Non-Interest Income and Bank Performance: Does Ring-Fencing Reduce Bank Risk?

The optimum scope of bank activities is central to many proposals for banking system reform. For example, a core component of the Dodd-Frank Act (2010) and regulatory proposals in the UK and the EU has been the concept of “ring-fencing” – i.e., restricting banks’ activities to their core retail and wholesale financial intermediation functions. It is argued that limiting the scope of bank activities reduces the likelihood of failure related to business lines that are highly risky. However, an alternative view holds that diversification of banks across traditional interest generating business and non-traditional businesses enhances bank profitability and reduces risk. Based on a sample of 368,006 quarterly observations on 10,341 US banks during the period 2002-2013, we find that a higher ratio of non-interest income (derived from fees and non-core activities such as investment banking, venture capital and trading) to interest income (associated with deposit-taking and lending to retail and commercial clients) is associated with higher profitability as well as lower failure probability. This finding is stronger during the crisis period than in either the pre- and post-crisis periods. We find similar results using trading (and investment banking) income instead of non-interest to interest income. Our results generally hold across bank size groups and are robust to the inclusion of bank fixed effects, bank size, and various measures of leverage and asset quality in the regressions. We find similar results in a sample of bank holding companies (BHCs), and in addition show that a higher fraction of non-traditional bank income is not associated with a higher contribution to systemic risk. Overall, our results question the benefits of “ring-fencing” bank activities.