A Tale of Different Capital Ratios: How to Correctly Assess the Impact of Capital Regulation on Lending

For almost two decades, quantifying the effect of changes in bank capital and capital regulation on lending has been one of the most important research questions. Yet, the empirical literature has remained largely fragmented in terms of the estimated parameters. In this paper, we collect more than 1,600 estimates on the relationship between bank capital and lending and construct 40 variables that reflect the context in which researchers obtain such estimates. After accounting for potential publication bias, the effect of a 1 percentage point (pp) change to the capital (regulatory) ratio on annual credit growth is set at around 0.3 pp, while the effect of changes to capital requirements is about -0.7 pp. Using Bayesian and frequentist model averaging, we expose the additional layers of fragmentation observed in our results. First, we show that the relationship between bank capital and lending changes over time, reflecting the post-crisis period of increasingly demanding bank capital regulation and subdued profitability. Second, we find the reported estimates of elasticities to be significantly affected by the researchers’ choice of empirical approach.