In recent years, the banking system has (1) undergone major changes in the stress test regulatory framework and (2) weathered a bank run crisis not seen since 2008. This thesis examines these two aspects from empirical and theoretical perspectives, respectively. First, I empirically analyze whether 2018 regulatory changes to Dodd-Frank supervisory stress testing rules — the Federal Reserve’s main oversight tool — created significant changes in idiosyncratic risk and significant new market information around stress test results disclosure dates. I find the idiosyncratic risk of bank stock returns is significantly different for certain banks, and, critically, market reactions to stress test results disclosures may not reflect this change in idiosyncratic risk. This suggests current stress tests may not fully reflect banking system risks. Second, I modify and solve a model of demand deposit contracts in the context of social media and deposit insurance. I find the probability of a bank run increases and decreases in these two factors, respectively. Policymakers should carefully consider the implications of stress test reforms, social media, and deposit insurance as they chart a path forward to safeguard the banking system.