Creating Order out of Chaos: A Theoretical Meta-analysis of the
Corporate Social Performance and Corporate Financial Performance
Relationship
Master Thesis Finance University of Groningen Groningen, June 26, 2015
Abstract
This thesis investigates the relationship between corporate social performance and corporate financial performance and encompasses a sample of 192 studies on corporate social and financial performance from 1972 through 2009. We categorize corporate financial performance measures and corporate social performance measures to be able to link existing theories to specific measures. The aim is to explore all possible relationships between corporate social performance and corporate financial performance, and to examine how these theories relate to the empirical findings. We provide rankings to find which corporate social performance measures are most associated with these theories. The main conclusion is that corporate social performance measures and corporate financial performance measures are not clearly defined. We find heterogeneous results and due to lacking metrics we are unable to compare results appropriately. We confirm our assumption that there exist different effects when using different CSP and CFP measures.
Keywords: Corporate Social Performance, Corporate Financial Performance, Corporate Social Responsibility, Firm Value, Business Ethics
JEL-Codes: G00, G02, G10, M14
Author: Killian Westrik
Study: MSc Finance, Faculty of Economics and Business Student number: 1883879
2
1. Introduction
This thesis investigates the relationship between corporate social performance and corporate financial performance. The relationship between Corporate Social Performance (CSP) and Corporate Financial Performance (CFP) has been researched continuously during the past half-century. An underlying question to research concerning the relationship between Corporate Social Responsibility (CSR) and CFP is: Why does CSR exist? Dam et al. (2009) identify three reasons why firms conduct CSR (altruism, strategic reasons, and greenwashing) and provide evidence that supports the idea that strategic reasons is the predominant motive of CSR. A positive link between CSP and CFP would legitimize CSP based on economic grounds (Davis, 2009; Useem, 1996).
McWilliams et al. (2006) define CSR as a situation where a firm goes beyond compliance and engages in actions enhancing the social good without being required to do so. Dam et al. (2009) state that ‘’there are many definitions of CSR, but for most scholars CSR occurs when firms engage in activities that appear to advance a social, environmental or ethical agenda beyond that which is required by law (Siegel and Vitaliano, 2007; Lyon and Maxwell, 2008; Heal, 2008).’’ Orlitzky et al. (2003) state that the performance of business organizations is affected by their strategies and operations in market and non-market environments (Baron, 2000), and mention that one construct that might capture a major element of these non-market strategies is CSP.
According to Margolis et al. (2009), CSP has been theoretically defined in two basic ways, whereas ‘’one approach casts social performance as a multidimensional construct, encompassing a company’s efforts to fulfill multiple responsibilities – economic, legal, ethical, and discretionary (Carroll, 1979, 1999) – or encompassing a company’s principles, processes of response to rising issues, and observable practices and outcomes (Wartick and Cochran, 1985; Wood, 1991).’’ A second approach, according to Margolis et al. (2009), ‘’casts social performance as a function of how a company treats its stakeholders (Campbell, 2007; Clarkson, 1995; Cooper, 2004; Post et al., 2002).’’ Margolis et al. (2009) mention that CSP measures vary widely and tend to capture either a single specific dimension or broad assessments of CSP.
3 challenges in studying CSR are caused by a lack of effort to empirically test definitions, propositions, and concepts (Aupperle et al., 1985). Although this study was published 30 years ago, it seems that not much has changed. Allouche and Laroche (2005), for example, state that ‘’it is quite possible that the wide range of conceptualization and operationalization of CSP has resulted in varying strengths of relationships between CSP and CFP.’’ We do not find a clear answer to what CSP and CSR exactly are and how they interact.
Concerning the measurement of CFP, Allouche and Laroche (2005) follow Orlitsky et al. (2003) and distinguish between three broad subdivisions of CFP measures: market-based, accounting-based and perceptual measures. Allouche and Laroche (2005) state however, that these conceptual definitions embrace a wide range of operational definitions of CFP. According to Allouche and Laroche (2005): ‘’The choice between employing an accounting-based measure and a market-based measure carries theoretical and empirical implications even if both measures may be suited to answer questions about the CSP-CFP interaction’’. When concerning the measurement of CSP, we also do not find clarity.
Several studies thus indicate the wide range of conceptualization and operationalization of CSP and also the wide range of operational definitions of CFP (Aupperle et al., 1985; Allouche and Laroche, 2005; Margolis et al., 2009). Clearly there is a need for structure, as the main conclusion during the past half-century is that there is a lack of consistency and a wide range of conceptualization and operationalization concerning CSP and CFP. This thesis attempts to present overview concerning the CSP-CFP relationship. We create order out of this chaos concerning CSP and CFP. Also, we actively take a step forward and categorize CSP and CFP measures in a coherent way.
In this thesis, both CSP measures and CFP measures are categorized in such a manner that overlap between categories is impossible. Previous categorizations of CSP consist of considerable overlap between categories that explicitly name the area or dimension of CSP and categories that are methods of collecting data. Margolis et al. (2009) for example develop nine categories of CSP1. An example - to demonstrate the difference between this categorization of Margolis et al. (2009) and our categorization - is a study that uses the CEP2 pollution index to measure CSP. When using the categories of CSP of Margolis et al. (2009) the CEP pollution index is categorized as Third-party audit3. The information that the CEP pollution index measures pollution (and thus environment) would be lost using this category. In this thesis the CEP pollution index is categorized as Environment and
1
Margolis et al. (2009) categorize types of CSP as follows: (1) Corporate policies, (2) Disclosure, (3)
Environmental performance, (4) Philanthropic donations, (5) Revealed misdeeds, (6) Self-reported social performance, (7) Observers’ perceptions, (8) Third-party audits, (9) Screened mutual funds.
2
Council on Economic Priorities measures environmental performance in terms of indices, which we will refer to as CEP pollution index.
3
4 subcategorized as Pollution. The metric is added as additional layer, which would be Third-party audit for the CEP pollution index. Our categorization allows more sophisticated analysis and reflects more information.
Third-party audits as a CSP category could consist of both pollution measures and social measures, because it does not specify the dimension of CSP. This thesis attempts to make distinctions between the dimension and the methods of collecting data, and creates several layers to categorize CSP measures, which enables us to make comparisons on a more detailed level. After all, much research has been done about the CFP-CSP relationship but still there is a lack of comparable and generalizable results.
In this thesis, the underlying method to measure firm value is the enterprise Discounted Cash Flow (DCF) method, as this is the method underlying most research. Firm value will be discussed in more depth in the literature review. CFP measures are proxies for economic performance, whereas economic performance has impact on a company’s value. The assumption in this thesis is that these economic performance proxies measure different effects on economic performance. Therefore, CFP measures are categorized to expose these different effects. CSP measures are proxies for CSR. The assumption in this thesis is that these CSR proxies measure different aspects of social responsibility. Therefore, CSP measures are categorized to expose these different aspects. This thesis investigates the relationship between firm value and CSR, where firm value is assessed by CFP measures and CSR is measured by CSP measures. The assumption is that different effects when using different CFP measures and CSP measures exist. The connection between these CSP and CFP measures is described by theories. These theories and their hypotheses are presented in the literature review. This thesis attempts to link existing theories to CFP measures and CSP measures. The theories are mapped based on two characteristics: the direction and the causality of the relationship between CFP measures and CSP measures. After that, this thesis examines how these theories relate to the empirical findings.
The final sample in this thesis consists of 252 results from 192 studies. The data set contains studies from 1972 through 2009. We recognize more than 60 different CSP measures and more than 30 different CSP measures in our sample. Ten theories connecting these CSP and CFP measures are mapped and related to the empirical findings in this sample.
5
2.
Literature review
The first part of this chapter describes the results of the meta-analyses that were used as input for this thesis’ sample. Then, we discuss firm value. After that, the CSP categories of Margolis et al. (2009) are presented in more detail, because it is one of the few categorizations made regarding CSP and it is used as starting point for an improved categorization of CSP measures. Then, we look into CFP measures. Lastly, the existing theories and their hypotheses are presented.
Meta analyses
This thesis uses five meta-analyses as starting point and as input for the sample. The studies used as input are: Aupperle et al. (1985), Griffin and Mahon (1997), Orlitzky et al. (2003), Allouche and Laroche (2005) and Margolis et al. (2009). Table 1 displays information regarding these meta-analyses.
Orlitzky et al. (2003) state that they intend to show that the general conclusion – that there is little generalizable knowledge concerning CSP and CFP - is built on shaky grounds. Orlitzky et al. (2003) show there is a positive association between CSP and CFP across industries and across study contexts conducting a meta-analysis of 52 studies. Allouche and Laroche (2005) provide some support for the conclusions reached by Orlitzky et al. (2003) using 82 studies. One of their aims is to provide a statistical integration of the existing research on the relationship between CSP and firm financial performance (Allouche and Laroche, 2005). Their results show conclusively that CSP has a positive impact on CFP.
Margolis et al. (2009) ask the following question underlying their research: ‘’In an era of rising concern about financial performance and social ills, companies’ economic achievements and negative externalities prompt a common question: Does it pay to be good?’’ Margolis et al. (2009) find a mean effect size of r =.133, and this mean was inflated by large effects sizes from a small number of studies that used relatively smaller samples of companies. Margolis et al. (2009) find that 59% of the results reveal a non-significant relationship, 28% a positive relationship, and 2% a negative relationship between CSP and CFP.
6
Table 1 – Overview of author(s) and year, amount of studies used in meta-analyses, the time period, the method and main conclusion/issues of the meta-analyses used as input for this thesis’ sample Author(s) and Year Amount of studies meta-analyses part / time period
Method Main conclusion/issues
Aupperle et al. (1985)
10 studies from 1972 through 1977
Theoretical examination major research efforts concerning CSP and CFP.
The relationship between CSR and profitability has frequently reflected either an ideological bias or limited methodological procedures.
Griffin and Mahon (1997)
51 studies from 1972 through 1994
Review articles for population tested, data source(s), methodologies employed, control variables, CFP and CSP measures, results and significance level, findings, and
reliability/validity testing.
Three key issues emerged: the focus on multi-industry samples, the multiple dimensions of CFP, and the need for multiple measures to assess CSP. Orlitzky et al. (2003) 52 studies from 1972 through 1997 Effect-size (r) meta- analysis, which calculates population parameter estimates (𝜌) by
correcting for the lack of validity of narrative reviews.
Findings suggest that
corporate virtue in the form of social responsibility and, to a lesser extent, environmental responsibility is likely to pay off, although the
operationalization of CSP and CFP also moderate the positive association. Allouche and Laroche (2005) 82 studies from 1972 through 2003 Meta-Significance Testing, Sample-size weighted average partial correlation.
They conclude that the Meta-Significance Testing supports strongly the finding of a positive association between CSP and CFP. Margolis et al. (2009) 214 studies from 1972 through 2007 Correlation r as common metric for the pool of studies, effect sizes all studies must be summarized using common metric.
The overall effect is positive but small (mean r =.13, median r =.09, weighted r = .11), and results for the 106 studies from the past decade are even smaller.
7 the assumption is for instance that accounting measures and market measures assess the same performance. This, however, is not the case. Dam (2006a) states that, with externalities, changes in financial performance measures must be interpreted differently. He names the negative effect of lower profits on the stock market value of the firm, but also a positive effect on stock market value when the internalization of external effects is valued by (socially responsible) stockholders (Dam, 2006a). Meta-analyses, such as Wu (2006), Orlitzky et al. (2003), and Margolis et al. (2009), try to capture the CFP-CSP relationship in one common metric. In this thesis, however, the assumption is that there are different effects when using different CSP measures and CFP measures.
Firm value
‘’A company’s value is driven by its ability to earn a healthy Return On Invested Capital4 (ROIC) and by its ability to grow’’ (Koller et al., 2010). Firm value can be measured in several ways. Koller et al. (2010) and Hillier et al. (2012) present several methods to measure firm value5. This thesis assumes the DCF approach as underlying model to measure firm value. The enterprise DCF model discounts free cash flows, meaning the cash flow available to all investors – equity holders, debt holders, and any other nonequity investors – at the weighted average cost of capital, meaning the blended cost for all investor capital (Koller et al., 2010). Following Koller et al. (2010), value is the sum of the present values of future expected cash flows, a point-in-time measure, and value creation is the change in value due to company performance. According to Hillier et al. (2012), with riskless cash flows, the Net Present Value (NPV) method is equivalent to the DCF method, which obtains the NPV by discounting all cash flows at the rate(s) of return prevalent in the securities markets and adding the discounted cash flows together. Koller et al. (2010) state that companies’ values can also be expressed as price-to-earnings ratios (P/Es). Koller et al. (2010) present evidence that investors do indeed value long-term cash flow, growth, and return on invested capital, and companies that perform well on those measures perform well in the market. The basis of valuing a company is to use the well-known cash flow perpetuity formula (Copeland and Weston, 1988):
𝑉𝑎𝑙𝑢𝑒 =
𝑊𝐴𝐶𝐶−𝑔𝐹𝐶𝐹𝑡=1 (1)Koller et al. (2010) adjust this formula to their key-value driver formula, which is the equation underpinning discounted cash flow (DCF) valuation in both theory and practice:
𝑉𝑎𝑙𝑢𝑒 = 𝑁𝑂𝑃𝐿𝐴𝑇𝑡=1(1− 𝑔 𝑅𝑂𝐼𝐶) 𝑊𝐴𝐶𝐶−𝑔 (2) 4
A simple definition of return on invested capital is after-tax operating profit divided by invested capital (working capital plus fixed assets) (Koller et al., 2010).
5
8 The fundamental drivers of economic value can be observed from the key-value driver formula: growth, ROIC, and cost of capital. Koller et al. (2010) identify revenue growth and ROIC as value drivers, which lead to cash flows. These cash flows then are discounted to estimate value. The cost of capital is defined as ‘’the price charged by investors for bearing the risk that the company’s future cash flows may differ from what they anticipate when they make the investment’’ (Koller et al., 2010). They state that the standard model for measuring differences in costs of capital has been the Capital Asset Pricing Model (CAPM).
Koller et al. (2010) state ‘’there is no universal agreement on value and that many still treat accounting earnings and value as one and the same, and focus almost obsessively on improved earnings.’’ They say that while earnings and cash flow are often correlated, earnings don’t tell the whole story of value creation. This overlapping nature of firm value, earnings, and growth and disagreement about value is embedded in the studies in this sample. Additionally, adjusting accounting performance to operating performance in order to use the key-driver formula is time-consuming. Koller et al. (2010) state that using accounting rates of return is incorrect and that accounting profits often are very different from cash flows generated by a project. Most studies don’t explain in which perspective they analyze. In this thesis, both the cash flow perpetuity formula and key-driver formula are assumed to be good proxies of firm value.
CSP categories
Orlitzky et al. (2003) measure CSP in several ways: CSP, CSP without CSP reputation, CSP without environmental performance, and corporate environmental performance. Allouche and Laroche (2005) use these same conceptualizations and add one more way to measure CSP: CSP with philanthropic donation. They mention that the narrow definition of social performance excludes environmental performance. Wu (2006) creates CSP subgroups for analysis, which are Social Concern, Social Action, Reputation Rating, and Asset Utilization.
9
Table 2 – Overview of CSP categories as used in Margolis et al. (2009) with their short definitions
CSP Category Short definition
Corporate Policies These studies examine a range of corporate policies
Disclosure The release of information by a company itself in
publicly available documents
Environmental Performance Measures of impact on the environment, whether objective or self-reported
Philanthropic Donation Cash and in-kind donations or the establishment of a philanthropic foundation
Revealed Misdeeds The public announcements of arrests, fines, guilty verdicts in lawsuits, involuntary recalls, and other actions that indicate socially irresponsible behavior
Self-reported Social Performance Survey that ask companies to report their own conduct in response to journalists’ or researchers’ inquiries
Observers’ Perceptions Relies upon observers’ intuitive impressions of a company’s CSP
Third-Party Audits Involves the systematic assessment of data by investigators who evaluate a company along a set of criteria
Screened Mutual Funds Performance of mutual funds that use screens to limit the companies included in the funds to these meeting certain criteria of social performance
10
CFP categories
Margolis et al. (2009) and Wu (2006) categorize CFP measures into two broad categories: market-based and accounting-based measures. Allouche and Laroche (2005) follow Orlitsky et al. (2003) and distinguish between three broad subdivisions of CFP measures: market-based, accounting-based and perceptual measures.
Allouche and Laroche (2005) state that ‘’the choice between employing an accounting-based measure or a market-based measure carries theoretical and empirical implications, even if both measures may be suited to answer questions about the CSP-CFP interaction.’’ Margolis and Walsh (2002) state that ‘’without a clear causal theory linking CSP and CFP, the prudent approach is to use both sets of measures and let the empirical evidence inform our theoretical underpinning.’’ The categorization of CFP is thus more clear, but the conceptual definitions embrace a wide range of operational definitions of CSP (Allouche and Laroche, 2005).
Gregory et al. (2014) state ‘’accounting measures are backward looking and their objectivity and informational value can be questioned (Benston, 1982). Stock market measures, by contrast, are forward looking with expectations of future cash flows embedded within the stock price, and they are more relevant for considering the implications of CSR for investors.’’
McWilliams, Siegel and Wright (2006) argue that stock price only relates to financial stakeholders whereas non-financial stakeholders are also affected by CSR activities. Scholtens (2008), however, argues that ‘’market-based measures are less susceptible to accounting rules and managerial manipulation because they refer to investors’ evaluations and expectations of firm performance.’’ Accounting measures are thus equally relevant to most stakeholders, whereas market measures are mainly relevant to investors.
Dam and Scholtens (2015) assess the relationship between (indicators of) financial performance and social responsibility using their theoretical framework6. Dam and Scholtens (2015) propose that from the demand side (investors) social responsibility is to be associated with higher market-to-book ratios and that from the supply side they find that more responsible firms have a higher return on their assets. Dam and Scholtens (2015) conclude that when firms announce social responsibility, this may lead to lower stock market returns, but not necessarily to a lower firm value. ‘’From the investor perspective, social responsibility is to be associated with a higher market value in relation to the book value of the assets of the firm’’ (Dam and Scholtens, 2015). And lastly, from the corporate perspective, they show that more responsible firms have a higher return on their assets
6
‘’Key assumptions are that socially responsible investors value taking account of social damage associated with production. As a result, they accept a lower financial return on responsible stock compared to
11 (Dam and Scholtens, 2015). They clearly show different effects using different performance measures.
Concluding, the literature above shows many different CSP measures and categories, and also different CFP measures and categories. Whereas other meta-analyses such as Orlitzky et al. (2003) and Margolis et al. (2009) find one overall size effect, we do not agree with the method to calculate one overall effect. This thesis categorizes both CSP and CFP measures to assess the CSP-CFP relationship on a more detailed level. We aim to link existing theories and hypothesis to empirical findings, using our sample of 192 studies.
2.1 Existing theories and their hypotheses
The theories concerning the relationship between CSP and CFP describe the possible connection between CSP measures and CFP measures. We use the name of the theory same as stated in the literature, which leads to view/perspective/theory all being used as synonyms.
These theories concerning the relationship between CSP and CFP can be classified into two dimensions. The first dimension is the direction of the relationship between CSP measures and CFP measures, which could be positive, negative, mixed, or non-existent. The positive direction and negative direction are both linear relationships and the mixed direction indicates a non-linear relationship. The second dimension is the causal relationship between CSP measures and CFP measures involved. As Preston and O’Bannon (1997) state: ‘’Does social performance influence financial performance; does financial performance influence social performance; or, is there a synergistic relationship between the two?’’ This leads to the following causalities: social performance could precede financial performance, financial performance could precede social performance, or there could exist a synergistic7 relationship between social and financial performance. Preston and O’Bannon (1997) lay the groundwork for these theories and Orlitzky et al (2003) and Allouche and Laroche (2005) build further on this. The hypotheses that have an opposite direction due to another theory stating the opposite are divided into for example H1a and H1b. Hypothesis that only have a number, for example H2, thus do not have an theory with a hypothesis stated in the opposite direction. In this section the theories concerning the relationship between CFP and CFP will be presented and their hypotheses will be stated.
7
12
Stakeholder theory/Good management theory
Stakeholder theory8 suggests that CSP is positively associated with CFP because it enhances the satisfaction of various stakeholders - and consequently the firm’s external reputation – and leads to better financial performance (Allouche and Laroche, 2005). This theory is equivalent to the social impact hypothesis discussed in Preston and O’Bannon (1997). According to Preston and O’Bannon (1997) ‘’the ‘social impact’ version of the stakeholder theory implied a lead-lag relationship between social and financial performance; external reputation (favorable or unfavorable) develops first, then financial results (favorable or unfavorable) follow.’’ According to Orlitzky et al. (2003) managers can increase the efficiency of their organization’s adaption to external demands by addressing and balancing the claims of multiple stakeholders. Donaldson and Preston (1995) state that the widely believed notion is that stakeholder management contributes to successful economic performance, but they add that this is insufficient to stand alone as a basis for the stakeholder theory. They state that ‘’studies have tended to generate implications suggesting that adherence to stakeholder principles and practices achieves conventional corporate performance objectives as well or better than rival approaches’’ (Donaldson and Preston, 1995).
Another theory, primarily derived from stakeholder theory, is the good management theory (Jones, 1995; Waddock and Graves, 1997). This theory states that ‘’high CSP bolsters a company’s competitive advantage by weighing and addressing the claims of various constituents in a fair and rational matter’’ (Orlitzky et al, 2003).
Donaldson and Preston (1995) demonstrate the stakeholder model with the following stakeholders: Governments, Investors, Political Groups, Customers, Communities, Employees, Trade Associations, and Suppliers. Preston and O’Bannon (1997) state the hypothesis as follows: ‘’Higher (lower) levels of social performance lead to higher (lower) level of financial performance.’’ Orlitzky et al. (2003) combine both stakeholder theory and good management theory by stating the following hypothesis: ‘’Corporate social performance and financial performance are generally positively related across a wide variety of industry and study contexts.’’ We state the hypothesis of Orlitzky et al. (2003) in this thesis:
Hypothesis 1a: Corporate social performance and financial performance are generally positively related across a wide variety of industry and study contexts9.
8
Freeman, 1984; Donaldson and Preston, 1995; Freeman and Evan, 1990; Cornell and Shapiro, 1987, Mitchell et al, 1997; Jones, 1995; Aupperle, Carroll and Hatfield, 1985; Barton, Hill, and Sundaram, 1989; Cochran and Wood, 1984; McGuire, Sundgren, and Schneeweis, 1988; Preston and Sapienza, 1990; Preston, Sapienza and Miller, 1991
9
13
Trade-off perspective
According to Preston and O’Bannon (1997), ‘’the trade-off perspective asserts that social performance is the independent variable and that social accomplishments involve financial costs.’’
Allouche and Laroche (2005) state that ‘’because social accomplishments involve financial costs, social responsibility may siphon off capital and other resources from the firm as suggested by Friedman (1970), putting it at a relative competitive disadvantage compared to other firms that are less socially active (Preston and O’Bannon, 1997).’’ According to Preston and O’Bannon (1997), ‘’a firm’s higher levels of social performance may thus lower its financial performance as compared to competitors and/or other norms.’’ This hypothesis stated is the opposite of stakeholder theory and good management theory. We thus state the opposite of hypothesis 1a:
Hypothesis 1b: Corporate social performance and financial performance are generally negatively related across a wide variety of industry and study contexts.
Internal resources theory
A different theory is the internal resources theory (Barney, 1991; Russo and Fouts, 1997; Wernerfelt, 1984; Majumdar and Marcus, 2001; Shrivastava, 1995). According to Orlitzky et al. (2003), ‘’this perspective suggests that investments in CSP may help firms develop new competencies, resources, and capabilities which are manifested in a firm’s culture, technology, structure, and human resources (Barney, 1991; Russo and Fouts, 1997; Wernerfelt 1984).’’ Orlitzky et al. (2003) say ‘’CSP may help build managerial competencies because preventive efforts necessitate significant employee involvement, organization-wide coordination, and a forward thinking managerial style (Shrivastava, 1995).’’ Orlitzky et al. (2003) conclude that ‘’CSP can help management develop better scanning skills, processes, and information systems, which increase the organization’s preparedness for external changes, turbulence, and crises.’’ We state the hypothesis of Orlitzky et al. (2003) in this thesis:
Hypothesis 2: CSP is positively correlated with CFP because CSP increases managerial competencies, contributes to organizational knowledge about the firm’s market, social, political, technological, and other environments, and thus enhances organizational efficiency.
Reputation perspective
14 better employees, and increase current employees’ goodwill (Orlitzky et al, 2003). ‘’An organization’s communication with external parties about its level of CSP may help build a positive image with customers, investors, bankers, and suppliers’’ (Fombrun and Shanley, 1990). According to Fombrun and Shanley (1990), ‘’by signaling consumers about product quality, favorable reputations may enable firms to charge premium prices (Klein and Leffler, 1981; Milgrom and Roberts, 1986b), attract better applicants (Stigler, 1962), enhance their access to capital markets (Beatty and Ritter, 1986), and attract investors (Milgrom and Roberts, 1986a).’’ In the reputation perspective, reputation is interpreted as the outcome of a competitive process in which firms signal their key characteristics to constituent to maximize their social status (Spence, 1974). Fombrun and Shanley (1990) present a social responsiveness factor in their model10 of reputation building. ‘’Firms high in CSP may use CSR disclosures as one of the information signals upon which stakeholders base their assessments of corporate reputation under conditions of incomplete information’’ (Fombrun and Shanley, 1990). Fombrun and Shanley (1990) find support for their hypothesis that publics assign higher reputations to firms that have foundations and give proportionally more to charity than other firms. We state the following hypothesis, as stated by Orlitsky et al. (2003):
Hypothesis 3: CSP is positively correlated with CFP because CSP helps the firm build a positive reputation and goodwill with its external stakeholders.
Slack resources theory
Slack resources theory (McGuire et al., 1988; Kraft and Hage, 1990; Preston and O’Bannon, 1997; Brammer and Millington, 2004; Seifert et al., 2003; Seifert et al. 2004), also called available funding hypothesis, suggests a positive relationship between CSP and CFP, in which firms follow the normative rules of good corporate social depending on the financial resources available (Allouche and Laroche, 2005). ‘’Profitability in one period may increase a firm’s ability to fund discretionary social performance projects’’ (Allouche and Laroche, 2005). Slack resources theory implies existence of a causal relationship from financial to social performance. Prior high levels of CFP may provide the slack resources necessary to engage in CSR and responsiveness (Ulmann, 1985; Waddock and Graves, 1997). According to McGuire et al. (1988), ‘’because CSP often represents an area of relatively high managerial discretion, the initiation or cancellation of voluntary social and environmental policies may, to a large extent, depend on the availability of excess funds.’’ We state the following hypothesis, as stated by Preston and O’Bannon (1997):
10
15 Hypothesis 4a: Higher (lower) levels of financial performance lead to higher (lower) levels of social performance11.
Managerial opportunism theory
Another negative relationship hypothesis is the managerial opportunism theory. This hypothesis states that ‘‘when financial performance is strong, managers may attempt to cash in by reducing social expenditure in order to take advantage of the opportunity to increase their own short-term private gains’’ (Allouche and Laroche, 2005). This is a form of agency costs. It also works the other way around: when financial performance weakens, managers might engage in social programs to offset or justify their disappointing results. This theory follows agency theory. Here, one believes a manager, when possible, has an incentive to put private gains first. When financial performance is strong, managerial opportunism expects less CSR. Thus, this theory assumes that CFP precedes CSP. We state the opposite of hypothesis 4a:
Hypothesis 4b: Higher (lower) levels of financial performance lead to lower (higher) levels of social performance.
Virtuous circle
Waddock and Graves (1997) introduce the possibility of what they call a virtuous circle. They find that better financial performance may lead to improved CSP. Also, better CSP may lead to improved financial performance, ceteris paribus (Waddock and Graves, 1997). When stakeholder theory holds, CSP leads to improved CFP. And when slack resources theory holds, improved CFP leads to improved CSP. Thus, when both stakeholder theory and slack resources theory hold, a virtuous circle exists (Waddock and Graves, 1997).
According to Orlitzky et al. (2003): ‘’if effect sizes are highly similar across all three meta-analytic subgroups (prior CSP related to subsequent CFP; prior CFP related to subsequent CSP; and contemporaneous (cross-sectional) associations), Waddock and Graves’s (1997) argument about a virtuous circle between CSP and CFP would be supported irrespective of study context, sampling error, and measurement error.’’
The virtuous circle is confirmed when both stakeholder theory and slack resources theory hold. Orlitzky et al. (2003) expect that these two constructs (stakeholder theory and slack resources theory) are related to each other reciprocally and they state the following hypothesis: ‘’There is bidirectional causality between corporate social performance and financial performance. ‘’
11
16 We want to allow for both a positive synergy (virtuous circle) and a negative synergy and thus follow the hypothesis stated by Preston and O’Bannon (1997):
Hypothesis 5a: There is positive synergy between financial performance and social performance12.
Negative synergy
The opposite hypothesis of the virtuous circle follows from negative synergy (Preston and O’Bannon, 1997). Preston and O’Bannon (1997) state this hypothesis as a possibility, but do not provide additional arguments. As it is the opposite of the virtuous circle, which holds when
stakeholder theory and slack resources theory hold, negative synergy will hold when both the trade-off perspective and managerial opportunism theory hold. Again, we follow the hypothesis stated by Preston and O’Bannon (1997):
Hypothesis 5b: There is negative synergy between financial performance and social performance.
Non-linear relationship
Another possibility besides an assumed linear relationship is a curvilinear relationship or other non-linear relationships between CSP and CFP. This results in a mixed direction. Barnett and Salomon (2006) find that ‘‘as the number of social screens used by an SRI funds increases, financial returns decline at first, but then rebound as the number of screens reaches a maximum’’. Bowman and Haire (1975) find a statistically significant inverted U relationship between CSP and CFP. These are results, but not theories. We state this hypothesis to allow for the possibility of a non-linear relationship between social performance and financial performance.
Hypothesis 6: There is a non-linear relationship between social performance and financial performance in any direction.
No relationship
There is also the possibility that there is no relationship at all between CSP and CFP. The possibility of no relationship states that there is no reason to observe any relationship as a number of other variables can mediate or moderate the relationship between CSP and CFP. According to
12
17 McWilliams and Siegel (2001) ‘’there are too many intervening variables to detect any relationship between social and financial performance.’’ They identify, for example, that R&D and CSP are highly correlated, and that, when R&D intensity is included in their equation, CSP is shown to have a neutral effect on profitability (McWilliams and Siegel, 2001). The following hypothesis is stated:
Hypothesis 7: There is no relationship between social performance and financial performance.
Categorizing Hypothesis
In order to analyze the hypotheses following from the different theories, we need to categorize these hypotheses. Table 3 presents the overview of theories with their causality and direction.
Table 3 – Overview of theories describing the relationship between CSP and CFP with their hypotheses number, categorized into a causality and direction of the CFP – CSP relationship
Hypothesis Name theory Causality
relationship Direction relationship 1 Stakeholder theory/Good management theory CSP precedes CFP Positive
1b Trade-off perspective CSP precedes CFP Negative
2 Internal resources theory CSP precedes CFP Positive
3 Reputation perspective CSP precedes CFP Positive
4a Slack resources theory CFP precedes CSP Positive
4b Managerial opportunism CFP precedes CSP Negative
5a Virtuous circle Synergetic Positive
5b Negative synergy Synergetic Negative
6 No relationship None None
7 Non-linear relationship CSP precedes CFP Mixed
18 In the next section the data and methodology are presented. Specifically, the data selection is explained, both CFP measures and CSP measures are categorized, the theories concerning the relationship between CSP and CFP are linked to possible CSP and CFP measures, and lastly the method of relating theories to results are presented.
3. Data and methodology
This section is arranged as follows: Section 3.1 presents the data selection, Section 3.2 shows the categorization of CFP measures, Section 3.3 presents the categorization of CSP measures, and lastly, Section 3.4 links existing theories concerning the relationship between CSP and CFP with broad CSP categories and broad CFP categories.
3.1 Data selection
The sample selection for a meta-analysis is a crucial aspect, especially in trying to avoid biases as much as possible. This thesis uses a sample selected from several meta-analyses of corporate social and financial performance. The studies used as input are Aupperle et al. (1985), Griffin and Mahon (1997), Orlitzky et al. (2003), Allouche and Laroche (2005) and Margolis et al. (2009). Table 4 shows the process of arranging the sample.
Table 4 – Overview of the process of data selection
Data selection process Step 1 Step 2
Author(s), Year Sample size used by articles –
first input
Publicly available studies
Aupperle et al., 1985 10 8
Griffin and Mahon, 1997 51 40
Orlitzky et al., 2003 52 45
Allouche and Laroche, 2005 82 64
Margolis et al., 2009 214 178
Total 409 335
Step 3
Overlap studies Count
Studies used in 5 articles 4
Studies used in 4 articles 15
Studies used in 3 articles 18
Studies used in 2 articles 46
Studies used in 1 article 109
Total remaining sample 192
19 The first step is to collect all original sample lists, which results in a total sample size of 409 studies. When a working paper in the original sample lists is published in the mean time, we use the published version. The second step is to find publicly available studies, which results in 335 available studies. Step 3 is the removal of overlap in the sample, resulting in a total remaining sample of 192 studies. The final sample contains 192 studies, involving 252 results. The results of studies are divided into several types of studies. Some studies involve multiple types of studies; these multiple results were then counted as separate results. Table 5 shows a count of the sample results. The division into types of studies is done due to the fact that results within these types of studies are not comparable at first sight and thus require a different approach. It also provides overview on the sample.
Table 5 – Count of results for each type of study
Number type of study Type of study Results Count
1 CSP and CFP Firms 112
2 Event Study 57
3 Firm Characteristics and CSP/CFP 21
4 Model/Construct Testing 29
5 Funds 29
6 SRI Indices 3
7 Banks 1
Sum 252
The first type of study, CSP and CFP Firms, consists of studies researching the relationship between measures of CSP and measures of CFP for firms. These studies do not indicate clearly which theory - describing the possible relationship between CSP and CFP - they test. The second type of study, Event Study, contains event studies, which is a specific form of study type one. Within this method the CFP measure is always the same, namely abnormal returns or an equivalent of abnormal returns. The third type of study is Firm Characteristics and CSP/CFP, wherein certain firm characteristics such as leverage, size, ownership and cash are compared to CFP and/or CSP. The fourth type of study is Model/Construct Testing, which consists of studies that test certain empirical models or created constructs. The fourth type of study Model/Construct Testing includes studies specifically testing certain theories. Study type 1-4 are thus methods used to assess the relationship between CSP and CFP (and characteristics) for firms.
20 the relationship between CSP and CFP specifically for banks. Study type 5-7 are thus methods used to assess the relationship between CSP and CFP for certain objects.
3.2 Categorizing CFP measures
In order to compare studies in a detailed manner, CFP measures need to be categorized.This categorization of CFP measures is done by allowing two layers: all CFP measures are categorized into a CFP subcategory and a broad CFP category. Appendix A presents the full categorization of CFP measures.
Broad CFP categories
21 CFP Subcategories
The broad CFP category Accounting Measure has three CFP subcategories: First of all Accounting Ratios, which contains measures such as Return on Assets (ROA), Return on Equity (ROE), Return on Sales (ROS), Return on Investment (ROI), Return on Capital Employed (ROCE), and Earnings Per Share (EPS). The second CFP subcategory is Accounting Profits, which contains measures such as sales growth, net income growth and accounting profits. The third CFP subcategory is Other Accounting Measures, which contains other accounting measures, such as total asset turnover and output. The broad CFP category Market Measure is split into five subcategories: The first CFP subcategory is Stock Returns, which contains measures such as Jensen’s Alpha, share price, and excess returns. The second CFP subcategory is Risk and Return Ratios, which includes the Sharpe ratio and the Treynor ratio. The third CFP category is Market-to-book Measure, which contains Tobin’s Q and market-to-book value as measures. The fourth CFP subcategory is Market Value, which contains MVA, Market value and P/E ratios as measure. The fifth and last CFP subcategory is Risk, which measures risk using beta or standard deviation. The broad CFP category Perceptual Measure has a similar CFP subcategory called ‘Perceptual Measure’, as all these measures are based on perception and thus all fall into the same CFP subcategory.
The CFP measures are subcategorized because there is much difference in measures within these categories. It allows to present results in a more detailed manner. Table 6 gives an overview of the broad categories and subcategories of CFP measures.
Table 6 – Categorizing CFP measures - Overview of broad CFP categories and their CFP subcategories
Broad CFP categories CFP Subcategories
Accounting Measure Accounting Ratios
Accounting Profits
Other Accounting Measures
Market Measure Stock Returns
Risk and Return Ratios Market-to-book Measure Market Value
Risk
Perceptual Measure Perceptual Measure
Accounting & Market Accounting & Market
Accounting & Perceptual Accounting & Perceptual
22
3.3. Categorizing CSP measures
In order to compare studies in a more detailed manner, CSP measures need to be categorized. We recognize more than 60 different CSP measures within this sample of 192 studies. And even within the same measures, mostly there are minor differences in attaining the information and slightly different prerequisites. The categorization of CSP measures is done by allowing three layers: all CSP measures are categorized into a broad CSP category, a CSP subcategory, and a CSP metric is added. Appendix B presents the full categorization of CSP measures.
Broad CSP categories
For categorizing CSP three broad CSP categories are recognized. In existing literature, for example Margolis et al. (2009), CSP categories overlap in what they measure and how they are measured. An example is the CSP category of disclosure, which could be environmental disclosure but also social disclosure. Therefore disclosure should be a CSP subcategory within broader CSP categories (dimensions). This enables having environmental disclosure and social disclosure as separate groups. We try to improve the CSP categories of Margolis et al. (2009) by creating multiple layers and we use these CSP categories as starting point and, whenever it is possible, the same definitions. CSP measures are categorized in broad CSP categories, CSP subcategories and also an additional layer stating the CSP metric used. The results can then be analyzed on these different layers. In this thesis, the broad CSP categories are based on what aspect/dimension of CSR they measure. These broad CSP categories are Environment, Social, and CSR. Environment and Social are mutually exclusive CSP categories and used when clearly mentioned the measure falls into the Environment or Social category. Allouche and Laroche (2005) mention that the narrow definition of social performance excludes environmental performance. The broad CSP category Social used here follows this definition of Allouche and Laroche (2005), containing studies specifically excluding environment, and thus also contains corporate governance. The broad CSP category Environment is used when specifically environment is mentioned as dimension for the CSP measure. CSR is used as broad CSP category when there is no specific mention about whether the environment or social part of CSR is measured, but when CSR is measured as a whole or left undefined.
CSP subcategories
23 Misdeeds (Environmental), and Pollution. The broad CSP category of CSR contains the following subcategories: Performance (CSR), Reputational (CSR), Disclosure (CSR), and Multiple Measures.
Performance (Social) measures attempt to assess social performance, involving measures such as a social index or perceptual measures of social performance. Reputational (Social) measures try to assess social performance based on reputation/comparisons with other companies, such as a social responsibility ranking. Disclosure (Social) is defined as ‘’the release of information by a company itself in publicly available documents, such as annual reports, is used as indicator of a company’s CSP‘’ (Margolis et al., 2009). In this case it specifically is an indicator of the social dimension. Revealed Misdeeds (Social) includes the public announcements of arrests, fines, guilty verdicts in lawsuits, involuntary recalls, and other actions that indicate socially irresponsible behavior (Margolis et al., 2009), in which the social dimension is specifically measured here. Philanthropy includes cash and in-kind donations or the establishment of a philanthropic foundation (Margolis et al., 2009). Philanthropy and Revealed Misdeeds can be seen as specific events of respectively (extremely) good and bad social behavior. Corporate Policies ‘’includes studies that examine a range of corporate policies, such as companies that divested from apartheid South Africa, firms that did business in apartheid South Africa and signed the Sullivan Principles for fair treatment of citizens, banks that offered low income loans, and defense contractors that agreed to a code of ethics’’ (Margolis et al., 2009).
Performance (Environmental) measures attempt to assess environmental performance and include measures of impact on the environment, whether objective or self-reported (Margolis et al., 2009). Reputational (Environmental) measures try to assess environmental performance based on reputation/comparisons with other companies. There are no examples of measures found in the sample. Disclosure (Environmental) is ‘’the release of information by a company itself in publicly available documents, such as annual reports, is used as indicator of a company’s CSP’’ (Margolis et al., 2009). In this case it specifically is an indicator of the environmental dimension. Revealed Misdeeds (Environmental) includes the public announcements of arrests, fines, guilty verdicts in lawsuits, involuntary recalls, and other actions that indicate socially irresponsible behavior (Margolis et al., 2009), in which the environmental dimension is specifically measured here. Pollution is a specific CSP subcategory of the environment dimension, including measures such as pollution indices and emission levels.
24 (Margolis et al., 2009), and in this case for the CSR dimension as a whole. Multiple Measures contains studies using multiple measures of at least two different subcategories.
CSP Metrics
The possible metric categories we find are: Third-Party Audit, Content Analysis, SRI vs. Conventional, Perceptual, Quantifiable/Measurable, Rating/Ranking/Index, Announcement, Governmental Agency, Regulations/Compliance, Convictions, and Multiple Measures.
Third-Party Audit is defined as ‘’a method that uses observers to assess CSP and involves the systematic assessment of data by investigators who evaluate a company along a set of criteria’’ (Margolis et al., 2009). Content Analysis is ‘’the content analysis of annual reports, where content analysis is a technique for gathering data that consists of codifying qualitative information in anecdotal and literary form into categories in order to derive quantitative scales of varying levels of complexity’’ (Abbott and Monsen, 1979). SRI vs. Conventional is used as metric when a study compares financial market performance of SRI portfolios with conventional portfolios. The metric Perceptual contains surveys used to assess CSP, based on the perception of the respondent. Quantifiable/Measurable is used when the metric is quantifiable, such as the amount of cash donations. Rating/Ranking/Index consists of measures such as the Fortune rating or an environmental performance index, and it excludes ratings/rankings/indices from governmental agencies. Announcement is the metric used by event studies. Governmental Agency is used when a measure provided by a governmental agency is used, such as the TRI pollution index. Regulations/Compliance is used for a range of corporate policies, such as compliance to the Sullivan principles or commitment to ethics. The metric Convictions is used for corporations that are convicted for illegal corporate acts. The metric Multiple Measures is used when at least two metrics from different categories are used. All metrics in this sample fit into these metric categories, but likely there exist other metrics. The metric layer adds the possibility to compare results based on the metric used, without considering other broader categories. It can reflect on whether some methods are inappropriate or not.
25
Table 7 – Categorizing CSP measures - Overview of broad CSP categories and their CSP subcategories
Broad CSP category CSP Subcategories
Social Performance (Social)
Reputational (Social) Disclosure (Social)
Revealed Misdeeds (Social) Corporate Policies
Philanthropy
Environment Performance (Environmental)
Reputational (Environmental) Disclosure (Environmental)
Revealed Misdeeds (Environmental) Pollution
CSR Performance (CSR)
Reputational (CSR) Disclosure (CSR) Multiple Measures
3.4 Linking theories to CSP and CFP measures
In this section the theories stated in the literature are presented and we suggest which CFP measures and CSP measures to use as proxies for CFP and CSP respectively. We do this using the broad categories13 as possible options. There are not enough empirical arguments to link them to more specific measures, which is due to the wide range of operationalization and conceptualization of CSP and the wide range of operational definitions of CFP (Aupperle et al., 1985; Allouche and Laroche, 2005; Margolis et al., 2009).
The broad categories for CFP measures are: Accounting Measure, Market Measure, Perceptual Measure, Accounting & Market, Accounting & Perceptual, and Market & Perceptual. These broad categories for CSP measures are: Social, Environmental, and CSR.
Stakeholder theory/Good management theory
As mentioned in the literature section, the stakeholder theory and good management theory as derivation of stakeholder theory, suggest that CSP is positively associated with CFP because it enhances the satisfaction of various stakeholders and leads to better financial performance (Allouche and Laroche, 2005).
Whereas stakeholder theory concerns various stakeholders, measuring financial performance by stock returns would focus only on investors. ‘’Market measures reflect the notion that
13
26 shareholders are a primary shareholder group whose satisfaction determines the company’s fate’’ (Cochran and Wood, 1984). This is not appropriate to use for stakeholder theory/good management theory, because it would ignore other stakeholders. We suggest assessing stakeholder theory by using a broad measure of financial performance or firm value, which would include as much stakeholder groups as possible. Good management theory is measured by a proxy of competitive advantage. Competitive advantage leads to better financial performance and higher firm value, thus all proxies attempt to measure firm value directly or indirectly. Competitive advantage, financial performance, and firm value are broad concepts and highly correlated. Therefore there is not one specific measure to assess these concepts, but a variety of measures could be used as proxy to measure them. This results in a broad operationalization when assessing the stakeholder theory.
Resulting from above information, we suggest the following categories as proxies for firm value and/or economic performance: Accounting & Market, Accounting & Perceptual, Market & Perceptual, Accounting Measure, and Perceptual Measure. We thus allow for all measures, except for the broad CFP category Market Measure as single measure. The stakeholder theory and good management theory do not specify CSP, thus we suggest all possible proxies as CSP measure: Environment, Social, and CSR.
Trade-off perspective
According to Preston and O’Bannon (1997), ‘’a firm’s higher level of social performance may lower its financial performance as compared to competitors and/or other norms.’’ The trade-off perspective suggests that CSP and CFP are generally negatively related. This hypothesis is the opposite of stakeholder theory and good management theory and can thus be measured using the same CFP measures and CSP measures as these theories. The categories of these CFP measures are: Accounting & Market, Accounting & Perceptual, Market & Perceptual, Accounting Measure, and Perceptual Measure. The CSP measures are: Environment, Social, and CSR.
Internal resources theory
27 The internal resources theory states that CSP leads to more efficient utilization of resources. This suggests CSP subcategories as Philanthropy and Criminal Behavior are incorrect CSP measures when assessing this theory, because these events have no effect on the utilization of resources. Both are costs that do not change the internal resources. Orlitzky et al. (2003), for example, say ‘’there is no theoretical causal mechanism between CSP disclosures (such as Philanthropy and Criminal Behavior) and internal (that is, accounting) CFP measures.’’ We will investigate this when assessing the results on the level of subcategories. We suggest all broad CSP categories as CSP measures to assess the internal resources theory: Social, Environmental, and CSR.
Reputation perspective
‘’The reputation perspective views upon CSP as an information signal to improve relations with bankers and investors, to attract better employees, and to increase current employees’ goodwill’’ (Orlitzky et al, 2003). We assess this theory from a narrow point of view and focus on investors. We suggest the following CFP measure: Market Measure. The market measures allow us to see how investors respond to for example CSR disclosures. We expect subcategories Philanthropy and Criminal Behavior as the most likely to support the reputation perspective, as these events directly affect the reputation of a company. The reputation perspective does not specify CSP however, and thus we suggest all possible CSP measures: Environment, Social, and CSR.
Slack resources theory
Slack resources theory implies there exists a causal relationship from financial to social performance. Prior high levels of CFP may provide the slack resources necessary to engage in CSR and responsiveness (Ulmann, 1985; Waddock and Graves, 1997). According to Mc Guire et al. (1988), ‘’because CSP often represents an area of relatively high managerial discretion, the initiation or cancellation of voluntary social and environmental policies may, to a large extent, depend on the availability of excess funds.’’ The relationship can thus be measured by excess funds, financial resources, or cash available. We suggest the following CFP measures: Earnings, Cash, Retained Earnings and financial resources available. The slack resources theory does not specify CSP, thus we suggest all possible proxies as CSP measure: Environment, Social, and CSR.
Managerial opportunism theory
28 We thus suggest the following CFP measures: Earnings, Cash, Retained Earnings and financial resources available. We suggest the following CSP measures: Environment, Social, and CSR.
Virtuous circle
Waddock and Graves (1997) introduce the possibility of what they call a virtuous circle. The virtuous circle is confirmed when both stakeholder theory and slack resources theory hold. Orlitzky et al. (2003) expect that these two constructs (stakeholder theory and slack resources theory) are related to each other reciprocally and they state the following hypothesis: ‘’There is bidirectional causality between corporate social performance and financial performance.‘’ This theory should thus be measured by confirming stakeholder theory and slack resources theory, and therefore the measures of these theories should be used.
Negative synergy
The opposite hypothesis of the virtuous circle follows from negative synergy (Preston and O’Bannon, 1997). As it is the opposite of the virtuous circle, which holds when stakeholder theory and slack resources theory hold, negative synergy will hold when both the trade-off perspective and managerial opportunism theory hold. Again, the same measures as trade-off perspective and managerial opportunism should thus be used.
Non-linear relationship
Another possibility besides an assumed linear relationship is a curvilinear relationship or other non-linear relationships between CSP and CFP. As measures are not specified, we suggest all measures as possible proxies. We suggest the following CFP measures: Accounting & Market, Accounting & Perceptual, Market & Perceptual, Accounting Measure, Market Measure, and Perceptual Measure. And we suggest the following CSP measures: Environment, Social, and CSR.
No relationship
29 Table 8 – Overview of theories concerning the relationship between CSP and CSP with their hypothesis and their suggested CFP proxies and suggested CSP proxies
Theory Hypothesis Suggested CFP proxies Suggested CSP
proxies Stakeholder
theory/Good management theory
Corporate social performance and financial performance are generally positively related across a wide variety of industry and study contexts.
Accounting & Market, Accounting & Perceptual, Market & Perceptual, Accounting Measure, Perceptual Measure Environment, Social, CSR Trade-off perspective
Corporate social performance and financial performance are generally negatively related across a wide variety of industry and study contexts.
Accounting & Market, Accounting & Perceptual, Market & Perceptual, Accounting Measure, Perceptual Measure Environment, Social, CSR Internal resources theory
CSP is positively correlated with CFP because CSP increases managerial competencies, contributes to
organizational knowledge about the firm’s market, social, political, technological, and other environments, and thus enhances organizational efficiency.
Accounting Measure Environment, Social, CSR
Reputation perspective
CSP is positively correlated with CFP because CSP helps the firm build a positive reputation and goodwill with its external stakeholders.
Market Measure Environment, Social, CSR
Slack resources theory
Corporate social performance and financial performance are generally negatively related across a wide variety of industry and study contexts.
Earnings, Cash, Retained Earnings or financial resources available Environment, Social, CSR Managerial opportunism theory
Higher (lower) levels of financial
performance lead to lower (higher) levels of social performance.
Earnings, Cash, Retained Earnings or financial resources available
Environment, Social, CSR Virtuous circle There is positive synergy between
financial performance and social performance.
Confirmed when both stakeholder theory and slack resources theory hold.
Confirmed when both stakeholder theory and slack resources theory hold
Negative synergy
There is negative synergy between financial performance and social performance.
Confirmed when trade-off perspective and managerial opportunism theory hold. Confirmed when trade-off perspective and managerial opportunism theory hold Non-linear relationship
There is a non-linear relationship between social performance and financial
performance in any direction.
Accounting & Market, Accounting & Perceptual, Market & Perceptual, Accounting Measure, Market Measure, Perceptual Measure Environment, Social, CSR No relationship
There is no relationship between social performance and financial performance.
30 Table 8 shows that most theories suggest multiple measures and are not specifically defined, especially the CSP proxies. We observe that the internal resources theory and the reputation perspective are the most specified (clearly defined) theories. The measures used for the internal resources theory and the reputation perspective could also be used to assess the stakeholder theory. This creates an overlap between these theories. To avoid this overlap in our analysis, we exclude the CFP categories used by more clearly defined theories from the broader theories. Specifically: we link broad CFP category Accounting Measure to internal resources theory exclusively, and we link broad CFP category Market Measure to the reputation perspective exclusively. Table 9 shows the categorization of the hypotheses following the categorization of Preston and O’Bannon (1997). It also links the broad CFP categories to the theories. Hypothesis 6 and 7 are not included in this table. We use the distribution shown in Table 9 to assess how much theories are supported in the empirical results.
Table 9 – Categorization of hypotheses following categorization of Preston and O’Bannon (1997), and additionally linking CFP categories to theories with causality CSP precedes CFP
Causality
relationship CFP – CSP
Category of CFP
Result direction linear relationship
Positive Negative
CSP precedes CFP Accounting & Market Stakeholder theory, good management theory (H1a)
Trade-off perspective (H1b)
Accounting Measure Internal resources theory (H2)
-
Market Measure Reputation perspective (H3)
-
Perceptual Measure Stakeholder theory, good management theory (H1a)
Trade-off perspective (H1b)
CFP precedes CSP Prior financial
performance/cash (no specific category of CFP)
Slack resources theory (H4a) Managerial opportunism hypothesis (H4b) Synergetic relationship CFP – CSP Prior financial
performance and prior social performance (No specific category CFP)
Positive synergy hypothesis (H5a)
Negative synergy hypothesis (H5b)
The Data and Methodology section presented the data selection, the categorization of both CSP and CFP measures and we suggested which suggested CFP proxies and CSP proxies to use to assess theories. This was done at the categorization level of broad categories14. After that, the hypotheses were categorized. The next section presents the results.
14
