Who is still in line? How bank beliefs drive fragility under runs

During a bank run, the bank observes massive withdrawals but cannot distinguish between genuine liquidity-needs and panic withdrawals. Optimal policy responses—e.g., rescheduling or suspending convertibility—depend on the bank’s beliefs about the liquidity needs of those yet to be served. We revisit Ennis and Keister (2009), relaxing their assumption that the share of impatient depositors still in line mirrors the population share, allowing arbitrary feasible beliefs. Our analysis shows that greater bank pessimism—i.e., a higher perceived share of impatient depositors still in line—increases fragility, defined as the likelihood that a bank-run equilibrium coexists with the efficient no-run outcome, regardless of the policy tool employed. This suggests that managing banks’ beliefs—by reducing undue pessimism—before a crisis may mitigate run incentives. In an era of real-time data and AI-driven monitoring, banks may strive to infer more precise information on remaining depositors—potentially pivotal for preventing panic-driven runs.