Last updated 10th July 2024: Online ordering is currently unavailable due to technical issues. We apologise for any delays responding to customers while we resolve this. For further updates please visit our website https://www.cambridge.org/news-and-insights/technical-incident
We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
Published online by Cambridge University Press: 19 October 2009
A number of studies have recently attempted to explain cross sectional variations in selected characteristics of commercial bank operations through least squares regression analysis.1 Generally, whether the particular study objective was to explain inter bank differences in costs, revenues, profits, or loan rates, several explanatory variables were isolated which varied systematically among banks, and which at least partially explained differences in the dependent variable. The explanatory variables which were shown to be theoretically relevant, and statistically significant, were then interpreted literally and designated as having causal characteristics.
1 See, for example, Schweiger, Irving and McGee, John S., “Chicago Banking,” The Journal of Business, July 1961, pp. 203–366Google Scholar; Gramley, Lyle E., A Study of Scale Economies in Banking, Research Department, Federal Reserve Bank of Kansas City, Kansas City, Missouri, 1962Google Scholar; Flechsig, Theodore G., Banking Market Structure and Performance in Metropolitan Areas, Board of Governors of The Federal Reserve System, Washington, D.C., 1965Google Scholar; Edwards, Franklin R., Concentration and Competition in Commercial Banking: A Statistical Study, Research Report No. 26, Federal Reserve Bank of Boston, Boston, Massachusetts, 1964Google Scholar; Mayne, Lucille Stringer, The Cost of Federal Reserve System Membership, Research Paper No. 2, Department of Economics and Research, The American Bankers Association, New York, 1967Google Scholar; Morrison, George R. and Selden, Richard T., Time Deposit Growth and The Employment of Bank Funds, Association of Reserve City Bankers, Chicago, Illinois, 1965Google Scholar; and Benston, George James, “Economies of Scale and Marginal Costs in Banking Operations,” The National Banking Review, June 1965, pp. 507–549.Google Scholar
2 The banks were stratified according to deposit size.
3 For a general discussion of principal component analysis and some examples of its use see Kendall, M. G., A Course in Multivariabe Analysis (New York: Hafner, 1957), pp. 10–36Google Scholar, 70–75 or Morrison, Donald F., Multivariate Statistical Methods (New York: McGraw-Hill, 1967), pp. 221–258Google Scholar. For other examples of the application of component analysis in economics, see Meyer, John R. and Kraft, Gerald, “The Evaluation of Statistical Costing Techniques as Applied in the Transportation Industry”,: American Economic Review, Vol. LI, No. 2 (May 1961), pp. 313–334Google Scholar, and Thurow, Lester, “The Causes of Poverty”, The Quarterly Journal of Economics, Vol. LXXXI (February 1967), pp. 39–57CrossRefGoogle Scholar. For a relatively simple and lucid discussion of the general factor analysis technique, see Rummel, R. J., “Understanding Factor Analysis”, The Journal of Conflict Resolution, Vol. XI, No. 4 (December 1967), pp. 444–480.CrossRefGoogle Scholar
4 All variables were standardized before introduction into the dimension or principal component analysis.
5 An identical principal component analysis was carried out using bank data for the year 1963. However, the results were virtually the same, so the 1961 analysis is presented here in order that the exogenous variables for which only 1959 and 1960 data are available correspond more closely in time to the bank variables.
6 The employee salaries variable might also have been included in this group. Since for any given factor the more highly interrelated variables are those with the relatively largest loadings there is room for judgment with regard to where to begin excluding variables, especially when the highly loaded variables are going to be introduced into a separate component analysis as is the case here. In this instance the employees salaries variable was excluded (.2664) partly because the other salary level variable, officers salaries (.3887), was already in the set.
7 Much has been written suggesting that factor costs tend to be relatively lower outside large urban areas. See, for example, Hoover, Edgar M., The Location of Economic Activity (New York: McGraw-Hill, 1948), pp. 105–111Google Scholar; Losch, August, The Economics of Location (New Haven, Conn.: Yale University Press, 1954), pp. 455–461Google Scholar; Thompson, Wilbur R., A Preface to Urban Economics (Baltimore, Md.: The Johns Hopkins Press, 1965), pp. 87–89Google Scholar; and Nourse, Hugh O., Regional Economics (New York: McGraw-Hill, 1968), pp. 71–75.Google Scholar
8 The constraint that the variables not be highly loaded on the first two 21-variable components helps assure that the third subset of variables will berelatively independent from the first two subsets previously identified.
9 The service charge variable does not reflect exogenous economic forces as explicitly as the other three variables. It is included here because it is closely interrelated with the three exogenous influence variables showing that sample banks in the relatively low income-high unemployment areas tend to levy relatively low service charges on deposit accounts.
10 See footnote one.
11 See Kendall, op. cit, pp. 10–36.
