Reconsidering the degree of inefficiency of financial institutions

There is an extensive existing literature on the cost efficiency of banking institutions that estimates cost inefficiencies of the order of 15 to 30%. The persistence of such staggering inefficiencies is surprising given the competitive nature of the industry. Using a novel estimation framework that complements traditional cost function estimation with a revenue analysis, we find that much of what the prior literature classified as inefficiency is actually productive expenditure. This expenditure effectively increases the quality of the banking institution?s products and results in higher revenues. True ?waste? is much lower, and averages only 9-12% of mean costs. By using broader measures of bank output we document that much of the unobserved product quality reflects omitted output. Our estimates of bank inefficiency are robust to the inclusion of these broader measures of bank output, while the inefficiency estimates that are based on the existing approaches are not robust. This robustness of our approach to variations in output definition forms a test of its internal validity, since our approach is designed to account for productive expenditures that are not directly attributed to observed output.