THE LITERATURE ON CORPORATE SOCIAL RESPONSIBILITY AND PROFITABILITY  企業社會責任和收益文獻 

The first issue of Business and Society Review gave us the initial impetus for examining the relationship between corporate social responsibility and profitability. In that issue, Editor Milton Moskowitz suggested that socially responsible firms were good investment risks even though "there is at this point no real evidence that capital markets will be materially affected by social performance" (1972: 71). While Moskowitz made no explicit claim that such firms were good investment risks, he clearly implied it, and, in addition recommended 14 firms as potential investments because of their social performance: ". . . the securities are being suggested here on the basis of corporate behavior that can be considered socially responsive" (1972: 72). However, he never revealed what criteria he used in selecting these 14 firms.

第一期《商業和社會評論》為我們展開對企業社會責任和收益的調查提供了源動力。其中，Milton Moskowwitz主編認為：有著社會責任感的企業是一項不錯的投資風險---“即便沒有明確的證據表明社會績效對資本市場有著較大的影響”。（1972：:71）然而，Moskowitz并未聲稱這些公司有著良好的投資風險，但他明確的暗示了這一點，此外，因社會績效良好而推薦了14家潛在投資企業，“這些證券，是在企業行為是衡量社會責任的基礎上建立的”（1972:72）但是，他從未透漏選擇這14家企業的標準的什么。

The next issue of Business and Society Review observed that the 14 socially responsible firms identified by Moskowitz had registered a stock price increase of 7.28 percent over the previous six months, in contrast to a 4.4 percent rise for the Dow-Jones, a 5.1 percent increase for the New York Stock Exchange, and a 6.4 percent gain for Standard and Poors Industrials during that period. This finding was used to support the notion that socially responsible firms were good investment risks. Stanley Vance challenged the findings and claims of Moskowitz and Business and Society Review in a 1975 Management Review article. Vance examined the market performance of the 14 Moskowitz-recommended firms from 1972 to 1975 and found that stock in all of the firms had declined in price and had performed far below the Dow-Jones, the New York Stock Exchange Index, and the Standard and Poors Industrials (1975: 19). To support his tentative conclusion that socially responsible firms are not good investment risks, Vance extended his analysis to looking at the performance of firms identified as having high and low levels of corporate social responsibility in surveys reported by Business and Society Review. He compared the financial performance of the firms rated highest in corporate social responsibility with that of those rated lowest and found that the latter outperformed the former. Although he performed no statistical test to determine whether differences were significant, Vance did correlate corporate social responsibility and financial performance, and, finding a negative relationship, concluded that socially responsible firms were not good investments (1975: 24). There are similar methodological problems with the exploratory work of both Moskowitz (1972) and Vance (1975). The initial 14-firm sample used by each was small and subjectively selected. In addition, both studies relied upon a performance criterion that considered only capital gains or losses. There was also no adjustment for risk; not all stocks reflect the same degree of risk. In addition, the performance time period was short in the Moskowitz study. The Vance study used reputational surveys that reflected a response rate of 11 percent, with the typical responder rating only 20 of 45 firms considered.#p#分頁標題#e#

第二期的《商業和社會評論》指出，由Moskowitz指定的有社會責任感的14家注冊企業的股票價格在過去的六個月內上升了7.28%，同期的道瓊斯增加了4.4%，紐約證券交易所上升了5.1%，以及標準普爾工業指數的漲幅為6.4%。這一發現是對有社會責任感的企業是有著良好投資風險的有力證明。然而，斯坦利-萬斯在1975年的《管理評論》一文中對Moskowitz的觀點與《商業和社會評論》的數據顯示發出了挑戰。他從1972到1975年間對Moskowitz指定的14家企業進行了調查，發現這些企業的股票價格都有所下降，并且遠低于道瓊斯，紐約證券交易所和標準普爾工業的股票價格（1975:19）。為了證明他的初步結論---有社會責任感的企業不具備良好的投資風險，萬斯進一步對《商業和社會評論》所指出的具有較高的社會責任感的企業和較低的社會責任感的企業進行了調查分析。他對最具社會責任和最不具備社會責任的企業的財務績效進行了對比，發現后者優于前者。盡管沒有數據調查說明這些差異是否有意義，萬斯做了相關的企業社會責任和財務績效研究，發現它們之間存在著反比關系，從而說明有社會責任感的企業不具備良好的投資風險（1975-24）。Moskowitz（1972）和Vance（1975）的調查工作存在相似的方法論問題。兩者最初選用的14家企業規模都不大且是主觀性的。此外，兩項研究以只考量資本得失作為其行為準則。同樣，風險也沒有一定的標準，并不是所有的股票都能反映相同程度的風險。另外，Moskotiwz的調研時間太短，而萬斯使用的是調查為11% 的反應率，是典型的只包括了20到45家公司在內的信息回饋。

The article provides a chronological review of the major research efforts in this area; some of the more important of these research efforts will be briefly examined. Alexander and Buchholz (1978) did an important study that followed up on the efforts of Moskowitz and Vance. The major methodological difference between the Vance study and that of Alexander and Buchholz lies in the area of risk adjustment for the firms identified in the reputational survey. The latter study utilized the betas1 of each firm to adjust performance. When examining the issue of risk, Alexander and Buchholz were able to conclude that "there seems to be no significant relationship between stock risk levels and degree of social responsibility. These findings suggest that the interpretations of both Moskowitz and Vance are invalid" (1978: 485). Since the data were derived from both 3-year and 5-year assessment periods and were adjusted for risk, this study was a considerable improvement over its predecessors. However, as with the Vance study, the nature of the sample of firms remains a potential problem in that it relies upon the reputational studies and ratings provided by the Businessand Society Review. In addition, the use of stock prices as the criterion for performance is undesir- able given their inability necessarily to reflect on a firm's profitability.#p#分頁標題#e#
Bowman and Haire (1975) conducted a study that used a different approach in investigating the issue of the relationship between corporate social responsibility and profitability. The researchers, in identifying firms as low or high in social responsibility on the basis of the number of lines devoted to the topic of social responsibility in their annual reports, point out that: In searching for a readily available surrogate measure for actual activities in the area of corporate citizenship, we chose to measure the proportion of lines of prose in the annual report devoted to social responsibility. The annual report is a kind of projective test that allows a firm to express its goals and motives in much the same way that a Rorschach or TAT does for an individual .... A critic could immediately scoff at this measure. It is at least a popular belief that "everybody that talks about heaven ain't going there," that talk is cheap, and that talk about socially desirable behavior is not necessarily a predictor of such behavior (1975: 49-50). To validate this line-count method, the researchers cross-validated it by applying the method to Moskowitz's 14 firms having high levels of social responsibility and found them to have much more line space devoted to the topic of social responsibility than the 14 other randomly chosen firms (1975: 51). Using this line-count procedure, the authors classified 82 firms into high, medium, and low social-responsibility categories, and then evaluated each category on the basis of 5-year return on equity (ROE). The researchers found that the firms with medium ratings for degree of corporate social responsibility performed the best and the firms with low ratings performed the worst, indicating a U-shaped relationship between corporate social responsibility and firms' financial performance (1975: 51-53). Bowman and Haire's (1975) study exhibits numerous methodological problems. First, what is or is not a sentence or comment on corporate social responsibility can be difficult to ascertain, as the researchers themselves implicitly demonstrate (1975: 50). Second, the issue of validity also arises when assessment of corporate social responsibility is based on simple line count and cross-validated by 14 other firms whose level of social responsibility is also indeterminant, as we earlier observed. Third, the study included more (51) firms having low social responsibility than firms with moderate (18) or high (13) levels of social responsibility. Fourth, reliance on ROE as a measure of firm performance could be misleading since that return is a function not only of profitability, but also of a firm's financial leverage.

Finally, the researchers performed no significance tests, nor did they adjust performance for risk.A subsequent study by Abbott and Monsen (1979) employed a similar but more sophisticated methodology. The researchers used a content analysis of Fortune 500 annual reports performed annually by the accounting firm of Ernst and Ernst. This content analysis involves 28 items monitored in the annual reports; the content analysis is then used to construct a Social Involvement Disclosure (SID) scale that Abbott and Monsen used as a surrogate for corporate social responsibility. They divided 450 firms from the Fortune 500 into high and low groups on the basis of this scale and then examined each group for profitability. They discovered little difference in investment yield between firms in the two groups, even when controlling for size. They concluded that: "Being socially involved does not appear to increase investor's total rate of return. Nor does it appear that being socially involved is dysfunctional for the investor" (1979: 514-515). Some methodological problems exist with this study, as well. The annual report method used to assess corporate social responsibility may be superior to that used by Bowman and Haire, but it is still subject to validity problems. In addition, there was no adjustment for risk, and the performance criterion of investor's yield is not necessarily an adequate surrogate for profitability: yield is a function of both capital gains and dividends, neither of which need be tied directly to profitability. Parket and Eilbirt (1975) conducted a study that took still another approach. In a previous study of corporate social responsibility the researchers had been able to get 96 firms from the Forbes 1971 Annual Directory to respond; they concluded that, since these firms had responded, they were clearly more oriented toward social responsibility than were nonrespondents. Parket and Eilbert point out that: The fact that all ninety-six of the replying forms identified them-selves as engaged in endeavors associated with social responsibility suggests that firms not actively undertaking such work are more heavily represented among our nonrespondents (1975: 6). They then compared  80 alleged socially responsible firms to the Fortune 500 firms (minus these 80 firms) on the performance criteria of dollar net income, profit margin, ROE, and earnings per share (EPS). The researchers conclude: "By all four measures, the 80 respondents who were considered to be the most socially active show up as more profitable" (1975: 8).However, no significance test was performed, and it appears that the differences in both ROE and EPS are insignificant between the firms identified as socially responsible and other firms. Other methodological limitations exist. One problem was the assumption that the 80 firms in the sample had demonstrated a socially responsible orientation because they responded to a previous survey. Also, the data analysis methods were incomplete; there was no risk adjustment, and the profitability measures employed are not definitive and cover only one year.#p#分頁標題#e#

Another major research effort in this area was based on the judgment of Moskowitz, who had classified 67 firms over time as essentially high, moderate, and low in corporate responsibility (Sturdivant & Ginter, 1977). The researchers note that: The study was based on the sixty-seven corporations that had been cited by business journalist Milton Moskowitz as exhibiting exceptional social responsiveness or lack thereof. While no claim can be made about the accuracy of these ratings, they had the advantage of consistency in that they came from a single source (1977: 30). Sturdivant and Ginter used this sample to derive yet a smaller 28 firm sample that they subdivided into four industrial groupings. They compared firms showing high, moderate, and low social responsibility in each grouping on the basis of 10-year EPS growth, and then normalized each firm by dividing growth by the industry average. They found that firms from the high and moderate groups outperformed those from the low group (1977: 38). However, Sturdivant and Ginter did not really mention the fact that firms in the moderate group were the best performers, a result similar to what Bowman and Haire had discovered earlier. A number of methodological problems exist in this study, the first of which was sample selection: Sturdivant and Ginter derived their sample from a single source whose judgment was used in classifying the various firms in terms of orientation to corporate social responsibility. No criteria were offered for this classification. Moreover, the four industrial groupings reflect inconsistencies: for instance, Weyerhaeuser was grouped with U.S. Steel, Giant Food with S. S. Kresge Company, and Ralston Purina with Campbell Soup. In addition, the final sample, having been reduced to 28 firms, was small; there was no adjustment for risk; and the performance measure of growth in earnings per share is not definitive. These studies reflect both varying methodologies and different degrees of rigor. Although reputational surveys and content analysis of annual reports do provide useful beginning points, other exploratory methods also exist. Also, it is surprising that so much research has been based on the value orientations of a single business critic, and that none of the studies used a financial performance measure, like return on assets, that is less susceptible to corporate manipulation. Only one study realized the critical importance of adjusting performance on the basis of risk. The two studies employing the most rigor (Abbott & Monsen, 1979; Alexander & Buchholz,1978) found no relationship between corporate social responsibility and financial performance. However, two studies employing different methodologies (Bowman & Haire, 1975; Sturdivant & Ginter, 1977) found a curvilinear relationship between corporate social responsibility and financial performance, with moderately socially responsible firms being best performers.