The Moderating Effect of a Golden Parachute on the Association between CSR and Firm Value: Does Gender-Driven Innovation Matter?

Abstract

We revisit the debate on whether a firm’s corporate social responsibility (CSR) activities enhance firm value. Research on related topics has produced mixed results suggesting a need to further investigate factors that directly or indirectly impact the CSR–firm value association. To this end, we examine if a firm’s adoption of a golden parachute (GP) moderates the relationship between CSR and firm value. We also investigate if diversity-based innovation as it pertains to the gender of executives reveals any difference in the CSR–firm value relation. Using a sample of 11,065 firm-year observations of publicly traded US firms from 2007 to 2016, we find that CSR activities are significantly and positively associated with firm value. More importantly, our study shows that for US firms that issue GPs, this severance pay strengthens this positive relationship, suggesting that CEOs with a GP engage in more value-enhancing innovative CSR projects than their counterparts without it. This finding supports the conflict resolution theory and the resource-based view of the firm. A test to examine if the gender of the corporate executives alters their behavior towards CSR when the GP protects them shows an inverse relationship between female executives and CSR–firm value association. This interesting finding lends credence to related theories suggesting that women in male-dominated fields may feel pressured to conform to the stereotype of women as less competent than men and may adopt traditionally masculine behaviors to counteract this stereotype. As they climb the corporate leadership ladder endowed with a GP, the stereotype threat may still prevail, adversely affecting the CSR–firm value outcomes. These results remain robust after a series of sensitivity tests.

1. Introduction

Corporate social responsibility (CSR) is a crucial component of contemporary business, and innovative CSR strategies can help organizations stand out from the competition and positively impact society. However, there is no consensus in the literature on the impact of CSR on firm performance. There is evidence suggesting a positive association between a firm’s CSR engagements and its financial performance [1,2,3,4]. In contrast, some empirical studies find a negative association [5,6,7]. Others see no association [8,9].

CSR is defined as the actions that a firm chooses to take that substantially enhance the well-being of its stakeholders, community, and environment [10]. The nonconsensus enumerated above has triggered skepticism surrounding CSR’s value-enhancing potential. It creates a void and gap in the literature, clogging the wheels of the CSR campaign. It has compelled some to posit that CSR is antithetical to sound business practice because it dilutes the firm’s focus on the shareholders’ wealth creation [11,12]. However, advocates of CSR argue that it is vital for sustainable business operations. Thus, firms must look beyond narrow economic returns and embrace CSR’s broader social impact [13]. CSR is paramount to society. The lack of unanimity in the related literature on the impact of CSR on firm value calls for further investigation examining how firms can better strategically align the interests of both the investing and non-investing stakeholders. This compelling need is a motivating factor that drives our research quest.

This study contributes to the literature on similar topics investigating the association between CSR and firm value. Focusing on the US corporate environment, we extend prior literature by examining CEOs’ behavior toward CSR when they have a golden parachute (GP) (A golden parachute is a contractual clause in the executives’ compensation package that guarantees a set of benefits to a CEO if a firm is acquired, employment is terminated, or the CEO remains with the firm through a recessionary cycle.). The effect of the executives’ gender on this relationship is also closely examined. Firms issue golden parachutes to attract the best talent for managerial positions and stay competitive in their respective industry. The uncertainty surrounding the value-adding potential of CSR poses some risk to both the firm and the executives. Is the manager protected from these risks? A golden parachute clause would somewhat insulate the manager from possible employment risks associated with the uncertainties of CSR investments. The provision of a golden parachute in the executive compensation contract requires that substantial benefits be given to the manager if their employment is terminated. The manager could become more willing to explore and promote innovative CSR investments with such a condition.

Corporate executives want to maximize the value of their golden parachutes. To achieve this, they cautiously and consciously approve those innovative projects with the highest likelihood of enhancing the firm value that, in return, should maximize the value of the golden parachute (According to Bloomberg, the largest portion of severance often comes from accelerated payments of unvested stock options and other stock units that would otherwise not be eligible for a payout—http://www.bloomberg.com/news/2013−06−06/golden-parachutes (accessed on 12 March 2023)). One can logically assume that this category of executives would only pursue and approve CSR projects that are undoubtedly value-enhancing.

Prejudice and gender bias can also stifle innovation in the workplace. Women may encounter barriers at work, for instance, that restrict their ability to take the lead on or participate in innovation projects [14]. Women may also be less likely to be acknowledged or given credit for their creative suggestions, which deters them from making further contributions [15]. What happens when the GP protects women in top leadership positions? How does this affect their behavior towards CSR?

Our findings show that the association between CSR and firm performance is positive. First, we test for the moderation role of the golden parachute. The results show that it magnifies this positive association because the CEOs with this clause in their contract engage in more value-enhancing CSR projects than their counterparts without it. Our results support the conflict resolution theory and the resource-based view of CSR. These related theories argue that top managers use CSR as a device to develop and build good relationships with the firm’s stakeholders, which ultimately leads to the enhancement of the firm value.

Second, we disaggregate the CSR measure to examine which types of CSR CEOs with golden parachutes believe are most likely to enhance firm value. The results of our analysis show that the golden parachutes positively moderate the value-adding power of three CSR components: community, human rights, and product.

Third, of distinct interest to this research, we classify the data into female and male groups. The results of the analysis conducted reveal that for firms with female executives who are protected by the GP, the relation between CSR and firm value is negative but not statistically significant, whereas for the male group, the strong positive moderating power of the GP is sustained. This contrasting finding could be explained by related theories on why women in male-dominated fields may feel pressure to adopt traditionally masculine behaviors, especially on their rise to the top of the corporate ladder.

Self-verification and social role theories suggest that women in male-dominated fields may feel pressure to adopt traditionally masculine behaviors because of societal expectations, the threat of being judged based on gender stereotypes, and a desire to maintain their sense of self. Drawing from the lens of stereotype threat theory, in the context of our study, the stigmatization of women in the corporate arena may cause them to experience anxiety or stress when they feel their performance could confirm a negative stereotype about their abilities. This anxiety or stress can impair their cognitive functioning and negatively impact their performance in managerial investment decisions that impact the CSR–firm value outcomes.

The rest of this study proceeds as follows: Section 2 provides the relevant literature review. Section 3 presents hypothesis development. Section 4 provides data and the research model. Section 5 presents the result analysis and endogeneity tests. Section 6 shows the discussion. Section 7 concludes the study.

2. Literature Review

2.1. CSR and Golden Parachutes

The literature on CSR and firm value has produced mixed and sometimes conflicting findings. Empirical evidence supporting a positive association between CSR and firm value argues that CSR could enhance a firm’s value in multiple ways. It could improve the firm’s reputation and reduce the frequency of investors’ explicit claims [16], reduce the cost of capital [17], and reduce an undesirable debt ratio [18]. There is also evidence to suggest that CSR increases the willingness of customers (investors) to pay premium prices for socially responsible products of the firm because a good reputation is usually considered a signal of quality products [19].

In contrast to the positive evidence enumerated above, there is also opposing evidence discrediting the positive association between CSR and firm performance. Some empirical studies find a negative association [5,6,7], and others see no association [8,9]. Reference [20] uses a meta-analysis of CSR and firms’ performance to show the literature’s ambiguity and inconsistent empirical findings. The authors find that 58% of the investigated studies show no significant relation, 27% show a positive relation, 2% show a negative relation, and 13% do not report a sample size.

Several compelling theories attempt to provide a better insight into why and how firms engage in CSR projects. The conflict resolution hypothesis stems from the stakeholder theory. It states that top managers use CSR as a device to develop and build good relationships with the firm’s stakeholders. A good relationship with the employees could yield productivity increases. A good relationship with the regulatory agencies and the community could save the firm the costs associated with litigation fees, consumer boycotts, reputation damage, and regulatory fines. Excellent product quality also increases the likelihood of satisfied return customers and decreases the costs associated with product defects.

The resource-based view theory is often cited as the process through which CSR engagement could generate financial benefits for the firm [21,22,23,24]. These benefits could be achieved through the improvement of stronger and better relationships with customers [25], enhancement of future revenue growth [26], motivation and enhancement of employees’ morale that could increase productivity [27], or increase in investor support [28] as well as providing “insurance-like” protection for the firm and its shareholders when an adverse event occurs [29,30]. In sum, the resource-based view posits that, by engaging in CSR, firms procure legitimacy in the public eye that they can strategically capitalize on to enhance their value.

The above two theories, conflict resolution and the resource-based viewpoint, explain CSR’s value-enhancing characteristics and how firms can efficiently and effectively use them. Thus, if CSR is value-enhancing, CEOs with a golden parachute would pursue more of these investments. As a result, this increased pursuit should also enhance the value of the CEOs’ golden parachute and long-term wealth.

The overinvestment hypothesis argues that top managers might overinvest in CSR projects. Managers could pursue CSR to satisfy a selfish desire to burnish their reputation as good global citizens at the shareholders’ expense [31,32]. That is, managers might seek to overinvest in CSR, regardless of its effect on the shareholders’ wealth, for their private gains. Consistent with the overinvestment hypothesis, some studies find a positive association between CSR and the CEO’s compensation [33,34].

The strategic-choice hypothesis posits that incumbent managers strategically select CSR projects in which to invest. The motivation is to generate likability and support from environmental and social activists to reduce the manager’s employment risk by decreasing the probability of future turnover [35,36]. The strategic-choice hypothesis argues that when the stakeholders’ protection relies on managers’ discretionary initiatives, the outcome could increase managerial opportunism. The authors of [36] assert that relations with social activists could become an effective entrenchment strategy for inefficient managers. That is, in a desperate attempt to build a good relationship with social activists, the manager could strategically choose to invest in selected CSR projects even if doing so reduces the shareholders’ wealth. Such CSR projects could be in the form of donations to charitable organizations that bring positive visibility to the manager, or the funding of a costly overelaborate campaign to promote diversity among the firm’s workforces.

The overinvestment and strategic choice theories both caution against the potential value-destroying characteristics of CSR engagements. Thus, CEOs with a golden parachute would likely frown on such CSR policies because they would most likely pose a risk to the value of their golden parachute and long-term wealth.

2.2. Executive Compensation, CSR, and Firm Performance

Several studies on related topics have explored the relationship between CSR, firm value, and various firm characteristics. They provide insights into the potential effects of factors such as ownership concentration and CEO incentives on the relationship between CSR and firm value [17,37].

Executive compensation can serve as a critical tool in aligning the manager’s interest with that of the shareholders. If well designed, it could also help monitor the activities and steer the manager’s decision making that affects the firm [38]. On the other hand, this monitoring tool could unintentionally restrict the manager’s decision making, negatively and adversely affecting the degree of CSR commitment. With the uncertainty surrounding CSR payoffs, the decisions or agreements on allocating resources to CSR projects could prove challenging for firms [39].

Presumably, firms design fixed and shorter incentive arrangements to focus the managers’ attention on projects geared towards immediate performance. Short-term performance values are usually captured using retrospective accounting numbers [40]. On the other hand, a longer-term compensation arrangement is more focused on the expectation that managers optimize their value in the long run. Thus, it should motivate managers to invest in CSR engagements [41]. Longer-term compensation arrangements are usually based on market valuation. When rightfully and efficiently implemented, CSR investments could drive the firm’s value upwards long-term.

Suppose a difference of opinion between the manager and the shareholders on the CSR commitment is significantly large. In that case, this difference could infringe on the flow and success of the CSR campaign. A CSR commitment above the shareholders’ willingness to commit could have negative consequences for the manager if earnings targets are not met due to the costs of the CSR investment. Thus, the manager might become reluctant to assume the risk associated with the uncertainty of the CSR payoffs. Job security and compensation risk are some of the risks that managers face.

Is the manager protected from these risks? The manager’s concerns about the repercussion of assuming higher risk by investing in CSR engagements could be alleviated by a type of compensation mechanism that insulates the manager from employment risk. The provision of a golden parachute in the executive compensation contract requires that substantial benefits be given to the manager if their employment is terminated. The manager could become more willing to explore and promote CSR investments with such a condition. Or the insulation that the GP provides could embolden CEOs to reduce CSR focus, knowing that the timing of its value-enhancing attributes may not be in alignment with the CEOs’ immediate personal financial interest.

Firms design long-term executive compensation schemes to focus their managers’ efforts on optimizing their long-term goals and value [42]. Reference [40] finds a relationship between long-term compensation and CSR weakness. That is, long-term executive compensation, if well designed, could reduce the risk associated with CSR noncompliance. Female executives are assumed to be more risk-averse than their male counterparts. Given that the CSR cost to the firm and the benefits, if any, are not always in alignment, this could create uncertainties that a risk-averse executive may prefer to avoid. Motivated by this line of reasoning, we examine female executives’ behavior towards CSR when they have a GP and when they do not have a GP.

2.3. Gender and Corporate Decision-Making

Gender diversity can influence creativity in a number of different ways. Research has demonstrated that diverse teams, including gender diversity, can produce value-enhancing innovations that drive corporate breakthroughs. Gender diversity brings to the table diverse viewpoints, rich and broad experiences, and creative problem-solving methods [43,44,45,46]. However, studies have shown that women tend to be more compassionate than men [47]. These tendencies are reflected in their business and corporate decision-making as it pertains to leadership style, consensus decision-making, and risk taking [48]. Firms must be willing to take optimal levels of risk to maximize both financial and non-financial performance. Women, being more risk-averse, could adversely impact corporate decision-making opportunities that enhance firm financial performance. In business environments where shareholders’ expectations are focused on value creation and management’s ability to drive stock appreciation, the female tendencies of being more compassionate and risk-averse may be viewed as weaknesses that hamper firm financial performance. So, this could explain why women who rise to positions of power in male-dominated fields may feel pressure to adopt traditionally masculine behaviors in order to be seen as competent and assertive leaders.

There are several related theories that explain why women in male-dominated fields may feel pressure to adopt traditionally masculine behaviors. Self-verification theory suggests that individuals have a need to be seen and treated in ways that are consistent with their self-concept [49]. Women in male-dominated fields may adopt traditionally masculine behaviors in order to verify their sense of self as competent and capable leaders, especially in a context where they may feel that their abilities are being questioned.

Social role theory posits that individuals adopt behaviors that are consistent with the gender roles they are expected to perform [50]. In male-dominated fields, women may feel pressure to adopt traditionally masculine behaviors in order to conform to societal expectations of what a competent leader should be.

Stereotype threat theory in social psychology refers to the feeling of being in danger of confirming unfavorable preconceptions about one’s group [51,52]. In the late 1990s, psychologists Claude Steele and Joshua Aronson made the initial presentation of it [53]. According to the theory, people who belong to a group that is stereotyped as possessing particular negative traits may feel anxious or apprehensive about supporting those stereotypes. Even when the person is otherwise capable of performing well, this can result in underperformance in tasks that are connected to the stereotype [53,54,55]. In the context of our research, if female executives are stereotyped as being less competent than their male counterparts, they may feel anxious when making investments or other related managerial decisions, which can, in turn, impair their performance.

There is also evidence in related literature that suggests that female and male executives differ systematically regarding their core values and attitudes towards risk, which differs from gender differences in the general population. Compared to their male counterparts, female executives are more security-oriented [56]. This line of thinking could shed light on the differences in behavior towards CSR when the executive is protected by a GP.

3. Hypothesis Development

If CSR is value-enhancing to the firm, then it should help maximize the value of the stock option component of the GP (According to Bloomberg, the largest portion of severance often comes from accelerated payments of unvested stock options and other stock units that would otherwise not be eligible for a payout—http://www.bloomberg.com/news/2013−06−06/golden-parachutes (accessed on 12 March 2023).). Executives endowed with a GP desire to maximize the value of their compensation package [39]. Therefore, the presence of a GP will moderate the association between corporate social responsibility (CSR) and firm value such that the positive relationship between CSR and firm value will be stronger for firms with a GP than those without a golden parachute.

Based on the above rationale, we propose the first hypothesis that the effect of CSR on firm value is contingent on the presence of a GP. A golden parachute is a compensation arrangement that provides executives with substantial financial benefits in the event of a change in control or a merger/acquisition. Corporate executives with a GP may be more inclined to pursue and engage in more CSR, especially if it enhances firm value, which ultimately helps maximize the value of their GPs. Therefore, we predict that the positive relationship between CSR and firm value, if it exists, will be amplified for firms with a GP compared to firms without one.

To test this hypothesis, we use a regression analysis with CSR, golden parachute, and their interaction term as independent variables and firm value as the dependent variable. The results of this analysis could indicate whether the presence of a GP moderates the association between CSR and firm value and to what extent.

Hypothesis 1: 

Ceteris paribus, US-based firms that issue a GP to their executives will show that it amplifies the positive relationship between CSR and firm value.

Research has shown that gender-diverse teams bring a range of perspectives and experiences to the table, which can lead to better financial outcomes for the firm. Some of the value-adding attributes include more creative and effective problem-solving, better decision-making, increased innovation, better talent recruitment and retention, and enhanced firm reputation [46,47,48]. Despite this evidence, women face challenges navigating the corporate workplace and upward mobility ladder.

There is empirical evidence such as the social role theory [50] and self-verification theory [49] that women adopt traditionally masculine behavior to counter these stereotypes that consider them less competent. The fear or anxiety that stems from stereotype threat theory [55] may induce or pressure female executives to conform to those stereotypes, which can lead to decreased performance in areas where the stereotype is relevant. There is an expectation that corporate executives make investment decisions that positively drives firm value. In the presence of stereotype threats, in high-stakes situations, female executives may experience anxiety and reduced performance due to the fear of confirming the negative stereotype.

Drawing from these frameworks for understanding the ways in which gender characteristics and social expectations shape women’s lives and experiences, we predict a difference in behavior when female executives are protected by a GP. They may feel insulated from these judgment-based social assessments and their associated consequences; ergo, they may exhibit different behavior toward CSR. It could also be that the stereotype threat prevails, consequently impacting the CSR–firm value relationship negatively.

Hypothesis 2: 

The behavior of female corporate executives in US-based firms towards CSR will change if they are protected by a GP.

This study contributes to the related research on CSR and firm value by introducing the golden parachute. We posit that the presence of a golden parachute alleviates the executive’s short-term concerns and incentivizes him or her to maximize the firm’s performance over the long term under a firm’s proactive CSR activities. Based on the arguments drawn from the previous studies, we predict that the interaction effect of both CSR and GP will be stronger than the individual effect of CSR on firm value. Accordingly, we expect the GP to amplify the positive relationship between CSR and firm value. Also, toeing the line of reasoning put forward by social role and stereotype theories that women may adopt traditionally masculine behavior to counteract the stereotypes of women as less competent than men, we predict that behavior toward CSR could change when female executives are insulated from these judgment-based social assessments.

4. Data and Research Model

4.1. Data

To examine our research question, we employed the sample period from 2007 to 2016. The CSR data were collected MSCI KLD social index. The index provides comprehensive CSR data covering about 3,100 public US companies. This CSR data cover approximately 98% of the listed United States companies [32] and show the largest comprehensive CSR performance database [57]. Given their validity for multi-dimensional CSR measurement [58], KLD CSR rating data are widely used in CSR research [39,59,60,61,62,63].

The KLD ratings have seven categories: community, governance, diversity, employee relations, environment, human rights, and product. However, researchers consider the governance category a different construct from other CSR activities and exclude it from CSR measurements [1,64,65].

A GP is a contractual clause included in some executive compensation packages. Originally intended as a hedge against takeovers, its rise and adoption in executive compensation package contracts reflect more than a three hundred percent growth [38]. A golden parachute remains a mainstay. GP data come from the institutional shareholder services database (IRRC). GP is a dummy variable, where 1 is an indicator that the executive has adopted a golden parachute.

We employ Tobin’s Q (TOBINQ) to measure firm value because Tobin’s Q is widely accepted as a firm’s long-term performance measure [63,66,67,68,69], and CSR is highly likely to pay off in the long run [70]. Moreover, management can easily manipulate accounting performance measures such as ROA by employing different accounting methods or policies.

We include control variables following prior studies to influence the CSR–firm value relationship in our model. We control for firm size (SIZE) [66,71,72] and sales growth rate (GROWTH) in that their possible effect on firms’ stock price. We expect a positive relationship between GROWTH and firm value. We also control financial leverage (LEVERAGE) [66,71]. A firm’s profitability can positively influence the return on assets (ROA) in the stock price. Then, we include a firm’s business risk (RISK) [16], R&D_INTENSITY [73,74], GOVERNANCE [75,76], and firm age (FIRM_AGE) [77,78]. Following prior studies, we also include CEOs’ characteristics, such as age [79], service period [80,81], and gender [82], affecting a firm’s financial performance. Lastly, we include industry effects and year effects to control for the possible influence on the relationship between CSR and firm value [71].

4.2. Research Model

To investigate the moderating role of GP in the relationship between CSR and firm value, we use the following model (1):

TOBINQ_i.t = α₀ + α₁ CSR_i,t +α₂ GP_i,t

+ α₃ CSR × GP_i,t + α₄ SIZE_i,t + α₅ LEVERAGE_i,t + α₆GROWTH_i,t + α₇ ROA_i,t +α₈ RISK_i,t

+ α₉ R&D_INTENSITY_i,t + α₁₀ GOVERNANCE_i,t + α₁₁ FIRM_AGE_i,t + α₁₂ CEO_AGE_i,t

+ α₁₃ CEO_TENURE_i,t + α₁₄ CEO_GENDER_i,t +∑IND + ∑YEAR + ε_i,t

(1)

where for a company i in period t, the variables are defined in the table below.

Variable	Definition
Dependent variables
TOBINQ	=(market value of equity + total liabilities)/book value of total assets, at fiscal year-end in the period;
Variables of interest
CSR	=summation of scores for the community, diversity, employee relations, environment, human rights, and product in KLD’s ratings;
GP	=a binary variable that equals 1 if the firms have adopted golden parachutes in that year and 0 otherwise;
Control variables
SIZE	=natural logarithm of total assets at the end of the fiscal year;
LEVERAGE	= total liability divided by total assets at the end of the fiscal year;
GROWTH	=net sales minus net sales at the prior year and divided by the net sales at the end of the fiscal year;
ROA	=operating income divided by total assets at the end of the fiscal year;
RISK	=standard deviation of EBIT (earnings before interests, taxes) from the past five years divided by total assets at the end of the fiscal year;
R&D_INTENSITY	=R&D expenses divided by net sales at the end of the fiscal year;
GOVERNANCE	=net score (strengths minus concerns) in KLD governance ratings;
FIRM_AGE	=natural logarithm of (1 + the number of years since the firm’s first appearance in the Compustat database);
CEO_AGE	=the age of the CEO;
CEO_TENURE	=how long the CEO has been serving in that capacity for the firm at the end of the fiscal year;
CEO_GENDER	=the gender of the CEO. A dummy variable that equals 1 if female and 0 if male;
IND	=industry indicator based on the two-digit SIC code;
YEAR	=year indicator.

5. Result Analysis and Endogeneity Test

5.1. Univariate Result

Table 1. Sample Distribution of Firm-Year Observations by Year.

Year	# of Firms	% of Sample
2007	622	5.62
2008	1099	9.93
2009	1135	10.25
2010	1133	10.24
2011	1174	10.62
2012	1183	10.69
2013	1136	10.27
2014	1195	10.80
2015	1169	10.56
2016	1219	11.01
Total	11,065	100.00

offers a firm-year description of our sample, showing even distribution except for 2007. Our sample consists of 11,065 firm-year observations. The number of firms is spread almost evenly across the years except for 2007. In 2007, relatively more companies were omitted than in other years while merging Compustat data and KLD CSR data, which accounts for 5.62% of the total sample. Finally, the average number of firms used in our sample is 1229 (11,065/9 years).

Table 2. Descriptive Statistics of Selected Variables.

Full Sample
Variables			n	Mean	Median	Std. Dev.		25th Percentile	75th Percentile
Dependent Variables
TOBINQ			11,065	1.842	1.481	1.154		1.128	2.117
Variable of Interest
CSR			11,065	0.793	0.000	2.628		−1.000	2.000
GP			11,065	0.742	1.000	0.438		0.000	1.000
Control Variables
SIZE			11,065	8.140	8.054	1.571		6.959	9.230
LEVERAGE			11,065	0.559	0.556	0.232		0.405	0.709
GROWTH			11,065	0.063	0.048	0.220		−0.024	0.130
ROA			11,065	0.090	0.081	0.088		0.042	0.130
RISK			11,065	0.032	0.021	0.053		0.010	0.037
R&D_INTENSITY			11,065	0.030	0.000	0.111		0.000	0.022
GOVERNANCE			11,065	−0.205	0.000	0.675		−1.000	0.000
FIRM_AGE			11,065	3.207	3.219	0.626		2.833	3.807
CEO_AGE			11,065	57.488	57.000	7.056		53.000	62.000
CEO_TENURE			11,065	11.043	8.000	9.065		4.000	15.000
CEO_GENDER			11,065	0.039	0.000	0.195		0.000	0.000
Descriptive Statistics Measured by Adoption of Golden Parachutes
		Adoption of Golden Parachutes			Non-Adoption of Golden Parachutes			Difference Tests: p-value
Variables		n	Mean	Median	n	Mean	Median	t-test	Wilcoxon Test
TOBINQ		8211	1.817	1.468	2854	1.912	1.523	<0.001	0.0789
CSR		8211	0.855	0.000	2854	0.615	0.000	<0.001	<0.001
SIZE		8211	8.162	8.089	2854	8.079	7.944	0.022	0.001
LEVERAGE		8211	0.569	0.568	2854	0.530	0.523	<0.001	<0.001
GROWTH		8211	0.061	0.046	2854	0.069	0.055	0.079	0.005
ROA		8211	0.087	0.078	2854	0.099	0.091	<0.001	<0.001
RISK		8211	0.031	0.020	2854	0.034	0.024	0.004	<0.001
R&D_INTENSITY		8211	0.031	0.000	2854	0.026	0.000	0.007	<0.001
GOVERNANCE		8211	−0.180	0.000	2854	−0.279	0.000	<0.001	<0.001
FIRM_AGE		8211	3.218	3.219	2854	3.177	3.219	0.002	0.002
CEO_AGE		8211	57.368	57.000	2854	57.836	57.000	0.005	0.051
CEO_TENURE		8211	10.231	8.000	2854	13.378	10.000	<0.001	<0.001
CEO_GENDER		8211	0.041	0.000	2854	0.034	0.000	0.071	0.084
Correlation Coefficients among Selective Variables (N = 11,065)
			1	2	3	4	5	6	7
1	TOBINQ		1.000
2	CSR		0.076 **	1.000
3	GP		−0.036 **	0.040 **	1.000
4	SIZE		−0.250 **	0.376 **	0.023 *	1.000
5	LEVERAGE		−0.182 **	0.121 **	0.073 **	0.515 **	1.000
6	GROWTH		0.161 **	−0.043 **	−0.017	−0.053 **	−0.079 **	1.000
7	ROA		0.550 **	0.058 **	−0.059 **	−0.149 **	−0.192 **	0.200 **	1.000
8	RISK		0.177 **	−0.067 **	−0.028 **	−0.240 **	−0.179 **	−0.011	0.035 **
9	R&D_INTENSITY		0.154 **	−0.007	0.022 *	−0.202 **	−0.186 **	0.042 **	−0.158 **
10	GOVERNANCE		0.024 *	0.151 **	0.064 **	−0.027 **	0.014	−0.013	−0.021 *
11	FIRM_AGE		−0.122 **	0.178 **	0.029 **	0.311 **	0.123 **	−0.136 **	−0.008
12	CEO_AGE		−0.078 **	−0.009	−0.029 **	0.080 **	0.037 **	−0.035 **	−0.023 *
13	CEO_TENURE		0.032 **	−0.100 **	−0.152 **	−0.127 **	−0.124 **	0.024*	0.006
14	CEO_GENDER		−0.003	0.086 **	0.016	0.014	0.033 **	−0.028 **	0.018
			8	9	10	11	12	13	14
8	RISK		1.000
9	R&D_INTENSITY		0.132 **	1.000
10	GOVERNANCE		−0.025 **	−0.017	1.000
11	FIRM_AGE		−0.149 **	−0.085 **	0.041 **	1.000
12	CEO_AGE		−0.020*	−0.054 **	0.016	0.140 **	1.000
13	CEO_TENURE		−0.017	0.003	0.037 **	0.030 **	0.499 **	1.000
14	CEO_GENDER		−0.01	−0.028 **	0.004	0.013	−0.052 **	−0.077 **	1.000

Pearson correlation coefficients are shown below the diagonal. * and ** represent statistical significance at the 5% and 1% levels, respectively, on a two-tailed test. Variables are defined in Appendix A. Variable definitions: TOBINQ = (market value of equity + total liabilities)/book value of total assets, at fiscal year-end in period; CSR = summation of scores for community, diversity, employee relations, environment, human rights, and product in KLD’s ratings; GP = a binary variable that equals 1 if the firms have adopted golden parachutes in that year and 0 otherwise; SIZE = natural logarithm of total assets at the end of the fiscal year; LEVERAGE = total liability divided by total assets at the end of the fiscal year; GROWTH = net sales minus net sales at prior year and divided by the net sales at the end of the fiscal year; ROA = operating income divided by total assets at the end of the fiscal year; RISK = standard deviation of EBIT (earnings before interests, taxes) from the past five years divided by total assets at the end of the fiscal year; R&D_INTENSITY = R&D expenses divided by net sales at the end of the fiscal year; GOVERNANCE = net score (strengths minus concerns) in KLD governance ratings; FIRM_AGE = natural logarithm of (1 + the number of years since the firm’s first appearance in the Compustat database); CEO_AGE = the age of the CEO; CEO_TENURE = how long the CEO has been serving in that capacity for the firm at the end of the fiscal year; CEO_GENDER = the gender of the CEO.

presents the univariate results of the firms in our sample. We begin by displaying a summary statistic for the sample, shown in panel A. Then, panel B offers a stylized overview of the sample by decomposing it into two groups: firm-year observation reflecting the adoption of a GP and firm-year observation with no adoption of a golden parachute. Panel C shows the correlation and significance of each variable.

In Table 2, Full Sample, we find that the mean of CSR’s variable is slightly positive (0.793), suggesting that the sample contains firms with more CSR strengths than CSR weaknesses. The mean of the GP variable (0.742) suggests that more than seventy-four percent of the firm-year observations have the GP clause for their executives.

In Table 2, Descriptive Statistics Measured by Adoption of Golden Parachutes, we offer a stylized summary and a univariate result comparing the differences between variables for firms that have adopted GPs and non-GP firms. We test the two categories using a parametric test and a non-parametric test. The evidence shows that the mean of TOBINQ is different for the GP and non-GP firms (1.817 vs. 1.912, p < 0.001). That is, firms without a GP have a slightly higher value than GP firms. Interestingly, the mean and median figures show that non-GP firms have a larger TOBINQ than GP firms. Given that we are simply comparing the mean and median, we should reserve any conclusions until we see the results of the regression analysis with control variables. The contrast in CSR and its breakdown shows a gleaning behavior for GP firms and non-GP firms. GP firms pursue CSR initiatives more than their counterparts.

In Table 2, Correlation Coefficients, the correlation matrix shows that all variables have a significant relationship with the explained proxy for firm value, TOBINQ, except the control variable, CEO gender, is not significant. As we expected, the firm value is positively correlated with CSR (0.076, p < 0.01), GROWTH (0.161, p < 0.01), ROA (0.550, p < 0.01), RISK (0.177, p < 0.01), R&D_INTENSITY (0.154, p < 0.01), GOVERNANCE (0.024, p < 0.05), and CEO_TENURE (0.032, p < 0.01), whereas the firm value is negatively correlated with GP (−0.036, p < 0.01), LEVERAGE (−0.182, p < 0.01), FIRM_AGE (−0.122, p < 0.01), and CEO_AGE (−0.078, p < 0.01). The negative correlation between GP and TOBINQ may allude to why some research findings suggest that the adoption of GP serves as an entrenchment clause [83,84]. Here, the correlation suggests that the GP’s adoption reduces a firm’s investment opportunities without considering the effects of control variables in the regression model. Moreover, we check whether the correlation coefficients are higher than 0.5, which may cause multicollinearity. We find the correlation coefficient between SIZE and LEVERAGE is 0.515. However, these two independent variables are necessary control variables shown in prior studies [1,64,65]. Further, we measure the variation inflation factor (VIF) in our regression and find no material multicollinearity issue in our regression models.

5.2. Regression Results

We display the regression results. We show the beta coefficient; thus, we can infer the key variables’ significant impact on the dependent variable. In

Table 3. Regressions of TOBINQ on CSR and GP (Golden Parachute).

						Adoption of GP	Non-Adoption of GP
		Dependent Variable: TOBINQ
		Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Independent variables	Predicted Sign	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.
		(t-stat)	(t-stat)	(t-stat)	(t-stat)	(t-stat)	(t-stat)
CSR	+	0.090 ***		0.089 ***	0.067 ***	0.108 ***	0.029 *
		(9.92)		(9.82)	(5.48)	(10.01)	(1.85)
GP	–		−0.044 ***	−0.043 ***	−0.047 ***
			(−5.58)	(−5.39)	(−5.73)
CSR × GP	+				0.027 ***
					(2.58)
SIZE	?	−0.141 ***	−0.099 ***	−0.144 ***	−0.143 ***	−0.178 ***	−0.064 **
		(−8.37)	(−6.61)	(−8.55)	(−8.52)	(−9.02)	(−2.49)
LEVERAGE	+	0.069 ***	0.068 ***	0.073 ***	0.072 ***	0.110 ***	−0.052 **
		(2.76)	(2.68)	(2.91)	(2.87)	(3.58)	(−2.29)
GROWTH	+	0.037 ***	0.032 ***	0.037 ***	0.037 ***	0.040 **	0.020
		(3.09)	(2.65)	(3.08)	(3.09)	(2.48)	(1.51)
ROA	+	0.530 ***	0.536 ***	0.529 ***	0.530 ***	0.508 ***	0.603 ***
		(12.50)	(12.55)	(12.49)	(12.48)	(9.76)	(16.81)
RISK	+	0.110 ***	0.113 ***	0.109 ***	0.109 ***	0.127 ***	0.041
		(4.31)	(4.34)	(4.27)	(4.27)	(4.21)	(1.16)
R&D_INTENSITY	+	0.172 ***	0.176 ***	0.172 ***	0.172 ***	0.166 ***	0.220 ***
		(4.20)	(4.23)	(4.22)	(4.23)	(3.72)	(4.28)
GOVERNANCE	+	−0.029 ***	−0.016 **	−0.029 ***	−0.030 ***	−0.032 ***	−0.031 **
		(−4.32)	(−2.38)	(−4.31)	(−4.39)	(−3.83)	(−2.56)
FIRM_AGE	–	−0.047 ***	−0.043 ***	−0.047 ***	−0.047 ***	−0.036 ***	−0.059 ***
		(−5.62)	(−5.08)	(−5.65)	(−5.64)	(−3.69)	(−3.41)
CEO_AGE	–	−0.067 ***	−0.066 ***	−0.066 ***	−0.066 ***	−0.052 ***	−0.084 ***
		(−6.95)	(−6.87)	(−6.85)	(−6.86)	(−5.04)	(−4.31)
CEO_TENURE	+	0.068 ***	0.056 ***	0.061 ***	0.061 ***	0.050 ***	0.056 ***
		(7.20)	(5.99)	(6.54)	(6.49)	(4.73)	(3.13)
CEO_GENDER	–	−0.020 ***	−0.013 **	−0.019 ***	−0.019 ***	−0.023 ***	0.003
		(−3.34)	(−2.25)	(−3.29)	(−3.24)	(−3.32)	(0.23)
Industry dummies		Included	Included	Included	Included	Included	Included
Year dummies		Included	Included	Included	Included	Included	Included
Mean of VIF		1.25	1.21	1.24	1.53	1.25	1.25
Adj. R2		0.47	0.47	0.48	0.48	0.46	0.55
F		71.97 ***	72.21 ***	71.13 ***	70.43 ***	52.73 ***	43.34 ***
n		11065	11065	11065	11065	8211	2854

*, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively, on a two-tailed test. All of the test statistics and significance levels are calculated based on robust standard errors adjusted for heteroskedasticity at both the firm and year levels.

, Model 1, we find that CSR is positively associated with TOBINQ (0.090, p < 0.01). Model 2 shows a negative relationship between GP and TOBINQ (−0.044, p < 0.01), which is consistent with prior studies [83]. GP is a dummy variable that takes the value of 1 if the firm’s executive has a GP and 0 if not. More importantly, in Model 4, we explore the interaction of GP with CSR (CSRXGP). The coefficient of the interaction variable (CSRXGP) is 0.027 (p < 0.01), implying that GP strengthens the positive relationship between CSR and firm value, TOBINQ. GP plays an important moderating role in the relationship between CSR and firm value. This result lends credence to our argument that the presence of a golden parachute alleviates the executive’s short-term concerns and incentivizes him or her to maximize the firm’s performance over the long term. Accordingly, this study supports our Hypothesis 1 that the GP amplifies the positive relationship between CSR and firm value.

Further, we divide the sample into two groups: adoption of GP and non-adoption of GP. By dividing our sample, we can measure the extent to which CSR activities affect the firm value in each different group to see the level of impact of the adoption of GP on the relationship between CSR and TOBINQ. In Model 5 and Model 6, we find that the coefficient and t value of the GP adoption group are much greater than those of the GP non-adoption group (0.108, t = 10.01 vs. 0.029, t = 1.85, respectively). This result supports the interaction effect of CSR and GP on the firm value.

In

Table 4. Regressions of TOBINQ on Individual Components of CSR and GP (Golden Parachute).

		Dependent Variable: TOBINQ
		Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8	Model 9	Model 10	Model 11	Model 12
Independent Variables	Predicted Sign	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.
		(t-stat)	(t-stat)	(t-stat)	(t-stat)	(t-stat)	(t-stat)	(t-stat)	(t-stat)	(t-stat)	(t-stat)	(t-stat)	(t-stat)
COMMUNITY	+	0.033 ***	0.019*
		(5.35)	(1.83)
COMMUNITYXGP	+		0.017 *
			(1.80)
DIVERSITY	+			0.047 ***	0.036 ***
				(5.49)	(3.14)
DIVERSITYXGP	+				0.014
					(1.40)
EMPLOYEE	+					0.062 ***	0.056 ***
						(6.84)	(4.32)
EMPLOYEEXGP	+						0.007
							(0.56)
ENVIRON	+							0.053 ***	0.062 ***
								(6.95)	(5.22)
ENVIRONXGP	+								−0.015
									(−1.47)
HUMAN_RIGHT	+									0.019 **	−0.004
										(1.96)	(−0.37)
HUMAN_RIGHTXGP	+										0.027 **
											(2.14)
PRODUCT	+											0.024 ***	−0.021 **
												(3.09)	(−2.07)
PRODUCTXGP	+												0.057 ***
													(5.23)
GP			−0.046 ***		−0.046 ***		−0.044 ***		−0.038 ***		−0.043 ***		−0.039 ***
			(−5.62)		(−5.72)		(−5.61)		(−4.45)		(−5.46)		(−4.73)
Control variables		Included	Included	Included	Included	Included	Included	Included	Included	Included	Included	Included	Included
Industry dummies		Included	Included	Included	Included	Included	Included	Included	Included	Included	Included	Included	Included
Year dummies		Included	Included	Included	Included	Included	Included	Included	Included	Included	Included	Included	Included
Adj. R2		0.47	0.47	0.47	0.47	0.47	0.47	0.47	0.47	0.47	0.47	0.47	0.47
F		72.42 ***	70.82 ***	72.41 ***	70.82 ***	72.22 ***	70.69 ***	75.34 ***	73.68 ***	72.45 ***	70.64 ***	72.42 ***	70.29 ***
n		11,065	11,065	11,065	11,065	11,065	11,065	11,065	11,065	11,065	11,065	11,065	11,065

** and *** indicate statistical significance at the 5% and 1% levels, respectively, on a two-tailed test. All of the test statistics and significance levels are calculated based on robust standard errors adjusted for heteroskedasticity at both the firm and year levels. Variable definition: COMMUNITY = a firm’s community activities such as charitable giving and innovative giving; DIVERSITY = a firm’s diversity activities such as CEO gender and the promotion of women and minorities: EMPLOYEE = a firm’s employee relations such as union relations and employee involvement; ENVIRON = a firm’s environmental activities such as pollution prevention and clean energy; HUMAN_RIGHT = a firm’s human rights activities; PRODUCT = a firm’s product quality activities. All variables are measured by strengths minus concerns. Other variables are defined in Appendix A.

, we report the results of the regression when six individual components of CSR are separately regressed on TOBINQ. Among the six interaction variables, the coefficients (p-value) of COMMUNITYXGP (Model 2), HUMAN_RIGHTXGP (Model 10), and PRODUCTXGP (Model 12) are positive and significant (0.017, p < 0.10; 0.027, p < 0.05; 0.057, p < 0.01, respectively). These results suggest that a firm’s GP adoption plays a more critical moderating role in the community, human rights promotion, and product quality activities enhancing firm value. This result confirms the robustness of our findings in Table 3.

5.3. Regression Results: Does Gender Play a Role?

Furthermore, in

Table 5. Regressions of TOBINQ on CSR and GP (Golden Parachute) in Female and Male Groups.

		Dependent Variable: TOBINQ
		Female CEO Group			Male CEO Group
		Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Independent Variables	Predicted Sign	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.
		(t-stat)	(t-stat)	(t-stat)	(t-stat)	(t-stat)	(t-stat)
CSR	+	0.105 ***		0.157 ***	0.089 ***		0.063 ***
		(3.33)		(2.79)	(9.61)		(5.11)
GP	–		−0.054 *	−0.006		−0.042 ***	−0.046 ***
			(−1.69)	(−0.14)		(−5.21)	(−5.50)
CSR × GP	+			−0.064			0.031 ***
				(−1.34)			(2.81)
SIZE	?	−0.111	−0.056	−0.126 *	−0.140 ***	−0.097 ***	−0.142 ***
		(−1.58)	(−0.92)	(−1.71)	(−8.07)	(−6.34)	(−8.21)
LEVERAGE	+	0.170 ***	0.180 ***	0.171 ***	0.065 **	0.063 **	0.068 ***
		(3.60)	(3.70)	(3.60)	(2.48)	(2.38)	(2.59)
GROWTH	+	0.075	0.064	0.072	0.037 ***	0.032 **	0.037 ***
		(1.41)	(1.30)	(1.34)	(2.99)	(2.56)	(2.99)
ROA	+	0.596 ***	0.598 ***	0.592 ***	0.526 ***	0.532 ***	0.525 ***
		(11.12)	(11.29)	(11.05)	(12.06)	(12.11)	(12.04)
RISK	+	0.145 **	0.139 **	0.143 **	0.109 ***	0.112 ***	0.108 ***
		(2.42)	(2.32)	(2.41)	(4.23)	(4.27)	(4.19)
R&D_INTENSITY	+	0.165 ***	0.171 ***	0.163 ***	0.171 ***	0.175 ***	0.171 ***
		(2.85)	(2.72)	(2.78)	(4.14)	(4.17)	(4.16)
GOVERNANCE	+	−0.006	0.016	−0.007	−0.031 ***	−0.018 ***	−0.032 ***
		(−0.24)	(0.62)	(−0.29)	(−4.46)	(−2.61)	(−4.54)
FIRM_AGE	–	−0.042	−0.048	−0.044	−0.048 ***	−0.044 ***	−0.048 ***
		(−1.09)	(−1.21)	(−1.13)	(−5.66)	(−5.10)	(−5.65)
CEO_AGE	–	−0.000	0.006	−0.004	−0.069 ***	−0.069 ***	−0.068 ***
		(−0.01)	(0.14)	(−0.09)	(−7.01)	(−6.94)	(−6.91)
CEO_TENURE	+	0.089 **	0.093 **	0.087 **	0.071 ***	0.059 ***	0.064 ***
		(2.40)	(2.39)	(2.34)	(7.37)	(6.18)	(6.70)
Industry dummies		Included	Included	Included	Included	Included	Included
Year dummies		Included	Included	Included	Included	Included	Included
Adj. R2		0.65	0.65	0.65	0.47	0.47	0.47
F		15.50 ***	15.13 ***	14.99 ***	71.10 ***	71.30 ***	69.32 ***
n		436	436	436	10629	10629	10629

*, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively, on a two-tailed test. All of the test statistics and significance levels are calculated based on robust standard errors adjusted for heteroskedasticity at year levels.

, we divide the data into female and male groups to test Hypothesis 2: The behavior of female corporate executives in US-based firms towards CSR will change if they are protected by a GP. The results show that for firms with female CEOs who are protected by GPs, the relationship between CSR and firm value is negative but not statistically significant (−0.064, t = −1.34). However, for the male CEO group, the strong positive moderating power of the GP is sustained (0.031, p < 0.01), supporting Hypothesis 2. Related theories could explain this contrasting finding on why women in male-dominated fields may feel pressure to adopt traditionally masculine behaviors, especially on their rise to the top of the corporate ladder.

Self-verification theory and social role theory both suggest that women in male-dominated fields may feel pressure to adopt traditionally masculine behaviors because of societal expectations, the threat of being judged based on gender stereotypes, and a desire to maintain their sense of self. Stereotype theory can lead to underperformance in tasks that are related to the stereotype, even when the individual is otherwise capable of performing well. The effects of stereotype threat can be significant and can contribute to persistent disparities in achievement and performance between different social groups.

In

Table 6. Regressions of TOBINQ on CSR and GP in Full, Female, and Male Group Propensity Score Matching Method.

		Dependent Variable: TOBINQ
		Full Sample	Female Group	Male Group
		Model 1	Model 2	Model 3
Independent Variables	Predicted Sign	Coeff.	Coeff.	Coeff.
		(t-stat)	(t-stat)	(t-stat)
CSR	+	0.028	0.157 ***	0.001
		(1.13)	(2.79)	(0.04)
GP	–	−0.083 ***	−0.006	−0.079 ***
		(−4.10)	(−0.14)	(−3.52)
CSR × GP	+	0.043 *	−0.064	0.058 **
		(1.81)	(−1.34)	(2.05)
CEO_GENDER	–	−0.034 **
		(−2.33)
SIZE	?	−0.099 ***	−0.126 *	−0.087 ***
		(−3.78)	(−1.71)	(−3.15)
LEVERAGE	+	0.069 ***	0.171 ***	0.047*
		(2.87)	(3.60)	(1.71)
GROWTH	+	0.020	0.072	0.020
		(0.98)	(1.34)	(0.88)
ROA	+	0.652 ***	0.592 ***	0.662 ***
		(21.85)	(11.05)	(18.89)
RISK	+	−0.000	0.143 **	−0.020
		(−0.01)	(2.41)	(−0.53)
R&D_INTENSITY	+	0.161 ***	0.163 ***	0.169 ***
		(5.37)	(2.78)	(5.94)
GOVERNANCE	+	−0.013	−0.007	−0.023
		(−0.96)	(−0.29)	(−1.56)
FIRM_AGE	–	−0.046 **	−0.044	−0.047 **
		(−2.32)	(−1.13)	(−2.05)
CEO_AGE	–	−0.069 ***	−0.004	−0.082 ***
		(−2.96)	(−0.09)	(−3.11)
CEO_TENURE	+	0.067 ***	0.087 **	0.081 ***
		(3.26)	(2.34)	(3.52)
Industry dummies		Included	Included	Included
Year dummies		Included	Included	Included
Adj. R2		0.59	0.65	0.59
F		34.79 ***	14.99 ***	28.95 ***
n		2035	436	1599

*, **, and *** indicate statistical significance at the 10%, 5%, and 1% levels, respectively, on a two-tailed test. All of the test statistics and significance levels are calculated based on robust standard errors adjusted for heteroskedasticity at year levels.

, we employ the propensity-score-matching model to mitigate the endogeneity issue. The propensity-score-matched samples seem effective in obtaining a balanced set of samples [85]. We impose a caliper distance of 0.001 using SIZE and ROA variables to calculate the propensity scores to obtain a sample of 1599 for the male CEO group out of 10629 firm-year observations to match with a sample of 436 for the female CEO group. In Table 6, Model 1, we show the result using the full sample of 2035 after applying the PSM methodology. The coefficient of CSR × GP is statistically significant (0.043, p < 0.10), supporting Hypothesis 2. We also report in Models 2 and 3 that for firms with female executives protected by the GP, the relation between CSR and firm value is negative but not statistically significant (−0.064, t=−1.34). However, for the male executives’ group, the strong positive moderating power of the GP is still maintained (0.058, p < 0.05), supporting Hypothesis 2.

5.4. Endogeneity Tests

In the CSR–firm value study, we need to consider the endogeneity problem to explain cause-and-effect relationships because of reverse causality issues, omitted variables, and measurement errors. For example, we may expect highly profitable firms to allocate more capital and other resources toward CSR activities. It could also be that proactive CSR firms are more successful in maximizing firm value. Therefore, it is unclear if CSR is the cause or the effect. To address this endogeneity issue, we conduct a two-stage least-square regression (2SLS) [3,4,86,87] using an instrumental variable (IV) for CSR [88,89,90].

The proper IV needs to be highly correlated with CSR but less correlated to the firm value variable (TOBINQ). We use the two-digit SIC industry average CSR score as IV following a prior study [88]. The industry average CSR score can be positively correlated with a firm’s CSR score yet does not influence its value. The minimum eigenvalue statistic or F statistic is 1735.65, which is much higher than 10, implying that our IV seems to be a proper instrument. The correlation coefficient between CSR and IV is 0.404 (p < 0.01), but the correlation coefficient between TOBINQ and IV is 0.102 (p < 0.01). Thus, this correlation supports the validity of our IV.

Table 7. 2 SLS Regressions of TOBINQ on CSR Using Instrument Variable.

Instrument Variable: CSR_INDUSTRY_AVG
	Model 1	Model 2
Independent variables	Coeff.	Coeff.
	(t-stat)	(t-stat)
	First Stage	Second Stage
	Dep. variable	Dep. variable
	CSR	TOBINQ
CSR_INDUSTRY_AVG	0.365 ***
	(33.31)
CSR_HAT (instrumented)		0.075 ***
		(4.81)
SIZE	0.411 ***	−0.164 ***
	(33.99)	(−7.68)
LEVERAGE	−0.087 ***	0.059 ***
	(−8.50)	(2.62)
GROWTH	−0.055 ***	0.041 ***
	(−6.25)	(3.25)
ROA	0.114 ***	0.551 ***
	(10.45)	(13.39)
RISK	0.033 ***	0.110 ***
	(4.75)	(4.66)
R&D_INTENSITY	0.055 ***	0.196 ***
	(3.54)	(4.32)
GOVERNANCE	0.118 ***	−0.039 ***
	(10.16)	(−4.90)
FIRM_AGE	0.050 ***	−0.061 ***
	(6.07)	(−7.45)
CEO_AGE	−0.021 **	−0.073 ***
	(−2.43)	(−7.39)
CEO_TENURE	−0.046 ***	0.065 ***
	(−5.35)	(7.20)
CEO_GENDER	0.068 ***	−0.019 ***
	(7.80)	(−3.02)
Year dummies	Included	Included
Mean of VIF	1.92	2.15
Adj. R2	0.32	0.43
F	189.93 ***	80.24 ***
n	11065	11065

** indicate statistical significance at the 5% levels, *** indicate statistical significance at the 1% levels, respectively, on a two-tailed test. All of the test statistics and significance levels are calculated based on robust standard errors adjusted for heteroskedasticity at year levels. Variable definition: CSR_INDUSTRY_AVG = an industry average score of CSR score based on the same SIC two-digit code. CSR_HAT (instrumented) = a predicted value of CSR score, which is a dependent variable using independent variables including an instrument variable, CSR_INDUSTRY_AVG. Other variables are shown in Appendix A.

shows the results of the 2SLS regression. In the first stage, in Model 1, we predict the value of CSR using IV of CSR_INDUSTRY_AVG and label that predicted value of CSR as CSR_HAT (instrumented). Then, in the second stage of Model 2, we use CSR_HAT to explore the relationship between CSR_HAT and firm value. We find that the coefficients of CSR_INDUSTRY_AVG and CSR_HAT variables are significantly positive (0.365, p < 0.01; 0.075, p < 0.01, respectively).

5.5. Alternative Measurement: CSR Dummy Variable

Table 8. Regressions of TOBINQ on CSR_DUMMY and GP (Golden Parachute).

						Adoption of GP	Non-Adoption of GP
		Dependent Variable: TOBINQ
		Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Independent variables	Predicted Sign	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.	Coeff.
		(t-stat)	(t-stat)	(t-stat)	(t-stat)	(t-stat)	(t-stat)
CSR_DUMMY	+	0.077 ***		0.078 ***	0.055 ***	0.095 ***	0.022
		(9.00)		(9.16)	(3.74)	(9.35)	(1.45)
GP	–		−0.044 ***	−0.046 ***	−0.057 ***
			(−5.58)	(−5.83)	(−5.65)
CSR_DUMMY × GP	+				0.029 *
					(1.89)
SIZE	?	−0.128 ***	−0.099 ***	−0.132 ***	−0.131 ***	−0.165 ***	−0.058 **
		(−8.00)	(−6.61)	(−8.25)	(−8.19)	(−8.69)	(−2.38)
LEVERAGE	+	0.066 ***	0.068 ***	0.070 ***	0.070 ***	0.108 ***	−0.054 **
		(2.63)	(2.68)	(2.80)	(2.78)	(3.54)	(−2.41)
GROWTH	+	0.036 ***	0.032 ***	0.036 ***	0.036 ***	0.039 **	0.020
		(3.01)	(2.65)	(3.00)	(3.01)	(2.41)	(1.47)
ROA	+	0.532 ***	0.536 ***	0.531 ***	0.531 ***	0.509 ***	0.604 ***
		(12.50)	(12.55)	(12.49)	(12.49)	(9.75)	(16.91)
RISK	+	0.111 ***	0.113 ***	0.110 ***	0.110 ***	0.127 ***	0.042
		(4.34)	(4.34)	(4.30)	(4.29)	(4.21)	(1.18)
R&D_INTENSITY	+	0.173 ***	0.176 ***	0.173 ***	0.173 ***	0.167 ***	0.220 ***
		(4.22)	(4.23)	(4.24)	(4.24)	(3.74)	(4.27)
GOVERNANCE	+	−0.022 ***	−0.016 **	−0.022 ***	−0.022 ***	−0.022 ***	−0.028 **
		(−3.27)	(−2.38)	(−3.30)	(−3.36)	(−2.73)	(−2.38)
FIRM_AGE	–	−0.044 ***	−0.043 ***	−0.045 ***	−0.045 ***	−0.032 ***	−0.058 ***
		(−5.31)	(−5.08)	(−5.34)	(−5.32)	(−3.34)	(−3.38)
CEO_AGE	–	−0.068 ***	−0.066 ***	−0.066 ***	−0.066 ***	−0.053 ***	−0.083 ***
		(−6.98)	(−6.87)	(−6.88)	(−6.88)	(−5.12)	(−4.27)
CEO_TENURE	+	0.066 ***	0.056 ***	0.059 ***	0.058 ***	0.046 ***	0.055 ***
		(6.99)	(5.99)	(6.28)	(6.21)	(4.44)	(3.07)
CEO_GENDER	+	−0.020 ***	−0.013 **	−0.019 ***	−0.019 ***	−0.024 ***	0.004
		(−3.34)	(−2.25)	(−3.32)	(−3.28)	(−3.44)	(0.30)
Industry dummies		Included	Included	Included	Included	Included	Included
Year dummies		Included	Included	Included	Included	Included	Included
Adj. R2		0.47	0.47	0.47	0.47	0.46	0.55
F		71.23 ***	72.21 ***	70.34 ***	69.61 ***	53.11 ***	43.32 ***
n		11065	11065	11065	11065	8211	2854

**, and *** indicate statistical significance at the 5%, and 1% levels, respectively, on a two-tailed test. All of the test statistics and significance levels are calculated based on robust standard errors adjusted for heteroskedasticity at year levels. Variable definition: CSR_DUMMY = an indicator variable that takes a value of 1 if the CSR score is greater than the median number of CSR score, and 0 otherwise. CSR_DUMMY × GP = an interaction variable measured as CSR_DUMMY times GP. Other variables are shown in Appendix A.

shows the results of the regression analyses using the CSR_DUMMY variable to test if the results of Table 3 are supported when we use a CSR score dummy variable instead of CSR scores itself. CSR_DUMMY variable takes the value of 1 if a firm’s CSR score is greater than the median number of CSR scores and 0 otherwise. In Model 1, we report that the coefficient (0.077, p < 0.01) of CSR_DUMMY is positively and significantly related to TOBINQ as expected. Moreover, the coefficient of the interaction variable, CSR_DUMMY × GP, is positively and significantly associated with TOBINQ (0.029, p = 0.059), implying that GP plays a positive moderating role in the positive relationship between CSR and TOBINQ. Further, we examine whether the adoption of GP can affect the relationship between CSR_DUMMY and TOBINQ. To perform this test, we divided the sample into two groups: adoption of GP and non-adoption of GP. In Models 5 and 6, we report that the positive relationship between CSR_DUMMY and TOBINQ in the firm group of adopting GP is much greater than that in the group of not adopting GP (0.095, p < 0.01 versus 0.022, t = 1.45 (not significant)). Thus, we conclude that the results of Table 3 are fully supported by the results of Table 8.

6. Discussion

Does the GP moderate the association between CSR and firm value for US-based firms? Does the gender of the executive make any difference in these associations? This study empirically examines these questions. Our findings shed much-needed light and contribute to the CSR–firm value debate, which has produced mixed results [1,4,5,7,8,9].

The results of our findings demonstrate the important role that gender plays in the GP adoption policy when assessing a US-based firm’s CSR investment. This study presents novel findings that contribute to CSR literature; specifically, these findings are related to the alignment of CEOs’ interest (GP) and the society’s campaign for CSR in the US-based environment. We show how firms’ long-term value, CEOs’ interests, and CSR practices can be optimally achieved. A CEO with a GP will pursue and engage in value-enhancing CSR that benefits both shareholders and the society from which they extract profit. We also provide empirical evidence suggesting that in contrast to the findings for the general population, women executives endowed with a GP do not perform well on the CSR–firm value-adding assessment spectrum. The implication of this suggests that corporate CSR strategies should be adjusted based on the CEO’s gender. Society benefits from CSR.

Our findings on the CSR–firm value associations support the conflict resolution and resources-based view theories. The resource base view and conflict resolution theories of CSR highlight the importance of CSR activities as a strategic tool for firms to create and sustain competitive advantages. From these theories’ perspectives, CSR activities can be seen as a way to develop and leverage strategic resources and capabilities. These theories emphasize the importance of businesses as key actors in society and call for a collaborative approach to addressing social and environmental issues. For example, CSR activities such as environmental sustainability programs or philanthropic initiatives can help firms build positive relationships with stakeholders, enhance their reputation and brand image, and foster innovation and learning within the organization. Businesses can build stronger relationships and create shared value for themselves and their communities by engaging with their stakeholders and working together to resolve conflicts [20,22,91,92].

Investors and society largely frown upon firms issuing golden parachutes to CEOs. The author of [93] finds that Fortune 500 firms that adopted a golden parachute in the 1980s recorded negative stock returns. The authors of [83] also recorded negative abnormal stock returns due to firms adopting a golden parachute. Thus, investors and society seem to frown upon firms adopting a golden parachute. Nonetheless, US-based firms issue golden parachutes to attract the best talent for managerial positions and stay competitive. Despite the seemingly ever-present controversies surrounding golden parachutes, firms’ issuance of golden parachutes is a practice that is here to stay [18]. Therefore, finding ways for these interests to coexist benefits the firm, the CEO, and society. Our research presents strategic alliances that benefit the interests of these groups: society, firms, and corporate executives.

Stereotype threat theory describes how individuals’ performance in a task can be adversely affected by the awareness or anticipation of negative stereotypes about their group. Female executives may face a situation where their behavior or performance is likely to be evaluated based on a negative stereotype about their gender. Being consciously aware of these negative stereotypes, female executives may experience anxiety and self-doubt, leading to underperformance in the task at hand (e.g., strategically and carefully identifying value-enhancing CSR projects).

Firms should strive to create a more supportive, inclusive, and equitable environment. Measures should be put in place that actively facilitate a corporate culture that helps female employers in leadership roles to reframe their thinking about these stereotypes and their own abilities. Interventions such as affirming positive aspects of one’s identity and providing role models from similar backgrounds can mitigate the effects of stereotype threat on female executives and improve their CSR–firm value performance.

7. Conclusions

This study examines the association between CSR and firm value. More notably, we investigate whether GP moderates the relationship between CSR and firm value. We also test whether gender plays a role in these observed associations.

First, our findings show that CSR is significantly and positively associated with firm value. Second, more interestingly, we find that the GP amplifies the positive association between CSR and firm value. We further tested to examine if gender plays a role in this observed association. Third, consistent with the general population’s results, male CEOs with a golden parachute engage in more value-enhancing CSR projects than their counterparts without a golden parachute. In contrast with the results of the general population, women executives endowed with a GP do not perform well on the CSR–firm value assessment spectrum. The implication of this suggests that corporate CSR strategies should be adjusted based on the CEO’s gender. Our results support both of our hypotheses.

As discussed above, in conclusion, when strategically implemented, for US-based firms, CSR could benefit society, firms, and corporate executives, thereby aligning the interests of these groups. This sheds propitious light and contributes to the CSR–firm value debate in related literature. Examining the moderating role of the GP is critical and peculiar to our study. The GP is here to stay, and US-based firms will continue to issue GPs to their executives [39]. So, findings ways where corporate practices critical to firms’ success coexist and align with society’s interests should be considered a welcome development.

Another important aspect of the findings in our study is the role that gender plays in the moderating impact of the GP on the CSR–firm value association. Several studies have demonstrated that gender diversity can significantly boost organizational innovation and firm performance [46,94,95,96]. However, firms must strategically promote an innovative culture that fuels success and growth by actively supporting gender diversity and developing inclusive workplaces where employees feel empowered and equally valued. Like their male counterparts, women should be equally acknowledged or given credit for their creative suggestions, encouraging them to always make decisions and contributions that maximize the firm’s value.

Although we conducted extensive regression analyses and robustness tests to validate and strengthen our research results, this study has some limitations for which future studies may compensate. First, we used variables, including CSR and GP, that may be subject to measurement errors. Second, additional alternative and more extensive robustness tests could be employed to overcome the endogeneity issue in our regression analyses. One should be cautious in interpreting causal inferences in regression analyses. Third, our data are US-specific. Future research using non-US data could reveal even more interesting insights.

Despite this limitation, our study contributes two major practical implications. First, from the lens of social implications, the findings of this study show that society’s interests and firms’ long-term goals are achievable. As the head and driver of a firm’s investment corporate policies, the CEO can pursue and engage in value-enhancing CSR that benefits both shareholders and the society from which they extract profit. Second, when designing corporate CSR strategies that align with shareholders’ value, calculated adjustments based on gender tendencies should be made to better focus the CEO’s incentive on CSR performance.