Why individual bank decisions make all banks more vulnerable to runs

The 2023 collapse of Silicon Valley Bank (SVB), followed by Signature and later First Republic, was a reminder that bank runs are no relic of the Great Depression. The roots of these failures were more complex, of course, than a simple stampede for cash. But in each case, nervous depositors demanded their money in a rush and brought on the end—necessitating a rescue by regulators to keep panic from spreading further.

A bank run is a devastating outcome for shareholders and management. So why—with all the machinery of modern financial risk assessment at their disposal—do banks seemingly need regulations to protect them from themselves?

In a new Minneapolis Fed working paper, Monetary Advisors Manuel Amador and Javier Bianchi uncover one reason: Banks don’t internalize the collective dampening of asset prices in the economy caused by their separate decisions to take on leverage (Working Paper 804, “Bank Runs, Fragility, and Regulation”). Even if each bank calibrates its own assets and liabilities to hedge appropriately against a run, the sum of these individual choices makes all banks more susceptible to runs down the road.

Depositors queue up outside of a Silicon Valley Bank office in Santa Clara, California, on March 13, 2023. The bank failure was the second largest in U.S. history. Justin Sullivan/Getty Images

Amador and Bianchi build on their earlier work in which they develop a macroeconomic model where banks may suffer, and ultimately fail, from self-fulfilling runs—snowballing events in which depositors rush to withdraw based largely on the fear that others might get there first. In the model, even banks with fully liquid assets (such as the Treasury securities held by SVB) are vulnerable to a run, as a run prevents banks from investing deposits and obtaining profits.¹ Importantly, the model abstracts from certain features of the regulatory landscape, such as deposit insurance or a last-minute rescue by regulators, in the interest of illuminating the externality it identifies.

The economists find that a central planner who sees the whole field of play would impose some form of restraint on borrowing, calibrated such that banks’ incentives align with the societal optimum. The result of this regulatory action is a lower level of industry leverage for any given probability of a bank run (Panel A), and a lower share of defaulting banks (Panel B).

The externality arises from how banks’ borrowing today affects everyone’s asset prices tomorrow—and, in turn, banks’ capacities to stave off a run if it occurs. In Amador and Bianchi’s model, banks are aware of the potential for runs, as well as other “fundamental” risks to the business (such as loans or investments turning sour). Bank managers balance the desire to take risk—and thus build equity and profits—with the potential for default based on runs or fundamentals. Greater leverage increases expected returns, while the heightened risk lowers the price of the bank’s bonds. All banks settle into a competitive equilibrium that makes sense for each—on its own.

Banks do not internalize, however, that taking on leverage today lowers the aggregate demand for assets in the future (more borrowing by banks today means less net worth and capacity to buy assets tomorrow). This depresses the future prices of assets across the economy. Unfortunately, this lowers the value of the same assets banks would need to sell to satisfy depositors in the event of a run (and stop the run from becoming self-fulfilling). The mathematical relationships in Amador and Bianchi’s model reveal that the broadly depressed price of future capital makes it more likely that banks will default when confronted with a run.

A planner, recognizing these social costs, would implement policies to align bank choices with the societal optimum. The economists propose that this might take the form of a tax on borrowing or higher capital requirements that limit banks’ leverage below what they might otherwise choose for themselves.

The dynamics modeled by Amador and Bianchi “rationalize the type of macroprudential policies that have become a cornerstone of [the] Basel III” international regulatory framework, they write, in which “regulations are adjusted throughout the cycle with the aim of containing systemic risk.” Bank runs are still with us, as the lines of anxious depositors stretching from the doors of Silicon Valley Bank attest. Fostering a stable financial system means continuing to develop our understanding of why.

Read the Minneapolis Fed working paper: Bank Runs, Fragility, and Regulation

Endnotes

1 Any bank, by making loans, keeps much less money on hand than the liabilities it has to its depositors; this tension between illiquid loans and liquid, fickle deposits is at the heart of canonical research on bank runs by Douglas Diamond and Philip Dybvig. Amador and Bianchi show that the risk of runs applies to the modern reality of banks that invest in many other assets, some highly liquid. When the return on these assets is higher than the return on bonds, a bank can seek higher profits by issuing deposits and exploiting the interest rate spread. If such a typically leveraged bank faces a run, however, it experiences a loss in franchise value (the present value of expected future profits), which can leave it more prone to default and, therefore, vulnerable to a self-fulfilling run.