Wesleyan Senior Creates Economic Model to Analyze Digital Bank Runs

Thinking of a bank run—when a mass sector of a bank’s depositors withdraw money in a short period of time—an image springs to mind. Seemingly unending lines of worried civilians encircling a bank teller in the 1930’s clamoring to recoup their entrusted funds as financial panic grips the nation. But modern bank runs look different, happen much faster and are largely unpredictable, according to Jennie Ebihara ’24, who analyzed new problems created by digital bank runs in her senior thesis paper.

Ebihara maintains that current models theorizing the growth and speed of bank runs do not really address the problems that modern technologies pose, so she created one that does. With the rise of mobile banking and federal interest rates, banks can be susceptible to quick and devastating periods of high withdrawals, she said.

“With current technology thousands of people can withdraw at once,” said Ebihara, an economics and college of social studies major. “You can withdraw your money with a swipe of a phone, so that’s just made the speed and size of bank runs much faster and much larger.”

Her model reconsiders the sequential nature of current models like the 1988 Diamond-Dybvig and the 2006 Ennis and Keister models, both built around assumptions that one depositor withdraws at a time as if in a line, and factor in an uncertain number of withdrawals at a time. Her goal was to better model the uncertainty and pace of bank runs and determine the optimal point for regulators to intervene, she said.

“The rising speed of runs is a concern because it makes it more difficult for policymakers to intervene quickly enough to prevent bank runs while evaluating the welfare benefits and losses of such intervention,” Ebihara wrote.

Through her model and research, she found two main takeaways: banks are more fragile and susceptible to possible runs when they are amid a period of higher federal interest rates and when a bank run unfolds over a short time span. This means that a policymaker would be likely to delay intervention and allow more depositors to pull their money out of a bank, thus affecting the cash reserves and overall health of the bank, she said.

“Policymakers must strive to create infrastructure that supports faster policy responses in the wake of a digital bank run,” Ebihara said.

She researched the run on Silicon Valley Bank in March of 2023 to help inform her model—although the model can be calibrated to any individual bank, she said. Silicon Valley Bank experienced a swift run, losing a significant amount of its total deposits in a short period of time—25 percent of its deposits, or $42 billion, on the first day and an additional 62 percent, or $100 billion, was scheduled for withdrawal by the second day of the run before regulators put a deposit freeze in place, according to Ebihara’s thesis paper. For comparison, depositors withdrew less than 10 percent from Washington Mutual, which experienced a 10-day bank run in September of 2008, in before a deposit freeze was triggered.

Interventions can be costly for regulators, and therefore taxpayers, and intervening too fast can leave many people who require liquidity without access to their accounts, Ebihara said.

In the case of Silicon Valley Bank, around 94 percent of its accounts were not eligible to be fully insured since they over the Federal Deposit Insurance Corporation’s (FDIC) $250,000 per account threshold of deposit insurance. This is substantially more than the national average of 42 percent of accounts, she said. This means that a vast majority of depositors were unable to recoup their entrusted assets, until Silicon Valley Bank was bailed out by First Citizens Bank.

“This thesis was like that first time I really realized you can do so much with economics,” Ebihara said. “You just have to go a bit further and really push yourself to connect with your interests.”

Her research was sparked, in part, by her internship with a bank in the summer of 2023, she said. Following graduation, she will venture to New York to work for a bank once again. She is also an avid musician and reader of Japanese literature—she is a native of Tokyo, Japan. Ebihara is currently learning the Chinese bass.