We test whether output growth in European economic agglomeration regions depends on nancial development. To this end we suggest a relative measure of the quality of nancial institutions rather than the usual quantity proxy of nancial development. In order to measure the quality of nancial development we use pro t e ciency derived from stochastic frontier analysis. We show that more e cient banks spur regional growth while the typically used quantity measure of nancial development is negligible. Also, our results suggest an additional channel through which better banking can spur growth: the interaction of more credit with e cient banks.Keywords: Bank performance, regional growth, bank e ciency, Europe JEL: G21, O16, O47, O52 Non-technical summary The creation of an e cient nancial system is an explicit objective of the European Commission. Ongoing harmonization of regulation to foster an increasingly homogenous European banking system ultimately serves the purpose to enhance competition and thereby economic prosperity.At the same time, regional di erences continue to persist across European banking markets. Therefore, we investigate in this study whether regional bank e ciency has a positive impact on regional growth. We employ bank-speci c data to estimate pro t e ciency with stochastic frontier analysis for approximately 3,000 banks active in 160 European regions between 1997 and 2003.Our study thus analyzes if a micro-economic measure of banks' abilities to convert savings and deposits e ciently into loans suited to fund investments is positively a ecting regional economic growth.Our results con rm that higher mean pro t e ciency fosters regional output growth. In particular, in Europe's fairly mature economic regions this quality channel for growth appears to be of larger importance compared to the traditionally employed quantity channel . In contrast to regional bank pro t e ciency, neither the e ect of credit volume relative to gross domestic product nor the interaction between quality and quantity are individually statistically di erent from zero.The existence and economic signi cance of this quality channel of banking on growth is corroborated by our nding that it exhibits the largest positive magnitude when jointly speci ed with the quantity and interaction e ects of nancial development. Notably, in this parsimonious model all three measures a ect growth signi cantly positive. However, the e ect of improving regional pro t e ciency is approximately three times as large as the respective two alternatives. We conclude that future studies on the nance-growth nexus should try to accommodate explicitly these three distinct mechanisms through which nance may foster growth.