Research Highlights Podcast

March 12, 2024

A textbook bank run

Andrew Metrick discusses the failure of Silicon Valley Bank and the government’s response.

Source: Minh Nguyen, CC BY-SA 4.0

In the middle of the day on Friday, March 10, 2023, bank regulators swiftly shut down Silicon Valley Bank (SVB), arguably averting a wider panic. Compared to past financial crises, it was not especially economically significant, but it stands out as an important, illustrative example of the economics of banking.

In a paper in the Journal of Economic Perspectives, author Andrew Metrick explains the causes behind SVB’s failure and how the government responded. He says that understanding the collapse of SVB is a stepping stone to making sense of more complicated financial crises such as the Global Financial Crisis.

Metrick recently spoke with Tyler Smith about why Silicon Valley Bank failed and what policymakers can do to prevent financial crises.

The edited highlights of that conversation are below, and the full interview can be heard using the podcast player.

 
 

Tyler Smith: What kind of bank was SVB? And was it a bank that was doing anything really out of the ordinary?

Andrew Metrick: It wasn't doing things that were out of the ordinary, but it was certainly far out on the risk curve. Most of their money was given to them by depositors, which are effectively loans from depositors to the bank that the depositors can take back any time they want. And with that money, they made loans to businesses. They also bought government bonds and securities when they were just holding the money, waiting for a good loan to make. That's a pretty standard thing for banks to do. What made Silicon Valley Bank somewhat unusual was just how much of the deposits were large deposits, such that they were bigger than what the Federal Deposit Insurance Corporation (FDIC) is willing to insure. There's a limit of $250,000 per deposit. They had a lot of deposits that were more than that.

Smith: What do you see as the primary reasons for SVB’s failure from an economic point of view?

Metrick: When you are a bank, your main service that you are providing is convenience. They give you money and they leave it in your bank as a deposit. Generally, banks pay an interest on that deposit that's really low compared to what the government pays on their own debt, which is the safest debt in the world. But you get an even lower rate of interest when you loan something to a bank through a deposit. So why is that? You get convenience out of that. You can use it to pay your bills. You can use it to pay your taxes. And that's the main service that a bank sells—you never have to worry that your money is safe. As soon as you have to worry that your money is safe, the rational thing to do is to take it out and move it to another bank.

If I have money in a bank and that money is not insured because there's a lot of it, then at the first sign that this bank may not survive, I'm going to take my money out and that's going to be dangerous to the bank. How does a bank prevent that? The main way a bank tries to prevent me from worrying is by being able to show how much extra money they have. They point at the huge difference between their assets and their liabilities, their capital. What happened essentially to Silicon Valley Bank is that their assets, the things they bought with all of the deposits people gave them, fell a lot in value. They fell in value to the point that the depositors were not sure they could get paid back. Once that happened, they ran. So the proximate cause is SVB made investments with the money they received, which fell a lot in value. What's interesting about that and clean about that is those investments were in very safe things, government securities.

Smith: Why did the value of their capital decrease?

Metrick: Silicon Valley Bank approximately tripled in size from before the pandemic to the end of 2021. That's very fast growth for a bank. And in the early days of COVID, we were awash in liquidity from the government, particularly in Silicon Valley and in the tech sector. With that money, many of the customers of Silicon Valley Bank put it in the bank. That's part of the reason they grew so fast. Now, what do you do when you're a bank and all of a sudden your deposit base is tripling over a short period of time? Usually what you want to do with money that comes in is you want to loan it out to commercial real estate or mortgages or whatever. But of course, during COVID, there wasn't a lot of that happening. So where do you park the money? They decided to park the money in medium- to longer-term government securities. Now you can say, why didn't they buy the safest form of government debt, which would be really short term? From their perspective that doesn't seem like such a great deal. It's paying zero. That turned out to be a mistake. When interest rates then rose very rapidly in 2022 because of concerns about inflation, the value of those government securities decreased, even though you know that you're going to get paid back 100 cents on the dollar when it eventually matures. That is just an unavoidable facet of bond math and the way that bonds get priced. And because of that, SVB, which had a lot of those bonds in its portfolio, saw the value of its assets fall. And, indeed, they fell so much that there was no longer any capital cushion.

Smith: What happened during that really short window to cause it all to crumble?

Metrick: Why March 2023? They had a nice capital cushion going into all of this. And that capital cushion had been completely wiped out by the end of late September 2022. The fact that their capital cushion had been wiped out was known and public. But it was six months before people actually ran. Why not run as soon as you see that the value of the bank’s assets have fallen? From the perspective of somebody who believed in Silicon Valley Bank, they would say that it's true that the value of the bonds have fallen because interest rates have gone up a lot, but what the bank actually has to pay their depositors on average is not very high. They would also say look how good SVB is doing on the profit side of things. They're only paying a couple percentage points on deposits, and they can reinvest it at five. That's a great deal. And, furthermore, SVB optimists would say the bank has very special relationships with their customers. Their customers have been with them a long time and offer services that are unique in the Silicon Valley community. So, because of the fact that they’re paying their depositors a whole lot less than market interest rates, they’re going to earn a profit stream over all this time. Even though on paper you might think they’re not solvent, if you were properly measuring the fact that they have this incredibly valuable business, because of all these customers who love them, who still let SVB pay them a very low amount, they’re very solvent. 

But what happened during this time is that people needed their money because the times changed. Interest rates were high, the days of really easy money were over, and so many of their customers who had deposited the money earlier were taking it out just for natural causes. But if SVB wanted to give those people their money back, they had to sell some of the securities that they had on their asset side. And when they sold those securities, they had to recognize the fact that they were selling them for less than what they had bought them for. When this news gets out, customers could very reasonably get nervous. They could wonder whether this deposit base is not quite as stable as they thought. And that's all you need. The rational thing to do, if you're basically just getting a service that you could almost replicate elsewhere, is to take the time and effort to go and move your money. And once that begins, the reasoning that leads deposits to remain safe in the first place gets weaker and weaker, and it builds upon itself. Which is why any sign of weakness at a bank can very quickly become self-fulfilling. And that's what happened specifically when people learned that SVB had some issues and needed to sell some securities in March of 2023.

What happened essentially to Silicon Valley Bank is that their assets, the things they bought with all of the deposits people gave them, fell a lot in value. They fell in value to the point that the depositors were not sure they could get paid back. Once that happened, they ran.

Andrew Metrick

Smith: Once people are headed for the door, what options did the government have and what did they ultimately end up doing?

Metrick: The first thing that happens is the bank, because it is a bank, and because it's a member of the Federal Reserve System, can borrow money from the Federal Reserve if it puts up collateral. Now, the bottom line is that SVB did not have sufficient collateral, given that on the Friday of that week, almost all of the money in their bank was going to get taken out. They didn't have enough assets to be able to borrow off of at that moment because they truly were insolvent or close to insolvent on a pure assets minus liabilities basis. And then what we will typically see happen is that the regulators will close the bank. They like to do that and historically have always done that at the end of the day on a Friday. And then they promise all of the deposit holders on Monday that they will reopen the bank under a new name or new management, and all of the insured deposit holders will get their money back. And almost all of the time, what the FDIC and the United States is able to do is sell the bank to somebody else.

What made the SVB case unique, though, is there were so many uninsured depositors, and there wasn't enough value left in the shell of this thing, or enough time, because it was pretty large, to get it sold and cleaned up over the weekend. And that, of course, scared the heck out of everybody. Everyone who had uninsured deposits at other banks wondered whether there were going to be such significant failures that the FDIC wouldn't be able to protect them. So, two days later, the Sunday of that weekend, the FDIC recognized that this was going to be very destabilizing for the entire financial system. They were concerned that all over the banking system people would run, that everyone who had uninsured deposits would think there wasn't enough to keep them safe, and that that would be very damaging to the overall financial system. So they changed course. They used what's called the “systemic-risk exception.” That is an exception which enabled them to effectively promise the uninsured depositors that they would get paid back as well.

The Failure of Silicon Valley Bank and the Panic of 2023” appears in the Winter 2024 issue of the Journal of Economic Perspectives. Music in the audio is by Podington Bear.