Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
1
Corporate Social Responsibility and CEO Risk-Taking Incentives
Craig Dunbar, Frank Li, and Yaqi Shi
Current version: December 3, 2016
______________________________
* Craig Dunbar is at Richard Ivey School of Business, Western University (Email:
[email protected]); Frank Li is at Richard Ivey School of Business, Western
University (Email: [email protected]); Yaqi Shi is at Richard Ivey School of Business,
Western University (Email: [email protected]). We are grateful for comments from Tima
Bansal, Ann Peng, and Stephan Vachon. We thank Connor Fraser and Leo Han for
research assistance. All errors remain our own responsibilities.
2
Corporate Social Responsibility and CEO Risk-Taking Incentives
Abstract
In this paper, we explore how firms incentivize their CEOs subsequent to undertaking
risk-reducing corporate social responsibility (CSR) initiatives. Specifically, we focus
on the effect of CSR standing on CEO’s future risk-taking financial incentives. We
hypothesize that because firms possessing better social performance generate
insurance-like moral capital that reduces firm risk, they will have more risk-taking
capacity and should respond, therefore, by offering greater risk-motivating incentives
to managers. Employing a large sample of US firms from 1992 to 2010, we find
strong empirical evidence to support our hypothesis. Indeed, CSR standing is
positively related to CEO pay-risk sensitivity, demonstrating that firms whose
sustainable initiatives are viewed to be successful are more likely to offer their CEOs
greater risk-motivating financial incentives. Further, this association is driven by CSR
strengths rather than CSR concerns. Finally, we provide evidence that firm overall
risk and idiosyncratic risk negatively moderate the association between CSR and CEO
future risk-taking financial incentives.
Key words: Corporate social responsibility, Executive incentives, Vega, Risk
management
3
1. Introduction
An extensive literature examines the relation between corporate social
responsibility (hereafter CSR) and corporate financial performance (hereafter CFP).
Given the mixed evidence on this relation1, researchers have explored the channels
through which CSR can either destroy or add value. The focus of studies exploring a
negative CSR-CFP relation has been agency theory (e.g. Hong, Li, and Minor, 2016).
Theories leading to a positive CSR-CFP relation consider increased customer
awareness (Servaes and Tamayo, 2013), improved transparency (Dhaliwal, Li, Zhang,
and Yang, 2011), strategic CSR (Porter and Kramer, 2011), and the risk
management/insurance properties of CSR. Godfrey (2005) argues that moral capital,
resulting from CSR investments, provides insurance-like protection for a firm’s
intangible assets. Succinctly noting that “good deeds earn chits” (Godfrey, 2005, p.
777), he finds that firms with greater moral capital face a less negative stock market
reaction in response to negative events. An implication is that CSR standing should
have a negative impact on measures of firm risk (both systematic and unsystematic).
Evidence on the risk reducing effect of CSR is quite strong and consistent with the
insurance/risk management theory (see Orlitzky and Benjamin, 2001; Godfrey,
Merrill, and Hansen, 2009; Jo and Na, 2012; Oikonomou, Brooks, and Pavelin,
2012).2
If enhanced sustainable CSR initiatives generate insurance-like properties,
how do firms change corporate financial policies in response to their CSR standing?
1 Some studies empirically find a positive CSR-CFP relation (Hillman and Keim, 2001) with other finding a
negative relation (e.g. Brammer, Brooks, and Pavelin, 2006). Many studies find no significant relation between
CSR and CFP (e.g. Bauer, Koedijk, and Otten, 2005). Margolis, Elfenbein, and Walsh (2007) conduct a
meta-analysis of many such empirical studies and conclude that the relation between CSR and CFP is positive but
small. 2 Several studies provide indirect support for the risk reducing effect of CSR. Firms with higher CSR have better
access to finance (Cheng, Hong, and Shue, 2014) and face lower financing costs (Chava, 2014; El Ghoul,
Guedhami, Kwok, and Mishra, 2011; Dhaliwalet al., 2011; Goss and Roberts, 2011).
4
In this paper, we address this question by investigating how CSR status affects
subsequent CEO compensation contracts with a focus on the risk-taking incentives
created by CEO stock options. Specifically, we adopt Vega, the sensitivity of CEO
wealth to stock return volatility, as our proxy for CEO risk-motivating financial
incentives. Many studies show a significantly positive relation between Vega and
measures of firm risk (e.g., Guay, 1999; Coles, Daniel, and Naveen, 2006; Armstrong
and Vashishtha, 2012), suggesting that Vega is an effective tool for firms to encourage
their executives to take risks. As such, higher Vega compensation contracts mitigate
the agency problem of “risk-shirking” (Haubrich, 1994). However, the link between
other compensation incentives and firm risk is less clear. Guay (1999) notes that
higher Delta (the sensitivity of CEO wealth to changes in stock prices) exposes
managers to more personal wealth risk. While CEO compensation contracts with
higher Delta can encourage executives to work harder to increase shareholder wealth,
it can also discourage risk-averse executives from taking risky projects. Empirically,
most studies find no significant relation between CEO compensation Delta and firm
risk (Coles et al., 2006; Low, 2009).
If positive CSR standing creates insurance-like protection and reduces firm
risk, firms with high CSR standing should have more future risk-taking capacity. We
expect, therefore, that firms would respond to high CSR standing by giving CEOs
compensation contracts with higher Vega in order to encourage greater risk taking. It
is also plausible that managers tend to make decisions to avoid firm risks when
market discipline declines. Bertrand and Mullainathan (2003), for example, show that
plant openings and closings decline, reducing firm riskiness, in response to the
passage of anti-takeover laws. Research shows that CSR standing is similarly
inversely related to the probability of CEO turnover (Harjoto and Jo, 2011), so the
5
insurance-like benefit from CSR potentially protects CEOs from market discipline. To
counter this potential risk-shirking, firms should respond by giving CEOs
compensation contracts with greater Vega. Overall, the risk capacity and CEO
risk-shirking arguments both predict a positive CSR standing – Vega relation.
We examine the relation between CSR status and CEO risk-taking incentives
for US firms from 1992 to 2010. Annual CSR status is obtained from the MSCI ESG
Stats database (formerly the Kinder, Lyndenberg and Domini, or KLD, database).
Data for CEO incentives, accounting information, stock information, and institutional
ownership are obtained from Execucomp, Compustat, CRSP, and 13F filings,
respectively. Merging different databases yields 11,272 firm-year observations. We
find strong support for a positive association between CSR status and CEO risk-taking
incentives. In our base model, a one standard deviation change in CSR status results
in a 0.13 standard deviation change in CEO Vega. To put this number in perspective,
the average CEO Vega for this sample is 176.9 with standard deviation of 340.5. The
positive relation between CSR status and Vega is robust to various specifications,
including a Granger causality test and an instrumental variable approach, that address
endogeneity concerns.
Our base models follow the extensive literature using the MSCI ESG Stats
database in adopting a single measure to capture overall CSR status. The MSCI ESG
Stats database considers five dimensions of CSR: community activities, diversity,
employee relations, environmental policies, and product development. For each
dimension, it identifies whether this is an area of strength or weakness for the firm.
Our single measure adds the number of strength areas for a firm and subtracts the
number of weakness areas. We build on some recent studies that consider potentially
asymmetric effects of strengths and weaknesses (McGuire, Dow, and Argheyd, 2003;
6
Strike, Gao, and Bansal, 2006; Mahoney and Thorn, 2006). We posit that increasing
strengths has a more significant effect on building moral capital and firm “insurance”
than reducing weaknesses. Consistent with this prediction, we find the positive
relation between CSR status and CEO Vega only holds for measures of CSR strengths.
Finally, we consider whether the relation between CSR status and CEO Vega is
moderated by a firm’s initial financial risk. We find the CSR-Vega relation to be more
economically and statistically significant when a firm’s financial risk is lower. This is
consistent with our risk capacity argument. Firms with both low risk and the
insurance-like protection of high CSR standing have greater capacity to add risk and,
therefore, are more likely to encourage CEO risk taking by using high Vega
compensation contracts. This finding is also consistent with behavioral agency theory
which argues that stock options can induce less risk-taking when firm risk is already
high (Wiseman and Gomez-Mejia, 1998). Granting stock options results in the
creation of some expectation of future wealth for executives. Loss-averse executives
will then avoid risky projects to preserve anticipated wealth and this avoidance
becomes more significant when firm risk is higher. Given this behavior, firms should
not increase Vega in response to increased CSR standing when firm risk is already
high as this would not reduce risk-shirking but instead exacerbate it.
Overall, we believe this study makes three primary contributions to the
literature. First, we add to the literature that considers the relation between CEO
incentives and CSR. Most existing studies investigate the impact of CEO incentives
on CSR (McGuire et al., 2003; Mahoney and Thorne, 2005; Petrenko, Aime and
Ridge, 2016). The only existing study to explore the impact of CSR on CEO
compensation incentives is Cai, Jo, and Pan (2011). The focus in their study is on the
level of compensation, however. Our study is the first to show that CSR standing
7
affects the structure of executive compensation and future risk-motivating financial
incentives. Second, this study adds to the literature examining the relation between
CSR and firm risk, as most existing studies do not examine how firms respond to their
CSR standing in setting corporate policies that have an impact on firm risk. Finally,
our results contribute to practice. Firms pushing to enhance their social performance
should be mindful that CSR standing could influence CEO risk-taking financial
incentives going forward. Consistent with our findings, boards should set policies to
adjust compensation policies as CSR status evolves.
The remainder of our study is organized as follows. In the Theory and
Hypotheses section, we develop three hypotheses pertaining to the association
between CSR and CEO’s risk-taking incentives. In the Data and Measurement section,
we describe data and variables. Univariate and multivariate results are reported in the
Empirical Results section. Theoretical and practical implications are presented in the
Discussion section. The final section concludes.
2. Theory and Hypotheses
2.1 CSR and CEO risk-taking incentives (H1)
The concept of CSR generally refers to discretionary managerial activities that
serve people, communities, and the environment in ways that go beyond the interests
of the firms’ shareholders and that which is required by law (McWilliams and Siegel,
2000 & 2001). Much of the early research on CSR examines the effect of CSR
standing on shareholder wealth. Given mixed findings, researchers have examined the
channels through which CSR standing can either add to or destroy shareholder wealth.
One stream of work has focused on the impact of CSR standing on firm risk. Orlitzky
and Benjamin’s review paper (2001) documents that CSR standing is significantly
negatively associated with firm risk. Recently, Luo and Bhattacharya (2009) further
8
show that CSR reduces firm-idiosyncratic risk. How CSR affects CEOs’ subsequent
risk-taking incentives is under-investigated, however.
While Friedman’s shareholder story (Friedman 1962, 1970) suggests that CSR
results in an unnecessary cost and is a misallocation of corporate resources that should
be spent on giving back to shareholders, Freeman’s stakeholder theory (1984) argues
instead that companies are accountable to the interests of a broader group of
stakeholders, such as employees, customers, local communities, and society. The
stakeholder theory has subsequently been strengthened in several ways. For example,
Jones (1995) advances the instrumental stakeholder theory, asserting that CSR
activities are efforts to benefit stakeholders with the ultimate goal of benefiting
shareholders. In line with instrumental stakeholder theory, the risk management
theory posits that CSR may engender positive relationship-based intangible assets, or
moral capital, among stakeholders, which provides the firm with insurance-like
protection (Godfrey, 2005). Smith and Stulz (1985) and Stulz (2002) show that risk
management adds value to shareholders when the theoretical perfect capital market
assumption is violated in the real world. Specifically, reducing any risks that would
result in deadweight costs for a firm that cannot be diversified away by investors (e.g.
bankruptcy costs) adds to firm value.
In the context of CSR, positive moral capital protects firm wealth against loss
by alleviating negative stakeholder reactions and related sanctions when undesirable
events transpire. Empirically, Godfrey et al. (2009) employ an event study of 178
negative legal/regulatory events and find that institutional CSR activities yield an
“insurance-like” benefit. The market reaction to a negative event is much less
negative for firms having a stronger CSR standing before the event. Likewise, Husted
9
(2005) relies on real option theory to argue that CSR is used as a strategy for risk
management.
If CSR standing generates an insurance-like benefit, how should a firm
optimally respond to increases or decreases in that standing? Since a positive CSR
standing reduces firm risk, risk-taking capacity should increase. We would expect,
therefore, that firms would respond to a positive CSR standing by offering CEOs
compensation contracts with features that encourage risk taking.
Complementing this argument, we also expect that in the absence of any
corporate policy change, managers would naturally make decisions to reduce risks
given higher CSR standing. The insurance-like benefit from CSR standing is also
likely to protect CEOs from market discipline and therefore entrench risk-shirking
CEOs. Consistent with this, Harjoto and Jo (2011) show that CSR engagement
reduces the likelihood of CEO turnover in the subsequent time periods. To counter
this natural incentive, firms should change policies by offering CEO compensation
contracts that encourage risk-taking.
In this study, we use Vega as a proxy for CEO risk-taking incentive. Vega, the
sensitivity of managerial wealth to firm risk, can offset the risk-shirking inclinations
introduced by manager-shareholder incentive alignment (Haubrich, 1994). Prior
studies suggest that Vega is an important and effective incentive for managerial risk
taking (Anantharaman and Lee, 2014; Coles et al., 2006; Low, 2009; Hayes, Lemmon,
and Qiu, 2012; Kim and Lu, 2011). Due to its correspondence to stock return volatility,
high-Vega compensation schemes make risk more valuable to CEOs (Guay, 1999).
Taken collectively, the above discussion suggests that superior CSR standing helps
firms to build moral capital and goodwill, but this also reduces market discipline for
CEOs, which would typically result in executives taking less risk. Firms should
10
respond by offering CEOs compensation contracts with higher Vega. Overall, based
on the risk capacity and risk-shirking arguments, we propose our first hypothesis:
H1: All else being equal, CSR standing is positively associated with CEO pay-risk
sensitivity (Vega).
2.2 CSR strength/weakness and CEO risk-taking incentives (H2)
CSR strengths and concerns capture strong social performance and weak
social performance respectively. Most prior studies have taken CSR as a single
construct where CSR standing is measured as the sum of strength areas subtract the
sum of weakness areas. We continue a development in the literature toward modeling
CSR strengths and weaknesses separately (Hillman and Keim, 2001; Godfrey et al.,
2009; Bansal, Gao, and Qureshi, 2014). Only limited studies have examined the
different perspectives of strong versus weak performance (McGuire et al., 2003;
Strike et al., 2006; Mahoney and Thorn, 2006). For instance, McGuire et al. (2003)
find a significant positive association between CEO compensation of salary and CSR
concerns and a positive association between CEO long-term incentives and CSR
concerns. This line of research highlights the importance of considering CSR
strengths and concerns separately. Truly, firms can simultaneously demonstrate both
strong and weak social performance (Strike et al., 2006). For example, Exxon-Mobil
has arguably the poorest environmental reputation and is under investigation for lying
over climate change risk.3 Nevertheless, Exxon-Mobil also contributed $227 million
in 2015 to help improve education, combat malaria, develop employment
3 For more information on the investigation, please refer to
http://www.wsj.com/articles/exxon-fires-back-at-climate-change-probe-1460574535.
11
opportunities for women, and advance workforce initiatives. 4 Therefore,
distinguishing socially responsible and irresponsible firms is challenging.
The risk management theory contends that strong social performance
generates positive moral capital among communities and stakeholders. When harmful
events occur, stakeholders can moderate their negative judgements and sanctions
toward the company because of this goodwill (Godfrey, 2005; Godfrey et al., 2009).
In contrast, improving weak social performance likely will not generate the
above-mentioned moral capital and relationship-based intangible assets. Therefore,
we conjecture that changes to CSR strengths and not changes to CSR weaknesses
affect firm moral capital. In this case, only CSR strengths should significantly affect
risk-taking incentives, or the Vega of option contracts, given to CEOs. Given this
reasoning, we posit the following hypothesis:
H2: All else being equal, CSR strength is positively associated with CEO pay-risk
sensitivity (Vega).
2.3 The moderating effect of financial risk (H3)
We expect that the relation between CSR standing and Vega would be
moderated by firm and situational factors. In particular, we focus on the effect of firm
risk. Firms with both low initial risk and high CSR standing should be most likely to
increase CEO risk-taking incentives using compensation contracts with higher Vega
because these firms have the greatest risk-taking capacity.
Behavioral agency theory also suggests that situational characteristics
moderate manager’s risk-seeking incentives (Wiseman and Gomez-Mejia, 1998;
Sanders, 2001; Sanders and Carpenter, 2003). The theory combines classical agency
4 For detail on Exxon-Mobil’s worldwide CSR activities, please refer to the company 2015 Worldwide Giving
Report:http://corporate.exxonmobil.com/en/community/worldwide-giving/worldwide-giving-report/overview?pare
ntId=b5244b73-b229-417c-a17e-a7c932d0d2ca.
12
theory, which considers situations where choices of principals and agents can diverge
(Jensen and Meckling, 1976), with prospect theory, which considers situations where
decision maker preferences exhibit loss aversion (Kahneman and Tversky, 1979).
Prospect theory posits that decision makers frame problems in ways that define
subsequent outcomes as either gains or losses. Decision makers are loss averse and
would avoid prospects that involve significant potential loss. Such preferences change
the conflict of interest between principals and agents regarding how much risk a firm
should take and also changes how the conflict should be addressed.
Under classical utility theory, risk averse managers are likely to avoid taking
risky projects, even those that can add to firm value. A solution is to compensate
managers with options possessing high Vega, as this encourages risk taking. The
situation is more complex given prospect theory. Wiseman and Gomez-Mejia (1998)
argue that granting options creates some expectation of future wealth for managers.
When options mature, managers can realize both gains and losses relative to this
expectation. When risks are high, loss-averse managers may avoid taking further risks
given fear of loss. The theory suggests that increasing Vega when firm risk is high
may discourage rather than encourage risk taking. Based on both risk capacity and
behavioral agency theories, we put forward the following hypothesis:
H3: The association between CSR and Vega is negatively moderated by firm risk.
3. Data and Measurement
3.1 Sample selection
We gather our data from various sources. We collect CSR data using the most
comprehensive database in the literature, MSCI ESG Stats (formerly known as the
Kinder, Lyndenberg and Domini (KLD) database). MSCI ESG Stats has been broadly
used in scholarly research (e.g., Deckop, Merriman, and Gupta, 2006; Chava, 2014;
13
Flammer 2015; Flammer and Luo, 2016; McGuire et al., 2003; Servaes and Tamayo,
2013; Werner, 2016). We collect data for CEO incentives, accounting information,
stock information and institutional ownership from Execucomp, Compustat, CRSP
and 13F schedules, respectively. Merging different databases yields 11,272 firm-year
observations for the period 1992-2010. The Appendix provides more detailed
descriptions of all variables employed in this paper.
3.2 Variable Measurement
3.2.1 Corporate Social Responsibility
Following previous research (Flammer, 2015; Flammer and Luo, 2016; Cheng,
Hong and Shue, 2014), we focus on five dimensions of CSR: community activities,
diversity, employee relations, environmental policies, and product development.
MSCI ESG Stats reports ratings of strengths and concerns for each firm across these
five categories. We define an aggregate CSR score by summing the total number of
CSR strengths and subtracting the total number of CSR concerns across these five
categories. To mitigate reverse causality, we use lagged CSR scores in all models. To
test our first hypothesis, our primary model is as follows:
VEGAt+1 =f (CSRt, Control Variablest) (1)
Where CSRt is the aggregate CSR score in period t and Control Variablest are defined
in section 3.2.4.
Our model to test the second hypothesis is as follows:
VEGAt+1 = f (STRENGTHt, CONCERNt, Control Variablest ) (2)
Where STRENGTHt is the sum of total CSR strengths and CONCERNt is the sum of
CSR concerns for the firm in period t.
14
To test the third hypothesis on the moderating effect of firm risk, we estimate
the following model:
VEGAt+1=f (CSRt, RISKt, CSRt*RISKt, Control Variablest) (3)
Where RISKt is a measure of firm risk in period t, and is defined in section 3.2.3.
3.2.2 Vega and Delta
Following Guay (1999) and Core and Guay (2002), we use the Black-Scholes
(1973) option valuation model to calculate Vega. This is consistent with many recent
papers such as Anantharaman and Lee (2014), Coles et al., (2006), Low (2009), Hayes
et al. (2012), and Kim and Lu (2011). Vega is defined as the change in the dollar value
of the CEO’s wealth for a 0.01 change in the annualized standard deviation of stock
returns. Here, Vega is a proxy for CEO pay-risk sensitivity, and thus captures the
executive’s risk-taking incentive.
3.2.3 Risk Measures
We adopt two measures for firm risk: 1) Idiosyncratic risk
(IDIOSYNCRATIC), which is the standard deviation of the residuals from the
Fama-French three-factor market model; 2) firm total risk (VOLATILITY), which is
the annualized monthly standard deviation of a firm’s return series.
3.2.4 Control Variables
We also include other control variables that are shown in the literature to have
influence on Vega: 1) TENURE, defined as the length of time that the CEO has been
at his or her position (Coles et al., 2006; Hayes et al., 2012); 2) INSTHOLD, defined
15
as the percentage of institutional share ownership5; 3) ROA, which is the operating
income deflated by total assets (Coles et al., 2006; Low, 2009); 4) CAPEX, which is
defined as capital expenditure expenses over total assets (Coles et al., 2006; Low,
2009); 5) Q, which is Tobin’s Q and is computed as the sum of the book value of total
assets plus the market value of common stock less the book value of equity over the
book value of assets (Kim and Lu, 2011)6; 6) SIZE, defined as the log of total assets at
the end of the fiscal period (Coles et al., 2006; Anantharaman and Lee, 2014); 7)
LEVERAGE, defined as total liabilities over total assets (Coles et al., 2006; Hayes et
al., 2012); and 8) DELTA, defined as the dollar change in the value of CEO’s annual
equity-based compensation for a 1% change in the stock price using Black-Scholes
(1973) option pricing theory (Coles et. al., 2006, Low, 2009).
4. Empirical Results
4.1 Descriptive Statistics and Correlation
Table 1 reports means and standard deviations for the primary variables of
interest in our sample of firm-year observations. The mean value of CSR is 0.201 and
the standard deviation is 2.401, suggesting that significant variation exists among
firms in their CSR standings. Our CSR scores are comparable to those in other studies
(e.g., Cai et al., 2011; El Ghoul et al., 2011; Hong, Kubik, and Scheinkman, 2012).
For instance, the mean score is 0.19 with a standard deviation of 2.22 in El Ghoul et al.
(2011). The means of STENGTH and CONCERN are about 1.699 and 1.494
respectively (similar to Hong et al., 2012). The mean values for community
involvement (COMS) and diversity (DIVS) are positive, whereas the average scores
5 Prior studies (Kim and Lu, 2011; Hayes et. al., 2012) suggest that corporate governance factors influence
managers’ risk-taking incentives; therefore, we add the percentage of institutional ownership as a control variable.
For our sensitivity checks, we also try to exclude this variable. The robustness test results are not sensitive to this
correction. 6 Most prior studies (e.g., Coles et al., 2006; Low, 2009) adopt the ratio of market-to-book as a control variable.
Our proxy Tobin’s Q is similar to market-to-book in capturing firms’ investment and growth opportunities.
16
for employee relation (EMPS), environmental policies (ENVS) and product
development (PROS) are negative, indicating that concerns, on average, outweigh
strengths in these dimensions.
The average value for VEGA is 176, which is similar to statistics reported in
Coles et al. (2006). In addition, our descriptive statistics reveal that the average
Tobin’s Q (Q) is 1.978 for our sample firms. On average, the CEO has been at his or
her position for 8.6 years. The average percentage of institutional share ownership is
71.57 percent; the average financial leverage is 21.7 percent. In sum, all variables
appear to be in sensible ranges and are comparable to those of prior studies (e.g.,
Coles et al., 2006; Hayes et al., 2012).
Insert Table 1 here
Table 2 provides Pearson correlation coefficients across variables. As our
hypothesis 1 predicts, firms with higher CSR scores are associated with higher VEGA.
Similarly, higher STRENGTH scores are positively related to higher VEGA.
Furthermore, firms with longer-tenured CEOs, higher return on assets, higher Tobin’s
Q, and larger size are more likely to offer compensation with greater risk-taking
incentives (i.e., higher Vega). Conversely, firms with higher percentage of institutional
owners and more spending on capital expenditure are less likely to incentivize their
CEOs to take on risks.
Insert Table 2 here
4.2 Multivariate Tests
4.2.1 Overall CSR Measures and Vega
Table 3 presents estimated results for model (1) which is related to our first
hypothesis. In order to mitigate the concerns of endogeneity, we employ three models:
17
namely, industry-fixed effect model, firm-fixed effect model, and lagged dependent
variable model. 7 Year-fixed effects are controlled for in all the models. The
R-squares for all models range from 30% to 63%, showing that these models are
significant in describing variation for CEO’s risk-taking incentives, Vega. Column (1)
presents results from industry-fixed effect model. The significantly positive
coefficient on CSR supports H1, indicating that in firms with higher social
performance, and therefore more moral capital, higher Vega is granted to incentivize
CEOs to take on more risk. Columns (2) and (3) show results from the firm-fixed
effect model and lagged dependent model, respectively. The coefficients on CSR in
predicting VEGA remain positive and significant. Taken together, our results indicate
that firms having better CSR standing are more likely to provide CEOs with greater
risk-taking incentives.8
Insert Table 3 here
4.2.2 CSR Strength Score and Vega
Table 4 summarizes the results related to the second hypothesis (H2). As
indicated in model (2), we split the CSR variable into two variables, STRENGTH and
CONCERN. Similar to Table 3, we also adopt industry-fixed effects (Column (1)),
firm-fixed effects (Column (2)), and lagged dependent variable model (Column (3))
approaches. The results from all models indicate that the coefficients on STRENGTH
are positive and significant. By contrast, the coefficients on CONCERN are not
7 Motivations to include industry-fixed effect model, firm-fixed effect model and lagged dependent variables are
discussed in Section 5.1 Control Variables and Fixed Effect Models. 8 We also test the impact of CSR on CEO’s pay-performance sensitivity, Delta, which is defined as the change in
the dollar value of the CEO’s wealth for a one-percentage point change in stock price. The impact of CSR on Delta
is not significant. While high Delta can align the interests of executives and shareholders, its relation to firm risk
is ambiguous. Empirically, prior studies demonstrate that there is no significant relation between CEO
compensation Delta and firm risk (Coles et al., 2006; Low, 2009). Our test also indicates there is no significant
relationship between CSR and Delta.
18
significant in all three models. Taken together, our results support H2 that the positive
effect of CSR on VEGA is driven by CSR STRENGTH rather than CONCERN.
Insert Table 4 here
4.2.3 The Moderating Effect of Risk
The results for testing the moderating effect of firm risks are reported in Table
5. In Column (1), the proxy for risk is idiosyncratic risk; in Column (2), the proxy for
risk is volatility (i.e., firm total risk). In each column, we add an interaction term
between CSR and the risk measure. Column (1) shows that the effect of the
interaction between CSR and idiosyncratic risk is negative and significant, indicating
that idiosyncratic risk negatively moderates the association between CSR and VEGA.
Similarly, Column (2) shows that the interaction between CSR and stock return
volatility is also significantly negative, consistent with the prediction that firm overall
risk plays a negative role in moderating the association between CSR and VEGA. The
above results support H3 that the positive effect of CSR on firms’ risk-taking
incentives are more pronounced in firms with lower risk. Economically, a one
standard deviation change in CSR and in firm total risk (idiosyncratic risk) leads to a
0.25 (0.07) standard deviation change in VEGA. Additionally, it is worth noting that
the coefficients on CSR continue to be significantly positive in all these models.
Insert Table 5 here
4.2.4 CSR Ratings in Five Categories
Some forms of CSR activities are more likely to build reputation-related
intangible assets and offer insurance-like protection than other forms (Godfrey et al.,
2009). For instance, the higher the level of consistency between CSR activities and a
community’s ethical values, the more likely the community will provide positive
19
moral evaluation toward that firm. CSR endeavors related to product development
may be less likely to produce moral capital because these activities normally aim at
maximizing profits, and thus are self-serving. Overall, CSR activities in categories
such as community, employee relation, diversity, and environmental protection are
more likely to produce moral capital because they are viewed as serving social
goodness (Godfrey et al., 2009). For instance, Chava (2014) finds that firms with
environmental concerns fact substantially higher cost of equity/debt capital. We,
therefore, build on our initial analysis by considering separate measures of CSR
standing for the different categories.
In Table 6, we partition our CSR score into five dimensions: community
involvement (COMS), diversity (DIVS), employee relation (EMPS), environmental
policies (ENVS) and product development (PROS). Each variable takes the value 1 if
the area is a strength for the firm, -1 if the area is a weakness, and 0 otherwise. In
column (1), we report the results for year/industry-fixed effect model, whereas in
column (2) we report findings for the year/firm-fixed effect model. In both cases, the
coefficients on product development (PROS) are significantly negative, while the
coefficients on community involvement (COMS), diversity (DIVS), and
environmental policies (ENVS) are significantly positive. This indicates that CSR
investments in product development are different from those in other dimensions and
do not generate goodwill or offer insurance-like protection. We conjecture that firms
need to offer CEOs lower risk-taking incentives (Vega) to compensate the enhanced
risk due to the uncertainty related to product development.
Insert Table 6 here
5. Endogeneity
20
5.1 Control Variables and Fixed Effects Model
In our analysis, thus far we have used different control variables and a fixed
effects models to address the omitted variable problem. The results are robust to
controlling for various observable firm and manager characteristics and unobservable
time, industry, firm, and manager fixed effects. We also include a lagged dependent
variable (i.e., lagged VEGA) as a right-hand side variable as a robustness check.
Lagged Vega is important because it contains all the information that determines Vega
until the point of year t. As shown in Tables 3 and 5, even after controlling for Vega at
year t, CSR at year t still provides incremental explanatory power to explain Vega at
year t+1.
5.2 Granger Causality Analysis
While the focus of our analysis has been on the effect of CSR standing on the
choice of Vega, reverse causality is possible where the executive risk incentives drive
a firm’s social performance. Executives with greater risk-taking incentives could
invest to enhance CSR if these CSR projects are risky, and decrease CSR if the CSR
projects are on average less risky than the firm’s other ongoing projects.
To study which direction of causality dominates, we conduct the Granger
Causality tests (Granger, 1969) to examine the nature of relations between CSR and
Vega and the direction of causality. Given the time series of the data on two variables
X and Y, X is said to “Granger cause” Y if the lagged values of X are significant
predictors of Y incremental to lagged values of Y. In Table 7, we use the following
specifications to test the significance of the coefficients on the lagged values of CSR
in equation 4 and the lagged values of Vega in equation 5:
it
n
i iit
n
i it CSRVegaVega 11 (4)
21
it
n
i iit
n
i it VegaCSRCSR 11 (5)
To determine the optimal lag lengths n, we refer to the Bayesian information
criterion (BIC) (Schwarz, 1978; Rissanen, 1978) and the Hannan-Quinn information
criterion (QIC) (Hannan and Quinn, 1979) and conclude the appropriate lengths
should be 4 years.9
Consistent with our hypothesis that a firm’s CSR standing influences its
executive contracting of risk incentives, the evidence in Table 7 suggests the causality
from CSR to Vega is much stronger than the reverse causality. Based on the computed
Chi-squares and their marginal significance level, Model 1 confirms that CSR
Granger causes or leads Vega and Model 2 suggests that Vega leads CSR with only
marginal significance. The evidence indicates the causality running from CSR to Vega
dominates the reverse causality.
Insert Table 7 here
5.3 Instrumental Variable Approach
To mitigate further reverse causality and endogeneity concerns, we use an
instrumental variable approach to provide reasonable exogenous variation to identify
the impact of CSR on Vega. We construct two instruments by calculating the average
CSR score for each state-year pair and industry-year pair. The first instrument is the
average CSR rating of all the firms, except the firm itself, in the state where the firm
is located. The rationale behind this instrument is that regional practices of CSR
influence a firm’s social performance (Goss and Roberts, 2011). Likewise, the second
instrument is based on industries because industry characteristics also determine CSR
performance (Ioannou and Serafeim, 2012; Cheng, Ioannou, and Serafeim, 2014).
9 For robustness, we also test 1, 2, 3 year lags and obtain similar results.
22
Meanwhile, it is unlikely that industrial CSR would directly affect a specific firm’s
compensation structure (after adjusting for industry and year fixed effects). In the
same spirit, Goss and Roberts (2011) and Cheng, Ioannou, and Serafeim (2014) also
use these instrumental variables in their studies. Table 8 shows results for Two-Stage
Least Square instrumental variable models (2SLS). We estimate three models in
which the endogenous regressors are STRENGTH, CONCERN, and CSR, respectively.
The first-stage results are reported in Column (1), (3), and (5), indicating that the
instruments significantly explain our CSR regressors. Columns (2), (4), and (6) show
that adopting two-stage models produce similar results as those of our primary
specifications, thus providing further support for our H1 and H2. We statistically test
the instruments for their relevance and validity. The first-stage F statistics all surpass
the usual rule of thumb of 10, the over-identification test (Basman’s test) cannot reject
the null hypothesis that the instruments are valid and orthogonal to the regression
residuals, and the Hausman test rejects exogeneity of the endogenous variable CSR.
These results suggest that these instruments are exogenous under the usual assessment
of instrumental variables, and therefore 2SLS is more efficient than OLS in this
setting.
Insert Table 8 here
6. Discussion
Do shareholders benefit when a firm’s strategies include allocating corporate
resources through CSR? Recent surveys indicate that a large majority of companies
believe that CSR “not only helps the environment and society but help create
goodwill for their reputations and contribute to their brands’ health and performance”
23
(The Nielsen Global Survey of Corporate Social Responsibility 2014). 10 However,
despite the growing importance of CSR investment, the channels through which CSR
affects a firm’s long-term financial success are still subject to much debate. Academia
holds that enhanced disclosure transparency (Dhaliwal et al., 2011), improved
customer awareness (Servaes and Tamayo, 2013), shared value (Porter and Kramer,
2011), and risk management related to CSR (Godfrey, 2005) all contribute to
improved corporate firm performance. We document that firms with better CSR
performance realize insurance-like benefits that can insulate managers from external
discipline. Firms respond by adjusting CEO contracts to enhance their risk-taking
incentives. This increased risk-taking incentive in turn mitigates the risk-averse
predisposition of most CEOs. By accepting more risky but positive NPV projects,
firm value should increase.
The separation between ownership and control potentially causes moral hazard
because managers have little to lose if they focus on their own interests rather than
shareholder wealth maximization (Jensen and Meckling, 1976; Shleifer and Vishny,
1997). Various forces and mechanisms are suggested to mitigate the concern of
agency cost, which include takeover threats, institutional shareholders, board of
directors, large creditors, legal protection of minority shareholders, and incentive
contracts (summarized in Shleifer and Vishny, 1997). Among these mechanisms,
effective incentive contracts are of enormous importance. Managerial risk aversion is
an important determinant of optimal incentive contracts (Ross, 1973). Our findings,
that firms with better CSR standing generally respond by granting managers more
incentives to mitigate their risk aversion, have important practical implications.
10 For details on this survey, please refer to
http://www.nielsen.com/content/dam/corporate/us/en/reports-downloads/2014%20Reports/global-corporate-social-
responsibility-report-june-2014.pdf.
24
Pursuing CSR enhancing activities alone may not drive improvements to shareholder
wealth because managers tend to become more risk-shirking when faced with less
market discipline. Firms need to adjust contracts, as CSR ratings improve, to ensure
managers remain willing to pursue risky but value enhancing projects. In other words,
firms ought to consider a combination of CSR policy and correspondent
compensation policy to improve firm value.
7. Conclusion
The literature strongly and consistently documents a negative relation between
CSR standing and firm risk. In this paper, we ask a natural question: do firms respond
to changes in CSR standing by adjusting executive incentives? More specifically, we
attempt to understand how a firm’s CSR initiatives influence risk-taking incentives
arising from the structure of compensation contracts with CEOs. We find that firms
with better CSR standing in one period have significantly higher Vega in a subsequent
period, implying that these sample firms tend to grant higher risk-taking incentives to
CEOs when sustainable initiatives are viewed to be successful. This is consistent with
the risk management theory-based (Godfrey, 2005) view that CSR diversifies risks
because it creates reputational intangibles that provide insurance-like protection
towards uncertainty. Firms possessing this insurance-like benefit due to higher CSR
standing should have more risk-taking capacity going forward, and therefore respond
by designing compensation contracts that encourage more risk taking in the future. In
instances where managers are prone to take less risk, offering enhanced incentives
through option contracts could prove helpful in mitigating CEO risk-shirking.
We also explore the asymmetric impact of CSR strengths and concerns; we
conclude that the insurance-like effect of CSR is driven by CSR strengths instead of
concerns. Finally, relying on risk capacity and behavioral agency theories, we test the
25
moderating effect of firm risk on the association between CSR and CEO risk-taking
incentives, Vega. We find that the negative association between CSR and Vega is more
salient for firms with lower idiosyncratic risk and firm total risk.
Our research adds to the debate on whether CSR activities are value enhancing
or destroying. Our work, the first in the literature to study the effects of CSR on
executive risk incentives, indicates that the CSR standing of a firm drives executives’
compensation contracts (beyond other corporate governance and firm factors
considered in the literature). Firms pursing CSR and managing executive incentives
are arguably in a better position also to enhance shareholder value.
While we provide evidence on the effect of CSR on CEO pay-risk sensitivity,
several other topics are worth investigating. For instance, in a socially responsible
firm, is it essential to provide high pay-performance-sensitivity (i.e. Delta) to its
executives? If one believes CSR is value enhancing, then high Delta may not be
necessary. Alternatively, if CSR merely exacerbates agency cost, high Delta may be a
good way to mitigate this cost. We leave these questions for future research.
26
References
Anantharaman, D., Lee, Y. G. 2014. Managerial risk-taking incentives and corporate
pension policy. Journal of Financial Economics 111, 328-351.
Armstrong, C., Vashishtha, R. 2012. Executive stock options, differential risk-taking
incentives, and firm value. Journal of Financial Economics 104, 70-88.
Bansal, P., Gao, J., Qureshi, I. 2014. The extensiveness of corporate social and
environmental commitment across firms over time. Organization Studies 35,
949-966.
Bauer, R., Koedijk, K., Otten, R. 2005. International evidence on ethical mutual fund
performance and investment style. Journal of Banking and Finance 29,
1751-1767.
Bertrand, M., Mullainathan, S., 2003. Enjoying the quiet life? Corporate governance
and managerial preferences. Journal of Political Economy 111, 1043–1075.
Black, F., Scholes, M. 1973. The pricing of options and corporate liabilities. Journal
of Political Economy 81, 637-654.
Brammer, S., Brooks C., Pavelin, S. 2006. Corporate social performance and stock
returns: UK evidence from disaggregate measures," Financial Management 35,
97-116.
Cai, Y., Jo, H., Pan, C. 2011. Vice or virtue? The impact of corporate social
responsibility on executive compensation. Journal of Business Ethics 104,
159-173.
Chava, S. 2014. Environmental externalities and cost of capital. Management Science
60, 2223-2247.
Cheng, I-H., Hong, H., Shue, K. 2014. Do managers do good with other people’s
money? Working paper
Cheng, B., Ioannou, I., Serafeim, G. 2014. Corporate social responsibility and access
to finance. Strategic Management Journal 35, 1-23.
Coles, J. L., Daniel, N., Naveen, L. 2006. Managerial incentives and risk-taking.
Journal of Financial Economics 79, 431-468.
Core, J., Guay, W. 2002. Estimating the value of employee stock option portfolios and
their sensitivities to price and volatility. Journal of Accounting Research 40,
613-630.
Deckop, J., Merriman, K., Gupta, S. 2006. The effects of CEO pay structure on
corporate social performance. Journal of Management 32, 329-342.
Dhaliwal, D., Li, O., Zhang, A., Yang, Y. 2011. Voluntary nonfinancial disclosure and
the cost of equity capital: The initiations of corporate social responsibility
27
reporting. The Accounting Review 86, 59–100.
El Ghoul, S., Guedhami, O., Kwok, C.C.Y., Mishra, D. R. 2011. Does corporate social
responsibility affect the cost of capital? Journal of Banking and Finance 35,
2388-2406.
Flammer, C. 2015. Does corporate social responsibility lead to superior financial
performance? A regression discontinuity approach. Management Science 61,
2549-2568.
Flammer, C., Luo, J. 2016. Corporate social responsibility as an employee governance
tool: evidence from a quasi-experiment. Strategic Management Journal,
forthcoming.
Freeman, R.E. 1984. Strategic management: A stakeholder approach (Pitman, Boston,
MA).
Friedman, M. 1962. Capitalism and freedom (University of Chicago Press, Chicago,
IL).
Friedman, M. 1970. The social responsibility of business is to increase its profits.
New York Times Magazine, 13 September: 32-33.
Godfrey, P. 2005. The relationship between corporate philanthropy and shareholder
wealth: a risk management perspective. Academy of Management Journal 30,
777-798.
Godfrey, P., Merrill, C. B., Hansen, J. M. 2009. The relationship between corporate
social responsibility and shareholder value: an empirical test of the risk
management hypothesis. Strategic Management Journal 30, 425-445.
Goss, A., Roberts, G. S. 2011. The impact of corporate social responsibility on the
cost of bank loans. Journal of Banking and Finance 35, 1794-1810.
Granger, C. W. J. 1969. Investigating causal relations by econometric models and
cross-spectral methods. Econometrica 37, 424–38.
Guay, W. 1999. The sensitivity of CEO wealth to equity risk: an analysis of the
magnitude and determinants. Journal of Financial Economics 53, 43–71.
Hannan, E.J. and Quinn, B.G. 1979. The determination of the order of an
autoregression. Journal of the Royal Statistical Society, Series B, 41(2), 190-195.
Harjoto, M., Jo, H. 2011. Corporate governance and the CSR nexus. Journal of
Business Ethics 100, 45-67.
Haubrich, Joseph. 1994. Risk aversion, performance pay, and the principal-agent
problem, Journal of Political Economy 102, 258-276.
Hayes, R., Lemmon, M., Qiu, M. 2012. Stock options and managerial incentives for
risk-taking: evidence from FAS 123R. Journal of Financial Economics 105,
174-190.
28
Hillman, A. Y., Keim, G. D. 2001. Shareholder value, stakeholder management, and
social issues: what’s the bottom line? Strategic Management Journal 22, 125-139.
Hong, B., Li, Z., Minor, D. 2016. Corporate governance and executive compensation
for corporate social responsibility. Journal of Business Ethics 136, 199-213.
Hong, H., Kubik, J. D., Scheinkman, J.A. 2012. Financial constraints on corporate
goodness. Princeton University, working paper.
Husted, B. W. 2005. Risk management, real options, and corporate social
responsibility. Journal of Business Ethics 60, 175-183.
Ioannou I., Serafeim G. 2012. The consequences of mandatory corporate
sustainability reporting. Harvard Business School Working Paper.
Jensen, M., Meckling, W., 1976, Theory of the firm: Managerial behavior, agency
costs and ownership structure. Journal of Financial Economics, 3, 305-360.
Jones, T. 1995. Instrumental stakeholder theory: a synthesis of ethics and economics.
Academy of Management Review 20, 404-437.
Kahneman, D., Tversky, A., 1979, Prospect theory: An analysis of decisions under
risk, Econometrica, 47, 262-291.
Kim, E. H., Lu, Y. 2011. CEO ownership, external governance, and risk-taking.
Journal of Financial Economics 102, 272-292.
Low, A. 2009. Managerial risk-taking behavior and equity-based compensation.
Journal of Financial Economics 92, 470-490.
Luo, X., Bhattacharya, C. B. 2009. The debate over doing good: corporate social
performance, strategic marketing levers, and firm-idiosyncratic risk. Journal of
Marketing 73, 198-213.
Mahoney, L. S., Thorne, L. 2005. Corporate social responsibility and long-term
compensation: evidence from Canada. Journal of Business Ethics 57, 241-253.
Mahoney, L. S., Thorne, L. 2006. An examination of the structure of executive
compensation and corporate social responsibility: a Canadian investigation.
Journal of Business Ethics 69, 149-162.
Margolis, J., Elfenbein, H., Walsh, J. 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., Dow, S., Argheyd, K. 2003. CEO incentives and corporate social
performance. Journal of Business Ethics 45, 341-359.
McWilliams, A., Siegel, D. 2000. Corporate social responsibility and financial
performance: correlation or misspecification? Strategic Management Journal 21,
603-609.
29
McWilliams, A., Siegel, D. 2001. Corporate social responsibility: a theory of the firm
perspective. Academy of Management Review 26, 117-127.
Oikonomou, I, Brooks, C, Pavelin, S. 2012. The impact of corporate social
performance on financial risk and utility: a longitudinal analysis. Financial
Management 41, 483-515.
Orlitzky, M., Benjamin, J. D. 2001. Corporate social performance and firm risk: a
meta-analytic review. Business & Society 40, 369-396.
Peterenko, O. V., Aime, F., Ridge, J. 2016. Corporate social responsibility or CEO
narcissism? CSR motivations and organizational performance. Strategic
Management Journal 37, 262-279.
Porter, M., Kramer, M. 2011. Creating shared value. Harvard Business Review,
89(1–2), 62–77.
Risannen, J. 1978. Modeling by shortest data description. Automatica, 14(5),
465-471.
Ross, S. A. 1973. The economic theory of agency: the principal’s problem. American
Economic Review 63, 134-139.
Sanders, W. 2001. Behavioral responses of CEOs to stock ownership and stock option
pay. Academy of Management Journal 44, 477-492.
Sanders, W. G., Carpenter, M. 2003. Strategic satisficing? A behavioral-agency theory
perspective on stock repurchasing program announcements. Academy of
Management Journal 46, 160-178.
Schwarz, G. 1978. Estimating the dimension of a model. Annals of Statistics, 6(2),
461-464.
Servaes, H., Tamayo, A. 2013. The impact of corporate social responsibility on firm
value: the role of customer awareness. Management Science 59, 1045-1061.
Shleifer, A., Vishny, R. W. 1997. A survey of corporate governance. Journal of
Finance 52, 737-783.
Smith, C. W., Stulz, R. M. 1985. The determinants of firms’ hedging policies. Journal
of Financial and Quantitative Analysis 20, 391-405.
Strike, V. M., Gao, J., Bansal, P. 2006. Being good while being bad: social
responsibility and the international diversification of US firms. Journal of
International Business Studies 37, 850-862.
Stulz, R. M. 2002. Risk management and derivatives. Thomson/South-Western
Publishing (Mason).
Werner, T. 2015. Gaining access by doing good: the effect of sociopolitical reputation
on firm participation in public policy making. Management Science 61,
1989-2011.
30
Wiseman, R., Gomez-Mejia, L. R. 1998. A behavioral agency model of managerial
risk-taking. Academy of Management Journal 23, 133-153.
31
Table 1
Descriptive Statistics This table provides descriptive statistics for the variables used in our analysis over the sample
period. The sample consists of 11, 272 observations over fiscal years 1992 to 2010. All variables
are defined in the Appendix.
N Mean Median
Standard
deviation
Minimum Maximum
CSR Variables
CSR 11,272
0.201
0
2.401
-9 15
STRENGTH 11,272
1.699
1
2.293
-1 21
CONCERN 11,272
1.494
1
1.729
-4
13
COMS 11,272
0.154
0
0.655
-2
5
DIVS 11,272
0.360
0
1.351
-3
7
EMPS
11,272
-0.047
0
0.925
-5
5
ENVS 11,272
-0.067
0
0.857
-5 6
PROS 11,272
-0.206
0
0.708
-4 3
CEO incentive
Variables
DELTA
l
11,272
1564.471
312.532
13636.930
0 709829.705
VEGA
L
L
LOG
LOG
11,272
176.928
71.388
340.523
0 11344.740
Other Variables
Q 11,272
1.978
1.525
1.386
0.348 19.823
VOLATILITY 10,268 0.375 0.328 0.198 0.083
3.505
IDIOSYNCRATIC 11,233 0.57 0.36 0.83 0.003 2.55
TENURE 11,272
8.577
6
7.767
0 59
INSTHOLD 11,272
71.57
73.42
19.119
0.00007 100
ROA 11,272
0.138
0.131
0.100
-1.475 0.965
LEVERAGE 11,272
0.217
0.202
0.181
0
2.151
CAPEX
11,272
0.031
0.015
0.048
-0.372
0.484
SIZE 11,272
7.792
7.657
1.798
1.964
14.468
32
Table 2
Pearson Correlation This table reports the Pearson correlation coefficients among variables for 11,272 observations for the period 1992-2010. See Appendix for variable definitions. The
superscripts *, **, and *** indicate significance at the 10%, 5%, and 1% levels (two-tailed test), respectively.
DELTA VEGA CSR STRENGTH CONCERN TENURE INSTHOLD ROA LEVERAGE CAPEX Q SIZE
DELTA 1
VEGA 0.213*** 1
CSR 0.064*** 0.210*** 1
STRENGTH 0.062*** 0.356*** 0.731*** 1
CONCERN -0.006 0.178*** -0.416*** 0.313*** 1
TENURE 0.099*** 0.020** -0.048*** -0.115*** -0.087*** 1
INSTHOLD -0.075*** -0.028*** -0.115*** -0.135*** -0.020** 0.017* 1
ROA 0.071*** 0.038*** 0.060*** 0.045*** -0.024*** 0.012 0.074*** 1
LEVERAGE -0.035*** 0.032*** -0.041*** 0.072*** 0.150** -0.089*** -0.023*** -0.105*** 1
CAPEX 0.001 -0.048*** 0.001 -0.055*** -0.074*** 0.057*** 0.014 0.276*** -0.041*** 1
Q 0.214*** 0.084*** 0.126*** 0.042*** -0.118*** 0.023*** 0.015* 0.524*** 0.524*** 0.162** 1
SIZE 0.066*** 0.378*** 0.175*** 0.474*** 0.383*** -0.131*** -0.190*** -0.208*** 0.287*** -0.176*** -0.274*** 1
33
Table 3
The Effects of CSR on VEGA This table presents the results for the effects of corporate social responsibility (CSR) on CEO incentive,
i.e., VEGA. The dependent variable in each regression is the leading Vega (i.e., VEGAt+1), where VEGA
is measured as the dollar change in the value of CEO’s annual equity-based compensation associated
with a 0.01 change in the annualized standard deviation of the firm’s returns. CSR is the net score of
CSR rating (total strengths subtracting total concerns), based on five categories of KLD rating data, i.e.,
community, diversity, employee relations, environment, and product. All other variables are defined in
the Appendix. Coefficient estimates (p-values) are provided in the top (bottom) row. The superscripts *,
**, and *** indicate significance at the 10%, 5%, and 1% levels (two-tailed test), respectively.
Industry-fixed effect
model (1) Firm-fixed effect model
(2) Lagged Dependent
model (3)
Intercept -0.750***
(0.00) -0.240***
(0.00)
VEGA 0.704***
(0.00)
CSR (H1) 0.018*** (0.00)
0.009*** (0.00)
0.004*** (0.00)
DELTA 0.005***
(0.00)
0.006*** (0.00)
0.002*** (0.00)
TENURE 0.002*** (0.00)
-0.0003 (0.73)
0.0002 (0.51)
INSTHOLD 0.001*
(0.08)
-0.0003 (0.41)
0.0003* (0.09)
ROA 0.100**
(0.03)
0.024 (0.69)
0.003 (0.93)
LEVERAGE -0.119***
(0.00)
-0.032 (0.38)
-0.030** (0.05)
CAPEX 0.026
(0.73)
-0.213** (0.04)
0.042 (0.46)
Q 0.034***
(0.00)
0.021*** (0.00)
0.017*** (0.01)
SIZE 0.101***
(0.00) 0.151***
(0.00) 0.031***
(0.00)
Fixed Effects Year&Industry
Year&Firm
Year&Industry
Adjusted R2 0.30 0.66 0.63
# Observations 11,272 11,272 11,272
34
Table 4
The Effects of CSR Strength and Concern on VEGA This table presents the results for the effects of CSR strength and concern on CEO incentive, i.e.,
VEGA. The dependent variable in each regression is the leading Vega (i.e., VEGAt+1), where VEGA is
measured as the dollar change in the value of CEO’s annual equity-based compensation associated with
a 0.01 change in the annualized standard deviation of the firm’s returns. STRENGTH is the sum of all
strength scores, and CONCERN is the sum of all concern scores, based on five categories of KLD
rating data, i.e., community, diversity, employee relations, environment, and product. All other
variables are defined in the Appendix. Coefficient estimates (p-values) are provided in the top (bottom)
row. The superscripts *, **, and *** indicate significance at the 10%, 5%, and 1% levels (two-tailed
test), respectively.
Industry-fixed effect
model (1) Firm-fixed effect
model(2) Lagged Dependent
model (3)
Intercept -0.696***
(0.00) -0.230***
(0.00)
VEGA
0.702*** (0.00)
STRENGTH(H2) 0.027*** (0.00)
0.015*** (0.00)
0.006*** (0.00)
CONCERN (H2) -0.003 (0.16)
-0.002 (0.62)
-0.001 (0.42)
DELTA
0.006*** (0.00)
0.006*** (0.00)
0.002***
(0.00)
TENURE 0.003*** (0.00)
-0.0003 (0.70)
0.0003 (0.40)
INSTHOLD 0.001***
(0.00)
-0.0004 (0.23)
0.0003** (0.05)
ROA 0.088**
(0.05)
0.026 (0.65)
0.001 (0.99)
LEVERAGE -0.106***
(0.00)
-0.033 (0.36)
-0.028** (0.07)
CAPEX 0.064
(0.40)
-0.192* (0.06)
0.049 (0.38)
Q 0.031***
(0.00)
0.021*** (0.00)
0.016*** (0.00)
SIZE 0.087***
(0.00) 0.142***
(0.00) 0.028***
(0.00)
Fixed Effects Year&Industry
Year&Firm
Year&Industry
Adjusted R2 0.30 0.66 0.63
# Observations 11,272 11,272 11,272
35
Table 5
The Moderating Effect of Risk on the Association between CSR and Vega This table tests the moderating effect of risk on the association between CSR and Vega. The dependent
variable in each regression is the leading Vega (i.e., VEGAt+1), where VEGA is measured as the dollar
change in the value of CEO’s annual equity-based compensation associated with a 0.01 change in the
annualized standard deviation of the firm’s returns. In model (1), our proxy for risk is idiosyncratic risk.
In model (2), our proxy for risk is volatility (i.e., total risk). All variables are defined in the Appendix.
Coefficient estimates (p-values) are provided in the top (bottom) row. The superscripts *, **, and ***
indicate significance at the 10%, 5%, and 1% levels (two-tailed test), respectively.
(1) (2)
CSR
0.012*** (0.00)
0.026***
(0.00) IDIOSYNCRATIC -0.018***
(0.00)
CSR*INDIOSYNCRATIC (H3)
-0.008*** (0.00)
VOLATILITY
0.397 (0.40)
CSR*VOLATILITY(H3)
-0.061***
(0.00)
DELTA 0.006*** (0.00)
0.010*** (0.00)
TENURE -0.0002
(0.80)
0.001 (0.40)
INSTHOLD -0.0004
(0.21) 0.001** (0.04)
ROA -0.023
(0.71) 0.201***
(0.00)
LEVERAGE -0.022 (0.55)
-0.102 (0.13)
CAPEX -0.213** (0.03)
-0.254 (0.12)
Q 0.022***
(0.00)
0.027*** (0.00)
SIZE 0.152***
(0.00) 0.230***
(0.00)
Fixed Effects Year&Firm
Year&Firm
Adjusted R2 0.66 0.66
# Observations 11,233 10,268
36
Table 6
Results for the CSR Five Categories This table presents the results for the effects of CSR five categories, i.e., community (COMS), diversity
(DIVS), employee relations (EMPS), environment (ENVS), and product (PROS) on CEO incentives
(DELTA and VEGA). The dependent variable is the leading Vega (VEGAt+1). All variables are defined
in the Appendix. Coefficient estimates (p-values) are provided in the top (bottom) row. The superscripts
*, **, and *** indicate significance at the 10%, 5%, and 1% levels (two-tailed test), respectively.
(1) (2)
COMS 0.024***
(0.00)
0.006***
(0.00)
DIVS
0.036***
(0.00)
0.028***
(0.00)
EMPS -0.0001
(0.99)
0.008*
(0.07)
ENVS 0.017***
(0.00)
0.014**
(0.02)
PROS -0.010**
(0.05)
-0.007***
(0.00)
DELTA 0.005***
(0.00)
0.006***
(0.00)
TENURE 0.002***
(0.00)
0.0006
(0.94)
INSTHOLD 0.0004*
(0.06)
-0.0006*
(0.08)
ROA 0.086*
(0.06)
0.041
(0.49)
LEVERAGE -0.121***
(0.00)
-0.038
(0.30)
CAPEX 0.065
(0.40)
-0.161
(0.11)
Q 0.032***
(0.00)
0.021***
(0.00)
SIZE 0.091***
(0.00)
0.138***
(0.00)
Fixed Effects Year/Industry Year/Firm
Adjusted R2 0.30 0.67
#Observations 11, 272 11,272
37
Table 7
Granger Causality Test This table presents the results of Granger causality test applied to the VAR residuals corresponding to
CSR rating and VEGA. The optimal lag length is set to 4 based on the BIC and the QIC. Lagged values
of CSR ratings and lagged values of VEGA are included as control variables. All variables are defined
in the Appendix.
H0: CSR Do Not Cause Vega
Dependent Variable = VEGA
Chi-square P-value
H0: Vega Do Not Cause CSR
Dependent Variable = CSR
Chi-square P-value
CSR 36.62 0.00
VEGA
8.68 0.07
All 36.62 0.00 All 8.68 0.07
38
Table 8
Results Using Instrumental Variables for CSR This table presents the results for the effects of CSR on CEO incentive using instrumental variables for CSR. The endogenous regressor for model (1) is STRENGTH, for
model (2) is CONCERN, and for model (3) is CSR overall score. In first stage, we employ two variables: (1) The first instrument is the average STRENGTH/CONCERN/CSR
score for each state per year; 2) The second instrument is the average STRENGTH/CONCERN/CSR score for each industry per year, basing on two-digit SIC score. In second
stage, we adopt the predicted STRENGTH/CONCERN/CSR values from the first stage. The dependent variable is the leading Vega (VEGAt+1) for all second-stage models.
Coefficient estimates (p-values) are provided in the top (bottom) row. The superscripts *, **, and *** indicate significance at the 10%, 5%, and 1% levels (two-tailed test),
respectively. Model (1) - Strength Model Model (2) – Concern Model Model (3) – CSR Overall Score Model
First Stage (1) Second Stage (2) First Stage (3) Second Stage (4) First Stage (5) Second Stage (6)
Intercept -5.775*** (0.00)
-0.829*** (0.00)
-3.386*** (0.00)
-0.753*** (0.00)
-2.014*** (0.00)
-0.773*** (0.00)
State-Year Mean of CSR
(First Instrument)
0.731***
(0.00)
0.548***
(0.00)
0.727***
(0.00)
Industry-Year Mean of CSR
(Second Instrument)
0.607***
(0.00)
0.717***
(0.00)
0.872***
(0.00)
STRENGTH 0.023*** (0.00)
CONCERN -0.024***
(0.00)
CSR 0.034***
(0.00)
DELTA -0.004** (0.02)
0.005*** (0.00)
-0.003*** (0.00)
0.005*** (0.00)
-0.001 (0.79)
0.005*** (0.00)
TENURE -0.013***
(0.00)
0.002***
(0.00)
-0.008***
(0.00)
0.002***
(0.00)
-0.005*
(0.08)
0.002***
(0.00) INSTHOLD -0.010***
(0.00)
0.002***
(0.00)
-0.005***
(0.00)
0.0002
(0.26)
-0.004***
(0.00)
0.0003
(0.15)
ROA 1.160*** (0.00)
0.117*** (0.01)
0.020 (0.90)
0.114*** (0.00)
1.292*** (0.00)
0.134*** (0.00)
LEVERAGE -1.088***
(0.00)
-0.137***
(0.00)
-0.444***
(0.00)
-0.126***
(0.00)
-0.692***
(0.00)
-0.133***
(0.52) CAPEX 0.060
(0.89)
0.075
(0.33)
-1.875***
(0.00)
0.051
(0.52)
1.813***
(0.00)
0.050***
(0.00)
Q 0.147*** (0.00)
0.036*** (0.00)
0.033** (0.02)
0.038*** (0.00)
0.093*** (0.00)
0.035*** (0.00)
SIZE 0.659***
(0.00)
0.103***
(0.00)
0.404***
(0.00)
0.109***
(0.00)
0.252***
(0.00)
0.106***
(0.00) Fixed Effect Year/Industry Year/Industry Year/Industry Year/Industry Year/Industry Year/Industry
Adjusted R2 0.42 0.29 0.43 0.28 0.29 0.29
# Observations 11,272 11,272 11,272 11,272 11,272 11,272
39
Appendix
Variables Definition and Data Sources Variables Definition Data Sources
CSR Variables
CSR Net score of CSR rating (total strengths subtracting
total concerns), based on five categories of KLD
rating data, i.e., community, diversity, employee
relations, environment, and product;
KLD database
STRENGTH The sum of all strength scores, based on five
categories of KLD rating data, i.e., community,
diversity, employee relations, environment, and
product;
KLD database
CONCERN
COMS
DIVS
EMPS
ENVS
PROS
The sum of all concern scores, based on five
categories of KLD rating data, i.e., community,
diversity, employee relations, environment, and
product;
Net score of CSR rating (total strengths subtracting
total concerns), based on one category of KLD rating
data, i.e., community;
Net score of CSR rating (total strengths subtracting
total concerns), based on one category of KLD rating
data, i.e., diversity;
Net score of CSR rating (total strengths subtracting
total concerns), based on one category of KLD rating
data, i.e., diversity;
Net score of CSR rating (total strengths subtracting
total concerns), based on one category of KLD rating
data, i.e., environment;
Net score of CSR rating (total strengths subtracting
total concerns), based on one category of KLD rating
data, i.e., product.
KLD database
KLD database
KLD database
KLD database
KLD database
KLD database
CEO Incentive Variables
DELTA Dollar change in the value of CEO’s annual
equity-based compensation for a 1% change in the
stock price (in $000s);
ExecuComp Database
VEGA Dollar change in the value of CEO’s annual
equity-based compensation associated with a 0.01
change in the annualized standard deviation of the
firm’s returns (in $000s).
ExecuComp Database
Other Variables
Q The firm–year Tobin’s Q, which is computed as the
sum of the book value of total assets plus the market
value of common stock less the book value of equity
over the book value of assets;
Compustat
VOLATILITY The annualized monthly standard deviation of a
firm’s return series;
CRSP
IDIOSYNCRATIC Idiosyncratic risk is the standard deviation of the
residuals (slope coefficient) from the Fama-French
three-factor market model;
CRSP
TENURE The length of time that the CEO has been at his or
her position;
ExecuComp Database
INSTHOLD
ROA
LEVERAGE
CAPEX
Percentage of institutional share ownership;
Return on assets;
Total liabilities over total assets;
Capital expenditure expenses over total assets;
Thomson Reuters
(CDA/Spectrum 13(f)
filings)
Compustat
Compustat
Compustat
SIZE Log of total assets at the end of the fiscal period; Compustat
40
YEAR Year dummies for the period from 1992 to 2010; Compustat
INDUSTRY Industry dummies, petroleum (SIC codes 13, 29),
consumer durables (SIC codes 30, 36, 37, 50, 55,
57), basic industry (SIC codes 8, 10, 12, 14, 24, 26,
28, 33), food and tobacco (SIC codes 20, 21, 54),
construction (SIC codes 15, 16, 17, 32), capital goods
(SIC codes 34, 35, 38, 39), transportation (SIC codes
40, 41, 42, 44, 45, 47), textiles and trade (SIC codes
22, 23, 51, 53, 56, 59), services (SIC codes 7, 73, 75,
80, 82, 83, 87, 96), leisure (SIC codes 27, 58, 70, 79),
unregulated utilities (SIC code 48), regulated utilities
(SIC code 49), and financials (SIC codes 60, 61, 62,
63, 65, 67).
Compustat