A Comparison and Test of the Use of Accounting and Stock Market Data in Relating Corporate Social Responsibility and Financial Performance

Abstract

A Comparison and Test of the Use of Accounting and Stock Market Data in Relating Corporate Social Responsibility and Financial Performance Management and finance can be enriched tremendously by research that crosses boundary lines[9, 28, 33, 38, 40]. Each field, however, may occasionally suffer from discipline bias and misunderstanding of the other. One discipline, for example, may use the tools and techniques of the other without comprehending their full ramifications. The relationship between corporate social responsibility (CSR) and financial performance is a topic that merits careful interdisciplinary examination. In this paper we explore the problems inherent in employing accounting data as a measure of financial performance, and we test the appropriateness of using accounting data in CSR research. We compare the results of a recent stock market based study employing an event test[18] with accounting returns computed from the same sample and show that the accounting returns are not sensitive to CSR performance. PAST CSR RESEARCH Researchers have reached no clear consensus on the relationships between corporate social responsibility (CSR) and subsequent financial performance. Of the 25 published studies examined, 14 reported a positive association[3, 8, 10, 12, 15, 16, 25, 26, 37, 39, 41, 44, 45, 46]. Seven studies[1, 2, 6, 21, 22, 30, 35,] reported no real link between these variables. Four studies[23, 27, 47, 50] found a negative relationship, and two[11, 48] described a U-shaped performance curve where the highest performing firms were those found in the middle range of CSR. Reviews of these contradictory findings (Arlow and Gannon[5], Cochran and Wood[16], Ullmann[49] suggest three major explanations for the lack of consensus on this issue: (1) the use of questionable indexes of social responsibility, (2) inadequate sampling techniques, and (3) poor measures of financial performance. The latter observation is particularly noteworthy. There appears to be a false belief in the minds of many that the methodological complexities of developing financial performance measurements appropriate for CSR studies are now fairly well appreciated[32, p. 3]. In the same year, Aupperle, Caroll, and Hatfield[6, p. 446] assert that profitability is a relatively clear-cut process, but assessing social responsibility is not. The purpose of this study is to demonstrate the fallacy of this assumption in the case of accounting-determined profitability measures that have been used in the majority of the studies relating CSR to financial performance. Specifically, Abbott and Monsen[1], Aupperle, Carroll, and Hatfield[6], Bowman[10], Bowman and Haire[11], Bragdon and Marlin[12], Chen and Metcalf[15], Cochran and Wood[16], Freedman and Jaggi[22], Fry and Hock[23], Heinz[25], Ingram and Frazier[27], Kedia and Kuntz[30], McGuire, Sundgren, and Schneeweis[35], Parket and Elibirt[39], Preston[41], Spencer and Taylor[45], Spicer[46], and Sturdivant and Ginter[48] all use net income, ROE, ROA, EPS, P/E ratios and/or variations of them as at least one measure of financial performance. Accounting measures of financial performance are inadequate for researchers making large cross-sectional comparisons across industries and across time. The problems with using accounting measures of financial performance in CSR studies fall into two categories. The first category includes general problems of accounting measures, and the second includes particular problems with specific measures of profitability. GENERAL PROBLEMS WITH ACCOUNTING DATA There are numerous general problems associated with the use of accounting data, particularly when cross-sectional comparisons are made. Some of these problems are discussed more completely in Merchant and Burns[36]. These problems include industry and regulatory differences, accounting and demographic differences, risk, leverage, inflation, and timing. …

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 105 citations exist upstream.