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:

cdunbar@ivey.ca); Frank Li is at Richard Ivey School of Business, Western

University (Email: fli@ivey.ca); Yaqi Shi is at Richard Ivey School of Business,

Western University (Email: yshi@ivey.ca). 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