Abstract
This study examines the relationship between board co-option and corporate social responsibility (CSR) among U.S. firms. The results indicate that higher levels of board co-option are linked to weaker CSR performance, particularly in terms of environmental initiatives and diversity-related social programs. The negative association between board co-option and CSR is more pronounced in firms in which CEOs hold substantial power and short-term managerial compensation is prioritized and in industries characterized by higher information asymmetries. In contrast, independent, non-co-opted directors appointed before the incumbent CEO enhance engagement in CSR, highlighting their important role in fostering socially responsible corporate behavior.
Introduction
This paper examines the relationship between board co-option and corporate social responsibility (CSR). Board co-option refers to the proportion of directors appointed after the incumbent Chief Executive Officer (CEO) took office relative to the total number of directors on the board (Coles et al., 2014). Investments in CSR initiatives that often are a trade-off between short-term profits and longer-term value creation are strategic decisions that necessitate creating a balance between managerial interests and the interests of the firm’s other stakeholders, including employees, customers, and the broader society (Sajko et al., 2020). The board of directors of a firm plays a key role in developing and maintaining the balance required between the interests of different stakeholders concerning CSR-related initiatives. In addition to the board’s other advisory and monitoring tasks, the board is responsible for establishing corporate policies, approving annual budgets for CSR, and creating committees that focus on CSR-related issues (Walls et al., 2012).
However, board composition can complicate these advisory and monitoring roles. As argued by Hermalin and Weisbach (2003), the independence of the board from the firm’s CEO is the most important factor determining the monitoring effectiveness of the board. Newly appointed directors, who may have benefited from the CEO’s support during the recruitment process, are more likely to share the CEO’s perspectives or have social ties with them (Finkelstein & Hambrick, 1989; Hwang & Kim, 2009a). The allegiance of these co-opted directors to CEOs may weaken the oversight role and the monitoring effectiveness of the board give more freedom to CEOs to decide on corporate investments (Coles et al., 2014; Lim et al., 2020; Zaman et al., 2021).
Previous studies have documented that co-opted directors may strengthen the CEO’s influence in strategic corporate decisions. For example, Baghdadi et al. (2020), Coles et al. (2014), and Zaman et al. (2021) document that board co-option is associated with higher CEO compensation, lower pay-performance sensitivity, increased likelihood of corporate misconduct, and more erratic and arbitrary decision-making that could intensify the firm’s default risk. Furthermore, the prior literature suggests that higher board co-option allows managers to make riskier investments which then, in turn, leads to higher stock return volatility and idiosyncratic risk as well as more inconsistent firm performance (Huang et al., 2021; Lee et al., 2021). Board co-option has also been documented to be negatively associated with dividend payouts and positively related to the firm’s cash holdings (Jiraporn & Lee, 2018).
CEOs may choose to invest in CSR for many reasons, for instance, to enhance social capital and trust that may further influence the firm’s profitability and sales growth, to attract, recruit, and retain skilled employees that, in turn, are likely to advance productivity and performance, to uphold a corporate culture fostering transparency and integrity, to decrease the cost of capital, or to improve corporate reputation (Agoglia et al., 2022; Dhaliwal et al., 2011, 2014; Edmans, 2011; Fauver et al., 2018; Greening & Turban, 2000; Lins et al., 2017; Linthicum et al., 2010; Wans, 2020). CSR investments typically strain short-term firm performance, as their financial benefits often require a longer time horizon to materialize. Although board co-option could, in principle, alleviate CEO concerns about dismissal due to weak short-term results and thus enable a stronger focus on long-term planning, several factors mitigate this effect.
Over the past few decades, increasing pressures to meet quarterly earnings targets, performance-based managerial compensation, shorter CEO tenures, and heightened career concerns have collectively driven a culture of short-termism in corporate decision-making (Mizruchi & Marshall, 2016). Consequently, CEOs with greater influence over the board may avoid long-term CSR initiatives to avoid immediate negative consequences, such as falling short of investor expectations or incurring personal financial penalties tied to short-term incentives (Fabrizi et al., 2013; Mizruchi & Marshall, 2016). Furthermore, as Hafenbradl and Waeger (2017) observe, some executives may downplay or even deny the existence of environmental and social risks, further reducing their willingness to address such issues through corporate strategies. This combination of short-term pressures and skepticism toward societal challenges undermines the likelihood of sustained investment in CSR initiatives.
In this paper, we contribute to the growing body of literature that has documented the negative externalities of co-opted directors on different corporate outcomes by empirically examining the effects of board co-option on firms’ engagement in CSR. Prior research suggests that the strength of board monitoring is positively associated with CSR involvement (see e.g., Harjoto & Jo, 2011; Rao & Tilt, 2016; Zhang et al., 2013). Thus, given that co-opted directors are likely to impede the board’s oversight and monitoring functions, we hypothesize a negative relationship between corporate social responsibility and the degree of board co-option.
In our empirical analysis, we use a sample of 1770 publicly listed U.S. firms between the years 2000–2018 to test the hypothesis that board co-option has an adverse influence on CSR engagement. We use four different measures to capture the degree of board co-option, and we employ the MSCI’s environmental, social, and governance (ESG) ratings to assess firm-level CSR performance. Our findings suggest that CSR is adversely affected by board co-option. Specifically, we document that firms with a higher proportion of co-opted directors and especially non-independent co-opted directors on their boards exhibit weaker overall CSR performance and have lower environmental and diversity performance scores. These results highlight the critical role of CEO-director relationships in shaping corporate engagement in CSR initiatives.
When CSR performance is decomposed into specific subcategories, we find that board co-option is negatively related to various aspects of environmental performance, such as carbon emissions, toxic emissions, waste management, energy efficiency, products’ carbon footprint, and weaker engagement in diversity-related social initiatives. These negative relationships may indicate that CEOs, with the support of co-opted directors, are better positioned to prioritize short-term objectives over environmental investments, which are often highly costly (Gillingham & Stock, 2018). They may also avoid appointing women to the board, who are generally stronger advocates for environmental initiatives (Post et al., 2015). Furthermore, the negative impact of board co-option on CSR is more pronounced in firms where CEOs exhibit greater decision-making authority, rely more heavily on short-term managerial compensation, or operate in industries with significant informational asymmetries.
To address potential endogeneity concerns, we employ propensity score matching (PSM) and a difference-in-differences (DiD) approach. In the PSM analysis, we match high board co-option firms (top quartile) with low co-option firms (lower three quartiles) based on all control variables and industry. The PSM results are consistent with our baseline findings, confirming that firms with higher levels of board co-option are linked to weaker CSR performance. In our DiD analysis, we follow Coles et al. (2014) and focus on the 2000–2006 period, during which the Sarbanes-Oxley Act (SOX) was implemented. SOX mandated that publicly listed firms maintain a majority of independent directors, leading to an exogenous increase in co-opted independent directors for non-compliant firms. The DiD results indicate that non-compliant firms experienced significantly lower CSR performance in the post-SOX period, driven by the increased presence of co-opted directors. Overall, the PSM analysis and DiD regressions provide support for our main conclusions.
Our study contributes to the literature on the linkage between corporate governance mechanisms and corporate social responsibility by examining how CEO-board interactions, captured through board co-option, affect CSR outcomes. While prior research reports both positive and negative implications of CEO-board connections (e.g., Adams & Ferreira, 2007; Cohen et al., 2012; Fogel et al., 2018; Hwang & Kim, 2009b; Sandvik, 2020; Westphal, 1999), our work focuses specifically on how this influence, measured via co-option, relates to CSR. We find that a higher proportion of co-opted directors is associated with weaker CSR performance, particularly in environmental and diversity-related initiatives. This adverse effect is more pronounced in firms in which CEOs hold substantial power, favor short-term compensation structures, or operate in settings characterized by high information asymmetry.
Our empirical findings are broadly consistent with earlier evidence that co-opted directors encourage short-termism and self-serving behavior (e.g., Baghdadi et al., 2020; Cassell et al., 2018; Coles et al., 2014; Lartey et al., 2021; Lim et al., 2020). While revising this study, we became aware of closely related independent studies by Develay and Virk (2024), Ghafoor and Gull (2024), and Gull et al. (2024), which examine the impact of board co-option on greenhouse gas emissions, ESG controversies, and waste management practices, respectively. In contrast to these studies, we examine how board co-option influences firms’ overall engagement in corporate social responsibility as measured by MSCI’s ESG ratings. Our study complements and extends the literature by taking a broader view of CSR performance and by documenting both the negative effects of board co-option on CSR engagement and the mitigating role played by non-co-opted independent directors. Nevertheless, the results of Develay and Virk (2024), Ghafoor and Gull (2024), and Gull et al. (2024) can be viewed as complementary to the empirical findings reported in this study. Taken together, these findings contribute to our understanding of how different corporate governance structures influence firms’ ability to fulfill their social and environmental responsibilities.
Related Literature and Research Hypothesis
CSR, the Board of Directors, and CEO-Board Ties
CSR encompasses a firm’s initiatives that extend beyond legal requirements, integrating environmental, social, and governance considerations into its business strategy and core operations (McWilliams & Siegel, 2001). However, stakeholders often hold conflicting views on CSR investments. Shareholders may support these initiatives, expecting them to enhance the firm’s long-term value. In contrast, managers with short-term priorities may be reluctant to pursue CSR, as it can negatively impact short-term performance and require a longer horizon for financial benefits to materialize. This classic agency conflict, arising from the separation of ownership and control, is intended to be mitigated by the board of directors (Jensen & Meckling, 1976). In addition to its advisory and oversight roles, the board is tasked with establishing shareholder-focused corporate policies, monitoring and limiting managerial opportunism, approving annual CSR budgets, and forming committees dedicated to CSR-related issues (Walls et al., 2012).
Given the focal role of board independence in effective monitoring (e.g., Chen et al., 2020; Hermalin & Weisbach, 2003; Hillman & Dalziel, 2003; Lanis & Richardson, 2018), the prior CSR literature has examined the role of board independence in firms’ CSR engagement and documented that independent directors promote socially responsible corporate behavior (see e.g., Harjoto & Jo, 2011; Shaukat et al., 2016; Zhang et al., 2013). The findings of de Villiers et al. (2011) suggest that more independent boards are more likely to have individual directors with a better understanding and knowledge of environmental opportunities, and thus, the advice and monitoring provided by these independent directors is reflected in CSR engagement. In addition to board independence, the monitoring effectiveness of the board with respect to CSR issues has been documented to be influenced by board size and CEO duality. Larger boards of directors may be more effective monitors, and consistent with this view, board size is found to be positively associated with firms’ CSR performance (e.g., Hillman & Keim, 2001; Kock et al., 2012). As argued by Hermalin and Weisbach (2003), the independence of the board from the firm’s CEO is the most important factor determining the effectiveness of the board. Consequently, CEO duality directly undermines the monitoring function of the board, and the findings of Surroca and Tribó (2008) indicate that CSR performance is negatively influenced by CEO duality. Taken as a whole, the existing literature generally suggests that corporate social responsibility is positively influenced by board characteristics that reflect more stringent monitoring and oversight by the board of directors (e.g., Ali Gull et al., 2022; Francoeur et al., 2019; McGuinness et al., 2017; Rao & Tilt, 2016; Shaukat et al., 2016).
The board’s ability to fulfill its monitoring and oversight role effectively can be compromised when the CEO develops strong ties with the board members. Close ties between the CEO and the board can limit the board’s capacity to implement CSR strategies by enabling the CEO to steer decisions toward personal goals, particularly when those goals prioritize short-term performance over long-term sustainability. One mechanism through which CEOs strengthen their ties with the board is by shaping its composition. Specifically, they may influence the recruitment process to appoint directors from their social or professional networks, increasing the likelihood that these directors will align with the CEO’s preferences (Finkelstein & Hambrick, 1989; Hwang & Kim, 2009a). By leveraging their connections and informal sway over the recruitment committee, CEOs can facilitate the selection of board members who are more inclined to support their agenda.
Board Co-Option
Coles et al. (2014) define board co-option as a measure of the level of interaction and ties between the CEO and the board. They classify co-opted directors as those appointed to the board after the CEO assumed office, suggesting that these directors may develop a stronger sense of allegiance to the CEO, who played a role in their appointment. Consequently, board co-option serves as a key indicator of the extent to which the firm’s board of directors fulfills its advisory and monitoring responsibilities (Baghdadi et al., 2020).
After the prominent corporate governance failures in the early 2000s, the Securities and Exchange Commission (SEC) and other regulatory bodies have introduced mechanisms to enhance board oversight and curb CEOs’ influence over the board and new directors’ nomination and selection processes. The Sarbanes-Oxley Act of 2002, for instance, includes specific guidelines on board composition and the nomination process of directors. The enactment of the SOX led both the New York Stock Exchange and NASDAQ to prevent CEOs from having formal influence over the nomination process of corporate directors by requiring that the nomination committees of listed firms be entirely composed of independent directors. Despite these developments, it is believed that CEOs continue to exert some influence over the nomination and selection of board members.
Consistent with the arguments of Coles et al. (2014), empirical studies have documented that co-opted directors generally have an adverse influence on the advisory and monitoring functions of the board. Recent studies suggest that a higher number of newly appointed directors, referred to as co-opted directors, is associated with higher CEO compensation, a lower likelihood of forced CEO dismissal, and a reduced probability of adopting clawback provisions that link executive pay to firm performance (Coles et al., 2014; Huang et al., 2019). Research also shows that board co-option leads to higher director compensation and increases the profitability of insider trading (Fedaseyeu et al., 2018; Rahman et al., 2021; Zaman et al., 2021). Moreover, Zaman et al. (2021) document that co-opted directors put forth fewer items on the board agenda, have poorer attendance records at board meetings, and increase the likelihood of corporate misconduct.
Lim et al. (2020) investigate the effects of co-option on external funding, especially covenant intensity and covenant violations. Their findings indicate that firms with higher degrees of board co-option are subject to more covenant restrictions by creditors as well as a higher likelihood of violating loan covenants. Related studies by Baghdadi et al. (2020) and Sandvik (2020) suggest that board co-option is associated with more erratic decision-making, leading to higher default risk, lower credit ratings, and larger credit spreads. Furthermore, these recent studies also show that the board’s allegiance to the CEO allows them to make riskier investment decisions that are reflected in higher stock return volatility and idiosyncratic risk as well as more inconsistent firm performance (Baghdadi et al., 2020; Huang et al., 2021; Lee et al., 2021). Board co-option has also been documented to be negatively associated with dividend payouts and positively related to the level of cash holdings (Jiraporn & Lee, 2018).
Hypothesis
Research increasingly highlights that co-opted directors reinforce ties between the CEO and the board, influencing the board’s capacity to balance stakeholder interests and advance CSR initiatives. These ties can undermine the board’s effectiveness in promoting CSR, particularly when directors prioritize loyalty to the CEO, who may focus on short-term objectives at the expense of long-term sustainability. Consequently, greater board co-option is expected to negatively affect CSR outcomes. Building on this rationale, we propose the following research hypothesis:
The proportion of co-opted directors is negatively associated with corporate social responsibility.
Data and Methodology
Data
Our empirical analysis uses a sample of publicly listed U.S. firms from 2000 to 2018. We collect director and board characteristics from BoardEx, corporate social responsibility data from MSCI, financial and balance sheet information from Compustat, and stock market data from CRSP. The final dataset includes 1770 firms and 16,177 firm-year observations.
Corporate Social Responsibility—The Dependent Variable
The dependent variable in our analysis is corporate social responsibility. Following the prior literature (e.g., Zhao et al., 2023), we employ the MSCI’s environmental, social, and governance (ESG) ratings to gauge firm-level CSR engagement. The MSCI’s ESG ratings are based on assessments of a firm’s strengths and concerns in terms of a wide range of environmental, social, and governance factors. The five main categories underlying the scores that are related to CSR performance are (i) environment, (ii) employee relations, (iii) diversity, (iv) community, and (v) human rights. Following prior literature (Li et al., 2021; Servaes & Tamayo, 2013), we construct the CSR measure in three steps. First, we divide the number of strengths and concerns for each category by the maximum possible number of strengths and concerns in that category and year, respectively. We then subtract scaled concerns from scaled strengths to obtain a CSR-scaled score for each category. Finally, we add the CSR-scaled scores over the five categories to obtain a measure of CSR for each firm. The resulting variable, CSR, ranges from −5 to +5, with higher values reflecting stronger CSR performance.
Board Co-Option—The Main Explanatory Variable
Following the definition of Coles et al. (2014), we calculate Co-option as the number of directors appointed to the firm’s board after the incumbent CEO assumed office divided by the total number of directors on the board in the given year. Thus, Co-option measures the proportion of the board that consists of co-opted directors for each firm-year, and by construction, the variable exhibits within-firm variation over time after new director appointments or CEO turnovers.
In addition to Co-option, we employ three alternative measures to gauge the degree of board co-option. First, because a longer tenure of directors may increase their influence on the board relative to newly appointed directors (Coles et al., 2014; Huang & Hilary, 2018), we use a tenure-weighted measure of board co-option. We define Co-option TW as the sum of the tenure of co-opted directors divided by the sum of the tenure of all directors on the board. Furthermore, given the key role of board independence, we distinguish between independent and non-independent co-opted directors. Using BoardEx data, we classify directors who are not current or former executives or employees of the company as independent directors. We then calculate Co-option Independence as the number of independent directors appointed to the firm’s board after the incumbent CEO assumed office divided by the total number of directors on the board in the given year. Finally, we also use a tenure-weighted version of the proportion of independent co-opted directors. Co-option TW Independence is defined as the sum of the tenure of independent co-opted directors divided by the sum of the tenure of all directors.
Control Variables
Following prior CSR literature (Bear et al., 2010; Harjoto et al., 2015; Jo & Harjoto, 2011; Zhang et al., 2013), we take into account the potential influence of firm-specific factors such as size, profitability, and risk in explaining CSR performance. Furthermore, given that our main focus is the influence of board co-option on CSR performance, we control for various other board characteristics that presumably reflect the strength of board monitoring.
Specifically, we use the following control variables in our regressions: (i) Firm size is the market value of equity, (ii) Leverage is the ratio of long-term debt to total assets, (iii) Return on assets is calculated as the income before extraordinary items divided by total assets, (iv) Tobin’s Q is calculated as the sum of the book value of debt and the market value of equity divided by total assets, (v) Sales growth is the annual change in sales, (vi) R&D intensity is the ratio of research and development expenditures to total assets, (vii) Capital intensity is the ratio of capital expenditures to total assets, (viii) Stock return volatility is the standard deviation of monthly stock returns over the previous five years, (ix) Cash flow volatility is the standard deviation of the cash flow from operations scaled by total assets over the previous five years, (x) Firm age is the number of years since the firm first appeared in the CRSP database, (xi) Board size is the number of directors on the board, (xii) Board independence is the number of independent directors divided by the total number of directors on the board, (xiii) Board qualifications is the average number of degrees held by the directors, and (xiv) Board tenure is the average number of years directors has served on the firm’s board. For convenience, all variables and their definitions are summarized in Appendix A.
The Empirical Model
To empirically test the hypothesis that board co-option is negatively associated with CSR performance, we estimate the following regression specification:
Results
Descriptive Statistics and Correlations
Descriptive Statistics.
Continuous variables are winsorized at the 1st and 99th percentiles. Variables are defined in Appendix A.
Regarding the control variables, Table 1 demonstrates the cross-sectional heterogeneity of our sample firms in terms of size, financial performance, growth, risk, and board characteristics. The average Firm Size is $11,953 million, with a mean Leverage ratio of approximately 21.4%. On average, the board comprises ten members, with 72% classified as independent directors. Their average tenure is eight years, and they typically hold about two academic or professional degrees.
Pairwise Correlations.
All continuous variables are winsorized at the 1st and 99th percentiles to adjust for potential outliers. * denotes statistical significance at the 0.01 level. All variables are defined in Appendix A.
Baseline Results
Board Co-Option and CSR.
The table presents estimates from equation (1). The dependent variable CSR is the sum of scaled scores over five different social responsibility rating categories, including community, diversity, employee relations, environment, and human rights. The four measures of board co-option are defined as follows: Co-option is the number of directors who have been appointed to the firm’s board after the incumbent CEO assumed office divided by the total number of directors on the board, Co-option TW is the sum of the tenure of co-opted directors divided by the sum of the tenure of all directors on the board, Co-option Independence is the number of independent directors who have been appointed to the firm’s board after the incumbent CEO assumed office divided by the total number of directors on the board, and Co-option TW Independence is the sum of the tenure of independent co-opted directors divided by the sum of the tenure of all directors. The t-statistics (in parentheses) are based on robust standard errors, adjusted for heteroskedasticity, and clustered by firm. ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels, respectively. All variables are defined in Appendix A.
In Model 1, the coefficient for Co-option is negative and statistically significant at the 1% level (β1 = −0.070, t-value = −2.95). This finding supports our research hypothesis, indicating that a higher proportion of co-opted directors on the board is negatively associated with corporate social responsibility. In terms of economic significance, the magnitude of the coefficient for Co-option indicates that a one standard deviation increase in the proportion of co-opted directors decreases the firm’s CSR score by about 15% relative to its mean [0.310* (−0.070)/0.142 = −0.153]. In Model 2, Co-option is replaced with Co-option TW, a tenure-weighted measure of the degree of board co-option. Similar to Model 1, the results indicate that board co-option is adversely related to CSR as the coefficient for Co-option TW is negative and significant at the 1% level (β1 = −0.085, t-value = −3.64).
Next, we employ Co-option Independence and Co-option TW Independence as the test variables in Models 3 and 4, respectively. While independent co-opted directors are generally considered more effective monitors than their non-independent counterparts, the regression results in Table 3 indicate that a greater representation of independent co-opted directors is negatively associated with CSR performance. Specifically, the coefficient for Co-option Independence in Model 3 is negative and statistically significant at the 5% level (β1 = −0.072, t-value = −2.52), whereas in Model 4, the coefficient for Co-option TW Independence is negative and significant at the 1% level (β1 = −0.100, t-value = −3.06). These findings indicate that irrespective of the board co-option measure employed, co-option is strongly and adversely related to CSR.
The coefficients for the control variables in Table 3 align broadly with prior literature. Regardless of the model specification, the results suggest that CSR performance is significantly and positively associated with Firm Size, Return on Assets, R&D Intensity, Capital Intensity, Board Size, Board Independence, and Board Qualifications. Conversely, CSR performance is negatively associated with Tobin’s Q, Sales Growth, and Board Tenure.
Endogeneity
Propensity Score Matching
Propensity Score Matching (PSM).
Panel A presents the means and mean differences of variables used to match the treatment and control groups before and after implementing the PSM method. The treatment group comprises firms with their board co-option value falling in the top quartile of their respective industry in the given year. In contrast, the control group consists of firms with their board co-option value falling in the lower quartiles in their respective industry in the given year. The dependent variable CSR is the sum of scaled scores over five social responsibility rating categories, including community, diversity, employee relations, environment, and human rights. The four measures of board co-option are defined as follows: Co-option is the number of directors who have been appointed to the firm’s board after the incumbent CEO assumed office divided by the total number of directors on the board, Co-option TW is the sum of the tenure of co-opted directors divided by the sum of the tenure of all directors on the board, Co-option Independence is the number of independent directors who have been appointed to the firm’s board after the incumbent CEO assumed office divided by the total number of directors on the board, and Co-option TW Independence is the sum of the tenure of independent co-opted directors divided by the sum of the tenure of all directors. The t-statistics (in parentheses) are based on robust standard errors, adjusted for heteroskedasticity, and clustered by firm. ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels, respectively. All variables are defined in Appendix A.
Our matching process results in 3,698 firm-year observations for both high and low co-option groups, for a total of 7,396 observations. Panel A of Table 4 presents descriptive statistics for both groups before and after PSM, where the insignificant differences in mean values confirm the success of the matching process. We then re-estimate our baseline regressions using this matched sample. The regression results reported in Panel B of Table 4 are consistent with our baseline results. Specifically, we find that high co-option is associated with lower CSR performance, confirming that this relationship is robust even after using comparable firms.
Difference-in-Differences
The estimates in our baseline model may be subject to endogeneity concerns arising from omitted variables or reverse causality. For instance, unobserved or omitted time-invariant firm characteristics could simultaneously influence both CSR performance and the number of co-opted directors, potentially leading to a spurious relationship between the two. To address these concerns, we next employ a DiD approach that utilizes the regulatory shock introduced by SOX. Following Coles et al. (2014), we focus on the period from 2000 to 2006, during which SOX was enacted. The reform required publicly listed firms to maintain a majority of independent directors on their boards, prompting a substantial increase in the number of co-opted independent directors among firms that were initially non-compliant. This regulatory change provides a quasi-experimental setting to isolate, at least partially, the impact of increased board co-option on CSR performance.
To capture this effect, we define two indicator variables. Non-compliance equals one for firms that had fewer than 50% independent directors prior to SOX and zero otherwise. Post-SOX equals one for observations in the post-reform period. The interaction term, Post-SOX × Non-compliance, serves as our key variable of interest and captures the relative change in CSR performance for non-compliant firms following the reform. To ensure a credible comparison between treatment and control groups, we implement PSM to construct a balanced sample. Each non-compliant firm is matched with a comparable compliant firm from the same industry, ensuring that the pairs are statistically similar in observable firm and board characteristics, apart from their initial board independence levels.
Difference-In-Differences Regressions (DiD).
Panel A presents the means and mean differences of variables used to match the treatment and control groups before and after implementing the PSM method. Panel B displays the estimates from DiD regressions. The variable Post-SOX is an indicator that takes the value of one for post-SOX years, representing the years 2003 and onwards. Another variable, Non-compliance, is an indicator that takes the value of one for firms where less than 50% of directors were classified as independent directors in the year 2000, following Coles et al. (2014). The firms categorized as non-compliant are considered the treatment group, while those with more than 50% independent directors are the control group. The dependent variable CSR is the sum of scaled scores over five social responsibility rating categories, including community, diversity, employee relations, environment, and human rights. The four measures of board co-option are defined as follows: Co-option is the number of directors who have been appointed to the firm’s board after the incumbent CEO assumed office divided by the total number of directors on the board, Co-option TW is the sum of the tenure of co-opted directors divided by the sum of the tenure of all directors on the board, Co-option Independence is the number of independent directors who have been appointed to the firm’s board after the incumbent CEO assumed office divided by the total number of directors on the board, and Co-option TW Independence is the sum of the tenure of independent co-opted directors divided by the sum of the tenure of all directors. The t-statistics (in parentheses) are based on robust standard errors, adjusted for heteroskedasticity, and clustered by firm. ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels, respectively. All variables are defined in Appendix A.
Figure 1 provides evidence supporting the parallel trends assumption, as CSR performance between treatment and control firms does not differ significantly before SOX. The pre-treatment coefficients (t−3 to t−1) are close to zero with overlapping confidence intervals, indicating similar patterns. The post-SOX decline in CSR among treatment firms suggests that the divergence is likely driven by increased board co-option rather than pre-existing differences. Parallel trends in CSR performance around SOX implementation. This figure illustrates the parallel trends analysis from our DiD regression, focusing on CSR performance before and after the implementation of SOX. The graph shows the coefficient estimates comparing non-compliant (treatment) and compliant (control) firms over time, with 90% confidence intervals indicated by error bars.
Subcategories of CSR
Board Co-Option and the Subcategories of CSR Performance.
Panel A investigates the association between board co-option and CSR subcategories. Panel B replicates the regressions from the preceding panel while incorporating additional controls in the form of the remaining CSR subcategories when analyzing a particular subcategory as the dependent variable. Co-option is the number of directors appointed to the firm’s board after the incumbent CEO assumed office divided by the total number of directors on the board. The t-statistics (in parentheses) are based on robust standard errors, adjusted for heteroskedasticity, and clustered by firm. ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels, respectively. All variables are defined in Appendix A.
For brevity, we report only the coefficient estimates for the main variables of interest and omit the estimates for the control variables. As can be seen from Panels A and B, the estimated coefficients for Co-option are negative and significant in the regressions with Diversity and Environment as the dependent variables, while being insignificant in the three other regressions. This suggests that the negative relationship between CSR performance and board co-option is primarily driven by the detrimental effects of co-opted directors on the firm’s environmental performance and diversity-related initiatives.
Further Analysis of CSR Performance Subcategories.
Panel A presents additional analysis on board co-option and CSR subcategories, focusing on strengths and concerns as the dependent variables. Panel B replicates the regressions from the preceding panel while incorporating additional controls in the form of the remaining strengths and concerns when considering the strength or concern of a particular subcategory as the dependent variable. Co-option is the ratio of directors appointed post-CEO to total board members. The t-statistics (in parentheses) are based on robust standard errors, adjusted for heteroskedasticity, and clustered by firm. ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels, respectively. All variables are defined in Appendix A.
The Role of Non-Co-Opted Directors
Coles et al. (2014) argue that co-opted directors, though formally independent, often align with the CEO and support managerial interests. In contrast, non-co-opted independent directors, appointed before the current CEO, retain greater autonomy and are less likely to endorse short-term managerial goals. We classify directors accordingly, assuming co-opted directors reflect stronger CEO-board ties, while non-co-opted directors are more likely to support CSR when it aligns with long-term firm sustainability and governance priorities.
We define Co-option Non-independence as the number of non-independent directors who have been appointed to the firm’s board after the incumbent CEO assumed office divided by the total number of directors. We also define Non-co-option Independence as the number of independent directors appointed to the firm’s board before the incumbent CEO assumed office divided by the total number of directors. We then estimate equation (1) with these two board co-option measures as the test variables of interest. Our prediction is that the coefficient estimate for Non-co-option Independence will be positive (i.e., β1 > 0).
Relationship Between Non-Co-Opted Independent Directors and CSR Performance.
This table presents regression results examining the relationship between non-co-opted independent directors and CSR performance. The dependent variable, CSR, represents the sum of scaled scores across five social responsibility rating categories: community, diversity, employee relations, environment, and human rights. Co-option Non-independence is the number of non-independent directors appointed to the firm’s board after the incumbent CEO assumed office, divided by the total number of directors. Conversely, Non-co-option Independence is calculated as the number of independent directors appointed to the firm’s board before the incumbent CEO assumed office, divided by the total number of directors. The t-statistics (in parentheses) are based on robust standard errors, adjusted for heteroskedasticity, and clustered by firm. ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels, respectively. All variables are defined in Appendix A.
Cross-Sectional Analysis
As the next step of our analysis, we investigate the roles of CEO power, managerial compensation incentives, external funding dependence, and informational asymmetries in influencing the observed negative relationship between board co-option and CSR engagement. We begin by examining how CEO power potentially influences the linkage between board co-option and CSR. The literature presents mixed findings on both positive and negative effects of CEO power on CSR (Fabrizi et al., 2014; Jiraporn & Chintrakarn, 2013; F. Li et al., 2016; Walls & Berrone, 2017). Despite these differences, it is widely acknowledged that CEOs derive power from multiple sources, including serving as the Chairperson of the Board, holding significant ownership stakes, or combining executive roles (Finkelstein, 1992).
Cross-Sectional Tests: CEO Power, Board Co-Option, and CSR.
Co-option is the number of directors appointed to the firm’s board after the incumbent CEO assumed office divided by the total number of directors on the board. The high and low subsamples correspond to firm-year observations in the top quartile and the bottom three quartiles, respectively. The t-statistics (in parentheses) are based on robust standard errors, adjusted for heteroskedasticity, and clustered by firm. ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels, respectively. All variables are defined in Appendix A.
Bebchuk et al. (2011) propose that the CEO Pay Slice (CPS)—defined as the proportion of the total compensation of the top five highest-paid executives allocated to the CEO—serves as a measure of the CEO’s value to the firm as well as their bargaining power and influence over the board. Their findings indicate that firms with a higher CPS often exhibit weaker corporate governance structures and lower firm value, as reflected in Tobin’s Q. Moreover, a higher CPS is linked to suboptimal firm performance, driven by misaligned incentives between the CEO and other executives and the CEO’s ability to extract rents. Accordingly, we next use CPS as an additional measure of CEO power. In Models 5 and 6 of Table 9, the coefficients for Co-option indicate a negative relationship between board co-option and CSR. This negative relationship is more pronounced in firms where the CPS is in the top quartile, suggesting that higher CEO influence over the firm exacerbates the adverse effect of board co-option on CSR.
To further capture the multifaceted nature of CEO power, we follow van Essen et al. (2015) and construct a composite variable, CEO Power, derived through principal component analysis (PCA). This measure aggregates five key indicators: CEO tenure, CEO ownership, CEO duality, the CEO holding dual roles as both Chair and President, and the proportion of stocks owned by institutional shareholders. These variables are considered to encompass critical dimensions of CEO influence. Longer CEO tenure allows the CEO to build expertise and secure the confidence of the board and other stakeholders, often leading to greater influence (Bebchuk & Fried, 2003, 2004). CEO ownership measures the share of the firm’s stocks held by the CEO, aligning their interests with those of shareholders and enhancing their leverage within the firm (Von Lilienfeld-Toal & Ruenzi, 2014). CEO duality and holding both Chair and President roles further amplify the CEO’s influence over board decisions and corporate strategy (Finkelstein, 1992). Conversely, institutional ownership moderates CEO power, as institutional investors actively monitor management to fulfill their fiduciary duty of maximizing returns (McConnell & Servaes, 1990). Using PCA ensures that the composite measure, CEO Power, accurately reflects variations in CEO influence across firms. Higher scores indicate a greater concentration of authority, highlighting the CEO’s capacity to influence corporate decisions.
Consistent with prior models, the coefficients for Co-option in Models 7 and 8 of Table 9 are negative, indicating an inverse relationship between board co-option and CSR. This relationship is stronger in firms in which CEO Power falls within the top quartile, as these firms exhibit larger negative coefficients in magnitude. The Wald χ2 test indicates that the difference between these coefficients is statistically significant. Overall, the results of these cross-sectional tests suggest that the negative impact of board co-option on CSR intensifies as CEO power increases.
We proceed by examining whether the linkage between board co-option and CSR is influenced by executive compensation, external funding dependence, and informational asymmetries. While executive compensation policies aim to align managerial interests with shareholders, they may also foster managerial short-termism and lead to incentive misalignment with other stakeholders. Short-term components, such as salary and bonuses, could discourage investment in CSR due to their potential drain on immediate firm performance (Fabrizi et al., 2013). In contrast, long-term incentives like option awards or restricted stocks may encourage CEOs to prioritize long-term value, including CSR initiatives, benefiting the firm’s future prospects (Deckop et al., 2016; Flammer & Bansal, 2017).
Further Cross-Sectional Tests.
This table presents the estimates of equation (1) based on different subsamples. The dependent variable CSR is the sum of scaled scores over five social responsibility rating categories, including community, diversity, employee relations, environment, and human rights. Co-option is the number of directors appointed to the firm’s board after the incumbent CEO assumed office divided by the total number of directors on the board. Short-term Incentives is the CEO’s annual salary, Option Awards is the value of stock options awarded to the CEO in the given year, External-finance Dependence is the industry-median proportion of investments not financed by cash flow from operations, and Information Asymmetry is the industry standard deviation of productivity growth as measured by the ratio of sales to the number of employees. The high and low subsamples correspond to firm-year observations in the top quartile and the bottom three quartiles, respectively. The t-statistics (in parentheses) are based on robust standard errors, adjusted for heteroskedasticity, and clustered by firm. ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels, respectively. All variables are defined in Appendix A.
Next, we assess the impact of External Funding Dependence on the co-option and CSR relationship by splitting the sample into two groups: industries in the top quartile and those in the bottom three quartiles based on average External Funding Dependence. Models 5 and 6 in Table 10 are the estimates of the regressions based on the external funding dependence subsamples. Once again, the estimated coefficients for Co-option are negative and statistically significant in both subsamples, with no significant difference between the coefficients.
Finally, we examine the potential role of informational asymmetries. Given that co-opted directors may impede the board’s oversight and monitoring functions, we expect informational asymmetries to exacerbate the negative relationship between co-option and CSR. We divide the sample into high and low Information Asymmetry groups based on industry averages. The estimates of equation (1) using these two subsamples are presented in Models 7 and 8 of Table 10. The estimated coefficients for Co-option are negative and statistically significant in both subsamples. The larger coefficient estimate in Model 8 suggests a stronger negative association between co-option and CSR in industries with higher informational asymmetries, and this observed difference in coefficients is statistically significant.
Additional Tests
Robustness Tests.
This table presents the estimates of various robustness checks. The dependent variable CSR is the sum of scaled scores over five social responsibility rating categories, including community, diversity, employee relations, environment, and human rights. Co-option is the number of directors appointed to the firm’s board after the incumbent CEO assumed office divided by the total number of directors on the board. The reported results remain robust when replacing Co-option with the three alternative measures of board co-option used in the previous tables. The t-statistics (in parentheses) are based on robust standard errors, adjusted for heteroskedasticity, and clustered by firm. ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels, respectively. All variables are defined in Appendix A.
Second, we address potential industry differences. In Model 3, we exclude financial firms due to their stricter regulations. Conversely, in Model 4, we focus solely on financial firms and replicate our baseline model. The results indicate that board co-option is negatively associated with CSR in both subsamples. However, the negative relationship is more pronounced in financial firms, and the difference is statistically significant. Third, to control for potential systematic variation in CSR performance across different states, we include state fixed effects in our regressions. Again, in Model 5, the coefficient estimate for the Co-option remains negative and highly significant after the inclusion of state fixed effects.
Fourth, we investigate whether our empirical findings are driven by firm-size effects. Because larger firms are generally associated with stronger CSR performance and firm size may correlate with unobserved factors influencing CSR investment, we exclude the largest quartile of firms and re-estimate our baseline model. The regression results based on this subsample are very similar to the estimates reported in Table 3. Fifth, recognizing that growth firms may face resource constraints and weaker incentives for CSR, we exclude firms in the top quartile of market-to-book ratios. Model 7 reports the regression results based on the sample without growth firms. Consistent with our main regressions, the estimated coefficient for Co-option is negative and highly significant.
Fifth, we examine the sensitivity of our findings to the sample period by re-estimating the regressions using two truncated periods: 2000–2008 and 2009–2018. The results of these additional regressions are presented in Models 8 and 9. In both subsamples, the coefficient estimates for Co-option are negative and statistically significant. Nonetheless, the coefficient for Co-option is more than twice as large in magnitude when the sample period is truncated to years 2009–2018 and the observed difference in the coefficient estimates between the two subsamples is highly significant. Thus, the regression results suggest that the negative influence of co-opted directors on CSR engagement has become more pronounced during the latter part of the sample period.
Robustness Tests Related to Omitted Variable Bias.
This table presents robustness tests related to potential omitted variable bias in the relationship between board co-option and CSR. The dependent variable CSR is the sum of standardized scores across five social responsibility dimensions: community, diversity, employee relations, environment, and human rights. Co-option is the number of directors appointed to the firm’s board after the incumbent CEO assumed office divided by the total number of directors on the board. The t-statistics (in parentheses) are based on robust standard errors, adjusted for heteroskedasticity, and clustered by firm. ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels, respectively. All variables are defined in Appendix A.
Conclusions
This paper investigates the relationship between board co-option and CSR performance. We empirically test the hypothesis that board co-option has an adverse influence on CSR engagement by utilizing a sample of 1770 publicly listed U.S. firms. Following Coles et al. (2014), we define board co-option as the proportion of directors appointed after the incumbent CEO assumed office. We also employ alternative measures of co-option that account for director independence and tenure. To gauge CSR performance, we employ the MSCI’s assessments of firms’ strengths and concerns in terms of five different aspects of corporate social responsibility. The association between board co-option and CSR is empirically examined with fixed-effects panel regressions in which we control for a range of firm attributes and board characteristics that are known to influence CSR engagement. We also utilize propensity score matching and difference-in-differences analysis to mitigate endogeneity biases.
Our findings indicate that board co-option is negatively associated with CSR performance. Specifically, firms with a higher proportion of co-opted directors, particularly non-independent co-opted directors, exhibit significantly lower CSR scores. A one standard deviation increase in the proportion of co-opted directors corresponds to an estimated 15% reduction in the firm’s CSR score. Conversely, independent, non-co-opted directors are positively associated with CSR performance, underscoring their role in fostering socially responsible practices. When examining different dimensions of CSR, we find that co-option particularly undermines environmental and diversity-related initiatives. Moreover, the negative relationship between co-option and CSR is more pronounced in firms with powerful CEOs, stronger short-term managerial compensation incentives, and higher informational asymmetries. These results suggest that CEO-board ties, evident in the presence of co-opted directors, may hinder the board’s ability to support corporate social responsibility initiatives effectively.
Our research offers important practical implications regarding the CEO’s interaction with board members. The findings suggest that when a CEO informally manipulates the recruitment process to appoint directors from their personal network, the board may become more aligned with the CEO’s interests rather than those of other stakeholders, leading to reduced focus on CSR. For investors, regulators, and other stakeholders, the informal influence of CEOs over board appointments is a critical indicator of potential governance risks. High levels of board co-option may signal vulnerabilities in CSR performance, particularly in key areas such as environmental initiatives and diversity programs. Therefore, policymakers could consider advocating for reforms that reduce the CEO’s control over board composition to ensure balanced governance for fostering sustainable and socially responsible business practices.
Footnotes
Acknowledgments
We thank two anonymous referees, Anna Agapova, Janis Berzins, Denis Davydov, Guanming He, Kim Ittonen, Gilad Livne, Dennis Sundvik, and conference and seminar participants at the 2023 Journal of Accounting, Auditing, and Finance (JAAF) Conference, the 12th Financial Markets and Corporate Governance Conference, the 2022 World Finance Conference, the 2022 International Workshop on ESG Values, and the University of Vaasa for valuable comments and suggestions. This work was supported by the Foundation for Economic Education, the OP Group Research Foundation, and Hanken Foundation. Part of this paper was written while A. Afzali was visiting the University of Toronto, H. Silvola was visiting the Weatherhead Center at Harvard University, and S. Vähämaa was visiting Bentley University and the University of Sydney.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Liikesivistysrahasto, Hanken Foundation, and OP Group Research Foundation.
