Abstract
A Data Envelopment Analysis (DEA) cost minimization model is employed to estimate the cost to thrift institutions of achieving a rating of ‘outstanding’ under the anti-redlining Community Reinvestment Act, which is viewed as an act of voluntary Corporate Social Responsibility (CSR). There is no difference in overall cost efficiency between ‘outstanding’ and minimally compliant ‘satisfactory’ thrifts. However, the sources of cost inefficiency do differ, and an ‘outstanding’ rating involves annual extra cost of $6.547 million or, 1.2% of total costs. This added cost is the shadow price of CSR since it is not an explicit output or input in the DEA cost model. Before and after-tax rates of return are the same for the ‘outstanding’ and ‘satisfactory’ thrifts, which implies a recoupment of the extra cost. The findings are consistent with CSR as a management choice based on balancing marginal cost and marginal revenue. An incidental finding is that larger thrifts are less efficient.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
All FDIC insured depository institutions are subject to CRA. Each is examined for CRA compliance by the regulatory agency under whose jurisdiction it falls. Credit Unions and independent mortgage companies are not subject to CRA. OTS regulates all federally chartered and many state chartered thrifts.
The large bank CRA examination consists of a lending test, an investment test and a service test. The institution is rated ‘outstanding’, ‘satisfactory’, ‘needs to improve’ or ‘substantial noncompliance’ in each category, which maps into an overall CRA rating. The size category ‘large’ includes institutions belonging to a bank holding company with assets over $1 billion, regardless of its own assets. The examination procedures put in effect July 1, 1997 emphasize flexibility and are meant to focus on outcomes rather than rigid rules and ratios in order to minimize the regulatory burden.
Some critics have charged that CRA ratings are too subjective. But Dahl et al. (2003) provide evidence that the scheduling of CRA exams and grading by regulators is significantly related to objective factors such as the volume of residential lending during the period 1990–1996.
The theoretical case for CRA-type regulations is asymmetric information: the potential borrower has better information about his credit risk than the lender. This may lead to adverse selection and credit rationing. But this outcome is contingent upon banks charging higher interest rates in response to imperfect information. If fears of bias override the concerns about information, banks may not raise rates and therefore not cause credit worthy borrowers to drop out of the market. There is widespread belief in financial markets that modern computerized data bases have largely eliminated asymmetric information.
The above cited industry estimate of $115,000 to comply with CRA is difficult to reconcile with such a low incidence of Outstanding performance. Either banks view the benefits of CRA ‘outstanding’, e.g. enhanced reputation, as trivial or, the compliance cost is much more significant—which is what we find.
This is a significantly higher proportion of outstanding banks than is typical. For example, Dahl (op.cit) report that only 11% of commercial banks were rated ‘outstanding’ and 81% rated ‘satisfactory’. Originally, our data set contained 140 large urban banks, 3 of which were rated ‘needs to improve’. These were dropped as being too few upon which to base any significant conclusions.
In response to charges that CRA ratings are too subjective more explicit examination guidelines were promulgated, effective July 1, 1997 for large banks. Since rating can persist for up to three years between examinations, some of those used here may straddle the changeover.
Efficiency of 0.75 means that firms could reduce their inputs to three-fourths of existing amounts without reducing observed output. Alternately, they are using 33% more inputs than necessary: (1 − 0.75)/0.75 = 0.33.
Loan risk is net loan charge-offs as a percent of gross loans as an annual rate for the period 1996–98 (Savings & Loans/Savings Bank Financial Quarterly 1998). This risk rate ranged from − 0.11% to 3.57%. Each loan risk value is transformed by multiplying by “ − 1” and adding 3.57. Only two observations had negative loan risk. The transformed risk variable now ranges from 0 (the highest risk) to 3.68 (the lowest risk). The input oriented VRS model is invariant to output (but not input) translation. See Cooper et al. (2000), 94 and 228, and Zhu (2003), 107.
Hermalin and Wallace (op. cit.) use the six-month Treasury bill rate to proxy the cost of equity capital and a set of inputs similar to those used here.
Net bank debt is defined as loans and investments minus equity capital and loan loss reserves. This piece of data is also purchased from IDC Financial. All the other input and output data is from OTS reports.
This is a two step procedure in which RTS is determined using a constant returns to scale and variable returns to scale input oriented production function. All DMUs with same efficiency scores are in the region of constant returns. For all others, if \(\sum\lambda>1\), increasing returns prevail, and if \(\sum\lambda>1\), the DMU is in the region of decreasing returns.
E c = 0.88 implies that actual cost is 0.136 above least cost.
The z-test statistic is 1.13 as compared the critical value of 1.96 (95%, two tailed test).
The Herfindahl Index (HI) is defined as the sum of the squared market shares (S i) of each firm, \(\hbox{HI}=\sum(S_{\rm i})^{2},\) summed over i firms. Since market shares are expressed as decimals, a pure monopoly corresponds to HI = 1. In the second step regression, 131 of 135 banks are located in metropolitan areas with HI < 0.20. Two observations had to be dropped because it was not possible to identify them in the FDIC data. They may have changed names, merged or gone out of business.
The marginal effect measures the change in the expected value of the dependent variable from going from CRA = 0 to CRA = 1, holding all other variables at their mean values.
ROE = 11.66+0.407(CRA) (t = 13.35) (t = 0.25); After Tax ROE = 11.11+0.540(CRA) (t = 14.68) (t = 0.38). ROE data from IDC(1998).
References
Berger AN, Humphrey DB (1997) Efficiency of financial institutions: international survey and directions for future research. Eur J Oper Res 98:175–212
Besanko D, Dranove D, Shanley M, Schaefer S (2004) Economics of strategy. John Wiley and Sons, New York
Cebenoyan AS,Cooperman ES, Register CA, Hudgins S (1993a) The relative efficiency of stock vs. Mutual S & Ls: a stochastic cost frontier approach. J Financ Serv Res 7:151–170
Cebenoyan AS, Cooperman ES, Register CA (1993b) Firm inefficiency and the regulatory closure of S & Ls: an empirical investigation. Rev Econ Stat 75:540–545
Coelli T, Rao Prasada DS, Battese GE (1998) An introduction to efficiency and productivity analysis. Kluwer Academic Publishers, Boston
Cooper WW, Seiford LM, Tone K (2000) Data envelopment analysis. Kluwer Academic Publishers, Boston
Dahl D, Evanoff DD, Spivey MF (2003) The timing and persistence of CRA compliance ratings. J Financ Serv Res 23:2:113–132
Federal Deposit Insurance Corporation. Statistics on Deposits. http://www.fdic.gov/SOD
Greene WH (2000) Econometric analysis, 4th edn. Prentice-Hall, Upper Saddle River, N.J.
Hermalin BE, Wallace NE (1994) The determinants of efficiency and solvency in savings and loans. Rand J Econ 25:361–381
Independent Community Bankers of America. (2002) The high cost of community bank CRA compliance (August). Washington, D.C. http://www.grantthornton.com
Joint Center for Housing Studies (2002). The 25th anniversary of the Community Reinvestment Act: Access to capital in an evolving financial services system. Harvard University, Cambridge, MA
McWilliams A, Siegel D (2001) Corporate Social Responsibility: a theory of the firm perspective. Academy of Management Review 26(1):117–127
Mester LJ (1993) Efficiency in the savings and loan industry. J Bank Financ 17:267–286
Office of Thrift Supervision. 2002 Fact Book. http://www.ots.gov
S&L-Savings Bank Financial Quarterly, Fourth Quarter (1998) Hartland. IDC Financial Publishing, Wisconsin
Thakor AV, Beltz JC (1993) An empirical analysis of the costs of regulatory compliance. Federal Reserve Bank of Chicago, Annual Conference (May)
Viscusi WK, Vernon JM, Harrington JE Jr (1995) Economics of regulation and antitrust, 2nd edn. The MIT Press, Cambridge, MA
Walker HM, Lev J (1953) Statistical inference. Henry Holt & Co, New York
Zhu J (2003) Quantitative models for performance evaluation and benchmarking. Kluwer Academic Publishers, Boston
Author information
Authors and Affiliations
Corresponding author
Additional information
An earlier version of this paper was presented at the North American Productivity Workshop 2004, University of Toronto, June 25, 2004, Session FB33.
Appendices
Data Appendix
Data sample
The 137 large urban savings banks and savings and loans analyzed in this paper are drawn from an OTS population of 305 thrifts with assets of $250 million or more. The percentage distribution of the sample by OTS defined asset categories is as follows, with the OTS population distribution in parentheses (OTS 2002):
Our sample over weights the largest thrifts, which is appropriate since they are fewer in absolute numbers and because CRA ‘outstanding’ is more common among the biggest thrifts. The OTS population includes non-urban institutions. Thrifts belonging to a holding company with assets in excess of $1 billion are treated as ‘large’ for purposes of CRA, even if its assets fall below the $250 million threshold. None of the banks in this data set is in that category.
Rights and permissions
About this article
Cite this article
Vitaliano, D.F., Stella, G.P. The cost of Corporate Social Responsibility: the case of the Community Reinvestment Act. J Prod Anal 26, 235–244 (2006). https://doi.org/10.1007/s11123-006-0018-2
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11123-006-0018-2