Corporate Governance and Executive Compensation for Corporate Social Responsibility

Abstract

We link the corporate governance literature in financial economics to the agency cost perspective of corporate social responsibility (CSR) to derive theoretical predictions about the relationship between corporate governance and the existence of executive compensation incentives for CSR. We test our predictions using novel executive compensation contract data, and find that firms with more shareholder-friendly corporate governance are more likely to provide compensation to executives linked to firm social performance outcomes. Also, providing executives with direct incentives for CSR is an effective tool to increase firm social performance. The findings provide evidence identifying corporate governance as a determinant of managerial incentives for social performance, and suggest that CSR activities are more likely to be beneficial to shareholders, as opposed to an agency cost.

Introduction

Over the last several decades, corporate social responsibility (CSR) activities have become an increasingly important investment by firms.¹ The growing significance of CSR as a phenomenon has raised a fundamental question: Does CSR enhance shareholder value, or is it an agency cost enjoyed by a firm’s managers at the expense of stockholders? While a substantial number of studies have examined this question from different perspectives, the evidence continues to be conflicting (Borghesi et al. 2014; Griffin and Mahon 1997; Kruger 2015; Margolis et al. 2009; Masulis and Reza 2015). In this study, we investigate the agency cost perspective of CSR using a novel empirical test that exploits variation in corporate governance across firms to predict the existence of executive compensation contracts linked directly to CSR activities. Our findings suggest that corporate governance is an important mechanism determining whether managers receive compensation linked to firm social performance outcomes, and that executive compensation for CSR leads to more CSR activities. To the best of our knowledge, our study provides the first quantitative empirical evidence identifying factors that lead to executive compensation for CSR.

In the debate among scholars about the effects of CSR on firm financial performance, one of the most prominent arguments against the financial benefits of CSR has been the agency cost prediction first made by Friedman (1970), who characterized CSR activities as self-interested behavior by individual managers at the expense of the firm’s shareholders.² Subsequent studies have found supporting evidence of CSR as a potential agency cost, finding that CSR may be used to advance personal interests over the interests of shareholders (Borghesi et al. 2014; Brown et al. 2006; Cheng et al. 2014; Jiraporn and Chintrakarn 2013; Kruger 2015; Masulis and Reza 2015), provide added job security to inefficient managers by pleasing stakeholders (Cespa and Cestone 2007), compensate for the negative consequences of engaging in earnings management (Prior et al. 2008), and enhance individual reputations of managers (Barnea and Rubin 2010). However, a number of studies have also found a positive relationship between CSR activities and firm financial performance (Orlitzky et al. 2003). Despite the considerable amount of academic attention, few definitive conclusions can be drawn from the collection of findings produced thus far.

In addition to the conflicting academic evidence, anecdotal evidence from firms with strong public commitments to CSR can also provide ambiguous conclusions. For example, consider the outdoor clothing company Patagonia, which donates 1 % of its revenues to environmental organizations. In a case study by Casadesus-Masanell et al. (2009, p. 209), a former senior manager at the company provides potentially conflicting views of the financial benefits of CSR, stating both that they seek to “dispel the myth that in order to have a high quality product you have to have something damaging to the environment,” and then later stating “There is some tension between the environment and product quality…the reality is that they don’t always go hand in hand.” These seemingly opposing statements raise more questions than provide answers in addressing whether CSR is truly beneficial for firm financial performance, or a net cost to the firm’s shareholders.

With respect to agency costs, a fundamental premise of the corporate governance literature within the field of financial economics is the notion that improved corporate governance ultimately leads to improved firm financial performance and value created for shareholders through the adoption of shareholder-friendly policies and the reduction of agency costs (Gompers et al. 2003, 2010). More shareholder-friendly (better) corporate governance is achieved through the implementation of rules, practices, and incentives to align the interests of a firm’s managers with shareholders. As a consequence, shareholders benefit economically by advocating for improved corporate governance.

In this paper, we link the corporate governance literature in financial economics and the agency cost perspective of CSR to derive theoretical predictions about the relationship between corporate governance and the existence of executive compensation contracts that provide incentives for firm social performance. The underlying theoretical logic of our test is a straightforward extension of the agency cost perspective: If CSR is truly an agency cost at the expense of a firm’s shareholders, then firms with better corporate governance should be less likely to compensate their managers for CSR outcomes. If, however, CSR increases shareholder value, better corporate governance should predict a higher likelihood of observing executive compensation contracts that provide incentives for firm social performance. Given that the structure of executive compensation contracts is a direct outcome of a firm’s governance process (e.g., Bebchuk and Fried 2004; Core et al. 1999; Yermack 1997), we are able to infer whether CSR activities are truly agency costs or beneficial for firm financial performance.

The relationship between corporate governance and CSR activities within firms has received increasing attention among scholars, motivated by the potentially important role of governance in influencing socially responsible firm behavior (see Walls et al. 2012 for a review).³ However, the literature has thus far produced contradictory results, finding evidence of negative (Coombs and Gilley 2005; David et al. 2007), insignificant (Schnatterly 2003; Waddock and Graves 1997), and positive (Coffey and Fryxell 1991; Jo and Harjoto 2011; Oh et al. 2011) relationships between corporate governance and CSR. We argue one possible reason for the inconsistency in previous findings may be the lack of empirical measures at the individual manager level of the channels through which corporate governance may influence CSR. Given that the purpose of corporate governance is to influence managerial decision making (Zeckhauser and Pound 1990), the effects of governance on firm social performance are likely to occur through its influence on individual managers. A more detailed examination of these channels may be necessary in order to better understand the nature of the relationship between corporate governance and CSR.

Similarly, a nascent literature has examined the effect of executive compensation on CSR, but has also found conflicting results (Mahoney and Thorn 2006). McGuire et al. (2003) find no significant relationship between incentives and firm social performance, while Deckop et al. (2006), Berrone and Gomez-Mejia (2009a),⁴ and Mahoney and Thorne (2005) find evidence of a positive relationship. However, as Berrone and Gomez-Mejia (2008) note, too few studies have been conducted thus far to reach any clear conclusions. Also, while prior studies have examined the relationship between executive compensation structure and CSR, no quantitative empirical study has yet examined the role of executive compensation contracts that explicitly incentivize managers for firm social performance.

We conduct our empirical test on hand-collected compensation contract data for the top five highest paid executives of firms in the Standard and Poor’s 500 Index in 2013, linked to corporate governance measures commonly used in the financial economics literature. The results provide consistent evidence that better corporate governance predicts a greater likelihood of observing executive compensation contracts with incentives linked to CSR outcomes. A one standard deviation increase in the percentage of the board of directors hired prior to a CEO’s arrival predicts a 13 % increase in the odds of compensation for CSR, while the incremental addition of a large institutional (block) shareholder is associated with an 8 % increase in the odds of observing CSR incentives.⁵ We also find that when managers have greater individual power within the firm and governance is less shareholder-friendly (weaker), they are less likely to have executive compensation contracts tied to CSR outcomes. A one standard deviation increase in the shares outstanding owned by an executive predicts an 8 % decline in the odds of observing incentives linked to CSR, and the odds decline 19 % if the executive is also a member of the board of directors. Taken together, the results provide evidence against the notion of CSR as an agency cost at the expense of shareholders, and suggest that CSR activities are likely to provide at least some form of economic benefit for firms. In addition, we find a positive relationship between the presence of explicit incentive compensation for CSR and actual firm-level CSR activities.

Our paper makes five main contributions. First, our findings establish explicit compensation for CSR as one important channel through which corporate governance may influence CSR outcomes, identifying a clear mechanism through which corporate governance may influence firm social performance. Second, by being the first to directly measure CSR contracting, we are able to demonstrate a clear positive relationship between a component of executive compensation structure (incentives for CSR) and firm social performance, contributing to the growing literature examining how executive compensation and CSR activities might be related. Third, we contribute more broadly to the literature examining executive compensation for non-financial performance measures, which finds that including incentives for value-relevant performance measures beyond purely financial performance metrics can improve managerial incentives (Chen et al. 2014; Davila and Venkatachalam 2004; Ibrahim and Lloyd 2011; Ittner et al. 1997; Schiehll and Bellavance 2009). Our results suggest that CSR is a non-financial performance measure that contains value-relevant information, and that corporate boards may implement CSR-linked incentives in order to increase shareholder value. Fourth, to our knowledge, our results provide the first empirical evidence of factors that predict the existence of compensation incentives linked to social performance outcomes. Surprisingly, although almost 40 % of the executives in our sample have compensation contracts that contain explicit incentives directly linked to CSR, little is known about the conditions under which such contracts are more or less likely to exist (Berrone and Gomez-Mejia 2009b). Our findings provide evidence of the role of corporate governance as one mechanism that may predict the emergence of such social performance-based incentives. Finally, our results provide clear supporting evidence of CSR as important for increasing shareholder value, complementing a number of studies that find a positive relationship between CSR and firm financial performance, and contrary to the agency cost prediction.

The remainder of this paper is organized as follows. In “Hypotheses” section, we develop our hypotheses linking corporate governance and executive compensation contracts tied to CSR. In “Data and Measures” section, we describe our data and measures, and our empirical analysis. In “Results and Discussion” section, we present and discuss the empirical results. In “Conclusion” section, we offer conclusions.

Hypotheses

Corporate governance is concerned with the mechanisms through which shareholders ensure a financial return on their investments (Shleifer and Vishny 1997). The fundamental dilemma of corporate governance is the imperfect alignment of incentives between shareholders and managers, which can lead to behaviors and decisions by managers that are not in the interests of the firm’s shareholders (Roe 1994).⁶ When managers act in their own personal interest at the expense of shareholders, this results in agency costs for the firm’s owners, reducing the level of shareholder value (Jensen and Meckling 1976). To assuage this cost, shareholders can use managerial incentives to help align the manager’s interests with those of shareholders.

The degree of agency cost within a particular firm is then determined by the balance of control between (independent) shareholders and managers—the strength of the corporate governance of the firm. As the balance of control increasingly favors shareholders, incentives are put in place such that managers are more likely to engage in shareholder-maximizing activities. By contrast, as managers have more control, they have greater discretion to engage in activities at the expense of shareholders. Thus, we develop our hypotheses by using a set of well-established indicators of the strength of corporate governance to predict the likelihood of CSR contracting.

The optimal contracting hypothesis (e.g., Grossman and Hart 1983; Holmstrom 1979) contends that boards of directors bargain at arms-length as shareholders’ loyal agents, and minimize agency costs and maximize firm value by actively monitoring the executives and optimally assigning executives with incentives and responsibility. However, the managerial power hypothesis (e.g., Bebchuk et al. 2002; Bertrand and Mullainathan 1999) argues that boards do not always bargain at arms-length, especially when board members are affiliated with the executives and the board is not independent from management. We begin by considering the average director tenure of the board to proxy for board independence. Social network theory suggests that board members develop and solidify their friendship or social ties with management as their tenure on the board increases, making them less independent (e.g., Boeker and Goodstein 1993; Harris and Helfat 2007; Vafeas 2003; Wade et al. 1990). Activists and governing institutions such as the National Association of Corporate Directors (1996), Council of Institutional Investors (1998), and U.S. Senate (2002) believe that longer service on the board does not bode well for monitoring management, and thus suggest tenure limits as a policy prescription.⁷

However, director tenure as a proxy for board independence could be ambiguous in its interpretation. An alternative theoretical perspective suggests a potential benefit of longer board service. The managerial talent paradigm posits that directors accumulate considerable experience and skill as their tenure on the board increases (Buchanan 1974; Salancik 1977; Vance 1983). Such directors are more confident and powerful, and more likely to challenge management when necessary to serve shareholder interests.

Therefore, following Harjoto et al. (2014) and Coles et al. (2014), we also consider another measure of board independence, the percentage of board members hired before the CEO, which we argue is less ambiguous. If a board member is hired after the CEO, they are more likely to be “sympathetic” because the CEO often exerts considerable influence in the board nomination process (Finkelstein and Hambrick 1988; Hwang and Kim 2009). To test our hypotheses of board independence, we consider both measures.

When considering executive compensation contracts, if CSR activities are an agency cost,⁸ contractual incentives for CSR should be less likely to exist when the firm’s board of directors has greater independence. However, if CSR activities maximize shareholder value, greater board independence should predict a greater likelihood of incentives explicitly tied to CSR. We present these two competing hypotheses formally here:

H1a

If CSR maximizes shareholder value, board independence makes it more likely that a firm contracts on CSR.

H1b

If CSR is an agency cost, board independence makes it less likely that a firm contracts on CSR.

In addition to board composition, shareholders themselves may serve as an important enforcement mechanism to reduce agency costs. Large institutional shareholders, by virtue of their significant ownership in firms, have both the incentives and power to monitor the decisions and activities of a firm’s managers (Shleifer and Vishny 1986). Consistent with this hypothesis, institutional shareholders have been found to play an important role in preventing the enactment of amendments harmful to shareholders (Brickley et al. 1988; Jarrell and Poulsen 1988), improving compensation practices by linking pay more directly with performance outcomes (Hartzell and Starks 2003; Matsumura and Shin 2005), and enhancing firm value as measured by Tobin’s Q (McConnell and Servaes 1990). As a result, the presence of large institutional shareholders suggests that agency costs should be reduced within firms, leading to a testable set of competing hypotheses about their relationship with the presence of incentives for CSR in executive compensation contracts.

H2a

If CSR maximizes shareholder value, the likelihood of contracting on CSR is increasing in the number of large institutional shareholders.

H2b

If CSR is an agency cost, the likelihood of contracting on CSR is decreasing in the number of large institutional shareholders.

In contrast to strong corporate governance, increased managerial power means boards do not always bargain at arm’s length because of management’s influence over them. In these cases, managerial power may be excessive compared to the efficient level suggested by optimal contracts (e.g., Bebchuk et al. 2002; Bertrand and Mullainathan 1999). When managers have excessive power and become “entrenched,” agency costs are much more likely to occur at the expense of shareholders (e.g., Bebchuk et al. 2002; Shleifer and Vishny 1989). Possessing higher degrees of managerial power also allows managers to have greater influence over how they are compensated, leading to overcompensation of managers through contract terms that are less transparent or more difficult to value (Bebchuk and Fried 2004). Consequently, managers with greater influence over the structure of their compensation contracts will be more likely to have incentives for CSR if social performance activities represent an agency cost. For example, if CSR outcomes are more difficult to measure, CSR-based compensation may be a means by which some managers can more easily extract additional compensation. Alternatively, if CSR is not an agency cost, greater managerial power should predict fewer CSR-based contracts.

To measure managerial power, we begin by considering the percentage of shares outstanding owned by individual executives at a firm, following previous studies in the financial economics literature considering the effects of managerial power on firm outcomes (e.g., Davila and Venkatachalam 2004; Linck et al. 2008; Moeller 2005). As the percentage of shares owned by an individual manager increases, the effectiveness of corporate governance mechanisms regulating their decisions becomes weaker, leading to greater influence on their part in determining the structure of their compensation contracts. We integrate the notion of managerial power with the perspective of CSR as an agency cost versus a shareholder value-enhancing activity with the following competing predictions:

H3a

If CSR maximizes shareholder value, the likelihood of contracting on CSR is decreasing in the level of top-manager ownership.

H3b

If CSR is an agency cost, the likelihood of contracting on CSR is increasing in the level of top-manager ownership.

As an additional measure of managerial power, we consider instances where executives are also members of a firm’s board of directors. In contrast to the percentage of board members who are independent of the firm, overlapping membership between the top management team and a firm’s board of directors can diminish the strength of a firm’s corporate governance in regulating the decisions of managers, since managers who are board members can directly influence the board’s decision-making process, including their compensation (e.g., Bebchuk et al. 2011; Finkelstein and Hambrick 1988; Li 2014). We consider this additional measure of managerial power here:

H4a

If CSR maximizes shareholder value, the likelihood of contracting on CSR is reduced when the manager is also a director.

H4b

If CSR is an agency cost, the likelihood of contracting on CSR is increased when the manager is also a director.

A necessary condition for the previous hypotheses to be valid is that providing incentives to managers to engage in CSR actually generates higher levels of CSR within firms, suggesting that the incentive should be effective in increasing the level of CSR engagement by firms. If instead incentive contracts tied to CSR did not result in greater social performance, then stronger corporate governance is unlikely to predict a higher likelihood of observing executive compensation contracts linked to CSR. Ultimately, CSR must provide sufficient returns for firm financial performance in order to be worth investing in, and executive compensation contracts providing incentives for CSR must lead to actual improved social performance within firms. Our final hypothesis captures this necessary condition:

H5

The existence of executive compensation incentives for CSR is associated with an increase in firm CSR.

Data and Measures

To test our hypotheses, we consider the sample of top five executives working at each firm in the Standard and Poor’s 500 Index (S&P 500) disclosed in public company filings to the Securities and Exchange Commission (SEC). The S&P 500, which comprises approximately 80 % of the entire US stock market in terms of market value (McGraw Hill Financial 2015), has been used widely for empirical analysis in related areas, including influential studies such as Hart and Ahuja (1996) and Stanny (2013) on emission reduction; Walls et al. (2012) on the relationship between corporate governance and environmental performance; Hall and Liebman (1997) on CEO compensation; and Anderson and Reeb (2003) on family ownership.

To conduct our analysis, we link relevant data from multiple sources. We begin by considering the universe of top managers available in the Execucomp database provided by S&P Capital IQ, and examine those executives working at firms comprising the S&P 500 index in 2012. All data from other sources are for the year 2012, except for CSR compensation contract data, which are for the year 2013. By lagging the independent variables by 1 year, we estimate how corporate governance predicts the subsequent existence of incentives for CSR. Execucomp includes a variety of personal information disclosed in public company filings for the top five executives at each firm. For our measure of CSR, we use data provided by Kinder, Lydenberg, and Domini (KLD), a third-party data vendor that provides CSR ratings for publicly traded firms. KLD ratings are among the most influential and widely studied measures of CSR used in prior literature (Berman et al. 1999; Chatterji et al. 2009; Di Giuli and Kostovetsky 2014; Flammer 2014; Flammer and Kacperczyk 2015; Servaes and Tamayo 2013; Werner 2014).⁹ Financial statement data are from Compustat, and data on board independence are from the Investor Responsibility Research Center‘s (IRRC) governance and director database. We obtain institutional ownership data from Thomson Financial, consisting of 13F filings reported to the Securities and Exchange Commission.

To obtain data on executive compensation contracts tied to CSR, we manually collected relevant compensation data from 2014 Proxy Statements, which describe 2013 compensation activities, for each firm where our sample of executives is employed. Specifically, in the Compensation Discussion and Analysis section, sources of variable pay and performance-based pay are identified. If compensation is based on non-financial performance and related to CSR activities (see Appendix 2 for a detailed description), an executive is coded as having incentives linked to CSR, and her firm is coded as offering CSR incentives. In addition, each company’s most recent annual report and CSR report were reviewed for any additional details, although these latter sources were not used to identify the existence of compensation linked to CSR. By following this approach, only firms that formally disclosed CSR-linked incentives to regulators and investors were recorded as providing compensation linked to CSR. For our analysis, we create a dummy variable equal to one (variable CSRCONTRACTING) if an executive’s compensation contract contains incentives linked to CSR. After excluding any observations without KLD, Compustat, and Execucomp data, the full sample consists of 2561 executive-level observations. Some specifications had fewer observations due to one or more missing data values.

As mentioned earlier, to measure the degree of independence the firm’s board of directors has from management, we calculate the average director tenure of the board (AVG TENURE), following previous studies (e.g., Harris and Helfat 2007; Vafeas 2003). As an additional measure of board independence, we follow Harjoto et al. (2014) and Coles et al. (2014), and use the percentage of board members hired before the CEO’s hire (% HIREBEFORE). In our results, we present our analysis for both AVG TENURE and % HIREBEFORE as distinct measures of board independence.

Following Ashbaugh-Skaife et al. (2006), we measure the number of large institutional shareholders by calculating the number of block shareholders (NUMBLOCKS), where a block shareholder is defined as an institution that owns more than 5 % of a firm’s outstanding voting shares. Jensen (1993) and Shleifer and Vishny (1997) argue that block shareholders that hold large equity positions in a company are important to a well-functioning governance system because they have the financial interest and independence to evaluate firm management and policies in an unbiased manner, and have the voting power to pressure management to make changes if they observe agency cost behavior. Consistent with this view, Gordon and Pound (1993) find that the structure of share ownership significantly influences voting outcomes on shareholder-sponsored proposals to change corporate governance structures. Block shareholders have also been found to improve financial performance through their efforts to improve corporate governance. For example, Nesbitt (1994) finds that firms targeted by the California Public Employees’ Retirement System (CalPERS) experience positive long-run stock returns, and Opler and Sokobin (1997) find that firms experience above-market performance the year after being targeted by the Council of Institutional Investors. These results suggest that active institutional shareholders lead to more efficient monitoring of management and less agency cost behavior.

For our measures of managerial power, we follow prior empirical studies in the financial economics literature examining executive compensation to create two distinct variables. For our first measure, we calculate the percentage of shares outstanding owned by executives at a firm (e.g., Davila and Venkatachalam 2004; Linck et al. 2008; Moeller 2005). For our second measure, we create a dummy variable equal to one if an executive is also a member of the firm’s board of directors (e.g., Bebchuk et al. 2011; Finkelstein and Hambrick 1988). Both of these variables, SHROWN and EXECDIR, are constructed based on Execucomp data.

Controlling for the level of firm social performance is important for our analysis, because it may be that firms with better corporate governance also engage in more CSR activities, leading to potential omitted variable bias in our regressions and incorrect inferences. To measure the level of CSR activities by firms, we calculate the total number of KLD strengths minus concerns over a set of categories measuring CSR (variable CSRLEVEL), following the convention used by previous empirical studies (Griffin and Mahon 1997; Johnson and Greening 1999; Mishra and Suar 2010; Waddock and Graves 1997). We exclude industry-specific categories and the corporate governance category, and consider the KLD categories of community, diversity, employee relations, environment, human rights, and product in our measure.

We also include several other control variables commonly used in the CSR and executive compensation literature in our analysis. To control for differences between the Chief Executive Officer role and other executives in our sample that might affect the probability of observing compensation contracts linked to CSR outcomes, we include a dummy variable equal to one if the executive is the Chief Executive Officer (variable CEO). Following previous studies, we control for the two most widely established determinants of executive compensation: firm size and financial performance (Berrone and Gomez-Mejia 2009a, b). We use (ROA) for the latter and logged sales for the former (Deckop et al. 2006; Mahoney and Thorne 2005). In addition to these important factors, we follow Jo and Harjoto (2012) and control for leverage (BOOKLEVERAGE), R&D intensity (R&D), advertising intensity (AD), and industry by including dummies for each Fama-French 48 industry classification (FF48 Industry). Appendix 1 provides a more detailed description of each variable used.

Table 1 presents summary statistics on the incentives for the top executives in our sample, executive characteristics, firm characteristics, and investment and financing measures.

Table 1 Descriptive statistics

Results and Discussion

Univariate Tests

To examine potential differences between firms with compensation contracts linked to CSR and those without contracts, we compare both types of firms in Table 2. In general, CSR contracting is more common in larger firms with more independent boards, less insider ownership, lower financial performance (measured by ROA), higher social performance (measured by KLD scores), and lower advertising intensity. More independent boards, less insider ownership, and fewer executives who are also board members in CSR-contracting firms are consistent with our hypotheses that CSR is likely to be financially beneficial to firms (H1a, H3a, H4a). By contrast, the negative correlation between the number of block shareholders and the presence of compensation for CSR is consistent with CSR being a wasteful activity borne as an agency cost (H2b). However, we note that these correlations are not adjusted for industry and do not account for the correlation of our independent variables of interest with our other control variables. Table 3 reports the correlations between all the key variables. As a next step, we conduct multivariate tests to address these potential issues.

Table 2 Univariate T test

Table 3 Correlation table

Multivariate Tests

Table 4 reports the results testing H1 and H2. In both cases, we find supporting evidence for hypotheses H1a and H2a. In particular, average director tenure is negatively related to the existence of CSR incentives (Column 2), while the percentage of board members hired prior to the CEO’s arrival and the number of institutional block shareholders are positively associated with the presence of CSR compensation contracts (Columns 4 and 6, respectively). In terms of magnitude, a 1-year increase in average director tenure is associated with 7 % lower odds of observing CSR incentives, a one standard deviation increase in the percentage of the board hired prior to the CEO’s arrival is associated with 13 % higher odds of CSR contracting, and the presence of one additional institutional block shareholder is associated with 8 % higher odds of observing CSR compensation. Taken together, the results suggest firms with better corporate governance are substantially more likely to provide compensation contracts linked to CSR.

Table 4 CSR contracting and corporate governance

Examining the potential influence of greater managerial power, Table 5 reports results that support hypotheses H3a and H4a. The coefficient estimate on the fraction of shares owned by the individual executive is negative and significant at the 1 % level for all specifications (Columns 1, 2, and 6). The estimate from Column 2 suggests that a one standard deviation increase in executive share ownership results in an 8 % decrease in the odds of CSR contracting. The coefficient estimate for the dummy variable of whether or not the executive is also a board member (i.e., EXECDIR) is modestly significant at the 10 % level. Results in Column 4, with all controls, suggest that if an executive is also a board member, there is a 19 % decrease in the odds of CSR contracting.¹⁰ However, although the magnitude is similar when including the fraction of executive shares owned in the same specification (Column 6), the coefficient estimate is no longer significant.

Table 5 CSR contracting and managerial power

Taken together, the coefficient estimates for all four of our hypotheses tests provide consistent evidence that both improved corporate governance and reduced managerial power predict a greater likelihood of observing compensation contracts tied to social performance. Our results suggest that CSR is not a form of managerial excess or agency cost, but may be beneficial to shareholders and firm financial performance.

Finally, we test whether the existence of compensation contracts tied to CSR is associated with higher levels of CSR activities at the firm level, with results shown in Table 6. The coefficient estimate for CSRCONTRACTING is positive and significant across specifications, providing support for Hypothesis 5 and suggesting that contracting on CSR does improve firms’ CSR level in the following year,¹¹ even after controlling for the firm’s current level of social performance. This model is used to mitigate the endogeneity problem if there is persistence in a firm’s level of CSR over time (Li 2015). The results suggest that providing incentives for CSR to managers may be effective in improving firm social performance.

Table 6 CSR contracting and CSR level

Conclusion

We investigated the link between corporate governance and executive compensation for CSR, and extended the agency cost perspective of CSR to develop our hypotheses. If CSR provides sufficient financial returns to shareholders, firms with stronger corporate governance should incentivize their managers to invest in social performance. However, if CSR represents managerial excess and is an agency cost, then firms with better governance should be less likely to provide incentives for CSR.

In our results, we found consistent evidence that CSR is likely to be financially beneficial for firms, and for shareholders. Firms with more shareholder-friendly governance are more likely to incentivize their managers to engage in CSR. Also, firms that provided compensation linked to CSR had greater levels of social performance on average, consistent with the notion that the provision of incentives for CSR leads to more CSR activities.

To our knowledge, our study is the first to document the prevalence of executive compensation contracts for CSR. Almost 40 % of our sample, which consists of the Standard and Poor’s 500 Index, includes some form of compensation for CSR to their top managers. Future research should consider whether our hypothesis test results might change in magnitude for smaller firms. There is some evidence that executives in smaller firms may play an even more influential role in CSR activities because they have more power and may wield more discretion in CSR-related decisions under conditions of weaker corporate governance (e.g., Palmon and Wald 2002). Thus, the effects of corporate governance and managerial power on CSR contracting and CSR performance may be even stronger in smaller firms, although this remains an area for future empirical work. Finally, better understanding exactly how and when firms choose to compensate executives for CSR provides fruitful opportunities for future research. By examining the role of corporate governance and agency cost explanations of CSR, our study aims to provide a first step in this direction, but the role of executive compensation for CSR remains largely unexplored.

Appendices

Appendix 1

See Table 7.

Table 7 Variable definitions

Appendix 2: CSR Compensation Terms

Based on an analysis of variable and performance-based pay, we identified those companies also incentivizing non-financial performance. We then analyzed the descriptions of non-financial performance to code the compensation as CSR-linked. The following were the most common such descriptions:

• Community

• Compliance with ethical standards

• Corporate social responsibility

• Diversity

• Employee well-being

• Energy efficiency

• Environmental compliance

• Environmental goals

• Environmental performance

• Environmental projects

• Greenhouse gas emissions reductions

• Health

• Performance relative to a corporate responsibility index (e.g., Dow Jones Sustainability Index)

• Product safety

• Reduced injury rates

• Safety

• Sustainability

For example, as discussed in its Proxy Statement, the Kellogg Company divided its Annual Incentive Plan for its CEO John Bryant into financial and non-financial incentives: “90 % of the annual incentive opportunity was based on performance against corporate financial metrics…and 10 % was based on performance against non-financial targets (people safety, food safety and quality, and diversity and inclusion).” Thus, John Bryant of Kellogg is coded as being offered CSR-linked incentives.

The compensation plans typically contract on a dimension of social performance with symmetric consequences. For example, one common area of contracting is Employee Diversity and Inclusion. In principle, a CEO could be credited for doing well, poorly or both well and poorly on this dimension. For an example of the latter, the firm may experience an increase in promotions for women but a reduction in promotions for minorities. In this setting, the CEO’s compensation would be based on the net results of these “strengths” and “concerns.” Additionally, the CSR performance compensation payment is usually based on the net result of “strengths” and “concerns” across multiple categories. Hence, it could be that a given category of CSR compensation contains only “strengths” or “concerns.” Nonetheless, the final payment to the CEO is based on the net result of all of the categories, which mirrors our empirical analysis.