Are government guarantees or financial regulation a more effective way to prevent banking crises? I study this question in the presence of a negative feedback loop between the fiscal position of the government and the health of the banking sector. I construct a model of financial intermediation in which the government issues, and may default on, debt. Banks hold some of this debt, which ties their health to that of the government. The government's tax revenue, in turn, depends on the quantity of investment that banks are able to finance. I compare the effectiveness of government guarantees, liquidity regulation, and a combination of these policies in preventing self-fulfilling bank runs. In some cases, a combination of the two policies is needed to prevent a run. In other cases, liquidity regulation alone is effective and adding guarantees would make the financial system fragile.

The 2023 banking turmoil highlighted how technological advancements have significantly accelerated the speed of bank runs. This paper investigates the impact of these faster bank runs on the effectiveness of policy interventions by interpreting them as a constraint on the relative speed of policy responses. Using a model of bank runs and ex-post policy responses, we examine how delays caused by this constraint affect financial fragility and welfare. We find that while delays exacerbate welfare loss by distorting allocations, they may also decrease fragility by making banks more cautious. We study the optimal level of structural delay, balancing the trade-off between distributional distortions and financial fragility. Furthermore, we extend this model to explore the roles of liquidity regulations and capital injections given such a delay. We show that regulation may be more desirable than a capital injection if the delay is substantial because the benefit of decreased fragility is particularly potent.

Should banks be transparent during a bail-in? Banks suffering losses may bail-in creditors to optimally allocate resources between early and late withdrawers. However, if losses are private information, then bail-ins may signal asset quality. In the absence of signaling, banks can sell assets at a pooled price, effectively insuring creditors against asset risks. However, when bail-ins signal quality, banks may delay bail-ins and sell assets at higher prices than otherwise, but this incentive can trigger inefficient bank runs. To prevent such runs, banks should choose to be either fully transparent or entirely opaque, ensuring asset quality is not private information.