it or I’ll give you your money back. That, in a sense, is one form of insurance or you bring in a third party who is a more formal insurer who’s overseeing this process.
Jerry:	I think we recognize that the credit card industry has usually benefited by the $50 limit on loss and so some thinking along that line where consumers are protected from loss as a result of AI injury might be worth considering. All looking way into the future. But Caroline, unlike you and I, will probably be solving all those problems.
Larry:	Of course. Of course she will. We’re counting. No, sorry. Sorry. Our great-grandchildren are counting.
Jerry:	They’re counting on you, Caroline. It’s all in your hands.
Larry:	You’re right. It’s all on your shoulders. On your shoulders.
Caroline:	I feel the weight of the burden. Thank you.
Jerry:	Thanks so much, Larry. It’s been very much fun having this conversation with you.
Larry:	Well, thank you. And it’s given me the opportunity to think more about these things, to read the acting controller’s presentation. It’s all good. Thank you very much for this opportunity.