Customer Complaints and Bank Deposit Outflows: Evidence from COVID-19 Pandemic and the SVB Collapse

<p dir="ltr">In this thesis, the extent to which customer dissatisfaction – measured by formal complaint volumes filed with the Consumer Financial Protection Bureau (CFPB) – affects deposit outflows and liquidity risk in the United States banking sector is examined. The study is motivated by the need to integrate behavioural signals into financial stability assessments and connects research relating to consumer behaviour with the literature on banking fragility. Drawing upon a balanced panel of bank-level financial data and complaint records that span from 2012 to 2023, fixed-effects regression analyses show that elevated levels of customer complaints have a significant association with deposit withdrawals, even after controlling for bank fundamentals and macroeconomic conditions. In order to capture crisis-specific dynamics, a novel measure of cumulative abnormal deposit outflow (CADO) is introduced and an event study is implemented that focuses on two major systemic shocks: the COVID-19 pandemic and the 2023 collapse of Silicon Valley Bank (SVB). The results show that banks with higher pre-crisis complaint volumes experienced deposit outflows that were more pronounced during periods of market stress. A dynamic interaction analysis further reveals that the marginal effect of complaints intensified during and following crises, which suggests that reputational concerns become more salient in influencing depositor behaviour under stress. Generally, the findings highlight the importance of customer relationship management in the preservation of liquidity stability and demonstrate the value of complaint data as an early indicator of depositor responsiveness and institutional vulnerability. The study makes a contribution to literature on depositor discipline, behavioural finance and financial risk surveillance, in addition to offering implications for policymakers and bank risk managers who wish to strengthen financial system resilience.</p>