A note on the “risk-adjusted” price–concentration relationship in banking

• Published in: Journal of banking & finance 2006-03, Vol.30 (3), p.1041-1054
• Main Authors: Brewer, Elijah, Jackson, William E.
• Format: Article
• Language: English
• Subjects: Bank operations; Banking; Banking industry; Banking system; Business conditions; Commercial banks; Consumer deposit pricing; Economic concentration; Efficient-structure; Empirical tests; Financial analysis; Hypotheses; Interest rates-deposits; Market analysis; Price policy; Prices; Risk assessment; Risk-adjusted; Structure-conduct-performance; Studies; U.S.A
• ISSN: 0378-4266; 1872-6372
• DOI: 10.1016/j.jbankfin.2005.06.006

Price–concentration studies in banking typically find a significant and negative relationship between consumer deposit rates (i.e., prices) and market concentration. This implies highly concentrated banking markets are “bad” for depositors. It also provides support for the Structure-Conduct-Performance hypothesis and rejects the Efficient-Structure hypothesis. However, these studies have focused almost exclusively on supply-side control variables and neglected demand-side variables when estimating the reduced form price–concentration relationship. For example, previous studies have not included in their analysis bank-specific risk variables as measures of cross-sectional derived deposit demand. We find that when bank-specific risk variables are included in the analysis the magnitude of the relationship between deposit rates and market concentration decreases by over 50%. We offer an explanation for these results based on the correlation between a bank’s risk profile and the structure of the market in which it operates. These results suggest that it may be necessary to reconsider the well-established assumption that higher market concentration necessarily leads to anticompetitive deposit pricing behavior by commercial banks. This has direct implications for the antitrust evaluations of bank merger and acquisition proposals by regulatory agencies. And, in a more general sense, these results suggest that any Structure-Conduct-Performance based study that does not explicitly consider the possibility of very different risk profiles of the firms analyzed may indeed miss a very important set of explanatory variables. And, thus, the results from those studies may be spurious.