Relative performance and systemic risk contributions of small and large banks during the financial crisis

During the financial crisis of 2007-2009, large banks (BHCs with assets exceeding $1 billion) suffered a sharper decline in performance and a greater increase in systemic risk, than small banks (BHCs with assets less than $1 billion). Performance is measured by return on assets (ROA) and EBIT/TA. Systemic risk (SRISK) is measured by the capital shortfall indicator proposed by Brownless and Engle (2017) and total risk is standard deviation of monthly returns. We test four hypotheses associated with the dissimilarity in behavioral patterns of large and small banks during the crisis: (1) the nontraditional banking activity hypothesis, (2) the organizational structure hypothesis, (3) the accounting risk hypotheses, and (4) the accounting conservatism hypothesis. We find that behavioral dissimilarities between large and small banks can be, at least, partially related to the first three hypotheses. Non-nested test procedures show that factors associated with nontraditional activity hypothesis fit the data for performance (ROA and EBIT/TA) and total risk the best. Organizational complexity factors are the most relevant set of variables related to systemic risk of banks (SRISK). The implications of our findings are that regulators should focus on factors related these two hypotheses to optimally alter bank performance and bank contribution to systemic risk, rather than limiting bank size and bank acquisitions as proposed by Section 622 of the Dodd-Frank Act (effective January 1, 2015). Our recommendation is in accord with attempts by the Trump administration to relax the Dodd-Frank Act.