Portfolio Substitution and the Effects of Bank Capital Regulation
Abstract
A large literature finds that bank capital requirements reduce lending because equity funding is costly for banks. But capital rules usually change equity requirements on several bank assets at once, so the decline in lending reflects both the direct effect of higher loan costs and the substitution effect from changes in other assets’ costs. For example, a 1-percentage-point increase in a bank’s risk-based capital requirement raises required equity on risk-weight-1 loans by 1 percentage point and on risk-weight-0.2 securities by 0.2 percentage points, encouraging banks to substitute from loans to securities even though both assets’ funding costs rise. To predict the effects of a counterfactual capital rule not observed in the data, regulators must separate these two responses. I develop a structural framework and empirical method to estimate each asset’s own-cost elasticity of supply, which governs both the asset’s direct and portfolio substitution responses, using differences in required equity across banks holding the same asset and across assets held by the same bank. I recover positive supply elasticities for every asset category, with securities more elastic than loans. Applying these elasticities to post-crisis requirements, I ask whether stronger capital regulation explains large banks’ shift from lending to securities since the financial crisis. Capital regulation accounts for about 28 percent of large banks’ aggregate shift from lending toward securities and 35 percent of the cross-sectional convergence in large banks’ average asset risk weights since 2011.
