THE EXISTENCE OF ECONOMIES OF STRUCTURE AND OF SCALE IN COMMERCIAL BANKING*

Abstract

A NUMBER OF EMPIRICAL STUDIES of branchand unit-bank costs have been undertaken in the past few years. These studies have sought to relate bank cost to bank with little attention being given in most of them to the definition of bank output. All but a few have recognized the relevance of a real-valued index of bank output, as opposed to a separate cost-analysis of each banking service, in a publicpolicy-oriented study of branch versus unit banking, but have chosen arbitrary proxies for that index (assets, deposits, loans, and investments, etc.). Stuart Greenbaum was the first to give detailed consideration to the construction of an output index. Greenbaum felt that a bank-output index should not reflect interbank interest differences. This paper establishes the necessity of including interbank interest differences in an output index that is intended to measure the service provided by a bank. An interest-insensitive index is appropriate for intrabank decision-making but inappropriate in a study intended to guide the formation of public policy on banking structure. An operational definition of bank output is defined and is used in the estimation of long-run average-cost curves for branch and for unit banks in the Seventh Federal Reserve District after extensive stratification by both bank size and product-mix. The index consists of current operating revenue and a measure of the bank's timedeposit output. Considerable work is done in the appendices to this study with alternative measures of bank output, the principal concern being the effect of removing interbank interest differences from the output index and of deleting the measure of time-deposit output. Although the extensive stratification prevents a definitive acceptance of the hypotheses that branch banks are less efficient than unit banks and that economies of scale exist for both branch and unit banks, there is strong supportive evidence. This evidence challenges, as does that of previous studies, the existence of an empirical foundation for the continuing policy bias in favor of branch banking.