Analyzing Corporate Social Responsibility, CEO Gender, and Compensation Structure: Evidence from U.S. Firms

Abstract

This article examines how CEO compensation structure and CEO gender were associated with corporate social responsibility (CSR) performance in U.S. firms in the period between 2003 and 2013. Building on prior research in economics, finance, accounting, and management, which suggests gender differences in commitment to CSR, this study provides empirical evidence that female CEOs were positively associated with higher CSR performance. The analysis further shows that a higher proportion of equity in CEO compensation was positively associated with CSR, whereas higher proportions of cash bonuses and long-term incentive plans were negatively associated with CSR. Notably, a higher proportion of a cash bonus in CEO compensation further reduced CSR in firms led by female CEOs. These findings offer valuable insights for firms seeking to design executive compensation packages that align CEO behavior with the firms’ CSR objectives. This study contributes to the growing body of literature on CSR by providing empirical evidence on the role of CEO gender and compensation structure.

1. Introduction

Corporate social responsibility (CSR) includes the economic and social engagement by firms that impacts the stakeholders’ quality of life through actions that improve community relations, protect the environment, and seek social progress (Hill et al., 2007). The emphasis on CSR is increasing in both the corporate world and academic literature (Kharabsheh et al., 2023). Stakeholders pursue higher standards of environmental, social, and governance (ESG) practices from firms and promote greater disclosure of CSR initiatives (Ferrell et al., 2016).

Academic debates on CSR center on whether it follows from well-governed management strategy that is in the interests of the firm (Deng et al., 2013; Ferrell et al., 2016; Albuquerque et al., 2019), or whether it reflects managers pursuing their own interests at the expense of the firms’ shareholders (Jensen & Meckling, 1976; Bénabou & Tirole, 2010; Masulis & Reza, 2015; W. Li et al., 2024; Harjoto et al., 2024). CSR has been shown to affect firms’ chief executive officer (CEO) compensation (Cai et al., 2011; Jian & Lee, 2015; Karim et al., 2018; Sarhan & Al-Najjar, 2023) and an increasing number of businesses include CSR measures in CEO pay (Qin & Yang, 2022; Liu et al., 2023).

In this article, we explore how a CEO’s compensation structure and gender are associated with measures of the firm’s CSR performance. We construct a data set that includes information on components of CEO compensation, firms’ financial variables, and CSR measures for over 2650 U.S. publicly traded firms covering the period from 2003 to 2013. Our empirical results confirm that a higher proportion of equity in a CEO’s total compensation was positively associated with CSR, while a higher proportion of performance-based incentive compensation including cash bonuses and long-term incentive plans (LTIPs) had a negative association with CSR.

We further investigate how CEO gender influences CSR. Our results confirm that firms with female CEOs had higher measures of CSR. This link was enhanced by CEO compensation. The basic relationships identified for CEO compensation components of equity, bonus, and LTIPs survive the addition of CEO gender controls to the empirical models. Furthermore, interaction variables indicate that female CEOs were more likely to reduce CSR initiatives when a higher proportion of their compensation came as a cash bonus.

Our analysis extends the literature in several aspects. First, we identify significant differences in the way various CEO compensation components affect CSR. We distinguish between salary, equity, and two forms of performance-based incentive compensation—cash bonuses which involve single year targets, and long-term incentive plans (LTIPs). We find no significant link between the proportion of salary and CSR in our data set. Meanwhile, we observe a positive relationship between the proportion of equity compensation and CSR, and a negative relationship between CSR and the proportion of both cash bonuses and LTIPs in total CEO compensation. Second, we confirm that female CEOs were associated with greater CSR performance, controlling for the differences in CEO compensation types. Third, our results suggest that boards of directors may strategically use CEO compensation contracts to align CEO behavior with the firms’ CSR goals. Specifically, a higher proportion of performance-based incentive pay, such as cash bonuses and long-term incentive plans, reduced CSR in our sample, while a higher proportion of equity compensation was associated with higher CSR. The results suggest that because performance-based plans typically require CEOs to meet certain accounting and financial metrics, the CEOs may have less incentive to invest in the firm’s social responsibilities. On the other hand, non-performance-based equity plans allow CEOs to seek diverse ways to enhance firm value and provide motivation to engage in CSR activities. This impact was stronger for female CEOs. In contrast, female CEOs with a higher share of LTIPs were less likely to be associated with lower CSR. Lastly, although LTIPs have become widespread in executive compensation packages, their impact is understudied in the literature (Z. Li & Wang, 2016). Illustrating the impact of using LTIPs on the firms’ CSR is another contribution of this study.

The article proceeds in the following order. Section 2 presents a review of the literature and frames the hypotheses for empirical testing. Section 3 describes the data sample and empirical approach. Section 4 outlines the results of the empirical analysis. Section 5 presents the discussion and conclusion.

2. Background and Hypotheses

There are two alternative views of CSR in academic literature—the stakeholder value maximization view and the agency theory view (Deng et al., 2013). The former view suggests that CSR activities positively impact stakeholders and therefore enhance their support of a firm, which contributes to shareholder wealth (Deng et al., 2013; Albuquerque et al., 2019). Alternatively, the agency view presents CSR activities as an expense that serves the interests of stakeholders instead of shareholders (Jensen & Meckling, 1976; Bénabou & Tirole, 2010; Masulis & Reza, 2015; Harjoto et al., 2024; W. Li et al., 2024).

Prior studies provide mixed evidence on the motivation of CEOs to invest in CSR activities. This motivation is linked to the role of CSR at a firm and has implications for CEO compensation. The agency view of CSR suggests that managers overinvest in CSR activities to promote their own reputation or disguise wrongdoing at the expense of shareholders (Barnea & Rubin, 2010; Hemingway & Maclagan, 2004). When CEOs overinvest in CSR to enhance their personal reputations, this could lead to higher compensation due to increased bargaining power (Cai et al., 2011; Ikram et al., 2020). On the other hand, the stakeholder value maximization view sees CSR activities as a tool to mitigate conflicts among stakeholders, improve stakeholder relationships, and promote support for the firm’s operations (Deng et al., 2013). Following this view, CSR activities are linked with better relationships between CEOs and stakeholders inside and outside the firm leading to conflict resolution (Karim et al., 2018). In the conflict resolution paradigm, firms engaging in CSR may pay their CEOs less, as these CEOs accept lower compensation to mitigate potential conflicts and promote fairness (Cai et al., 2011).

A typical CEO’s compensation package is composed of a base cash salary, equity compensation that includes stock and options, and performance-based incentive plans (Edmans et al., 2017). Firms, through their boards of directors, strive to craft executive compensation structure so that they align the executives’ behaviors with shareholder interests and thus improve firm performance (Edmans et al., 2017; J. Kim et al., 2013). The performance-based incentive plans include annual cash bonuses and long-term incentive plans (LTIP) and condition the payouts on achieving specific targets (Mishra et al., 2000). LTIPs, also known as multi-year accounting-based performance (MAP) plans, have seen a significant growth in executive compensation after 2000 (Z. Li & Wang, 2016).

Empirical evidence on the relationship between different types of CEO compensation and CSR is mixed and suggests that this relationship differs significantly among the compensation components. Cai et al. (2011) report that CSR adversely affects both the total compensation and cash compensation, controlling for firm and board parameters. Karim et al. (2018) model compensation as a function of CSR and report a negative relationship between a cash bonus and CSR, supporting the conflict resolution view. In contrast, Kharabsheh et al. (2022), in their analysis of a limited sample of Jordanian firms, report a positive relationship between cash compensation and CSR, with the effect being more pronounced under a weaker corporate governance. This finding supports the overinvestment view of CSR, as posited by agency theory. LTIPs have a longer time horizon of three to five years and may be linked to non-financial performance indicators, opening the potential for variation in their effect on CSR (Mishra et al., 2000; Mahoney & Thorne, 2005; Frye et al., 2006).

Prior studies also indicate that incentive-based compensation tends to focus CEOs on the specified short-term financial performance, which reduces their attention to CSR. However, with equity plans, such as restricted stock with longer vesting periods, CEOs have freedom to explore various ways to enhance the firm’s value and are more likely to invest in CSR (McGuire et al., 2003, 2019). Fixed compensation in the form of salary does not have a significant association with CSR. However, variable compensation tied to non-financial performance metrics, such as environmental or social targets, is positively associated with CSR (Mahoney & Thorne, 2005; Frye et al., 2006; Flammer et al., 2019).

The performance-based incentive components, such as cash bonuses and LTIPs, are paid out contingent upon achieving certain financial and non-financial goals each year. CEOs who have a higher proportion of cash bonus to total compensation tend to focus on achieving single-year targets, typically accounting metrics, and put a lower priority on CSR. LTIPs award cash or options to executives who meet certain performance targets over a three-, four-, or five-year period (Bebchuk et al., 2002). The performance targets can include share price increase, earnings per share, net asset value, or return on capital. While the intention of LTIPs is to align compensation to long-term performance, Bebchuk and Fried (2004) warn that poorly designed LTIPs may reward executives who myopically focus on achieving specific accounting performance metrics or who have luck with a strong stock market, potentially reducing sustainability or stakeholder welfare. This provides the following hypotheses H1 and H2 for empirical testing.

H1: 

The proportion of bonus to total CEO compensation is negatively associated with firms’ CSR.

H2: 

The proportion of LTIPs to total CEO compensation is negatively associated with firms’ CSR.

Equity compensation typically refers to time-vested plans that do not require the achievement of specific performance targets. Restricted stock grants are the most prevalent form. The primary purpose of equity compensation is to address agency problems by aligning the interests of the CEO with those of the shareholders (Edmans et al., 2017; H. Kim et al., 2023). In a study of executive compensation structure, Karim et al. (2018) find that equity-based CEO compensation is positively associated with CSR. We hypothesize that equity compensation motivates CEOs to invest in CSR as this may enhance the firm’s long-term performance and value.

H3: 

The proportion of equity to total CEO compensation is positively associated with firms’ CSR.

Research on the link between CEO gender and CSR suggests that female CEOs are associated with higher levels of CSR. Several studies attribute this trend to factors such as different risk preferences, ethical decision-making, and stakeholder orientation observed in female executives (Bear et al., 2010; Nicolò et al., 2021; Nekhili et al., 2017; J. Kim & Wang, 2023; Liu et al., 2023). Female CEOs diverge from their male counterparts in leadership styles in several ways. They are more likely to promote sustainability (Glass & Cook, 2018), prioritize CSR and ethical practices (Post et al., 2011), and invest in social capital (Gul et al., 2013). This literature provides the following hypothesis H4.

H4: 

Female CEOs are positively associated with their firms’ CSR.

3. Materials and Methods

To analyze the link between CSR and CEO compensation as well as CEO gender, we merge data on U.S. publicly traded firms from four databases—ExecuComp data focusing on top executives, Compustat financial data, CRSP (Center for Research in Security Prices) for stock-level variables, and the MSCI ESG Kinder, Lydenberg, and Domini (KLD) database for measures of CSR. Our methodology in creating the data set for empirical analysis follows such studies as Deng et al. (2013), Chulkov (2024), as well as Chulkov and Barron (2024). We employ Stata statistical software (version 18) for the data set creation and empirical analysis.

The KLD database has been widely used in studies focusing on CSR (Deng et al., 2013; Chen et al., 2020; Ho et al., 2022; Chulkov & Wang, 2023). KLD computes the ESG ratings by integrating multiple metrics and sources such as voluntarily disclosed corporate ESG information, corporate governance disclosures, regulatory agency data, and company characteristics. A CSR measure for each firm adds one strength (concern) point for each socially good (bad) action in each dimension of CSR. Following Lins et al. (2017), this study includes five dimensions of KLD scores: community, diversity, employee relations, environment, and product quality.

Our basic CSR index is constructed by taking the difference of total strength and concern scores across these five dimensions of CSR. This study further adopts the method used by Deng et al. (2013) to construct an adjusted CSR index because the basic CSR index may be biased due to the inconsistent numbers of strength and concern measures in different years. Following Deng et al. (2013) as well as Chulkov and Wang (2023), we compute weighted CSR strength and concern scores by dividing the individual CSR measures of each firm by the total number of CSR categories identified for each firm in each year. The weighted strength and concerns scores are then totaled by firm to produce the adjusted CSR index that becomes the dependent variable in our analysis.

The KLD database has had several structural changes over the years. It became available starting from 1991 for the MSCI KLD 400 Social Index companies, then in 2003 it was significantly expanded to incorporate the top 3000 U.S. companies by capitalization, and in 2013 the data changed again as it expanded to global companies. We focus our data set construction on the period from 2003 to 2013 to avoid structural issues associated with changes in the KLD data set. This sample aligns with the samples used by multiple studies on CSR including Deng et al. (2013), Karim et al. (2018), Ikram et al. (2020), Chulkov and Wang (2023, 2024), and Derchi et al. (2021).

We merge CEO compensation and demographic data, firm-level financial data, and CSR data to create our final sample, which is composed of 2,659 firms and 14,046 firm-year observations for the period between 2003 and 2013. Descriptive statistics for the data sample are presented in

Table 1. Summary statistics for the data sample.

Variable	Mean	Median	Standard Deviation	Min	Max
CSR index	−0.166	−0.2	0.452	−2.726	3.026
Female CEO	0.032	0	0.176	0	1
Ratio of bonus to total compensation	0.073	0	0.142	0	0.651
Ratio of LTIPs to total compensation	0.164	0.133	0.178	0	0.718
Ratio of equity to total compensation	0.434	0.487	0.272	0	0.926
Ratio of salary to total compensation	0.267	0.199	0.212	0.01	0.988
Log of CEO tenure	2.012	2.079	0.668	0	3.258
Ratio of market to book value of assets	1.272	0.928	1.184	0.042	6.464
Log of firm’s assets	8.052	7.946	1.696	4.553	12.704
Return on assets (ROA)	0.037	0.04	0.090	−0.401	0.261
Leverage	0.011	0.007	0.020	−0.051	0.134
Current ratio	2.022	1.612	1.461	0.389	8.608

. In order to minimize the influence of outliers, we Winsorize continuous variables at the 1 percent and 99 percent levels.

The dependent variable for our empirical models is the adjusted CSR index described above. The independent variables include proportions of CEO compensation components in total compensation, as well as CEO gender. Control variables include firm size in terms of assets, return on assets, market-to-book value ratio, leverage, and current ratio. We classify the firms into industry sectors based on the four-digit SIC code. The industry classification is used to control fixed effects for industry in all our empirical models. In addition, we include fiscal year controls and cluster the standard errors at the industry level in all model specifications. Appendix A 

Table A1. Variable Definitions.

Variable	Definition	Data Source
	Dependent Variable
CSR	Adjusted CSR index following Deng et al. (2013).	MSCI ESG KLD
	Independent Variables
Female CEO	Dummy variable equal to 1 if the CEO is female.	ExecuComp
Ratio of bonus to compensation	Ratio of cash bonus to total compensation for the CEO in a given fiscal year.	ExecuComp
Ratio of LTIP to compensation	Ratio of long-term incentive plans to total compensation for the CEO in a given fiscal year.	ExecuComp
Ratio of equity to compensation	Ratio of equity pay including stock and options to total compensation for the CEO in a given fiscal year.	ExecuComp
Ratio of salary to compensation	Ratio of salary to total compensation for the CEO in a given fiscal year.	ExecuComp
	Control Variables
Log of CEO tenure	The natural logarithm of the number of years in which CEOs stay in their position.	ExecuComp
Market-to-book value of assets	The ratio of the book value of firm’s total assets to the market value of assets.	Compustat
Log of firm’s assets	The natural logarithm of the firm’s total assets.	Compustat
Return on assets (ROA)	Net income scaled by total assets.	Compustat
Current ratio	Current assets divided by current liabilities.	Compustat
Leverage	Total debt scaled by total assets	Compustat

provides the definitions of the variables and describes the data sources.

4. Results

We start our empirical analysis by examining how CSR is associated with CEO compensation components. We estimate models designed to test hypotheses H1, H2, and H3 using panel regression analysis. The results are reported in

Table 2. CSR and CEO compensation components.

	Model (1)	Model (2)	Model (3)	Model (4)
Ratio of bonus to total compensation	−0.124 ** (0.0476)			−0.109 ** (0.0430)
Ratio of LTIP to total compensation		−0.0943 ** (0.0358)		−0.0808 * (0.0439)
Ratio of equity to total compensation			0.0956 *** (0.0244)	0.0609 ** (0.0273)
Log of CEO tenure	0.0131 (0.0098)	0.0141 (0.0099)	0.0169 * (0.0100)	0.0154 (0.0100)
Market-to-book value of assets	0.0486 *** (0.0128)	0.0481 *** (0.0129)	0.0456 *** (0.0125)	0.0461 *** (0.0127)
Log of firm’s assets	0.0916 *** (0.0195)	0.0928 *** (0.0195)	0.0864 *** (0.0203)	0.0887 *** (0.0204)
Return on assets (ROA)	0.101 ** (0.0501)	0.115 ** (0.0542)	0.107 ** (0.0512)	0.141 ** (0.0532)
Current ratio	−0.0129 * (0.00743)	−0.0132 * (0.00737)	−0.0130 * (0.00749)	−0.0137 * (0.00734)
Leverage	−0.00502 (0.397)	0.0228 (0.399)	0.0231 (0.398)	0.00870 (0.395)
Industry fixed effects	Included	Included	Included	Included
Year control variables	Included	Included	Included	Included
Observations	14,046	14,046	14,046	14,046

Robust standard errors are reported in parentheses. Statistical significance is denoted by *, **, and *** at the 10%, 5%, and 1% levels, respectively.

. Models (1), (2), and (3) test hypotheses H1, H2, and H3, respectively. The dependent variable is our CSR index, and the independent variables include the proportions of individual CEO compensation components—bonus, LTIP, and equity—in total CEO compensation. In unreported analysis, we also checked the link between the proportion of salary and CSR, and did not find a significant relationship. We perform an additional test in model (4), that includes all three compensation ratios from other models in the same specification with the proportion of salary being the excluded group. Firm-level control variables include CEO tenure, market-to-book value ratio, firm size, ROA, current ratio, and leverage.

Model (1) supports hypothesis H1 as the coefficient for the proportion of a bonus in total CEO compensation is significant at the five percent level. In our data sample, CEOs receiving greater bonus incentives were less likely to invest in CSR. Control variables also indicate that CSR was positively and significantly associated with firm size, ROA, and market-to-book value ratio and negatively associated with the current ratio of firms in our sample.

Model (2) provides support to hypothesis H2 as a higher proportion of LTIPs in total CEO compensation is found to be negatively associated with CSR. The coefficient is significant at the five percent level. Results for model (3) support hypothesis H3, as a higher proportion of equity in total compensation is found to be positively associated with CSR at the one percent level. Including the three separate components of CEO compensation in model (4) provides results that are consistent with the other specifications. While the design of earlier studies (Cai et al., 2011; Karim et al., 2018; Ikram et al., 2020; Kharabsheh et al., 2022) explored compensation as the dependent variable with CSR as a control, our models focus on firms’ CSR as the dependent variable.

We continue the analysis by investigating the impact of CEO gender on the relationship between CEO compensation and the firms’ CSR measures. We introduce a dummy variable that identifies female CEOs. The results are reported in

Table 3. Influence of CEO gender and compensation components on CSR.

	Model (5)	Model (6)	Model (7)	Model (8)	Model (9)
Female CEO	0.292 *** (0.0614)	0.290 *** (0.0614)	0.291 *** (0.0616)	0.289 *** (0.0621)	0.287 *** (0.0621)
Ratio of bonus to total compensation		−0.114 ** (0.0475)			−0.0965 ** (0.0428)
Ratio of LTIP to total compensation			−0.0889 ** (0.0366)		−0.0731 (0.0443)
Ratio of equity to total compensation				0.0917 *** (0.0234)	0.0607 ** (0.0255)
Log of CEO tenure	0.0163 (0.0100)	0.0159 (0.0099)	0.0168 (0.0101)	0.0194 * (0.0102)	0.0180 * (0.0101)
Market-to-book value of assets	0.0486 *** (0.0129)	0.0487 *** (0.0131)	0.0482 *** (0.0133)	0.0458 *** (0.0129)	0.0462 *** (0.0131)
Log of firm’s assets	0.0917 *** (0.0194)	0.0912 *** (0.0194)	0.0923 *** (0.0194)	0.0863 *** (0.0202)	0.0883 *** (0.0203)
Return on assets (ROA)	0.0856 (0.0513)	0.105 ** (0.0506)	0.119 ** (0.0544)	0.112 ** (0.0513)	0.143 *** (0.0535)
Current ratio	−0.0123 (0.00741)	−0.0128 * (0.00745)	−0.0131 * (0.00737)	−0.0129 * (0.00749)	−0.0136 * (0.00736)
Leverage	−0.0227 (0.396)	−0.0250 (0.399)	0.000581 (0.400)	0.00110 (0.399)	−0.0113 (0.397)
Industry fixed effects	Included	Included	Included	Included	Included
Year control variables	Included	Included	Included	Included	Included
Observations	14,046	14,046	14,046	14,046	14,046

Robust standard errors are reported in parentheses. Statistical significance is denoted by *, **, and *** at the 10%, 5%, and 1% levels, respectively.

. Model (5) explores the basic relationship between CEO gender and the firms’ CSR in order to test hypothesis H4. The results support the hypothesis as the coefficient for female CEOs is positive and significant at the one percent level. Female CEOs in our sample were associated with higher CSR measures at their firms. The control variables show patterns consistent with the ones reported in Table 2.

Models (6)–(9) replicate the analysis of Table 2 on the link between individual CEO compensation components and CSR introducing the female CEO variable. The results demonstrate that female CEOs were positively associated with CSR in the specifications that control for the various CEO compensation components. The coefficients for female CEOs remain positive and significant at the one percent level in all the models, further supporting hypothesis H4.

The coefficients for the proportion of bonus in total compensation (model 6) and the proportion of LTIP (model 7) are negative and significant at the five percent level. Meanwhile, the coefficient for the ratio of equity in CEO compensation (model 8) is positive and significant at the one percent level. These results provide further support for hypotheses H1–H3. Firms may employ these findings to inform the design of CEO compensation mechanisms aimed at providing incentives for achieving CSR objectives.

We continue the empirical analysis by examining the effect of interaction variables to identify how female CEO commitments to CSR may be motivated by the various types of compensation. The results are reported in

Table 4. CEO gender interaction effects on firm CSR.

	Model (10)	Model (11)	Model (12)
Female CEO	0.310 *** (0.0616)	0.225 *** (0.0727)	0.286 *** (0.100)
Ratio of bonus to total compensation	−0.107 ** (0.0460)
Female CEO × ratio of bonus compensation	−0.424 ** (0.208)
Ratio of LTIP to total compensation		−0.0986 *** (0.0368)
Female CEO × ratio of LTIP compensation		0.443 ** (0.213)
Ratio of equity to total compensation			0.0915 *** (0.0235)
Female CEO × ratio of equity compensation			0.00574 (0.139)
Log of CEO tenure	0.0157 (0.00993)	0.0170 * (0.0101)	0.0194 * (0.0102)
Market-to-book value of assets	0.0485 *** (0.0131)	0.0484 *** (0.0132)	0.0458 *** (0.0129)
Log of firm’s assets	0.0912 *** (0.0194)	0.0921 *** (0.0195)	0.0862 *** (0.0202)
Return on assets (ROA)	0.105 ** (0.0508)	0.121 ** (0.0542)	0.112 ** (0.0513)
Current ratio	−0.0127 * (0.00742)	−0.0132 * (0.00739)	−0.0129 * (0.00748)
Leverage	−0.0302 (0.400)	−0.00331 (0.400)	0.00101 (0.399)
Industry fixed effects	Included	Included	Included
Year control variables	Included	Included	Included
Observations	14,046	14,046	14,046

Robust standard errors are reported in parentheses. Statistical significance is denoted by *, **, and *** at the 10%, 5%, and 1% levels, respectively.

. Model (10) demonstrates that cash bonus incentives were more likely to motivate female CEOs to reduce CSR engagement. This may be due to the additional incentive for achieving financial goals created by the cash bonus. This conclusion does not hold for LTIPs in model (11) as the coefficient of the interaction term is positive. This indicates that LTIPs were less likely to motivate female CEOs to reduce CSR. Finally, we do not find significance in the interaction effect with equity compensation for female CEOs in model (12).

The analysis of interaction variables reveals an important distinction between the role of cash bonuses and LTIPs in our data sample. While both types of performance-based incentive compensation had an overall negative relationship with the firm’s CSR, for female CEOs, a higher proportion of cash bonus in total compensation was associated with a further reduction in CSR, while a higher proportion of LTIPs actually had a positive link with CSR. LTIPs have a longer time horizon than the annual cash bonus and they may be linked to achieving non-financial performance indicators, including ones related to CSR (Mishra et al., 2000; Mahoney & Thorne, 2005; Frye et al., 2006; Liu et al., 2023). This helps explain why the impact of LTIPs for female CEOs on the firms’ CSR differs from that of cash bonuses. This finding may be helpful to firms’ compensation committees in setting the incentives for the CEOs that are aligned with the firms’ goals related to both financial metrics and CSR.

5. Discussion

This article empirically examines the relationship between CEO compensation structure, CEO gender, and a firms’ CSR measures using a sample of 2,659 U.S. publicly traded firms for the period between 2003 and 2013. The analysis highlights distinctions in the links between specific components of CEO compensation and CSR performance. We explore major types of executive compensation, including salary, equity, and two performance-based incentive plans: cash bonuses and LTIPs. Additionally, the study investigates the role of CEO gender in CSR performance.

Our findings indicate a positive relationship between the proportion of equity compensation and CSR in our data sample. Restricted stock grants, the most common form of equity compensation, do not require CEOs to achieve specific performance metrics while aligning the interests of the CEO with those of the shareholders. This structure provides CEOs with the flexibility to pursue various strategies aimed at enhancing firm values, including investments in social responsibility initiatives.

The results demonstrate that the proportions of cash bonuses and long-term incentive plans (LTIPs) in total CEO compensation were negatively associated with CSR for the firms in our sample. The payouts of these performance-based plans are contingent upon achieving certain accounting targets (Edmans et al., 2017). These arrangements appeared to constrain CEOs to focus on the predetermined metrics, thereby discouraging investment in CSR activities.

This study also finds that firms led by female CEOs exhibited higher overall CSR performance. Analysis of the interaction terms reveals that a higher proportion of cash bonus in CEO compensation further reduced CSR in firms with female CEOs, whereas a higher ratio of LTIPs did not appear to influence female CEOs’ engagement with CSR negatively.

While our data sample is large and covers 2659 firms and 14,046 firm-year observations, a limitation of this study is the use of a specific time period between 2003 and 2013. While this aligns well with earlier studies that use the KLD data set including Deng et al. (2013), Karim et al. (2018), Ikram et al. (2020), Derchi et al. (2021), and Chulkov and Wang (2023), additional research with data from other time periods would complement our findings. Our data sample is limited to U.S. firms, and testing our hypotheses using international data would help consider the robustness of our results to other corporate cultures.

The empirical methodology of this study relies on regression analysis with fixed effects and includes a variety of control variables. However, unidentified external factors may influence both compensation structure and CSR performance and this has the potential for endogeneity concerns. One area for future research is the use of an instrumental variables technique to further clarify empirical results. An additional limitation of this study follows from the quantitative nature of the empirical analysis. While the data set is robust, additional qualitative analysis, perhaps with a smaller but more granular data set, could enrich the interpretation of differences between male and female CEOs in the context of CSR and executive compensation.

The findings of this study offer valuable insights for boards of directors in designing compensation packages that align with firms’ CSR objectives. Specifically, the results suggest that cash bonus incentives, with their short-term focus, may divert CEOs’ attention away from CSR initiatives and toward immediate financial performance. This effect is observed to be more pronounced among female CEOs in our sample. In contrast, female CEOs receiving a higher proportion of LTIPs are less likely to exhibit reduced CSR engagement. Furthermore, equity compensation demonstrates a positive association with CSR performance, irrespective of CEO gender, indicating its effectiveness in promoting long-term strategic goals beyond accounting-based targets. These findings underscore the importance of structuring executive compensation to foster behaviors that align with firms’ broader objectives, including CSR practices.