1
The Influence of Corporate Social Responsibilities on
Financial Performance
Abstract
In this paper, the effect of corporate social responsibility (CSR) on financial performance (CFP) is examined. The CSR-CFP relationship has been tested extensively already, but results are controversial. With the substantial growth of CSR, firms seem to appreciate the influence of engaging in CSR. This paper uses recent data, as retrieved from the MSCI-ESG database, to dive deeper into the effects of CSR. CSR is measured by an ethical rating: a
multidimensional index that takes account of multiple dimensions of CSR. A random effects OLS model is used to test effects of CSR on return on assets, return on equity and Tobin’s Q. Results indicate that effects on accounting-based performance are controversial, as return on assets is negatively affected and return on equity is positively affected. CSR does nevertheless have a clear influence on the market-performance, as Tobin’s Q is strongly positively
affected. CSR does therefore seem to benefit shareholders the most, while effects on the overall firm are two-sided.
Enzo Stilma 26-01-2018
10774424
Supervisor: dr. Liang Zou Bsc Economics & Business Track: Finance & Organization
2
Statement of originality
This document is written by student Enzo Stilma who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3 Table of contents 1. Introduction ... 4 2. Literature review ... 5 2.1 Defining CSR ... 5 2.2 Measuring CSR ... 6 2.2.1 Reputational measures ... 6 2.2.2 Ethical rating ... 7
2.2.3 One dimensional measures ... 7
2.2.4 Content analysis ... 7
2.2.5 Questionnaire surveys ... 8
2.3 Measuring financial performance ... 8
2.4 The CSR-CFP relationship ... 9
2.4.1 Theories on the influence of CSR ... 9
2.4.2 Empirical results ... 10 2.5 Hypothesis ... 11 3. Methodology ... 12 3.1 CSR variable ... 12 3.1.1 Database explanation ... 12 3.1.2 CSR score ... 13
3.2 Financial performance measures ... 13
3.3 Model ... 14
3.4 Control variables ... 15
3.5 Sample data ... 16
4. Results ... 17
4.1 Descriptive statistics ... 17
4.2 Model specifications and assumptions ... 17
4.3 Regression results ... 18 5. Conclusion ... 20 5.1 Summary ... 20 5.2 Discussion ... 21 References Appendix
4
1 Introduction
Firms have been implementing policies on social issues more and more. Speaking more specifically, an increasing number of firms are implementing a corporate social responsibility (CSR), as the popularity for CSR increases and corporate social responsibility has evolved into a major strategy (McPherson, 2017). The reasons behind implementing CSR are diverse. Behaving socially might influence a firm’s reputation and therefore affect financial
performance of a firm.
The relation between corporate social responsibility and financial performance is still uncertain, however. Results of research on the relationship between corporate social
responsibility and financial performance are mixed. Most researchers found a positive relation (Margolis, Elfenbeim, & Walsh, 2007), but some found a negative or non-significant one as well. As there is controversy on found results, but CSR is growing, it is interesting to dive deeper in the CSR field, and look at the recent impact of CSR on financial performance.
Observing this impact results in the following research question: ‘does the extent of engagement in a corporate social responsibility have an effect on a firm’s corporate financial performance (CFP)?’ Following the stakeholder’s theory, as conceptualized by Freeman (1984), and findings of reduced agency costs, it is expected that corporate social responsibility positively affects a firm’s financial performance.
Although the relationship between CSR and CFP has already been investigated
extensively, this paper contributes to the literature by using recent data. As the growth of CSR might be accompanied with a change in impact on the financial performance, examining recent data is of great relevance to the field of corporate social responsibility. In addition, multiple financial performances are measured. Most researchers have only investigated effects on one (kind of) financial performance. Examining effects on multiple kinds makes the
research more complete, and shows effects which might not be seen when using one measure only.
The effects of corporate social responsibility are measured by the return on assets (ROA), return on equity (ROE) and Tobin’s Q. In this way effects on both accounting and market performance can be analyzed. CSR is measured by an ethical rating. More
specifically, it is an overall score, which is based on involvements in social practices, a method earlier applied in the literature (e.g. Van der laan, Van Ees, & Van Witteloostuijn, 2008; Waddock & Graves, 1997). Data on the involvement in practices of CSR is collected from the MSCI-ESG database.
5
To complete the model, proxies for size, innovation and risk, and dummies for industries are added as control variables. Previous research motivates the relevance and intervening effect of these variables in the CSR-CFP relationship. Required financial data, other than data on CSR, is found on the Compustat database. Although results of the analysis confirm a positive relation between CSR and CFP as measured by market-performance, results on accounting performance are controversial, as ROA is positivity affected and ROE is negatively affected.
In the following section earlier research on both corporate social responsibility and its relation to financial performance are discussed in the literature review. After that the method of research and model setup are explained in the methodology. Then results of the empirical study are shown and analyzed. The paper ends with a conclusion, in which results are briefly summarized and discussed.
2 Literature review
Earlier research on CSR and financial performance shows that the meaning and measurement method of CSR are major issues in the field. As a consequence do many researchers explain the mixed results by a lack of consensus on measurement methods, for both CSR and
financial performance. Issues of defining and measuring CSR are therefore discussed first. Then the measurement of financial performance is described. At last is the CSR-CFP relationship discussed; both the theories on how CSR affects CFP, and previous empirical findings in the literature.
2.1 Defining CSR
Although CSR has been studied for a long time already, researchers have not yet agreed on a strong general definition of corporate social responsibility (McWilliams, Siegel, & Wright, 2006). This is illustrated by the review of Dahlsrud (2008), who identifies 37 used definitions of CSR in the literature. The lack of a consensus on the meaning of CSR causes problems in both measurement and theoretical development (McWilliams et al, 2006).
Some researchers, who follow the example of Friedman (1970), suggest that a firm should only focus on the value of the shareholders. The only social responsibility of a firm would therefore be to increase profits. However, recent studies on CSR show that most researchers agree, at least in some sense, with the argument of Davis (1973, p. 312). He defines CSR as a ‘consideration of, and response to, issues beyond the narrow economic, technical, and legal requirements of the firm.’ The European Union Commission defines CSR
6
in a similar way, as ‘actions by companies over and above their legal obligations towards society and the environment’ (European Commission, 2011, p. 3).
Implementing a CSR strategy can thus be seen as trying to achieve broader societal and environmental goals, next to maximizing the wealth of shareholders (Ferrell, Liang, & Renneboog, 2016). These societal and environmental goals can be categorized in seven relevant dimensions of corporate social responsibility, as is done in the MSCI-ESG database. The dimensions are: community relations, customers, diversity, employee relations,
environment, human rights and investors.
2.2 Measuring CSR
The variety of definitions of CSR has its implications for measuring corporate social
responsibility as well. In the literature many different methods have been used for measuring CSR, as there is a lot of discussion on what the best CSR measure is. This is illustrated by Graves and Waddock (1994), who argue that all measurement methods of CSR have their own concerns. However, Soana (2011) categorized the variety of measurement methods in five ways. In this subsection all methods are introduced and discussed. The methods of
measuring CSR are: reputational measures, ethical ratings, one dimensional measures, content analysis and questionnaire surveys.
2.2.1 Reputational measures
These measures (mostly indexes) are ratios that indicate a firm’s reputation based on a subjective definition of social performance (Soana, 2011). In this way, firms are ranked on their reputation. Engaging in CSR would then increase reputation, and cause firms to be ranked higher. In most cases the corporate reputational index (CRI) by Fortune is used (Soana, 2011). Reputational measures are popular, because of the huge data availability and their ability to easily compare companies ( McGuire, Sundgren & Schneeweis, 1988). The downside is however, that they are very subjective measures, as they are formed by private firms who don’t necessarily use scientific methods (Graafland, Eijffinger & Smid, 2004). The CRI has also been criticized for measuring financial performance just as much as measuring CSR (Graves & Waddock, 1994).
7
2.2.2 Ethical ratings Ethical ratings are a form of reputational measures. Ethical ratings are, however, more related to CSR practices, where reputational indices are more measures of the overall reputation of a firm. Measuring CSR by ethical rating is done by using a multidimensional index, in which several indicators of corporate social responsibility are selected and measured, each belonging to a different group of stakeholders (Soana, 2011). All indicators are eventually aggregated, leading to an overall score or ethical rating. This measure is increasing in popularity, as it gives the benefits of the data availability, but is also better than the reputational measures because of its more complete and extensive CSR measure (Graves & Waddock, 1994). Sometimes the aggregated value is taken as a fraction of the potential maximum score, as is done by Van der Laan et al. (2008).
The MSCI-ESG database is widely used for ethical ratings of CSR. Reason for that is the quality and comprehensiveness of the data on social performance (Coombs and Gilley, 2005). Graves and Waddock (1994) mention the objectivity of the database as another reason, as the screenings of social performance are applied consistently for each firm. The database collects data on multiple dimensions of CSR, thereby creating a multidimensional and transparent view on a firm’s social practices.
2.2.3 One dimensional measures
One dimensional measures measure a single aspect of corporate social responsibility. Just one indicator of social practices is examined, and the level of involvement in this practice
determines the CSR measure. Most one dimensional measures used in the literature belong to the following social practices: respecting the environment, the relationship with local
communities and philanthropy, the degree of involvement in criminal practices and the orientation towards clients (Soana, 2011). Although one dimensional measures are easiest to compare and find data on, Carroll (1979) mentions that they are no good measure for
corporate social responsibility, because CSR is a multidimensional concept.
2.2.4 Content analysis
In content analysis, the extent of the reporting of CSR is used as a measure of corporate social responsibility. This means that published documents of firms are analyzed, and the
appearance of CSR in these documents is being observed. Various methods are used in content analysis. Sometimes words, lines or sentences regarding (a dimension of) CSR are
8
just counted (e.g. Aras, Aybars & Kutlu, 2010), and sometimes a more advanced way is used. This can be quality analysis, in which multiple indicators or dimensions of CSR are chosen, and rated on their presence in the published documents (e.g. Karagiorgos; Yang, Lin & Chang, 2009). An advantage of content analysis is that the procedure of measuring is more objective than other measures (Cochran & Woods, 1984). Choosing what kind of CSR is observed is nevertheless a subjective procedure. Cochran and Woods (1984) also mention that content analysis only looks into the CSR that is reported, and not the actual CSR. That creates a bias. Ullman (1985) confirms this bias, as he didn’t find a significant relation between social performance and social disclosure in his research.
2.2.5 Questionnaire surveys
In questionnaire surveys researchers analyze responses of managers and directors about the social practices of a firm. CSR hasn’t been measured often in this way, as there are many drawbacks compared to the other discussed methods. The main drawback is the subjectivity of the responders. According to Soana (2011), responses only reflect a manager’s own perception of social responsibility. Cadez and Czerny (2016) mention on top that selective bias might occur, in which only the more socially responsible firms are willing to respond.
2.3 Measuring financial performance
The methodology of financial performance in CSR related studies is more agreed upon than measuring CSR itself. There is, however, no strong consensus in the matter of CFP either, as various meta-analyses have shown (e.g. Orlitsky, Schmidt, & Rynes, 2003; Margolis et al., 2007). According to Margolis et al. (2007), measures of financial performances broadly fall in two categories: accounting-based measures and market-based measures.
Accounting-based measures represent the historical performance of a firm, as described in the annual reports. These measures are used to show the effect of a change in some managerial policy (like CSR) on earnings or profitability (Cochran & Wood, 1984). Orlitsky et al. (2003) confirm this by explaining that accounting-based performance reflects internal decision making and managerial performance. They add that it differs from market-based performance, as accounting-market-based performance is less about responding to external market development. As earnings are most important in accounting-based performance methodology, often ratios concerning earnings, like return on assets or return on equity,
9
are used (Margolis et al., 2007). Accounting-based performance measures are also most affected by CSR, according to Margolis et al. (2007).
Market-based measures of financial performance are focused on the value of the firm from the perspective of investors, the shareholder’s value (Cochran & Wood, 1984). Usually share returns or similar measures, like returns including dividend, are used (Margolis et al., 2007). When measuring share returns, sometimes event studies are used, especially when an announcement of CSR policy is measured (e.g. Krüger, 2015). These studies show a bigger impact of CSR on financial performance than any other study involving market-based measures (Margolis et al., 2007). In the literature, Tobin’s Q is often used as a performance measure too (Rodgers, Choy, & Guiral, 2007). Tobin’s Q relates the market value of a firm to the book value of the assets. It measures the growth of a firm as expected by the market (Rodgers et al., 2007).
2.4 The CSR-CFP relationship
2.4.1 Theories on the influence of CSR
The impact of CSR on CFP can be explained in several ways. Most researchers explain a positive relation, but there is still some controversy, as there are findings of negative impact as well (Orlitsky et al., 2003). CSR impact is often explained through reputation
improvement, or a reduction in agency costs. In this section, the most important theories on the CSR-CFP relationship are discussed.
The stakeholder theory, as introduced by Freeman (1984), is the most used explanation for the use of CSR (Rodgers et al., 2013). The essence is that managers should not exclusively focus on shareholders. Demands of other stakeholders should also be taken into account, as it is an unavoidable cost of doing business (Ruf, Muralidhar, Brown, Janney, & Paul, 2001). These demands are not only important because of the doing good1 argument, but also because they can increase a firm’s value and performance. The resource-based view and transaction cost economics theory build on this.
The resource-based view argues that meeting demands of all stakeholders can be seen as a strategic investment, in which a firm develops important assets, like a social reputation or strong customer relationships. Improving the assets then leads to a competitive advantage and better financial performance (Barney, 1991). Mcguire et al. (1988) illustrate this by saying
1 Doing good refers to the argument that firms should promote, and help with, activities for external groups or factors, like the society or the environment, because they are able to do so.
10
that investing in employees or customers, through corporate social responsibility, can avoid employee problems and encourage customers to be of more value, as they are treated better under a policy of corporate social responsibility.
Another theory is the transaction cost economics argument. This theory explains that firms who take all stakeholders into account or show willingness to cooperate with all stakeholders, can avoid high (transaction) costs (Ruf et al., 2001). If firms ignore (some) stakeholders, more formalized contracts, like government regulation or union contracts, are required and these are likely to result in much higher costs. Therefore investing in
stakeholders, and so engaging in CSR can affect financial performance through reducing and avoiding transaction costs (Cornell & Shapiro, 1987).
The final theory explaining a positive effect of CSR on CFP is the good governance view, as mentioned by Ferrell et al. (2016). They explain that engaging in CSR often adheres to a value-maximization of corporate governance practices. This happens through a reduction of agency costs, as they find CSR to moderate and reduce the effects of managerial
entrenchment on a firm’s value. Management is therefore less bothered by agency costs, and can work more efficiently. Dimson, Karakas and Li (2015) find the reduction in agency costs, as a result of engaging in CSR, to specifically affect the market value of a firm.
There are also studies explaining a negative influence of CSR. Most important is the agency view, by Ferrell et al. (2016). In this theory, CSR is seen as some policy representing agency problems within the firm (Benabou & Tirole, 2010). An example of this is that managers engage in CSR to benefit themselves, as they just get value normally received by shareholders (Krüger, 2015). Another argument, by Jensen (2001), is that managers who spend much time on running a corporate social responsibility, lose focus on important other responsibilities, and therefore lose track of value-enhancing assets and projects. To
summarize, the general argument against CSR is that managers should always focus on shareholders only, as engaging in other activities only reduces value (Ferrell et al, 2016).
2.4.2 Empirical results
The findings in the CSR-CFP relationship are quite controversial, as already said. According to various meta-analyses on the CSR-CFP relationship, there is a small, but positive effect observed (Margolis et al. 2007; Orlitsky et al, 2003). However, the mixed results make it difficult to draw the conclusion that CSR positively affects CFP. In this section empirical results of the effect of CSR on financial performance are summarized and explained.
11
The conflicting results can be explained by six reasons, according to Ruf et al. (2001). Most of them are related to methodological issues, as they mention that there is no systematic measure for CSR, that applied models include different control variables, that financial performance measures differ and that the sample size is often small. Also the lack of
theoretical foundation is mentioned, as multiple theories are used to explain the relationship. On top is the mismatch between CSR and CFP suggested, but this is caused in particular by wrongly defining CSR and CFP (Wood & Jones, 1995).
Diving deeper in the research on the CSR-CFP relationship reveals interesting results. Margolis et al. (2007) performed a meta-analysis on 192 empirical studies of CSR-CFP relationships, in which they found a small, but positive average effect across all studies. They did nevertheless find that 57% of all studies have no significant result, compared to 29% with positive results and 2% with negative results. Margolis et al. (2007) add that CSR is stronger related to accounting-based performance than market-based performance.
Orlitsky et al. (2003) find the same positive relation between CSR and CFP. They add some notions. Firstly, they find that reputational indices, as a measure for CSR, predict financial performance the best. Secondly, they confirm the issues mentioned by Ruf et al. (2001) and explain that 15% to 100% of the variation in the effect of CSR is explained by errors in methodology, sample size and mismatching. Most importantly though, is that they suggest the existence of a bidirectional relation, in which financial performance also affects the social performance of a firm.
This reverse relationship has already been theorized by Waddock and Graves (1997). They argue that a firm’s financial performance has an influence on the extent of a firm’s corporate social responsibility, as a better performance allows the firm to make more investments, so in CSR as well. Waddock and Graves (1997) confirm this relation in their study, as they find accounting-based measures to have an effect on CSR.
2.5 Hypothesis
Based on the stakeholder’s theory and findings of reduced agency costs CSR would improve the value and performance of a firm. As these theories are also slightly backed up by previous empirical findings, it is expected that the extent of engagement in a corporate social
responsibility positively affects the financial performance. As CSR has been increasing extensively in popularity (McPherson, 2017), it is expected that firms are familiar with the positive impact. Therefore a larger effect of CSR, compared to previous results, is expected, with the timeframe used in this paper.
12 3 Methodology
The effect of CSR on financial performance is investigated by an empirical study. Therefore an OLS-regression of CSR and other model variables on CFP is done, as this can show the statistical significance of corporate social responsibility in explaining financial performance. In this section the method of doing the research is explained. First of all, an explanation of the CSR variable is given, in which the collected data from the MSCI-ESG database first has to be explained. Then the multiple financial performances used in this study are described, followed by an explanation of the model, with all its variables. At last the sample and treatment of data are discussed.
3.1 CSR variable
3.1.1 Database explanation
To measure corporate social responsibility, it is required to use some kind of social
performance. In this paper an ethical rating is used. This means that an index is used, in which multiple dimensions of CSR are investigated. To be able to investigate, data is collected from the MSCI-ESG database. This database collects data on seven dimensions of corporate social responsibility: customer relations ( in terms of the database: product), diversity, employee relations, environment, human rights and investors (corporate governance). All these
dimensions have several indicators2 of social performance in that dimension. Additionally, the database collects data on involvement in six socially harmful fields3 of CSR. As this category
of data is only about negative social performance, it is biased and also not of interest, as a potential positive effect is hypothesized. This data is therefore not included4 in the research.
All indicators are based on the involvement or activity of a firm in the CSR practice measured by the indicator. Each dimension has indicators of positive social performance and negative social performance, the strengths and concerns. The indicators are binary variables: when a firm is involved in a certain CSR practice it gets a score of 1 for that indicator; when the firm is not involved it gets a zero score. This mechanism works for both positive and negative indicators. All indicators are assumed to have equal weight, as there is no objective weighting scheme in the literature yet (Van der Laan et al, 2008).
2 Table 4 In the appendix lists all indicators used in this paper and matches them to the dimension they belong to. A description of each indicator can be found on http://www.msci.com.
3 These fields are: alcohol, civilian firearms, gambling, military weapons, nuclear power and tobacco. 4 The methodology of Van der Laan et al. (2008) is followed on this point.
13
3.1.2 CSR score
To get an eventual useful ethical rating of corporate social responsibility, first all strengths, the indicators of positive involvement, and all concerns, the indicators of negative
involvement, across all dimensions are aggregated. In this way there is a score for strengths and a score for concerns. These scores are still biased though, as the CSR activities of some firms have been investigated more than other firms. This means that some firms have not been rated on some indicator(s), and causes some firms to have more observed indicators than others. On top of that do some indicators have a limited time frame, which causes some indicators to be missing in one or multiple years. The consequence is that the number of observed indicators can differ from year to year. These two data issues create a bias, because firms with more observations can get a higher score, while these firms do not necessarily engage more in CSR.
This problem can be solved by scaling the aggregated positive and negative score to the total firms can score, as is done by Van der Laan et al. (2008). This total score is simply the total amount of observed indicators. This scaled measure, or ratio, enables an equal treatment of all firms, as there are no direct differences in getting a score due to data issues. Finally, the ratio of concerns is subtracted from the ratio of strengths, which results in the overall ethical rating for CSR. This rating will be further referred to as the CSR score. Below this method is summarized in formulas.
𝑆𝑡𝑟𝑒𝑛𝑔𝑡ℎ′𝑠 𝑟𝑎𝑡𝑖𝑜 = ∑ 𝑓𝑖𝑟𝑚′𝑠 𝑠𝑡𝑟𝑒𝑛𝑔𝑡ℎ𝑠 ÷ 𝑡𝑜𝑡𝑎𝑙 𝑖𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟𝑠 𝑜𝑓 𝑠𝑡𝑟𝑒𝑛𝑔𝑡ℎ𝑠
𝐶𝑜𝑛𝑐𝑒𝑟𝑛′𝑠 𝑟𝑎𝑡𝑖𝑜 = ∑ 𝑓𝑖𝑟𝑚′𝑠 𝑐𝑜𝑛𝑐𝑒𝑟𝑛𝑠 ÷ 𝑡𝑜𝑡𝑎𝑙 𝑖𝑛𝑑𝑖𝑐𝑎𝑡𝑜𝑟𝑠 𝑜𝑓 𝑐𝑜𝑛𝑐𝑒𝑟𝑛𝑠
𝐶𝑆𝑅 𝑠𝑐𝑜𝑟𝑒 = 𝑆𝑡𝑟𝑒𝑛𝑔𝑡ℎ′𝑠 𝑟𝑎𝑡𝑖𝑜 − 𝐶𝑜𝑛𝑐𝑒𝑟𝑛′𝑠 𝑟𝑎𝑡𝑖𝑜
3.2 Financial performance measures
The other variable of interest is the financial performance. Measures of financial performance are required to find a possible relationship. Multiple kinds of financial performance are investigated, as it gives a more complete view of the effects of CSR compared to using one measure. In this study, effects on return on assets (ROA), return on equity (ROE) and
14
Tobin’s Q are examined. These belong to the most used proxies for financial performance in examining the CSR-CFP relationship (Margolis et al., 2007).
ROA is defined as the net income (loss) divided by the total assets. It measures in what extent a firm is able to convert assets into profits. ROE is defined as the net income divided by the shareholder’s equity. It measures the extent in which the money of
shareholders is converted into profits. Observing the financial implications shows that ROE also differs from ROA in sensitivity to leverage: ROE is more sensitive to the leverage of a firm than ROA (Berk & DeMarzo, 2014). Both kinds of performances are accounting-based, which means that they are historical data, as found in the published document.
𝑅𝑂𝐴 = 𝑁𝑒𝑡 𝐼𝑛𝑐𝑜𝑚𝑒 / 𝑇𝑜𝑡𝑎𝑙 𝐴𝑠𝑠𝑒𝑡𝑠
𝑅𝑂𝐸 = 𝑁𝑒𝑡 𝐼𝑛𝑐𝑜𝑚𝑒 / 𝑆ℎ𝑎𝑟𝑒ℎ𝑜𝑙𝑑𝑒𝑟’𝑠 𝐸𝑞𝑢𝑖𝑡𝑦
Tobin’s Q is defined as the market value of the firm scaled by the book value of the assets. It measures the firm’s financial growth and therefore its future prospects, as expected by markets (Rodgers et al., 2013). In this paper, Tobin’s Q is used to measure the market-based performance of firms. Specifically, Tobin’s Q is measured as the sum of the market value of the shares and book value of debt, divided by the book value of the total assets. It is widely used in previous research on the CSR-CFP relationship (e.g. Garcia-Castro et al., 2010; Rodgers et al., 2013).
𝑇𝑜𝑏𝑖𝑛′𝑠 𝑄 =𝑚𝑎𝑟𝑘𝑒𝑡 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑒𝑞𝑢𝑖𝑡𝑦 + 𝑑𝑒𝑏𝑡
𝑏𝑜𝑜𝑘 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑎𝑠𝑠𝑒𝑡𝑠
3.3 Model
As the effect of CSR on CFP is empirically determined, a model is required. In this paper three relations are investigated (effect of CSR on ROA, on ROE and on Tobin’s Q), and therefore three model are required. The models, including control variables, are shown below. Control variables are: size, R&D intensity, leverage (or risk) and industry. In the next
subsection control variables of the model are discussed.
𝑅𝑂𝐴𝑖,𝑡 = 𝛽0+ 𝛽1𝐶𝑆𝑅𝑖,𝑡+ 𝛽2𝑆𝑖𝑧𝑒𝑖,𝑡+ 𝛽3𝑅&𝐷𝑖,𝑡+ 𝛽4𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑖,𝑡+ 𝛽5𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦𝑖+ 𝜀𝑖
𝑅𝑂𝐸𝑖,𝑡 = 𝛽0+ 𝛽1𝐶𝑆𝑅𝑖,𝑡+ 𝛽2𝑆𝑖𝑧𝑒𝑖,𝑡+ 𝛽3𝑅&𝐷𝑖,𝑡+ 𝛽4𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝑖,𝑡+ 𝛽5𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦𝑖+ 𝜀𝑖 𝑇𝑜𝑏𝑖𝑛′𝑠 𝑄
15 3.4 Control variables
Previous research has shown that the CSR-CFP relationship is influenced by multiple factors. Ullman (1985) showed the importance of size, risk and industry already, and the intensity of research & development has also proven to affect the CSR-CFP relationship (McWilliams & Siegel, 2000). It is important to control for these factors, as ignoring the intervening effects leads to an incomplete model, in which various biases, like omitted variables bias, take place. As already said, the model’s control variables are: size, R&D intensity, leverage and industry.
Size is measured by total assets. Total assets is often used as a proxy in the literature, next to number of employees and total sales. Total assets are nevertheless expected to be the best measure, as they represent the firm’s total value, as described in the published
documents. On top of that is data on total assets easiest obtainable. Size is proven to intervene, as many researchers (e.g. Ullman 1985; Waddock & Graves 1997) have shown. Reason for the effect on the CSR-CFP relationship might be that smaller firms are not able to invest in CSR as much as larger firms can do (Waddock & Graves, 1997).
R&D intensity is an important factor as well, because it measures the innovation level
of a firm, which is strongly connected to a firm’s engagement in CSR (McWilliams & Siegel, 2000). R&D intensity can therefore influence a firm’s performance through the CSR score. This causes an upward bias in the CSR-CFP link, thus R&D should be controlled for
(McWilliams & Siegel, 2000). R&D intensity is measures as the R&D expenses, scaled by the total sales. This is done by McWilliams and Siegel (2000), but the measurement method is used in later research as well (e.g. Garcia-Castro et al, 2010).
Leverage (the debt to assets ratio) is a proxy for risk. Orlitsky and Benjamin (2013) found risk to be affecting the relationship between CSR and CFP by performing a meta-analysis. They argue that risk is also a measure of financial performance, next to the returns widely used in literature. As CSR is expected to have an effect on financial performance, risk, as a form of financial performance, should also be accounted for. Orlitsky and Benjamin (2013) add that the effect of CSR on financial performance (as measured by earnings ratios) can be manifested through risk, and therefore, effects of risk should be controlled for.
Industry is a variable used to control for differences between industries. This is done by including dummies in the models, that match to the corresponding industries involved in the analysis. Industrial differences do influence the CSR-CFP relationship, as CSR practices and rules for reporting CSR activities can be different between industries (Margolis et al., 2007). McWilliams and Siegel (2000) add that differences in CSR practices can be explained
16
by differences in economies of scale and competitive intensity of industries, thus by industrial differences.
3.5 Sample data
As mentioned earlier, CSR data is collected from the MSCI-ESG database. Data on the six controversial involvements was removed, and CSR scores were calculated, based on the seven dimensions and their available indicators. Firms were identified by applying tickers, and in this way the required financial data (e.g. data on net income, total assets) was collected from the Compustat database. A timeframe of 2011-2015 was used, as this research was performed to show the recent relationship between CSR and financial performances.
After cleaning the data and matching the financial data to CSR scores, a number of 9,208 observations remained. Firms in the dataset that were only observed on positive or negative involvements, and not both, were excluded from the sample. The dataset consists of panel data, as firms in the sample were observed for 1-5 years. The firms were categorized in industries, as this was required to control for industrial differences. This was done by
applying SIC5 codes. The included industries in the dataset and the distribution of the sample in industries is shown in the table below.
Table 1
Distribution of industries in the sample.
SIC-code Industry Variable Frequency Percent
0100-0999 Agriculture, forestry and fishing 1 32 0.35%
1000-1499 Mining 2 584 6.41%
1500-1799 Construction 3 177 1.94%
2000-3999 Manufacturing 4 4,414 48.46%
4000-4999 Transportation, communication, electric, gas and sanitary service
5 693 7.61%
5000-5199 Wholesale trade 6 332 3.65%
5200-5999 Retail trade 7 760 8.34%
6000-6799 Finance, insurance and real estate 8 315 3.46%
7000-8999 Services 9 1,783 19.58%
9100-9729 Public administration - 0 0%
9900-9999 Non-Classifiable 10 18 0.2%
Variable numbers refer to the regression model, in which the numbers match with the dummies in the model.
5 SIC is short for Standard Industrial Classification. It is a system used for classifying industries by four digit codes, categorizing them in 11 groups of industries.
17 4 Results
In this section results from the statistical analysis are discussed. First descriptive statistics and model assumptions are covered, then the results from the regression are discussed. This is combined with an analysis of the implications of the results.
4.1 Descriptive statistics
The descriptive statistics of the key variables are shown in table 2, below. Means, standard deviations and minimum and maximum values of the variables included in the models are shown. Additional descriptive statistics on dimensions and indicators of CSR are shown in table 5 in the appendix. This table shows the descriptive statistics of the seven dimensions of CSR. Strengths and concerns of each dimension are shown in different panels. This is done to be able to identify them in detail. Means, standard deviations, minimum and maximum values are based on the number of social practices a firm is involved in. So the mean of strengths in the environment dimension for example, stands for the mean of positive environmental
practices firms are involved in. Besides the standard output, the total number of indicators and the mean6 of the number of indicators observed in each dimension is shown. In this way the standard output can be interpreted better, as the observed indicators were eventually scaled. All variables in both tables have 9,108 observations on them.
Table 2
Descriptive statistics of key variables.
Each variable has a number of 9,108 observations.
4.2 Model specifications and assumptions
The models used to investigate the CSR-CFP link are statistically tested by an OLS (ordinary least squares) regression. However, as panel data is involved, a model specifically meant for panel data is required. In this case a random effects model is required, in which variables are
6 As the number of observed indicators can differ for each firm and each year, it is not possible to work with a standard amount of observed indicators. Therefore dimension means of the observed indicators are shown.
Variable Mean Standard deviation Minimum Maximum
ROA 0.019 0.178 -3.738 2.462
ROE 0.071 2.976 -81.851 86.082
Tobin’s Q 1.925 1.835 0.064 30.814
CSR score 0.055 0.157 -0.368 1.000
Size 7402.826 26440.540 2.789 717242
R&D, scaled to sales 0.020 0.925 0 75.973
18
assumed to be random variables, in contrast to some fixed effects of variables. Variables are assumed to be random, as it is expected that firms differ in how their CSR policy influences CFP. Also yearly fixed effects are excluded, because indicators differ from year to year, as many indicators have limited timeframes.
Some variables (size and Tobin’s Q) were assumed and found to have non-normal distributions. To avoid the corresponding biases, logarithms of these variables were used. Heteroscedasticity was taken into account by using robust standard errors in the regression. Multicollinearity was checked by the correlation matrix as shown in table 3, below. No strong correlations were found, and therefore no issues caused by multicollinearity were expected.
Table 3
Correlation matrix of variables.
Variable ROA ROE Ln(Tobin’s Q) CSR score Ln(size) R&D Leverage
ROA 1.000 ROE 0.080 1.000 Ln(Tobin’s Q) -0.013 0.019 1.000 CSR score 0.084 0.029 0.030 1.000 Ln(size) 0.244 0.026 -0.279 0.462 1.000 R&D -0.035 -0.002 0.017 -0.023 -0.029 1.000 Leverage -0.091 0.011 -0.060 0.053 0.283 -0.011 1.000 4.3 Regression results
The results of the regression are shown below, in table 4. Three regressions are included. Model effects on return on assets, return on equity and Tobin’s Q are shown, as caused by CSR and the control variables. The results on the accounting-based performance (ROA and ROE) are two-sided, as the effect of CSR on the return on assets is significantly negative, while return on equity is positively affected by CSR. The negative effect on ROA differs from previous empirical research, but on the other hand does CSR show a significant small positive sign for return on equity, which is in line with previous results.
The found effect of CSR on market-performance is nevertheless plain; it is positively significant on the 1% level. This might be caused by the expected strong CSR-CFP link, as motivated by the popularity of CSR. CSR seems to affect market-performance more positively than accounting performance. The results implicate that shareholders benefit the most from engaging in CSR, while the overall firm, as measured by total assets, sees its performance fall.
19
Table 4
Results of the regression.
*** denotes significance at 1%, ** at 5% and * at 10%. Robust standard errors in parentheses. Size and Tobin’s Q are in logarithmic form.
The reason of the ambiguous findings on ROA and ROE is unclear. A potential reason is that the expectation that CSR improves financial performance through a management’s use and control of assets isn’t (entirely) right. A more direct link could be the case, in which direct investments (in CSR) of investors or shareholders are causing the better performance, which is more accurately measured by the return on equity. Another reason could be the possibility of some bias in the model, as there are findings of reverse causality in previous research. This would cause the financial performance to affect CSR as well, and therefore the model fails to control entirely for endogeneity.
Another remarkable result is that Tobin’s Q is significantly stronger affected by CSR than ROE (focusing on positive effects only). This is in contrast to previous results, as
OLS with random effects (1) ROA (2) ROE (3) Tobin's Q CSR score -0.056 (0.013)*** 0.414 (0.236)* 0.794 (0.050)*** Size 0.035 (0.003)*** 0.027 (0.038) -0.140 (0.005)*** R&D, scaled to sales -0.005
(0.006) -0.005 (0.011) 0.008 (0.005) Risk -0.140 (0.023)*** 0.077 (0.248) 0.169 (0.034)*** Industry 2 -0.101 (0.018)*** -0.224 (0.165) -0.208 (0.099)** Industry 3 -0.025 (0.016) -0.053 (0.050) -0.318 (0.099)*** Industry 4 -0.043 (0.016)*** -0.016 (0.068) 0.117 (0.097) Industry 5 -0.019 (0.016) 0.013 (0.133) -0.087 (0.098) Industry 6 -0.019 (0.016) -0.147 (0.160) -0.182 (0.100)* Industry 7 0.015 (0.016) 0.169 (0.109) 0.132 (0.099) Industry 8 0.012 (0.020) 0.109 (0.105) -0.093 (0.108) Industry 9 -0.023 (0.016) -0.102 (0.062)* 0.175 (0.097)* Industry 10 -0.168 (0.048)*** -0.728 (0.618) 0.246 (0.130)* Constant -0.169 (0.023)*** -0.140 (0.262) 1.283 (0.103)*** R-squared 0.1125 0.0395 0.1186 Observations 9,108 9,108 9,108
20
Margolis et al. (2007) find CSR to affect accounting performance more than market performance. However, as mentioned by Dimson et al (2015), the market-performance is affected by a reduction in agency costs, through the engagement in CSR. Theory does
nevertheless suggest an effect on accounting-based performance as well. A possible reason for accounting performance to stay behind could be that the reduction of agency costs isn’t
necessarily caught in the published documents of that year. As a result, effects of CSR engagement are observed in the market performance, but not in the accounting-based performance.
5 Conclusion
In this section concluding remarks about the research are given. First the performed research is summarized, followed by a discussion of limitations of the research and recommendations for further research.
5.1 Summary
In this paper the relationship between CSR and CFP was investigated. Although there is no consensus on the definition of CSR, corporate social responsibility was defined as combining targets of maximizing value for shareholders with broader environmental and social goals, which go beyond the direct obligations of the firm. Because of the multidimensional nature of CSR, CSR was measured as an ethical rating consisting of ratings on multiple dimensions, as provided by the MSCI-ESG database. The hypothesis was that engagement in CSR leads to a better financial performance. To test the hypothesis an OLS random effects model was used. The findings on accounting-based performance are two-sided, as the return on assets is found to be negatively influenced by CSR, while return on equity is found to be positively affected. Therefore no conclusions can be drawn on the effect CSR has on accounting-based performance. Findings on market performance are nevertheless plain, as CSR is found to have a strong positive effect on Tobin’s Q. Therefore results implicate that especially shareholders benefit from engaging in CSR, while effects for the overall firm are controversial. The
findings of such a strong relationship between CSR and any kind of financial performance are rare in the literature. The recent timeframe of the data might be involved in this large effect, as corporate social responsibility has grown extensively in the last years. Managers might have engaged more in CSR because they observed that the engagement in CSR helps firms to maximize their financial performance. The final conclusion is that engaging in CSR leads to a
21
better market-performance of firms, while the effect on accounting-based performance remains unclear.
5.2 Discussion
Due to the time limit and complications in CSR methodology, this paper has some limitations. First of all, the model used for regression possibly has issues of endogeneity, as CSR is
directly regressed on financial performance. Reason for the endogeneity is that previous research showed the existence of a reverse causality between CSR and financial performance, in which financial performance also affects the social performance of a firm. There is,
however, not an obvious solution to this problem. Some researchers (e.g. Waddock & Graves, 1997) tried a time-lagged approach in which effects of CSR were measured on later financial performance. Problem of this method is that CSR of previous years isn’t an accurate measure, because financial performance tends to be more affected by recent social performance. A better solution might be to perform an instrumental variables regression (as done by Ferrell et al., 2016), but there is no consensus on which instrument should be used to find effects of CSR. Since IV regressions seem to be best in investigating the CSR-CFP link, further research into the use of IV regressions is recommended.
Another limitation is the lack of weighting schemes of the several dimensions and indicators of social practices, because some practices are observed to be more important than others (Graves & Waddock, 1994). As there is no general weighting scheme in the literature (Van der Laan et al., 2008), all dimensions and indicators were treated equally. Therefore a suggestion for further research is to look into weighting schemes, as creating optimal weighting schemes makes the CSR measure more accurate.
This paper is also limited in measuring market-performance, as it does only measure effects on Tobin’s Q. Better is to investigate effects on share returns by using an event study. Event studies give more accurate results because they control for external factors that affect share price movements. Because of the time limit, it was decided to not include an event study or any other way to measure effects on share returns.
22 References
Aras, G., Aybars, A., & Kutlu, O. (2010). Managing corporate performance. International Journal of Productivity and Performance Management, 59(3), 229-254.
Barney, J. (1991). Firm resources and sustained competitive advantage. Journal of Management, 17(1), 99–120.
Berk, J., & DeMarzo, P. (2014). Corporate Finance, third edition. Harlow: Pearson.
Carroll, A.B. (1979). A three-dimensional conceptual model of corporate performance. The Academy of Management Review, 4(4), 497-505.
Cochran, P.L., & Wood, R.A. (1984). Corporate social responsibility and financial performance. The Academy of Management Journal. 27(1), 42-56.
Coombs, J. E., & Gilley, K. M. (2005). Stakeholder management as a predictor of CEO compensation: Main effects and interactions with financial performance. Strategic Management Journal, 26(9), 824– 840.
Cornell, B. & Shapiro, A. (1987). Corporate stakeholders and corporate finance. Financial Management, 16(1), 5–14.
Dahlsrud, A. (2008). How corporate social responsibility is defined: An analysis of 37 definitions. Corporate Social Responsibility and Environmental Management, 15(1), 1–13.
Davis, K. (1973). The case for and against business assumption of social responsibilities. Academy of Management Journal, 16(2), 312–322.
Dimson, E., Karakaş, O., & Li, X. (2015). Active Ownership. The Review of Financial Studies, 28(12), 3225–3268.
European commission. (2011). A renewed EU strategy 2011-14 for corporate social responsibility. Retrieved from http://www.eurlex.europa.eu
Ferrell, A., Liang, H., & Renneboog, Luc. (2016). Socially responsible firms. Journal of Financial Economics, 122(3), 585-606.
Friedman, M. (1970, September 13). The social responsibility of business is to increase its profits. The New York Times. Retrieved from http://www.nytimes.com
Freeman, R.E. (1984). Strategic management: a stakeholder approach. Boston: Pitman. Garcia-Castro, R., Ariño, M. A., & Canela, M. A. (2010). Does social performance really lead
to financial performance? Accounting for endogeneity. Journal of Business Ethics, 92(1), 107–126.
23
Graafland, J.J., Eijffinger, S.C.W., & Smid, H. (2004). Benchmarking of corporate social responsibility: methodological problems and robustness. Journal of Business Ethics, 53(1), 137-152.
Graves, S.B., & Waddock, S.A. (1994). Institutional owners and corporate social performance. The Academy of Management Journal, 37(4), 1034-1046.
Jensen, M. (2001). Value maximization, stakeholder theory and the corporate objective function. Journal of Applied Corporate Finance, 14(3), 8-21.
Karagiorgos, T. (2010). Corporate social responsibility and financial performance: An empirical analysis on Greek companies. European Research Studies, 13(4), 85–108. Krüger, P. (2015). Corporate goodness and shareholder wealth. Journal of Financial
Economics, 115(2), 304-329.
Margolis, J. D., Elfenbein, H. A., & Walsh, J. P. (2007). Does it pay to be good? A meta-analysis and redirection of research on the relationship between corporate social and financial performance. Unpublished working paper. Harvard University, University of California – Berkeley, and University of Michigan.
McGuire, J. B., Sundgren, A., & Schneeweis, T. (1988). Corporate social responsibility and firm financial performance. Academy of Management Journal, 31(4), 854–872. McPherson, S. (2017, January). 6 CSR trends to watch in 2017. Forbes. Retrieved from
http://www.forbes.com
McWilliams, A., & Siegel, D. (2000). Corporate social responsibility and financial
performance: correlation or misspecification? Strategic Management Journal, 21(5), 603-609.
McWilliams, A., Siegel, D.S., & Wright, P. M. (2006). Corporate social responsibility: strategic implications. Journal of Management Studies, 43(1), 1–18.
Orlitsky, M., & Benjamin, J.D. (2001). Corporate social performance and firm risk: a meta-analytic review. Business & Society, 40(4), 369-396.
Orlitsky, M., Schmidt, F.L., & Rynes, S.L. (2003). Corporate social and financial performance: a meta-analysis. Organizational Studies, 24(3), 403-441.
Rodgers, W., Choy, H., & Guiral, A. (2013). Do Investors Value a Firm’s Commitment to Social Activities? Journal of Business Ethics, 114(4), 607-623.
Ruf, B. M., Muralidhar, K., Brown, R. M., Janney, J. J., & Paul, K. (2001). An empirical investigation of the relationship between change in corporate social performance and financial performance: a stakeholder theory perspective. Journal of Business Ethics, 32(2), 143–156.
24
Soana, M. (2011). The relationship between corporate social performance and corporate financial performance in the banking sector. Journal of Business Ethics, 104(1), 133-148.
Ullmann, A. (1985). Data in search of a theory: a critical examination of the relationship among social performance social disclosure, and economic performance. Academy of Management Review, 10(3), 540–577.
Van der Laan, G., Van Ees, H., & Van Witteloostuijn, A. (2008). Corporate social and financial performance: an extended stakeholder theory, and empirical test with accounting measures. Journal of Business Ethics, 79(3), 299-310.
Waddock, S.A., & Graves, S.B. (1997). The corporate social performance–financial performance link. Strategic Management Journal, 18(4), 303-319.
Wood, D.J., & Jones, R.E. (1995). Stakeholder mismatching: a theoretical problem in empirical research on corporate social performance. The International Journal of Organizational Analysis, 3(3), 229-267.
Yang, F., Lin, C., & Chang, Y. (2009). The linkage between corporate social performance and corporate financial performance. African Journal of Business Management, 4(4), 406–413.
25 Appendix
Table 4
Variable list of all indicators included in this paper, matched to the dimension the indicators belong to.
Dimension Strengths (positive indicators) Concerns (negative indicators)
Community -charitable giving -community engagement -innovative giving -other strengths -community impact Corporate governance (investors)
-corruption & political instability -financial instability
-public policy strength -reporting quality
-business ethics
-controversial investments -governance structure -other concerns -public policy concern -reporting quality Diversity -board of directors - gender
-employment of underrepresented groups -gay & lesbian policies
- other strengths -promotion
-women & minority contracting -work-life benefits
-board of directors - gender -board of directors - minorities -non-representation
-workforce diversity
Employee -cash profit sharing
-compensation & benefits -controversial sourcing
-employee involvement -employee health & safety -employee relations
-human capital management -other strengths
-professional development -supply chain labor standards -union relations
-child labor
-employee health & safety -labor-management relations - supply chain
-union relations
Environment -biodiversity & land use -climate change -electronic waste -energy efficiency
-environmental management systems -environmental opportunities
-biodiversity & land use -climate change
-impact of products & services -operational waste
-other concerns -regulatory compliance
26
Environment (continued) -green buildings
-insuring climate change risk -natural resource use -other strengths
-packaging materials & waste -product carbon footprint -raw material sourcing -renewable energy -waste management -water stress
-supply chain management -toxic spills & releases -water management
Human rights -human rights policies & initiatives -indigenous peoples relations strength
-freedom of expression & censorship -human rights violations
-other concerns
-operations in Sudan -support for controversial regimes Product
(customers)
-access to finance
-access to communications -chemical safety
-financial product safety
-insuring health & demographic risk -opportunities in nutrition & health -quality
-privacy & data security -responsible investment -social opportunities
-anticompetitive practices -customer relations -marketing & advertising -other concerns product -privacy & data security -quality & safety
27
Table 5
Descriptive statistics of strengths and concerns of all categories of corporate social responsibility.
Each dimension has a number of 9,108 observations. Descriptive statistics on strength’s and concern’s scores are based on the aggregated value of all dimensions.
Panel A. Strengths of corporate social responsibility Dimension
Mean
Standard
deviation Minimum Maximum
Number of indicators Mean of observed indicators Community 0.091 0.340 0 4 4 0.853 Customer 0.125 0.353 0 2 10 1.092 Diversity 0.207 0.639 0 7 7 1.499 Employee 0.476 1.012 0 8 11 4.344 Environment 0.457 0.969 0 6 16 4.180 Human rights 0.057 0.290 0 2 2 0.661 Investor 0.084 0.298 0 2 4 0.709 Total 1.498 2.504 0 22 54 13.339 Strength’s score 10.14% 14.91% 0% 100% Panel B. Concerns of corporate social responsibility
Dimension
Mean
Standard
deviation Minimum Maximum
Number of indicators Mean of observed indicators Community 0.026 0.159 0 1 1 1.000 Customer 0.116 0.395 0 4 6 4.773 Diversity 0.508 0.716 0 3 4 2.353 Employee 0.108 0.381 0 5 5 4.622 Environment 0.115 0.425 0 5 9 7.628 Human rights 0.024 0.157 0 2 5 3.321 Investor 0.249 0.470 0 3 6 4.066 Total 1.146 1.414 0 16 36 27.763 Concern's score 4.67% 6.12% 0% 66.67%
