In this paper we consider some factors which are of potential importance in the debate concerning the sources of performance for intermediaries. Using data from depository institutions (banks and savings and loans), we find that the distributional intensity (provided by standardized number of offices in a market) is consistently important in explaining cross-sectional profitability. This result implies that the number of offices in a market is at least as important as more traditional measures of efficiency and concentration in determining returns in this sector of the financial services industry. Indeed, when pooled data are used, there is a strong quadratic relationship between return on assets and the number of offices in a market. We show that this relationship can be viewed as coming from spatially differentiated markets as opposed to collusion or efficiency per se. Finally, we provide evidence that results concerning the rule of efficiency versus market concentration are themselves sensitive to the implicit assumption that there are no close substitutes for the services provided by a sub-set of the industry. In particular, results from 'pooled' bank and thrift data often provide conclusions which are different from those which include only banks.