George Selgin on the New Deal and the History of Banking Crises

George Selgin was a senior fellow and director emeritus at the Center for Monetary and Financial Alternatives at the Cato Institute and professor emeritus of economics at the University of Georgia. George returns to the show to discuss the New Deal’s record on economic recovery, the Fed’s balance sheet and operating system, the Main Street Lending Program, and the history of banking crises, and much more.

Subscribe to David's Substack: Macroeconomic Policy Nexus

Read the full episode transcript:

This episode was recorded on July 28th, 2026

Note: While transcripts are lightly edited, they are not rigorously proofed for accuracy. If you notice an error, please reach out to [email protected]. 

David Beckworth: Welcome to Macro Musings, where each week we pull back the curtain and take a closer look at the most important macroeconomic issues of the past, present, and future. I am your host, David Beckworth, a senior research fellow with the Mercatus Center at George Mason University, and I’m glad you decided to join us. 

Our guest today is George Selgin. George is a senior fellow and director emeritus of the Center for Monetary and Financial Alternatives at the Cato Institute. George is also returning as a guest to the show, so please check out his past episodes. George, welcome back to the program.

George Selgin: Thank you, David. I’m glad to be back.

Beckworth: It’s great to have you on. Now, today, we’re going to cover a number of topics with you. We’re going to catch up on the Fed’s balance sheet. You’re one of the earliest guests discussing this. You, in fact, introduced me to Bill Nelson, who’s been on the show many times since then, but you’re the reason I got on this topic, and some folks know I spent a lot of time discussing the Fed’s balance sheet, and we’ll talk about that, Kevin Warsh’s task force on it, and your thoughts on that area.

Also, George, we’re going to talk about Main Street Lending facility that was used during the pandemic, one of the facilities the Fed set up, and your assessment of it. You gave an assessment during the actual run of it, but now you’ve looked back and have some thoughts. We’ll share that. 

Then finally, George, I know you have some desire to address some fallacies and myths that really have you bothered, worked up, so we’re going to give you a chance to do that as well. 

George’s Book on the New Deal

Before we do that, George, and get into all those great conversations, you had a book out titled False Dawn: The New Deal and the Promise of Recovery, 1933–1947. We did a previous podcast on it, but now it’s been out for a while. It’s been received. It’s been discussed. Can you give us an update of where it stands?

Selgin: Well, the one thing I know is that my publishers, University of Chicago, they’re happy. They say the book is selling well. I actually don’t have any statistics or anything like that, so I can’t say how well. It was well reviewed, though the reviews have petered out recently, though there is a recent one from the Journal of Economic Literature by Joshua Hausman. That was the latest. In general, I think the reception’s been good from my perspective as well.

Beckworth: I hope this book is widely read. I understand Ben Bernanke is working on a new book on the Great Depression as well. David Wessel was on the podcast recently, and he said Ben Bernanke’s working on the role that borrowers played during the early stage of the Great Depression. Maybe he will reference your book and cite it. Ben Bernanke, if you’re listening, make sure to get a copy of George Selgin’s book as well.

Selgin: Thank you. I will say that I think the borrowing aspect—of course, Bernanke, some years ago, wrote about the debt deflation angle of the Great Depression. I talk about that a lot in my book. I think the most successful New Deal recovery program was the one that involved the refinancing of residential mortgages. I put that very much on the plus side of the New Deal ledger as far as what it did that was helpful, as opposed to some programs that weren’t.

Fed’s Balance Sheet Policy Task Force

Beckworth: All right. Well, let’s move on to some more recent developments that have been happening. I mentioned at the top of the show, George, that you’re the one that introduced me to these issues surrounding the Fed’s balance sheet, its operating system. You had a book, was it 2014, that was called Floored? Was that the publication date or sometime around that?

Selgin: It’s called Floored! with an exclamation mark, as in the economy getting floored. I don’t know if it was 2014. I think it was a couple of years later. I got to Cato in 2014 and wrote the book early on while I was there. I think it was the first full-length critical analysis of the ample reserve system. I do consider myself an early critic of that arrangement.

Beckworth: Yes. You were the avant-garde, leading the way. As you stepped to the side in your semi-retirement, George, Bill Nelson and I, maybe Paul Kupiec at AEI, we all carried the torch. Now, it’s mainstream conversation. Right now, people are talking about it. Even champions of ample reserves are talking about how can we reduce the size of the balance sheet. Of course, the first way of doing that is to reduce the structural demand for reserves by banks. There’s been a lot of conversations surrounding this.

In particular, Chair Kevin Warsh has introduced this task force, one of them being a task force on the Fed’s balance sheet. There’s several things. He’s going to look at the composition of its assets, also the size, as well as the operating system itself. In recent testimony before Congress he said that there are several alternatives that are “sustainable equilibriums that one could use.” He also said, however, that you can’t go back to pre-2008. I think what he means by that is not that you can’t go back to scarce reserves, but you can’t go back to the asymmetric type of corridor system that existed back then.

In fact, interestingly, George, I had Don Kohn on the podcast recently. We were talking about that period because Don Kohn was instrumental in bringing about the legislation, at least shifting the conversation about having interest on reserves. One of the big takeaways from our conversation was that you could argue that demand for reserves was inordinately low, artificially low before 2008, because we didn’t pay interest on reserves, because we had all these crazy rules that had there been a standard corridor framework and a scarce reserve system, probably there would have been more demand for reserves. Yes, we can’t go back to that if we value the symmetric setup that we have now.

Selgin: There’s nothing sacrosanct about the pre-2008 corridors, so-called corridor system, which, as you said, was not an orthodox corridor, it was asymmetric, it had zero interest on reserves. There’s no reason why the question of a corridor versus a floor should be confused with that of interest on reserves versus no interest on reserves. You can have interest on reserves and have a corridor, and not just with a tiered system either. You can have interest on reserves with an orthodox corridor, and some countries had that.

Only, of course, the interest rate you pay on reserves in that case has to be an interest rate that’s somewhat below the policy rate. It can’t be at, let alone above, your policy target. When Don Kohn and others got the 2006 legislation through that originally authorized interest on reserves, their intent was to create a corridor system with interest on reserves that would be more symmetrical than the old one. They were not trying to create legislation that would usher in an ample reserve system.

Indeed, if you read the legislation, what it says—the 2006 legislation, which was not changed in 2008, though it was implemented more rapidly—you’ll see there that it says the interest rate on reserves cannot be, has to be at or below, the going equivalent market rate for short-term risk-free funds. The Fed has been violating that requirement, that legal requirement, at least the strict letter of it, since implementing interest on reserves in October 2008, and I complained about that in the Wall Street Journal back then. What they are operating right now is a system that, in my opinion, is against the law, apart from its other drawbacks.

Beckworth: Interesting. Okay. What do you think about the new task force that Chair Kevin Warsh has set up?

Selgin: Well, I’m all for it, first of all. I don’t see what harm can come from having people think seriously about whether this system should be maintained or not, and what other alternatives might replace it, and how to get from it to one of those. These are all extremely important questions. I am delighted that a Federal Reserve chair is not taking the superiority of the floor system for granted, and he shouldn’t. Because if you go back and look at the history of this setup, first of all, let’s be clear, it was a mistake to introduce interest on reserves in the first place.

It was done to prop interest rates up in October 2007 at a time when we now know, in retrospect, the Fed should have been thinking about how to get the darn thing down, how to get interest rates down. They should have been thinking about whether they could have negative interest rates, not how to keep interest rates from falling to zero. That’s the plain truth.

I was just reading a paper recently about supposedly a critical assessment of the floor system, and the authors say in that paper, just a passing remark, “It was also arguably necessary, in the heat and wake of the Great Financial Crisis, given the urgency of the moment and the tools available at the time,” meaning the interest on reserves. No, it wasn’t. It was a mistake, a clear mistake.

The other thing is that all of the promises that were made about the advantages of the floor system, every last one of them has turned out to be something that wasn’t delivered on. It was supposed to be possible to make the thing work with, originally, a couple hundred billion dollars of reserves. Well, that wasn’t true. It was supposed to simplify the implementation of monetary policy and keep interest rates reliably under control.

It wasn’t able to do that, certainly not without a number of supplemental contraptions that the Fed has had to introduce, including overnight repurchase agreements, the sub-floor under the interest rate floor, the standing liquidity facility, and so on. It’s become a Rube Goldberg contraption. It was supposed to save effort and resources. We should be able, therefore, to look at the staff at the New York Fed and see how many people they’ve managed to lay off. I think the answer is that they’ve got more people working on monetary policy implementation than ever. How come?

It seems to have had some crowding-out effect on bank lending, according to several studies now, which is something I worried about in my book. I must say my arguments weren’t very good, but I’ve been vindicated by people who knew better how to check this and so on. The other thing they said was, when we pump in a lot of extra reserves, we’ll just take them out afterwards, most of them. Now they know that they don’t know how to do that. I could go on.

This thing needs to be critically studied very badly, and I’m very glad that the new Fed chair has decided that they’re going to do that.

Beckworth: Now, just to be clear, George, you have articulated elsewhere that you’re not opposed to having a temporary floor system or temporary ample reserves in the context of a sharp recession when you hit the zero lower bound, but the key is to reverse it completely, to go back and return to some scarce reserve system. You’re more in line with a state-contingent operating system. Is that fair?

Selgin: Yes and no, David, because it’s very simple. If you have a corridor system and you run into the zero lower bound, automatically, it turns into a floor system at that point. Automatically, at that point, the interest reserves cease to have an opportunity cost. You don’t have to actually consciously have a state contingent thing where you say, “Oh, we’re switching to a floor system now.” You switch to a floor system when interest rates hit the lower bound. Then you have no choice, and then you can inject reserves with all of the usual consequences that doing so would have in a floor arrangement. That is, you can engage in quantitative easing. 

With a corridor system, you can’t normally engage in quantitative easing when you’re not constrained by the zero lower bound. Guess what? That’s when you shouldn’t do it. The only reason to engage in quantitative easing is at a binding zero lower bound, but a corridor system automatically allows that. You don’t need to have a system that changes from a corridor system that’s consciously changed. You just have to have a corridor system, and you’re done.

Beckworth: Okay. A lot of the conversation now, as I mentioned earlier, is about shrinking the demand for reserves as one way to shrink the balance sheet. Even advocates of ample reserves say, okay, there’s still more progress we could make. They may still want to preserve the spreads where interest on reserves is at the bottom or equal to the policy rate, but at least they can agree to shrink the balance sheet. I think there’s consensus whether you support ample reserves, scarce reserves, or tiered reserves. Whatever system you support, I think there is this growing consensus that there’s more we could do to reduce the structural demand for reserves. 

One of the possibilities in doing that is to integrate liquidity regulations into the discount window facility. That means, as you know, George, simply count the collateral. You’ve got park there toward your liquidity coverage ratio or your internal liquidity stress tests. Other proposals, make the discount window business as usual, so bring term options facilities. Do everything you can to make the discount window a place where you get liquidity on a regular basis, so shrink the stigma. That would reduce the structural demand for reserves. On many fronts, it’d be a way to fight this battle, but there’s other suggestions as well. 

Let’s stick with this one because I’ve championed this as well. As someone at the Mercatus Center, we love markets and all that good stuff. I’ve gotten push back. David, how can you support this? Because you’re encouraging moral hazard. That stigma plays an important role. The way I’ve responded is, well, look, the flip side of that is we have a bloated Fed balance sheet. We’ve lost the overnight interbank market, so we’ve lost price discovery. We’re missing a market right now. 

The Fed also has a large footprint in Treasury markets and repo markets. We also have the challenge outlined by Raghu Rajan that as the balance sheet grows, there’s this ratchet effect, this lock-in. He argues that it can actually reduce aggregate liquidity if banks begin to assume that it’s there, and so you have to continue to grow it. 

There’s all these other distortions that I see that if we could reduce those, and yes, if we get a little bit more moral hazard, and I’m not even willing to concede that. For the sake of argument, let’s say we do get more moral hazard along the way. I think that’s a reasonable tradeoff. How would you respond to someone who says, how could a market-loving person want to make the discount window more accessible?

Selgin: Well, my view here is a little bit in between, David. What I would say is that I understand the arguments you’re making for trying to make the discount window more effective, a substitute for ample reserves or for trillions of dollars of ample reserves. I get that, but as you alluded to, the fact is that this solution is a very poor substitute for reviving the interbank lending market, which ought to be, which used to serve as the lending source of first resort for banks that needed a little extra liquidity at the last minute and did not involve any moral hazard problem.

While we might make the discount window work better than it has or be taken advantage of more often and reduce demand for reserves that way, I think that doing that, rather than taking steps that could help revive unsecured interbank lending, is a second-best solution. I think we really should think about alternative arrangements, corridors, and tiered systems that would ultimately serve to encourage banks to once again start lending and borrowing from each other for their short-term liquidity needs. That’s my answer.

Beckworth: Well, George, this is your lucky day because guess what? You can have your cake and eat it, too. You can have both. Many of the proposals would say, “Yes, let’s make the ceiling facilities more accessible in terms of the long-term growth.” How do we grow liquidity over time? How do we grow reserves as the economy grows? Well, let’s rely more on the discount window or the standing repo operations to be the place to get it long term. Short term, you want to also resurrect that interbank market so banks can reallocate reserves among themselves depending on needs, short-term perspective, but over the long term.

This actually gets me into a related point I want to raise because I know you’ve thought about these issues much more than I have. Part of my other pushback to this critique is, one way or the other, the Fed’s going to be injecting liquidity into the economy as the economy grows. Given we have a fiat monetary system, given we have a central bank, it’s going to be the institution that injects liquidity or reserves into the economy as the economy grows. Some people seem okay with him doing it through open market operations as if that’s neutral versus doing it through the discount window.

To me, both of them imply some sense, some government intervention. In fact, originally, the Fed was set up to work through the discount window. In your mind, is there anything purer or market-friendly when you compare open market operations versus a discount window injection of liquidity?

Selgin: I think that the argument that open market operations are more reliable or a better way to inject liquidity into the system itself depends on the reality so far that discount operations carry a stigma. For that reason alone, they’re not as good a substitute. Open market operations, of course, have their own problems, one of which is that they limit severely, so far, the assets that can be traded for reserves. They also limit the direct participation in the operations or have done so to so-called primary dealers.

Of course, I’ve written a paper you know about where I proposed a more diverse open market framework where you have these multi-asset auctions where basically anybody can participate in open market operations who has assets, where you have competitive bids, et cetera, where the assets aren’t limited to US government securities. That’s a complicated story there that we can’t go into in detail. I think that an auction mechanism can be broadened so that it gives you the best of both worlds. It gives you participation directly.

Any bank that’s short of liquidity can go into this auction with the Fed, where the idea is that the Fed isn’t going to change the total liquidity. It’s adding to the system just based on the individual bank’s involvement, but this auction mechanism is going to allow banks to all bid for their pieces of what the Fed is injecting in a way that should result in an efficient allocation of those bits of liquidity to where they’re best provided.

Beckworth: What I’m hearing you say is, in principle, there’s no reason to favor one over the other. It all depends on the institutional setup, how things work.

Selgin: That’s right.

Beckworth: For listeners and watchers of the video, we’ll provide a link to George’s proposal. It’s called, if I recall correctly, “Reforming Last-Resort Lending: The Flexible Open-Market Alternative.” 

Selgin: Yes, “Flexible Open-Market Alternative.” It’s only about three-quarters baked, in my opinion. Maybe somewhere between half and three-quarters. I wish somebody would pick up on it and think it through the rest of the way. 

Beckworth: It might be a useful part of the conversation as we think about the Fed’s balance sheet and the operating system. I’m hopeful, too, that at least we’ll have a conversation. Sometimes, it’s useful to have this external review. Even if you know what people think ahead of time, at least they can sit down and have discussions. On the balance sheet panel itself, we’ve got Raghu Rajan, where we know he comes down strongly against QE and a large balance sheet. He thinks it creates this fragility, aggregate liquidity, and a ratchet effect.

Jeremy Stein, on the other side, at least in the past, he’s argued for ample reserves or at least abundant liquidity the Fed can provide. Karen Dinan, I’m not sure where she comes down. She might be the moderator between the two. At least we’re getting two perspectives. We’re going to sit down. They’re going to have a conversation. There’ll be others involved in the conversation as well. Maybe they’ll cite your work, Bill Nelson’s work. It’s going to be, I think, a very productive conversation. I think it’s important. I am hopeful for this.

At least let’s bring the conversation again to the forefront so we’re thinking through these issues. Some of us have been. There’s been thoughtful people within the Fed. Lorie Logan’s been thinking about this, Stephen Miran, Darrell Duffie, Bill Nelson, you, others. This conversation has been there, but I think having a broader conversation—look, there are many arguments one way or the other, but at the end of the day, I think one of the key, most compelling reasons to have this conversation is the Fed’s independence.

The balance sheet can become highly politicized. Even if the economics and the motivations are wrong or superficial, at the end of the day, no matter how people get there, they’ll look at the big balance sheet for any number of reasons and politicize it, whether from the right or the left.

Selgin: Absolutely. This was a whole theme of my little book before the Great Depression book, The Menace of Fiscal QE. This ample reserve regime very much lends itself to political abuse of the Fed’s balance sheet. I hope that danger will be part of what informs these discussions. I’m very upbeat based on what you and others are saying. 

We’ve come a long way from our little corridor club, which your listeners may not know about this, but way back when I was in DC in around 2017 or ’18, we had a little club. It was an internet club called the Corridor Club, which consisted of all the people on the planet who thought the floor system was a bad idea. Eventually, I think we got up to about 12, David, something like that. The point is now a lot more people are raising questions, if not outright criticizing the floor system. In my opinion, that’s progress.

Beckworth: To be clear, I want to manage our expectations here. I think people are very open to shrinking the Fed’s balance sheet, reducing the demand for reserves. I wouldn’t go as far to say everyone is willing and ready to abandon ample reserves. You can imagine a world where the Fed’s balance sheet is reduced more, but we still have something like an ample reserves in terms of where interest rates are placed, the spreads between them. I will take that win. I will take that win because I think it’s progress. It also would, I think, enhance the Fed’s independence. Again, just to go back to what I started off with, we’re never going to go back to the system pre-2008. That’s another way to manage our expectations. We want to go back to a robust operating system. 

The Fed’s Main Street Lending Program

Well, let’s transition, George, into another operational facility discussion. That surrounds an innovation the Fed brought onto the scene during the COVID pandemic, the inflationary period as well. That was the Main Street Lending facility. You wrote about it back then. I know we talked about it on the podcast back then. Recently, you had a nice thread of it on Twitter, now known as X. We’ll provide a link to that. Walk us through what was the Main Street Lending facility, and what is your assessment of it now, looking back since time has passed?

Selgin: Well, this is one of those I-told-you-so stories. Normally, as I think your listeners will know, the Fed doesn’t lend to ordinary businesses. Its lending is confined to financial institutions, usually just banks. Occasionally, under an exception known as 13(3), it has lent to other financial institutions. What’s very rare is for the Fed to lend to nonfinancial businesses. Even there, there have been precedents, one of which was in the 1930s. We’re under a separate authority that no longer exists for good reasons. This was Article 13b., not 13(3). The Fed was, for a while, allowed to lend to ordinary businesses. It was a Main Street Lending Program.

Another Main Street Lending Program under 13(3) was created during COVID. Now, what I did was to look back at that old Main Street Lending Program to see how that worked out. The argument then, as during COVID, was, well, you know these commercial banks are not making loans to all the worthy businesses that need help, so the Fed’s got to come in there and do it for them, take the place, lend to the sound businesses the banks refuse to lend to.

Back in the 1930s, it turns out, not surprisingly, in my opinion, that the banks weren’t overlooking good lending prospects, that the Fed couldn’t find very many worthy businesses to lend to. The program, therefore, ended up being much smaller than anyone had hoped, and yet that 1930s Main Street Lending Program ended up enduring loan losses equal to about 3% of the advances, which was more than three times what any ordinary banking system can tolerate.

It was generally concluded, at the time by people who studied that round of Main Street lending, that it was a flop, it was a bad idea, and there you had it. I wrote about this in the context of the new proposed Main Street Lending facility and basically said, “Good luck avoiding the outcome that you had last time because there was nothing to suggest that they had come up with ways to avoid the same problems that had been faced by their 1930s predecessors.”

Well, lo and behold, at first, uptake on the Main Street Loan facility was very limited because the Fed did try to stick to avoiding excessively risky loans to unworthy borrowers. But the outcome was so disappointing, such a tiny fraction of the amount of lending the facility was expected to do, that, sure enough, they started to lower their standards. They started to make the terms more appealing, the length of time for repayment and other aspects of the borrowing. They got more uptake, still disappointing compared to their original expectations.

They got a lot more uptake, but they did it by taking on a lot more risk. Well, because a lot of those loans didn’t start coming due, principal or interest, until some years after the loans were actually made, we’ve had to wait a while to get some grasp of just how much money this thing has been losing. We have had quarterly reports for some time, and the loans have come due, and the principal and the interest are now due. So far, as of last June, the actual losses of the facility have amounted not to 3%, which was the horribly high figure of the ’30s, but to over 12% of the advances, a 12% loss rate so far, and there will be more losses. 

The Fed has already provided for future losses, has provisions for future losses, but I expect the actual losses in the future to be greater than the provisions for them so far. We’ll probably end up with a facility that has lost 15% of its advances, an absolute disaster. What’s the moral? The moral is ordinary banks do a good job of lending to needful businesses that are worthy during crises, and there is no reason for the Fed to stick its mitts in. It’ll only end up wasting a lot of resources by doing so, and we now have two experiments, costly experiments, showing that that’s the case.

Beckworth: What would you recommend Congress do or create as an alternative for the next big crisis?

Selgin: Well, first of all, if they want to help businesses, they should keep the Fed out of it because we’re talking ultimately about giving a lot of money away, right? That’s fiscal policy. Let Congress take responsibility for bailing out individual nonfinancial businesses. That’s something that Congress has every right to try to do, subject to democratic approval. It still might not be the wisest idea. It still could squander public resources, but let’s keep the Federal Reserve out of it.

The Fed’s responsibility is monetary policy, and if it lends to financial institutions, it’s only in order to preserve monetary stability that could otherwise interfere with the proper conduct of monetary policy. The Fed isn’t supposed to rescue banks just to help the bankers. It’s supposed to do that to stabilize the financial system, and there’s no argument that it should bail out or lend to ordinary businesses because it has to do so to conduct monetary policy.

Myths and Fallacies

Beckworth: All right. Well, let’s move on to some of the myths and fallacies that you want to address. I want to bring one up that wasn’t on the list, George, that you made, but it’s very much in line with some of them. It ties into a previous guest we had on the show recently. I know you listened to the show. It was Stephan Luck. He’s from the New York Fed, and he has a paper with Emil Verner and Sergio Correia. What they did is they go back, this amazing dataset, used large language models to look and look at the context of each bank panic, and they drew upon 300 million articles. Mind-blowing. 

As I told Stephan Luck at the time, I’m like, “This is fantastic, but it also means the expectations of what counts as good research has just gone up many, many levels.” One of the things they show is, at the end of the day, for commercial banks, what really matters, whether a bank closes or not, is the fundamentals. It’s how well is it funded in terms of capital, and is it solvent or not. It has less to do with bank runs by themselves, which then raises questions about the appropriateness or the relevance of the Diamond–Dybvig model. I know you’ve been someone who has brought this point up in the past, but I want to hear from you. How does this research resonate with your own thinking about the Diamond-Dybvig model?

Selgin: Oh, well, it resonates plenty, David. I love this research because I love all research that says that I was right. You see, the thing is that before these guys did this wonderful research, truly very, very good, painstaking research, there were other economic historians that had done less exacting but still very serious research on the causes of bank failures and crises.

You could already discern from what they had done that most bank failures were not due to panicking customers or to mere liquidity problems, but rather reflected serious underlying pre-run insolvency problems, and that this was also true for the vast majority of bank failures during the Great Depression, for example, where a lot of our myths about banking crises come from. It has been true throughout US banking history, and it has been true throughout the history of banking in most countries. Banks fail because they make bad loans and investments. 

Now, what is also true, that does not come across so much from that particular research paper, is that a lot of times the fundamentals are bad because of stupid government regulations. The best-known example of this is, again, the 1930s. Many of the banks that failed were underdiversified unit banks in agricultural communities. Basically, didn’t have to be a bad harvest, it could be a fall in the international price of the crop involved, that would be that. The bank didn’t have enough diverse assets to withstand the shock.

Whereas with branch banking, it could have. As we know by comparing, for example, Canada and the United States, Canada had these big branch banks with diversified assets. Not one failed during the Great Depression. Some people say they were insolvent on a mark-to-market basis. I’m sorry, but mark to market isn’t how it works in banking traditionally. Those banks were able to see it through in Canada.

As for Diamond and Dybvig, yes, Diamond and Dybvig has long been a pet peeve of mine. Diamond and Dybvig is a model of a legend. They start with a belief about why banking crises happen. They start with the myth that everybody just panics. People think that that’s what happened in the 1930s, but it isn’t. They end up with a model of that myth. Not a model that is actually helpful directly for understanding what causes banking crises. It’s a model of the banking panic in Mary Poppins, to be precise.

By the way, it is not a good model of the banking panic in It’s a Wonderful Life, first of all, because that’s not a bank. It’s a building and loan. Second, because there’s a real loss in that bank. I think it’s George Bailey’s brother who misplaces the money or something. There’s actually a fundamental shock in the It’s A Wonderful Life bank story. Plus, the bank avoids, actually, any sort of run without any deposit insurance or anything. 

The Diamond and Dybvig model, yes, okay, it’s a wonderful intellectual exercise. Very clever. Look at this. The bells and whistles work properly. It’s self-contained, but it doesn’t describe why actual banking crises have happened in the past. It doesn’t even coherently explain in a completely coherent way how deposit insurance or a central bank could solve the imagined problem of a panic-based run while sticking to the assumptions that the model depends on asymmetric information and all that.

It’s a great intellectual exercise, but if you’re going to read it, you’ve got to read all the criticisms of it, and you’ve got to know some economic history, and only then can you decide whether this is a model that sheds light on actual banking crises. What annoys me, it’s not so much Diamond and Dybvig. It’s all the people who just say matter-of-factly whenever they come to talk about why we need deposit insurance or lenders of last resort, “see Diamond and Dybvig,” and they’re done. That’s it. No, you’re not done. You’re not done because what you’re citing doesn’t prove what you think it proves.

Beckworth: Related to that, let me go to one of your fallacies that you sent me in this list of fallacies and probably spend the rest of our time together discussing this. This is a fallacy where people claim that were it not for central banks and deposit insurance, many perfectly solvent banks would fall victim to sheer panic. We’ve touched on this a little bit, but I want you to discuss that and maybe invoke the Scottish free banking system. Just give us some concrete examples of why this is not the case.

Selgin: The basic story here is it’s an empirical question, right? We need to actually look at the evidence. We know that it’s conceivable that banks could fail because people just panic for no reason, try to get their money out. They could fail. Actually, even then, it’s not that likely if you have a bunch of banks that are independent, because usually solvent but illiquid banks can turn to other banks. They don’t need to turn to a central bank. If they’re solvent, they’ll have assets that they can use to secure loans from other banks. That’s happened all the time in history.

To get to your question, when we look at banking systems that didn’t have a lot of stupid regulations like restrictions on branching and other things, when we look at some of the freer banking systems—and I use the term literally; I’m not referring to the antebellum so-called free banking systems of the United States, which weren’t free at all—we find that they had very good records of stability. Not perfect, but very good. There were hardly any bank failures in Scotland between the failure of the Ayr Bank in 1772, which was a calamitous failure, by the way, though not entirely due to lack of government involvement, as Tyler Goodspeed has shown in his book on the subject. Between then and sometime in the early 1880s, you basically didn’t have an important Scottish bank failure. 

The Canadian record was similar between the failure, I think, in the 1880s, Prince Edward Island Bank. Then you had nothing until the Home Bank failure, which was a big one, in the 1920s. These are just individual bank failures that were rare. Systemic crises were unknown during the same periods.

Banking systems can be very, very robust. Banking systems don’t have to be constantly collapsing as a result of systemic bank failures. When we look now no longer at those successful systems, but at less successful ones, like the United States, I won’t repeat stuff about the 1930s, I’ve already said a little bit about that, but if we look at all of the pre-Fed crises—that is, 1907, 1893, 1884—if we dig into them, we find all sorts of stupid regulations. Particularly concerning are the conditions under which banks could issue bank notes that played a crucial part, not just an incidental part, but played a crucial part in the crises in question.

I used to pose a challenge to my students in money and banking and my graduates and say, “Look, you show me a banking crisis, and I’ll show you some stupid regulations that were absolutely essential to the crisis happening that you couldn’t tell the story right unless you talked about those regulations.” I’ve never had anyone beat me on that challenge by finding an example that didn’t fit it, and I don’t think anyone can, to be honest. Let’s talk about the real causes of financial crises instead of elaborating upon myths about how they might happen. If we do that, we’re going to take very seriously the thesis that banking crises are caused by dumb regulations and can be prevented just by avoiding dumb regulations.

Beckworth: Can you give us some more details on the dumb regulations, say, behind 1907, the late 1800s? What was making the US banking system so prone to challenges back then?

Selgin: There were several things, David. One is that unit banking remained the norm throughout that period, and that is going to make banks more vulnerable to any kinds of shocks. The other thing was the inelastic supply of currency. Now, everybody learns that the Federal Reserve was created to make for an elastic currency supply. What they don’t learn is why the supply of currency was inelastic before the Fed. They don’t learn that often enough.

It was inelastic because of stupid regulations. Mainly, when the national banking system was set up during the Civil War, one of the motives for setting it up was to create a captive market for Union government debt, Union bonds. Why? Because they wanted the Union to win and they needed money. They set up this system of new banks, and they would also soon after suppress notes issue by state-chartered banks in order to boost demand for the new national bank charters.

They set it up with a rule that said the national banks could only issue notes at least 100%, actually about 110%, backed by specified US government securities. Hey presto, you’ve got a nice new market for Union government bonds. Now, whether you think that legislation helped pay for the war or not, the long-run repercussion was that the stock of national bank currency, the quantity that could be issued profitably, was dependent on the availability of eligible US government bonds.

There was other currency around, by the way, the greenbacks, but the quantity of the greenbacks was absolutely fixed by legislation. There couldn’t be any elasticity there. State banks had been prohibited from issuing notes. Their notes were taxed out of existence, so they couldn’t help. The only source of flexible currency had to be national banknotes. As I said, the supply was tied to the availability of US government securities.

After the Civil War, for decades, the US government ran surpluses with which it retired its outstanding debt, including those eligible securities that became fewer and fewer. Their market prices rose to a very high premium over the course of the 1880s, 1890s. As that premium rose, as they got harder to come by, the stock of national currency declined. It was a perfectly negative relationship. The bond prices are going up, and the quantity of currency is going down. By 1890, there were half as many national banknotes, half as much paper currency of that form in the country as there was in 1880 in a rapidly growing economy.

Now, some people say, well, there were still some of those bonds out there. Why didn’t the banks buy? You see, the banks weren’t going to buy bonds just to satisfy temporary demands for currency, especially because the demand for currency would spike every harvest season. If the banks spent all the money to buy bonds to secure notes that would only go out for a couple months and then come back, then those notes would sit in their vaults, and they would be bearing the cost of acquiring the necessary collateral at a high premium, so they didn’t.

In the harvest season, in the fall in particular, the demand for currency would spike, the supply wouldn’t do a darn thing, and you would often get big interest rate spikes and other stringency, which, because of correspondent banking, another consequence of laws against branching, all that pressure would emanate into New York, and the New York money market would be where all the spiking of interest rates would happen, and it would cause all kinds of hell because in order to make up for shortages of credit, you would have the correspondents drawing on their New York balances.

Then in big New York banks, basically the call money market, that people borrowing money in the call money market come up with more collateral to maintain their loans with these changes in interest rates, that’s where a lot of these crises got started. It was all due to legislation. The Fed, of course, was a solution to this, but it was an inferior solution to just letting the banks issue notes on the basis of their general collateral. No special bond-backing requirements. That’s how Canadian banks work. That’s how Scottish banks work. That’s how most commercial banks work.

Whatever assets they could have backing their deposits, the same assets could back their notes. A farmer comes to his bank and says, “Hey, I want to cash my deposit.” The bank says, “Okay. We’ll take away this kind of liability, deposits, and we’ll give you this other kind of liability instead. It hasn’t affected our liquidity. It hasn’t caused any trouble for us. Go and pay your workers.” 

In the United States, the farmer comes and says, “Hey, I want some currency.” They say, “Well, we can’t afford to issue more currency of our own notes because we can’t afford to go out and get the collateral.” The farmer doesn’t walk away. He says, “Okay, I’ll have my gold then, or some greenbacks.” Now he’s dipping into the bank’s reserves, don’t you see? Now the bank’s going to call its New York correspondent and say, “Hey, we’re going to have to draw on your balance.” The correspondent’s going to say, “Okay, New York money market, sorry, we don’t have much money for your loans anymore.” 

Beckworth: It’s all tied together.

Selgin: What’s the Fed? The Fed is simply 12 banks exempt from the national bank bond collateral requirement. Now a national bank wants more currency; it goes to the Fed, with its little bowl in its hand and says, “Please, may I have some currency? Here are some of my assets that the goddamn law won’t let me use myself to back my notes. Here they are, discount them, and give me some cash.” The Fed says, “Yes, we can do that because we’re exempt.” There’s no magic there.

Beckworth: That is interesting.

Selgin: That’s all it was. Yes. It was all done to protect the big New York banks that were keeping correspondent balances, and that was their source of profit. The idea was, if they let other banks be free to issue notes and to branch, it would have killed the big New York banks’ correspondent business. Nelson Aldrich and his gang made sure that we got a different kind of reform that wouldn’t do that. That didn’t work very well, as the next couple decades would prove quite clearly, but it kept the correspondent banks happy in New York. That’s that.

Beckworth: George, you’ve actually written a paper on this. We’ll provide a link to it in the transcript. One last question before we wrap things up. During this time when there was a growing shortage or reduction in the supply of notes, deposits were growing. That’s the flip side. Why couldn’t the deposits meet the demands for liquidity? During the harvest season, why was it that there was a spike for currency and deposits couldn’t satisfy that need.

Selgin: We still have payments today for which deposits, checks, or transfers are not effective. Back then, that was much more the case for certain kinds of payments. Among those, most importantly, were the payments to the migrant workers who would be harvesting the crops. That’s why the harvest season was so important. These people didn’t keep bank accounts. They didn’t stay in one town long enough to have a bank account. Remember, there were no branch banks, right?

You don’t have to open an account, and then open another account; and it wasn’t that easy. These people want to be paid in cash. They have to be paid in cash. It’s going to be either the notes of national banks, or it’s going to have to be greenbacks, or it’s going to have to be coin, the last two of which are reserve assets. 

By the way, if you look at a chart of the supply of currency in these decades, in the US, as I told you see this shrinking supply and no seasonal variation, no sawtooth. If you look at Canada, it goes like this. The Canadian currency cost sawtooth, perfectly following changes in the demand for currency with no central bank, because Canada had no central bank until ’35. Market forces and sufficient liberty to issue currency were all that was needed to keep the supply changing along with demand.

Beckworth: On that note, our time is up. Our guest today has been George Selgin. George, thank you for coming back on the program one more time.

Selgin: You’re very welcome, David, and I’m always glad to do so.

Beckworth: Macro Musings is produced by the Mercatus Center at George Mason University. Dive deeper into our research at mercatus.org/monetarypolicy. You can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app. If you like this podcast, please consider giving us a rating and leaving a review. This helps other thoughtful people like you find the show. Find me on Twitter @DavidBeckworth, and follow the show @Macro_Musings.