�������

In a speech just a few days before Silicon Valley Bank failed, Federal Deposit Insurance Corporation (FDIC) Chairman Martin Gruenberg said that the market value of U.S. banks’ long-term assets had dropped $620 billion in 2022. Yet a sobering new analysis finds that the banking industry’s unrealized losses are now more than three times that — and many banks are facing the same kinds of troubles that brought down SVB.

The research, coauthored by Amit Seru, a professor of finance at Stanford Graduate School of Business and a senior fellow at the Stanford Institute for Economic Policy Research (SIEPR), looked at more than 4,800 U.S. banks to gauge their exposure to the risks that caused SVB to fail. Some of its key findings:

• As a result of rising interest rates, the market value of the average U.S. bank’s assets is about 9 percent lower than its value on paper. Overall, the U.S. banking system accumulated $2.2 trillion in these unrealized losses over the past year. Ten percent of banks have had larger unrealized losses than Silicon Valley Bank.
• While SVB had an unusually large share of uninsured depositors, other banks with high levels of uninsured depositors and large losses are also prone to solvency crises that could trigger bank runs. That, in turn, would require the FDIC to spend billions to save their insured deposits.

“These results support the view that preventing a bank run by uninsured depositors should be of central focus to bank regulators and policymakers,” Seru says.

SVB “Is Not an Outlier”

Working with of Chicago Booth, of Kellogg School of Management, and of Columbia Business School, Seru published their findings three days after SVB failed. Their analysis — which the also recently released — focused on the conditions that led to the one-two punch that felled Silicon Valley Bank: heavy unrealized losses from rising interest rates combined with a large base of uninsured depositors who could trigger a bank run.

The first part of this equation was interest rate risk — a well-known phenomenon where the value of mortgages, bonds, and other long-term assets declines when interest rates rise. These unrealized losses are not reflected on banks’ balance sheets, but they can be calculated by marking investments to their current market value. Between early March 2022 and early March 2023, the market value of Silicon Valley Bank’s assets declined by nearly 16 percent, or $34 billion.

To measure how interest rate increases had affected the rest of the banking industry, Seru and his coauthors deflated the long-term investments held in banks’ portfolios. They found that the average bank’s assets have lost around 10 percent of their value over the past year.

“Silicon Valley Bank is not an outlier,” Seru says. “Lots of banks have had a bad shock.” He and his colleagues found that 11 percent of U.S. banks had worse unrealized losses than SVB. “If SVB failed because of losses alone,” they conclude, “more than 500 other banks should also have failed.”

In addition to its losses, Silicon Valley Bank had an unusually large share of customers who were not covered by the Federal Deposit Insurance Corporation, which insures deposits up to $250,000. Around a quarter of the average bank’s debt is uninsured. Yet more than 78 percent of SVB’s assets were funded by uninsured deposits, putting it in the top 1 percent of banks with uninsured leverage.

This made SVB especially vulnerable to a bank run if its uninsured depositors got panicky. And that’s exactly what happened: After the bank revealed it was taking a major loss on the sale of Treasurys and mortgage-backed securities, its customers withdrew $42 billion in a single day, leaving it with a negative cash balance of nearly $1 billion. It failed a day later.

Many other banks could face a similar situation, according to Seru and his colleagues. “Typically, a regional bank doesn’t have as many uninsured depositors as Silicon Valley Bank. That was unusual,” he says. “But there are lots of other banks out there where the losses are pretty high and you do have a reasonable proportion of uninsured depositors.”

Take the Money and Run

Uninsured customers are crucial to maintaining — or undermining — the stability of banks with solvency issues. Customers whose deposits are FDC insured tend to be “sleepy” — they’re unlikely to withdraw their money because they know it’s safe even if a bank fails. (Which is the entire reason the federal government insures deposits.) But uninsured depositors have no guarantee that their money will be safe. That makes them more likely to run if there are signs that their bank doesn’t have enough assets to cover its liabilities.

If it looks like a bank’s equity may be underwater, Seru explains, customers with deposits over $250,000 will think, “OK, I better get out of there, like, yesterday.” If enough bolt at once, they can leave a distressed bank without enough funds to cover its insured deposits, forcing the FDIC to step in to help insured customers. “Insured deposits will start getting affected pretty quickly if a bank has a lot of uninsured depositors, because uninsured depositors start running away faster,” Seru says.

His paper calculates the impacts of several hypothetical scenarios where uninsured customers run. In the worst-case scenario, if all uninsured depositors withdrew all their money from U.S. banks, more than 1,600 banks would be left without enough funds to cover their insured deposits. Altogether, such a scenario would affect $2.6 trillion in total deposits and would require the FDIC to pay out $300 billion. The 10 largest insolvent banks would include small and large banks, including 2 with asset values of more than $250 billion.

The researchers consider such a scenario “likely too extreme, although not impossible once the news of a run spreads.” In a more plausible scenario, half of uninsured depositors would pull their money from U.S. banks. That would force the FDIC to spend $10 billion to protect deposits at nearly 190 banks. “Our calculations suggest these banks are certainly at a potential risk of a run, absent other government intervention or recapitalization,” the researchers write.

That scenario assumes that bankers could quickly raise some cash by selling off their long-term assets at market price. But even more banks would be at risk in a scenario in which mass withdrawals by uninsured customers sparked a “fire sale” of devalued assets.

The FDIC tried to stem the impact of the Silicon Valley Bank’s failure by taking the unusual step of “backstopping” its uninsured deposits. Seru says this move makes sense, though it could send the wrong signal to bankers dealing with heavy losses. “Backstopping is fine, but how are we going to prevent several banks that are underwater on the equity side from taking huge risks to gamble for resurrection?”

And backstopping alone can’t solve the deeper issues threatening the stability of the U.S. banking system that Seru and his colleagues have identified. Regulators, he says, “have to think further down the game tree.”

This story was March 20 by Stanford Graduate School of Business Insights.