Until recently, the only bank run many had ever experienced was in the movie “It’s a Wonderful Life.” But when the depositors of Silicon Valley Bank withdrew $42 billion in a single day—the largest bank run in U.S. history –the modern world discovered that even a regional bank default can tip the whole financial system onto the brink of a crisis.
Qi Chen and Rahul Vashishtha, professors of accounting at Duke University’s Fuqua School of Business, have long studied the role information plays in panic-induced bank runs. From their research, they argue that the fragility of the U.S. banking system is a feature inherent to a bank’s business model, and some transparency rules can increase the chance of panic-based deposit withdrawals (bank runs).
“We have known forever that banks are prone to runs,” Vashishtha said. “Runs are very natural in the banking system and connect to their very business model.”
Banks receive deposits and use them to make loans, transforming something liquid (cash) into something illiquid (loans), Vashishtha said. This is inherently risky, because if all the depositors were to withdraw the money at any point in time, banks wouldn’t have enough cash to meet their requests, he said.
But why do depositors decide to run to the bank—as happened with Silicon Valley Bank? Is it more because they have information about the bank’s bad fundamentals, or more because of panic?
The motivation behind the runs
In a National Bureau of Economic Research (NBER) working paper “Liquidity Transformation and Fragility in the U.S. Banking Sector,” Chen and Vashishtha--with co-authors Itay Goldstein from University of Pennsylvania’s Wharton School and Zeqiong Huang from Yale School of Management--show that even small drops in a bank’s performance can trigger larger deposit withdrawals than would be justified by a reasonable analysis of the bank’s fundamentals.
“Nobody really needs the cash here,” Vashishtha said. “Depositors are withdrawing purely because they expect others will do the same, and they want to be the first in line, because they know that banks don’t have enough cash.”
In a previous paper, the authors explained the cost of transparency for banks. Following the 2008 global financial crisis, the professors said, a new regulatory framework imposed new transparency (disclosure) requirements for banks. But these new requirements, their paper suggests, can hurt the banks in their basic function of providing liquidity and creating value through loans, by forcing the banks to offer higher deposit rates to keep uninsured depositors from fleeing, or by constraining banks to rely on their limited pool of internal funds to provide loans.
“In the US there is such a strongly ingrained view that transparency is always good,” Chen said. “But in the context of banking, the message may be different.”
The unintended signal behind runs
In a third working paper—with Fuqua Ph.D. candidate Shuyan Wang—the researchers dug deeper into the consequences of banks’ reporting the fair value (that is, the market value, or unrealized gains or losses) of their loans. As with the previous papers, they found that depositors disproportionately take away their money at any sign of decrease in fair value of the bank's assets.
“A one-standard-deviation decrease in fair values is associated with 10% lower uninsured deposit flows,” the authors write.
The outsized sensitivity of depositors to information about banks’ assets can possibly explain the collapse of Silicon Valley Bank, the authors pointed out.
“This occurred in response to information about unrealized losses on long-term bonds that (SVB) was holding,” Vashishtha said. But if SVB were to hold those bonds to maturity, he said, they would have had money to pay off their depositors. “We are not talking about a sudden decline in the cash-generating potential of banks’ assets, or a scenario where they made bad loans,” he said.
“It’s been known for a while that all banks are having trouble paying high interest to keep depositors from running away,” Chen added. “This same fundamental issue didn’t drive an overnight collapse of other banks. That’s why we think the case of SVB has an element of panic,” he said.
Why Silicon Valley was different
Silicon Valley Bank is an unusual bank, Vashishtha said, because it was mainly investing in long-term Treasury bonds and mortgage-backed securities. Usually, he said, the typical bank transforms deposits into loans, as is the case with the latest U.S. midsize bank to fall, First Republic Bank (the bank reported a loss of $100 billion worth of deposits during the March panic.)
Whatever investment portfolio different banks might have, bank runs are not always justified by the banks’ bad performance but can be panic-induced, the researchers said.
“Which is important to demonstrate with data like we did,” Chen said, “because if runs were purely driven by bad performance, it would be difficult to make the case for a role of governments and central banks in stopping the runs with rescue measures–including bailouts. On the other hand, bailouts are better justified when runs are panic-driven.”
Chen and Vashishtha said their findings suggest possible policy corrections, like extending deposit insurance to all depositors, or connecting banks’ capital requirements to the fair value of their assets. But the larger open question about transparency remains, they said.
“The striking part of this research for me,” Vashishtha said, “was confirming with data that opacity has benefits, in the banking system, and transparency can hurt.”
The disclosure requirements introduced after the 2008 financial crisis were supposed to reduce the fragility of the banking system, the researchers said. But some disclosures, such as the fair value of assets, are not always a reliable indicator of the health of the bank, and they can still trigger panic, they said.
“Are disclosures making banks less or more fragile?”, Chen wondered. “More research is needed to see whether depositors are better off with less information than looking at some points of data which don’t actually point to a bank’s health.”