Internal Governance and Corporate Social Responsibility: Evidence from Chinese Companies

Abstract

Stakeholder management researchers have recently put a lot of effort into figuring out why organizations facing extensive pressure respond differently to social responsibilities. In particular, ethics researchers believe that senior management must drive corporate social responsibility since their attitudes toward such issues are so important. In line with this sentiment, our study develops a framework of management power, composed of CEOs’ power and the organizations’ power, and explores how managerial power heterogeneity affects the corporate social responsibility (CSR) performance of a firm. Using sample data from the largest emerging market—China—for the period 2010–2018, we submit that CEOs with structural power and shareholders with the highest concentration tend to show a lower commitment to CSR activities. On the other hand, we recognize that the ownership, expertise, and prestige power of CEOs’, the supervision, monitoring, and political power of the board can improve a firms’ CSR performance. These results are also validated by using a fixed effect model, two stage least square (2-SLS) regression, and the propensity score matching (PSM) technique. Our results imply that the implementation of social policies fundamentally results not only from powerful CEOs, but also from powerful boards and shareholders. Moreover, our study provides useful implications with regard to the social outcomes of power authorized by CEOs and the organizations.

1. Introduction

Recently, social scholars have focused on understanding how managerial power determines corporate choices and outcomes. Power at the top management level is equally fundamental to research as it gives key decisions makers the freedom to exert their own will [1,2]. Prior scholars agree that corporate choices must be triggered by upper-level managers and decision makers whose orientation towards important matters play a meaningful role [3] because they use power to shape and influence corporate policies [4,5]. Although the majority of work suggests that as top managers, the power of Chief Executive Officers(CEOs) significantly influences corporate outcomes such as a firms’ productivity [6], strategic change [7], a firms’ financial performance [8], debt financing [9], corporate social responsibility (CSR) [10], and environmental performance [3]. However, this line of work has only examined the power of CEOs, focusing on their power dimensions, while the role of heterogeneous sources of power remains relatively ignored [11]. Given the importance of managerial power in corporate outcomes, this study raises an important question: Does power heterogeneity matter for a firm’s CSR performance?

Based on the fundamental notion of CSR, organizations have various groups of stakeholders such as employees, customers, community groups, investors, suppliers and governmental agencies. The main purpose of senior managers is to achieve or fulfill the goals of each of these stakeholders simultaneously. Hence, managers, particularly CEOs, play a significant role in managing and addressing the goals of multiple stakeholders. However, our research question is particularly based on earlier anecdotal evidence which reveals that in organizations where the CEOs wield dominant hold and power, only the CEO can produce sufficient information, but this may not be always the case [12,13]. Most organizations have dominant coalitions that usually empower a subset of leaders to set and implement policies [14]. For instance, shareholders with the largest ownership can be more powerful than CEOs [15]. Similarly, in some organizations, the directors play a powerful role in examining the behavior of the CEOs and thus can influence strategic decisions [16]. From the standpoint of the upper echelon theory of Hambrick and Mason [17], a coalition of upper echelons can provide better explanations as to corporate outcomes (i.e., CSR performance). Hambrick and Mason [17] argued that every member of the top management team plays an influential role in increasing the potential strength of management to make important corporate decisions. This perspective further postulates that managers’ decision making is reflected by managers’ personal and professional characteristics [17,18], which generally defines managerial controlling power [7,19,20].

Along these lines, several studies have attributed the major portion of organizational outcomes to managerial controlling power [21] which is predominantly a multidimensional concept and should not be considered a single indicator [19,22,23]. Thus, prior studies acknowledge the leadership of a firm as a shared responsibility whereby the strategic decisions are based on the collective interactions, capabilities, and cognitions of the entire management team [19,24]. On the other hand, few authors have an opposing view, acknowledging that top managers usually do not evenly share the controlling features (e.g., superiority and power) which tend to belong typically to CEOs [25,26] and which eventually restrain the influence of directors and other executives in corporate strategy [27].

Within the upper echelon theoretical framework, this study argues that a consideration of the power of a set of upper echelons for the sake of decision making, such as CSR investment, is necessary to capture the diversity of managerial orientations. Given the uncertainty and financial loss associated with CSR investment and the level of managerial discretion involved in such decisions, we believe that managerial power is a critical cognitive trait, helping managers and owners to interpret their own will regarding social investment. Thus, this research particularly investigates heterogeneous sources of power accumulated by top managers (CEOs), boards of directors, and the largest shareholders using Chinese listed companies. This study utilizes the Chinese setting because of its distinctive institutional setting where firms face different governance problems like, concentration of ownership, weak structure of corporate governance, weak stakeholders’ protection, and a less effective legal system. Overall, the governance structure of listed firms in China varies significantly from advanced countries in terms of an independent system, the prevalence of state intervention, and the strong managerial political ties at executive and board levels. Thus, using a data sample from this emerging market, our results confirm that the structural power of CEOs and the power of the largest shareholders adversely affects organizational social behavior. On the other hand, the ownership, prestige, and expert power of CEOs the board’s supervision power the directors’ political power, and the independent directors’ monitoring power positively influence the CSR performance of firms.

Our study attempts to make several contributions. First, the existing stream of research [3,20,28,29] considers an index of managerial power encompassing several power dimensions drawn from the study of Finkelstein [19]. This study endeavors to contribute by showing a broad picture of power dimensions accumulated by top managers (CEOs), boards of directors, and the largest shareholders. Second, the prior literature on power and CSR [10,30] has concentrated on executive’s power by exploring a few sources such as CEO tenure, pay slice, or CEO duality. However, in our case, we focus on power heterogeneity encompassing a variety of power dimensions of a set of decision makers in relationships with CSR. Our work shows that not only CEOs, but other corporate leaders can also, through power, significantly shape corporate social outcomes. Third, we contribute to the existing debate on internal drivers of CSR [31,32,33,34] by stressing the importance of each individual’s power in the team that affects the firms’ orientation towards socially responsible practices. Finally, this study adds to the notion of upper echelons theory by emphasizing top management teams and their powerful role in the interpretation of CSR performance. Moreover, from a practical standpoint, this research demonstrates an in-depth investigation of managerial power confirming which source of power is beneficial (or detrimental) in terms of societal welfare.

2. Literature and Hypotheses

We ground this study in the upper echelon theory (UET), which assumes that corporate outcomes—corporate options and performance levels—can be predicted by top manager’s values and cognitions [17]. As the cognitive traits of top managers are not apparent to measure, UET researchers advocate that the observable traits of top managers may instead be used as proxies. However, UET research assumes that managers’ observable characteristics are directly linked with corporate outcomes [35,36] In particular, previous studies examined several demographic traits of top managers (CEOs), including managers’ age, functional background, and educational level, and their impact on corporate options and performance [37,38,39].

As part of a broader research strategy, the top managers’ traits can explain the dimensions of management power and its ability to influence strategic behavior as explained by Finkelstein [19]. As the most influential member of the organization [40,41], a CEO’s power can be classified into structural, ownership, expert, and prestige power. However, this article expands on and divides the power dimensions into three major categories: (1) the power assumed by top managers (CEOs) themselves (structural, ownership, prestige, and expert power), (2) the power assumed by boards (including the board of directors’ supervision power, the board’s political power, the independent directors’ monitoring power, (3) and the largest shareholders’ power.

2.1. CEO Power and CSR

Though empirical findings have been discordant, the existing literature contends that managers’ power significantly influences corporate social behavior. For example, some sources of power can deter social performance, such as structural power that the CEOs derive from hierarchical positions (i.e., duality). Prior studies have claimed that serious corporate governance issues arise if CEOs hold more than single position (such as also holding the chairman position) because the fairness and effectiveness of the board at the time of taking significant decisions can be seriously affected [42,43]. The CEOs’ dual status can limit board independence and can lead to conflicting interests that may result in reduced attention to stakeholder’s interests and thus adversely affect the firm’s accountability for CSR [44,45]. Moreover, the CEOs, who enjoy the influence and supremacy bestowed by the duality, may take decisions at the cost of the environment and society in the wake of complying with the guidelines or instructions of regulatory bodies.

CEOs also grow more powerful when they own a higher fraction of ownership. Earlier research has established that it is obvious that higher ownership may lead to a situation where managers become entrenched [46]. An increase in the CEO’s entrenchment may enable them to overinvest in CSR to gain fame and a reputation [47,48]. Additionally, several studies have witnessed the positive effect of top management’s ownership on CSR which suggests that ownership enables and empowers executives to sustain the stakeholder’s support in the long run through the efficient allocation of resources among them [49,50].

On the other hand, the entrenched CEOs can subdue the board into making decisions about companies’ CSR investment. CEOs with higher ownership may grow egocentric, and thus negatively influence corporate decisions related to CSR [51,52]. Moreover, the CEOs who own a fraction of their companies’ shares usually align their incentives with those of the common shareholders [53], who prefer their economic objectives over the social objectives [54]. On the whole, the existing body of knowledge offers limited evidence on the linkage between ownership powers and CSR. Moreover, the existing literature is divergent in nature suggesting mixed evidence, both against and in favor, on the said relationship. Thus, the limited and divergent literature calls for an in-depth study.

Apart from the above power sources, long tenure has been proven to increase managerial expert power [55,56,57] which also affects managerial behavior. Longer tenures of CEOs point to a higher expertise, experience and achievement level [58] and assist CEOs in coping with a firm’s strategic dilemmas [20]; thus, they are thought of as experts in dealing with a firm’s complexities. Moreover, CEOs with more experience are more capable of realizing the potential benefits of CSR activities [59,60]. This is because they establish knowledge about the characteristics, powers, and situations of different stakeholders, and an association with the focal firm. Therefore, experienced CEOs are in a better position to safeguard the stakeholders’ interests. Moreover, their higher acquaintance with the stakeholders induces inter-organizational and/or interpersonal trust [61]. On the contrary, the less experienced CEOs are usually less familiar with the workplace environment and the managers and are too shy to speak out. Consequently, they may tend to follow rather than take the leading role, predominantly due to lack of expertise, when it comes to coping with the social responsibility and social responsiveness-related issues [47,62]. This shows that the tenure of a CEO may have multiple implications for the social outcomes of a firm.

As to prestige power, UET researchers claim that prestige power is the main feature and personal quality of decision-makers that has a marked impact on the outcomes of strategic behavior [21,36] which may derive from a manager’s educational background [63]. Finkelstein [19] has shown that, within dominant coalitions, the top managers who have superior educational background prove to be more influential. The CSR literature shows that the higher educational level of executives may impart CEOs with a prestige which eventually urges them to be socially responsible [31,64,65]. Different educational levels specify differences in individual’s traits and intellectual base [17,66]. More precisely, education may enhance the perspectives, knowledge and ability of CEOs to conceptualize abstract and technical concepts which in turn shape their strategic decisions including the ones related to CSR. Consequently, it seems that the prestige venues (such as education) of executives are likely to influence the CSR implementation within companies.

In summary, the above reasoning on managerial power concludes that the power of a CEO is a multidimensional concept and is likely to be a function of duality, ownership, tenure, and a high educational level. CEOs may not live up to the expectations of different stakeholders when CEOs grow more powerful through many hierarchical positions. We predict that they will curb social initiatives because those CEOs will arrogate corporate resources for self-interest rather than for the stakeholders’ interest. On the other hand, the above discussion suggests that CEOs with ownership, expert, and prestige power will invest in CSR to maintain their image among societal stakeholders. Thus, the differing nature of each dimension of managerial power and its role in CSR leads us to assume the following hypotheses:

Hypothesis 1(a).

The power (structural power) of CEOs is negatively associated with CSR performance.

Hypothesis 1(b).

The power (including ownership power, expert, and prestige power) of CEOs is positively associated with CSR performance.

2.2. Board Power and CSR

In spite of the importance of exploring the power features of CEOs, UET research admits that emphasis on the whole management team can produce higher outcomes for firms than the conventional emphasis on the CEOs alone [67]. Finkelstein [19] was the first to demonstrate that the power venues of an entire management team are more likely to affect strategic actions when differing amounts of the power of team members are concerned. His methodology for assessing managerial power reminds us that strategic actions (e.g., CSR policies) should not be determined only by the power characteristics of CEOs but also by other decision makers such as the board of directors and shareholders.

The board of directors is the main internal mechanism supervising the management and the implementation of regulations [68]. Boards can influence corporate governance, social behavior, and a firm’s operating efficiency [69]. Board size is one of the main characteristics that puts considerable pressure on the efficacy of the board’s supervision role, since a large board can have more expertise, information, and viewpoints from various sources. When directors are large in number they can be more powerful when controlling the supremacy of influential managers (i.e., CEOs), leaving them as the less dominant figures [70,71]. Besides, larger boards can be more conducive to better participation and are supposed to be more efficient at supervising corporate management, resulting in effective decision making [72]. Because of their larger size, boards may benefit from smooth communication and mutual coordination, as well as better commitment of the members [73,74], which in turn strengthens the firm’s belief in the value of CSR [75,76]. To sum up, the long-term outcomes of CSR are easier to achieve for the firms working under larger boards.

Independence is the most explored attribute of the board. Independent directors have monitoring power de jure [77], and their presence is vital to monitor strategic behavior [53,78]. Further, the anecdotal evidence suggest that the independent status of independent directors from top executives strengthens their monitoring power [12,79]. Moreover, generally they have a higher inclination towards compliance with the rules and the ethical conduct of the firm [47,80]. Consequently, such directors sensibly meet the needs of society [81] and are more concerned about the firm’s ethical aspects than inside directors [82]. Furthermore, most of the independent directors prefer a higher degree of voluntary disclosures and transparency [43,83,84]. Building upon these notions, it can be concluded that the presence of a higher fraction of independent directors on a firm’s board will result in comparatively efficient monitoring and control of management affairs that will significantly enhance the firm’s chances of becoming involved in CSR activities.

In addition to the attribute of monitoring, the board’s political connection is also considered crucial to influence organizational social behavior [85], especially in emerging economies such as China [86]. One of major arguments regarding this is that a top manager’s political connection makes them more powerful [87] and enables them to promulgate government policies [88]. Moreover, such boards are more liable to government since their firms usually enjoy relaxations and preferential treatments in several ways such as: relaxation in regulatory oversight, lighter taxation, privileged treatment by state owned enterprises (e.g., raw material suppliers or banks), and preference in the allotment of government contracts, etc. [89,90,91]. Specifically, larger boards with political connections can approach important information related to social policies and make decisions about social investment using political power [92], as well as enjoy a high institutional back up [93,94], which may also enhance the CSR performance of companies.

Given that boards of directors are major players in corporate governance, their power dimensions can be determined to have significant influence on CSR, which leads to the following hypotheses.

Hypothesis 2.

The board’s power (including the supervision and monitoring of the directors and political power) is positively associated with CSR performance.

2.3. Shareholdes Power and CSR

A critical source of pressure that the top management face regarding social and ethical issues is the presence of the largest shareholders [95]. Shareholders are the salient stakeholders due to their legitimacy and power, and they create urgent issues [96]. The largest shareholders maintain such mechanisms through which they raise their voice to obtain formal responses from the top management [97]. However, it is very likely that large controlling shareholders may use their power to expropriate wealth from minority shareholders. When controlling shareholders take full control of corporate boards and management teams, it is difficult for internal corporate governance to work effectively which may affect a company’s commitment to stakeholders, such as employees, customers, suppliers, and communities.

As to their role in CSR, their ability to expropriate the minority shareholders’ resources may result in corporate decisions that are not complaint with CSR because their tendency towards entrenchment generally increases with an increase in power. Such entrenchments may induce a conflict of interest among the minority and large shareholders which ultimately leads to a lower CSR performance [98]. Furthermore, the shareholders view suggests that the basic purpose of a firm is the realization of the shareholders’ ends provided that these are legitimate and essentially non-deceiving. Practically, those ends are nearly always to extract private benefits (profit maximization) [54,99] since CSR is driven primarily by political and social factors rather than economic considerations. In such a case, engaging in CSR may harm the interest of shareholders, and thus they are supposed to be less prone to CSR. Overall, this discussion leads us to make the following prediction.

Hypothesis 3.

The largest shareholders’ power is negatively associated with CSR performance.

3. Methodology

3.1. Sample

This study used a panel sample from Chinese listed firms. The secondary data, spanning from 2010 to 2018, was collected from the China Stock Market and Accounting Research (CSMAR) database for the estimation of the study variables. On the other hand, the CSR ratings were obtained from the Rankin’s agency. The data obtained from these databases were merged after eliminating the missing firm-year observations of the study variables. Moreover, following Reimsbach, et al. [100] and Zeng, et al. [101], the firm-year observations of the firms with H and B-shares were excluded. These observations were dropped because such firms are regulated by overseas rules and regulations. Consequently, the characteristics and nature of these firms are distinct from those of the Chinese A-share listed firms. Similarly, the financial firms were also excluded due to their different ownership structures, rules and regulations, and accounting standards. The initial sample of A-share listed firms contained 10,524 firm-year observations. However, after dropping the observations of missing values for the study’s variables, we applied a final sample of 9774 firm-year observations.

3.2. Measures

3.2.1. Dependent Variable

The Corporate Social Responsibility (CSR) performance was used as a dependent variable. It was measured on the basis of the CSR score of the Rankin (RKS) agency (an independent agency established in China). Following the approach of the Global Reporting Initiative (GRI), RKS computed a composite CSR score for all listed Chinese firms using 70 indicators from the CSR reports. These indicators were classified into three evaluation groups: (i) overall evaluation items, (ii) technical evaluation items, and (iii) content evaluation items.

The overall evaluation covered 14 items (having 30% weighted average) that predominantly focused on the CSR-related strategies of the firm, the extent and innovativeness of the firm’s social activities, and the external audit evaluations. The technical evaluation, on the other hand, covered 11 items (having 20% weighted average) that focused the availability, clarity, and consistency of the CSR information. The content evaluation covered 45 items (having 50% weighted average) that mainly focused on the special metrics of economic, social, and environmental performance and the role of leadership and the organizational system in implementing CSR. The total CSR score was computed using the weighted average scores of the three sub-groups. The total score ranged between 0 and 100, where a high score represented a superior performance in terms of the constitutions in CSR. The RKS CSR dataset has been used and validated by the prior studies [81,85,102].

3.2.2. Independent Variables

Management control corporate social behavior, and social policies are influenced by the various sources of management power. Therefore, motivated by the methodology used for power classification by Chai and Sikandar Mirza [11], Finkelstein [19], Wang and Liu [103], we used an appropriate measurement of power for best representation of an individual’s ability to have an influence on CSR., and divided the different power dimensions into three groups.

(1) CEO’s power: (i) Structural power (SP); Such a power in an organization is fundamentally determined by its hierarchical structure. Hence, we used a dummy variable for structural power which took a value of 1 as if the Chief Executive Officer (C.E.O.) and chairman of the board of directors were the same person and 0 otherwise. (ii) Ownership power (OP); CEOs possessing shares are more powerful since they are representative of shareholders and management. Ownership power was equal to 1 if a CEO’s equity ownership was more than the sample average and 0 otherwise. (iii) Expert power (EP); EP points to the manager’s capability to manage a firm in a complex environment effectively. We measured this construct as a CEO’s tenure by taking a value of 1 if the tenure of a CEO was longer than the sample average and 0 otherwise. (iv) Prestige power (PP); PP was measured as a CEO’s highest educational degree because with a high level of education, they are perceived to have a high level of prestige. We assigned a value of 5 if CEO had a doctoratee, 4 for a master’s degree, 3 for a bachelor’s degree, 2 for an intermediate level, and 1 for high schooling.

(2) Power accumulated by a board: (i) Board supervision power (BP); the supervision power of the board of directors was measured according to the size of the board. The board can play its supervisory role more efficiently if the size of the board is comparatively large. Hence, we used a dummy variable for BP, which took a value of 1 if the number of the directors on the board was greater than the sample average of directors and 0 otherwise. (ii) Independent director’s monitoring power (MP); MP refers to the capability of independent directors to monitor management behavior independently. The larger the number of independent directors on the board, the more powerful they will be; thus, we created a dummy for MP by assigning value of 1 if the number of independent directors was greater than the industry average and 0 otherwise. (iii) The board’s political power (B_PP); political power is one of essential determinants of CSR in emerging economies (i.e., China) because most of firm’s directors are appointed by a political setup that exerts influence over corporate social policies by using political support. Hence, we used a dummy variable for B_PP, which took a value of 1 if the number of politically connected directors in a firm was greater than the industry average and 0 otherwise.

(3) Shareholder’s power (Holders_P); the largest shareholder has more impetus to participate in a firm’s decision making; thus, we measured shareholders’ power as (the shareholding ratio of the first largest shareholder over the sum of the shareholding ratio of the second and third largest shareholders).

3.2.3. Control Variables

Besides the sources of managerial power, there are various factors which are known to affect CSR that have been established in the literature. For instance, we controlled for the manager’s age (Age), as older managers are considered more experienced to make important decisions related to CSR [47]. The equity ownership by board members may also persuade the firm to engage in social objectives to build the firm’s reputation, so we controlled for board ownership (B_Own), which equaled the proportion of shares owned by board members. Further, previous research has contended that the size of the firm (i.e., small and medium sizes of firms) matters for corporate social outcomes [104]. Some evidence has established that larger and profitable firms have sufficient resources to invest in CSR [105]. Therefore, we controlled for firm size (F_Size) which was measured by the natural log of assets, growth (Grow), estimated by change in total assets, and return on assets (ROA) measured as net profit divided by total assets. It has previously been noted that older firms tend to take less part in social responsibilities than younger firms do [85,102]; thus, our study added firm age (F_Age) estimated as the total years since the firm was established. Similarly, firms under financial stress may also behave differently and are more likely to reduce CSR to save resources. Thus, the ratio of total debt to total assets was used to control for firm leverage. Moreover, we considered the book to market value of shareholders’ equity (BTMA) to estimate the effect of growth opportunities [106].

3.3. Descriptive Statistics and Pairwise Correlation

Table 1. Descriptive Statistics (N = 9774).

Variable	Mean	Std. Dev.	Min	Max
CSR	29.707	19.023	0	90.84
SP	0.253	0.435	0	1
OP	0.42	0.494	0	1
EP	0.452	0.498	0	1
PP	3.523	0.833	1	5
BP	0.59	0.492	0	1
MP	0.421	0.494	0	1
B_PP	0.452	0.498	0	1
Holders_P	8.107	17.097	0.5	50.444
Age	51.112	3.771	36.846	66
B_Own	0.107	0.189	0	0.892
F_Size	3.091	0.059	2.622	3.35
Grow	0.838	1.911	−9.109	64.697
ROA	0.047	0.251	−6.776	22.005
F_Age	10.435	6.307	1	26
Lev	0.448	0.365	0.007	16.545
BTMA	0.527	0.253	0	1.464

highlights the descriptive statistics of this study’s variables with 9774 firm-year observations. The summary of statistics shows that CSR ratings varied from 0 to 90.84. The average CSR rating was 29.707 with the standard deviation of 19.023, which is indicative of substantial variations in CSR scores across Chinese firms.

Concerning power’s variables, the average of SP (0.253) showed that on average, 25% of the sample firms had CEOs with structural power that maintained a dual hierarchical position. The mean of OP (0.42) shows that 42% of CEOs hold ownership greater than the sample average. The mean of EP was (0.452), confirming that 45% of the CEOs in the sample firms acquired expert power by staying on in jobs for longer than the sample average. The mean of PP (3.523) specifies that on average, CEOs hold a Master’s degree. Regarding the other sources of power, the mean of BP (0.59) shows that 59% firms have a board size larger than the sample average representing the board’s supervision power. Besides, the mean values of MP and B_PP scored an average value of (0.421) and (0.452), which indicates that 42% and 45% firms have boards with independent directors that hold monitoring power and political power (respectively). Moreover, Holders_P secured a mean value of (8.107), demonstrating that on average, the first largest shareholders kept a ownership that was 8 times larger than the sum of the share ownership of the second and third largest shareholders in a firm, which highlights their power in decision making. The descriptive statistics of some power variables from which we originally created dummy variables (OP, EP, BP, MP, B_PP) are given in Appendix A (

Table A1. Descriptive statistics of power sources.

Variable	Mean	Std. Dev.	Min	Max
CEO_Own	0.042	0.108	0	0.805
CEO_Tenure	3.454	2.936	0	19
Board_Size	10.217	2.587	5	26
Ind_Dir	3.881	1.231	2	13
PC_Dir	2.609	1.875	0	16

Note: This table presents summary statistics for the continuous variables from which we created dummy variables. CEO_Own indicates ownership power (OP) and its average value (0.042) suggests that on average, a CEO holds 4% ownership in a firm. CEO_Tenure refers to expert power (EP) and the mean value (3.454) tells us that the average tenure of CEOs is about 3 years. Board_Size has been used to scale a board’s supervision power (BP), and the mean value (10.217) suggests that on average a firm maintains 10 directors on the board. Ind_Dir indicates the number of independent directors and has been used to measure monitoring power (MP). Its mean value shows that on average a firm has 4 independent directors on the board. Similarly, PC_Dir refers to the number of politically connected directors which has been used to create a dummy for indicating political power. The mean value (2.609) suggests that on average 3 directors on a board maintain political connections.

).

Table 2. Pairwise correlations.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)	(16)	(17)
(1) CSR	1.000
(2) SP	−0.087 *	1.000
(3) OP	0.016 *	0.231 *	1.000
(4) EP	0.059 *	0.080 *	0.164 *	1.000
(5) PP	0.135 *	−0.032 *	−0.070 *	0.020	1.000
(6) BP	0.041 *	0.083 *	0.082 *	0.073 *	−0.077 *	1.000
(7) MP	0.075 *	−0.052 *	−0.043 *	−0.034 *	0.077 *	−0.649 *	1.000
(8) B_PP	0.056 *	−0.023 *	−0.007	0.001	0.061 *	−0.214 *	0.224 *	1.000
(9) Holders_P	−0.029 *	−0.098 *	−0.153 *	−0.022 *	0.033 *	−0.017	0.019	−0.012	1.000
(10) Age	0.103 *	−0.106 *	−0.087 *	0.047 *	0.097 *	−0.064 *	0.128 *	0.178 *	0.073 *	1.000
(11) B_Own	−0.095 *	0.264 *	0.493 *	0.026 *	−0.143 *	0.126 *	−0.075 *	−0.069 *	−0.167 *	−0.196 *	1.000
(12) F_Size	0.394 *	−0.172 *	−0.105 *	0.006	0.175 *	−0.165 *	0.181 *	0.147 *	0.138 *	0.297 *	−0.297 *	1.000
(13) Grow	0.006	0.050 *	0.038 *	−0.053 *	−0.007	−0.023 *	0.024 *	0.013	−0.055 *	−0.062 *	0.059 *	0.041 *	1.000
(14) ROA	0.056 *	0.025 *	0.021 *	0.000	−0.007	0.012	−0.008	−0.008	−0.018	−0.021 *	0.024 *	−0.028 *	0.019	1.000
(15) F_Age	0.061 *	−0.228 *	−0.349 *	−0.022 *	0.116 *	−0.083 *	0.062 *	0.003	0.170 *	0.144 *	−0.554 *	0.265 *	−0.047 *	−0.024 *	1.000
(16) Lev	−0.002	−0.091 *	−0.149 *	−0.028 *	0.051 *	−0.070 *	0.060 *	0.028 *	0.063 *	0.036 *	−0.228 *	0.191 *	0.018	−0.160 *	0.271 *	1.000
(17) Grow_Opp	0.188 *	−0.166 *	−0.153 *	0.001	0.055 *	−0.110 *	0.091 *	0.099 *	0.162 *	0.174 *	−0.278 *	0.627 *	−0.071 *	−0.077 *	0.240 *	0.232 *	1.000

* indicates significance level at 5%.

shows the results of the pairwise correlation for all variables. The correlation’s results provide preliminary support to the main claim of this study by showing that all power sources are positively correlated to CSR except structural power (SP) and shareholder’s power (Holders_P). Overall, this initially recommends that SP and Holders_P may deter corporate social objectives.

3.4. Empirical Model

As this study used panel data, we applied the Ordinary Least Square (OLS) technique as a baseline method to estimate the effect of the heterogeneous dimensions of power on CSR. However, the one of limitation of OLS model is that it does not account for omitted factors. In such a case, one may argue that the estimated outcomes may not necessarily be due to power dimensions, they can be due to time-invariant factors. To address this bias, we also estimated the fixed effect model to validate the baseline estimations. In addition, we also utilized the Two Stage Least Square (2-SLS) method and propensity score matching technique as an appropriate remedy to tackle possible endogeneity issues.

We employed the following equation using the OLS model to investigate the link between power heterogeneity and CSR:

$${\mathrm{CSR}}_{\mathrm{it}}=\mathsf{\alpha }+{{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{n}}{\mathsf{\beta }}_{\mathrm{n}}{\mathrm{Power}}_{\mathrm{it}}+{{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{n}}{\mathsf{\beta }}_{\mathrm{n}}{\mathrm{Control}}_{\mathrm{it}}+{ \mathsf{\epsilon }}_{\mathrm{it}}$$

(1)

As illustrated in Section 3.2.1 and Section 3.2.2, CSR denotes the social performance measured by Rankin’s CSR ratings and Power represents different sources of managerial power accumulated by CEOs and organizations. In addition to power variables, Control indicates different firm-level factors associated with CSR.

4. Results and Discussion

4.1. Main Regression Analysis of the Power of CEOs

Table 3. OLS regression analysis of the power dimensions of CEOs (n = 9774).

Variables	Model 1	Model 2	Model 3	Model 4
SP	−0.850 **
	(−2.106)
OP		2.220 ***
		(5.623)
EP			1.909 ***
			(5.691)
PP				0.878 ***
				(3.356)
Age	0.050	0.054	0.041	0.119
	(1.039)	(1.126)	(0.860)	(1.531)
B_Own	1.981 *	0.796 *	1.552 *	3.899 ***
	(1.760)	(1.663)	(1.910)	(3.120)
F-Size	188.656 ***	186.884 ***	188.689 ***	195.485 ***
	(45.039)	(44.533)	(45.146)	(36.293)
Grow	−0.315 ***	−0.324 ***	−0.296 ***	−0.243 **
	(−3.532)	(−3.646)	(−3.328)	(−2.027)
ROA	3.074 ***	3.047 ***	3.064 ***	2.609 ***
	(4.561)	(4.529)	(4.553)	(3.699)
F_Age	−0.032	−0.004	−0.023	0.051
	(−0.929)	(−0.104)	(−0.685)	(1.160)
Lev	−3.033 ***	−2.966 ***	−2.986 ***	−4.578 ***
	(−6.121)	(−5.992)	(−6.034)	(−6.956)
Grow_Opp	−19.990 ***	−19.607 ***	−19.890 ***	−21.843 ***
	(−19.319)	(−18.953)	(−19.256)	(−16.659)
Constant	−548.147 ***	−544.151 ***	−548.923 ***	−575.496 ***
	(−43.979)	(−43.659)	(−44.180)	(−36.237)
R-squared	0.254	0.256	0.256	0.279
Industry	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes

Note: This table highlights the findings for the power dimensions of CEOs’. SP, OP, EP, and PP indicate structural, ownership, expert, and prestige power, respectively. *, **, *** show 10%, 5%, and 1% significance levels, respectively.

reports the estimation results of different types of power accumulated by top managers (CEOs) themselves. Model 1 presents the results for the structural power (SP) of CEOs in which the coefficient on SP (β = -0.850, p < 0.05) declares a negative and significant relationship with CSR. This confirms that CEOs with structural power (dual hierarchical position) devote less attention to social responsibilities. Model 2 highlights that the coefficient value of ownership power (OP) is positive and highly significant (β = 2.220, p < 0.01), suggesting that CEOs through ownership power may positively influence CSR performance. Similarly, the statistically significant and positive coefficient on expert power (EP) (β = 1.909, p < 0.01) in Model 3 suggests a positive association of EP with CSR. It submits that a longer tenure may enhance the expertise and abilities of CEOs to evaluate the benefits of being socially responsible. Finally, Model 4 shows the results for prestige power (PP), showing that PP is significantly and positively (β = 0.878, p < 0.01) related to corporate social performance, which confirms that high educational careers encourage CEOs to behave in socially responsible manner.

Overall, these findings advocate that firms wherein CEOs maintain structural power may show a lower level of social performance to satisfy a wider range of stakeholder interest. The prior finance literature has demonstrated that weaker corporate governance and agency conflicts primarily result from the dual job status of CEOs, which leads to undesirable outcomes [45,107]. Our findings also suggest that CEOs with ownership power, expert power, and prestige power demonstrate a high level of CSR performance. The possible reason could be that CEOs with such powers care much more about their reputation and future career opportunities. However, among the various important channels that powerful and entrenched CEOs can utilize, CSR activities may play significant role in building their career and reputation.

4.2. Main Regression Analysis of Board Power and the Largest Shareholder’s Power

Table 4. OLS regression analysis of board power and shareholder’s power (n = 9774).

Board Power Dimensions				Shareholder’s Power
Variables	Model 1	Model 2	Model 3	Model 4
BP	0.901 ***
	(2.604)
MP		0.716 **
		(1.973)
B_PP			0.623 *
			(1.725)
Holders_P				−0.024 **
				(−2.411)
Age	0.054	0.242	0.237	0.056
	(1.145)	(0.5044)	(0.544)	(1.176)
B_Own	1.488 **	0.593 ***	0.640 ***	1.563 *
	(1.991)	(4.843)	(4.706)	(1.698)
F_Size	19.272 ***	14.774 ***	16.796 ***	18.253 ***
	(45.175)	(29.818)	(30.100)	(45.213)
Grow	−0.317 ***	−0.039**	−0.041**	−0.331 ***
	(−3.555)	(−2420)	(−2.437)	(−3.711)
ROA	3.049 ***	4.159 ***	4.163 ***	3.041 ***
	(4.525)	(5.874)	(5.880)	(4.514)
F_Age	−0.025	0.002	0.006	−0.019
	(−0.727)	(0.064)	(0.180)	(−0.541)
Lev	−2.997 ***	−3.403 ***	−3.375 ***	−3.049 ***
	(−6.047)	(−6.521)	(−6.472)	(−6.154)
Grow_Opp	−19.962 ***	−3.837 ***	−3.899 ***	−19.799 ***
	(−19.300)	(−4.022)	(−4.083)	(−19.116)
Constant	−554.110 ***	−511.498 ***	−519.439 ***	−550.346 ***
	(−44.136)	(−37.140)	(−32.105)	(−44.227)
R-squared	0.254	0.275	0.257	0.254
Industry	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes

Note: This table highlights the findings for the organizations’ power dimensions. BP, MP, B_PP, and Holders_P suggest board supervision, monitoring, political, and shareholder power, respectively. *, **, *** show 10%, 5%, and 1% significance levels, respectively.

reports the analysis of the power sources assumed by an organization. Model 1 in Table 4 tests the relationship between a board’s supervision power (BP) and CSR performance, highlighting that board supervision power significantly and positively ((β = 0.901, p < 0.01) affects social performance. The coefficient of the monitoring power of independent directors (MP) also has a positive and significant (β = 0.716, p < 0.05) influence on CSR. It seems that a greater number of independent directors increases their monitoring power in a firm within which they effectively monitor corporate social behavior. Similarly, the coefficient of the board’s political power (B_PP) confirms that B_PP generates a positive impact (β = 0.623, p < 0.10) on CSR.

However, Model 4 represents the findings of the largest shareholder’s power (Holders_P) showing that it has a negative effect on CSR. The negative coefficient on Holders_P (β = −0.024, p < 0.05) is statistically significant at a 5% level of significance. This negative relationship may imply that it is very likely that the largest controlling shareholders may use their power to exploit the corporate resources for private interest rather than for broader social objectives.

Our findings on the power dimensions of the board suggest that the board’s power variables including the board’s supervision power, monitoring power, and political power are beneficial to achieve broader social objectives. In particular, boards that are formed of a larger number of directors have greater communication and supervision power (BP) [72,73], which leads to better governance according to society’s expectations. Therefore our results show that board supervision power (BP) contributes towards corporate social agendas and thus improves CSR performance. As to monitoring power (MP), it must be documented that a higher representation of independent directors boosts their monitoring power which signals the fulfillment of stakeholders’ demands. Highly independent boards with monitoring power effectively control management decisions by exercising greater surveillance that may also influence CSR performance. Similarly, a higher presence of politically connected directors on a board indicates a superior power (B_PP) to reflect state policies (i.e., CSR policies), especially in an economy with a socialist tradition such as China [85,108]. This is because the government provides strong backup to politically connected firms, but in return, such firms have to comply with government rules and regulations in general and ensure CSR objectives in particular.

Moreover, our findings suggest that the largest shareholder’s power (Holders_P) can discourage CSR investment. This may be due to the fact that they retain dominant voting and decision-making power as compared to minor shareholders, who lack the legal coverage to affect managerial decisions. Hence, the largest ownership encourages shareholders to seek personal benefits [109,110] that negatively impacts the organization’s commitment to social activities and thus can be detrimental to society at large.

Regarding the control variables, we found that all variables had a significant impact on CSR except the manager’s age (Age) and the firm age (F_Age). In particular, B_Own is positively associated with CSR as board owners face substantial pressure from the institutional environment. Similarly, firm size (F_Size) and profitability (ROA) have a positive relationship with CSR because larger and more profitable firms obtain more attention from stakeholders. Moreover, such firms are considered resourceful to invest in CSR. Looking at other firm-level variables, our study found that a firm’s growth (Grow), leverage (Lev), and growth opportunities (Grow_Opp) have a negative impact on social performance. Particularly, the coefficient of Lev suggests that with a higher level of leverage, firms will be less likely to make an investment in CSR. Similarly, the higher growth opportunities may allow top management to be less resistant to societal demands, preferring the allocation of resources in profitable projects.

4.3. Endogeneity

Our OLS model findings on the association of power sources and social performance may be biased due to possible endogeneity matters. To resolve such issues, this research used three alternative statistical approaches for the sake of checking the reliability and validity of our research findings.

First, following Conyon and He [111], Luo, et al. [112] and Ali, et al. [113], we used fixed effects estimations to address the issue of unobserved factors. For example, one of the major issues in the estimation of panel datasets is the possibility of unobservable time-invariant characteristics. In such a case, a fixed effects model has the advantage over other OLS estimations in terms of explicitly modeling features that are stable over time which may help in mitigating cross sectional variations.

However,

Table 5. Robustness test using the fixed effect model.

Column-1					Column-2
CEO Power Dimensions			Board Power Dimensions				Shareholder Power
Variables	SP	OP	EP	PP	BP	MP	B_PP	Holders_P
SP	--0.889 **
	(−1.986)
OP		1.851 ***
		(4.125)
EP			1.398 ***
			(4.035)
PP				1.233 ***
				(4.003)
BP					0.902 **
					(2.524)
MP						0.602 *
						(1.678)
B_PP							0.639 *
							(1.705)
Holders_P								−0.019 *
								(−1.764)
Age	0.041	0.037	0.029	0.105 *	0.045	0.229 ***	0.222 ***	0.053
	(0.859)	(0.783)	(0.615)	(1.722)	(0.943)	(4.597)	(4.414)	(1.106)
B_Own	2.377 **	1.262	2.187 *	5.116 ***	2.270 **	1.224	1.208	1.550
	(2.095)	(0.989)	(1.941)	(3.979)	(1.989)	(1.032)	(1.019)	(1.386)
F_Size	18.185 ***	15.587 ***	17.566 ***	12.718 ***	19.531 ***	4.714 ***	4.754 ***	18.105 ***
	(44.903)	(44.178)	(44.808)	(35.525)	(44.970)	(29.345)	(29.785)	(45.162)
Grow	−0.309 ***	−0.309 ***	−0.287 ***	−0.207 *	−0.305 ***	0.053	0.055	−0.330 ***
	(−3.472)	(−3.471)	(−3.231)	(−1.727)	(−3.426)	(0.572)	(0.590)	(−3.707)
ROA	3.103 ***	3.040 ***	3.060 ***	2.665 ***	3.091 ***	4.161 ***	4.153 ***	3.043 ***
	(4.606)	(4.522)	(4.551)	(3.782)	(4.589)	(5.882)	(5.870)	(4.516)
F_Age	−0.040	−0.024	−0.040	0.010	−0.046	−0.016	−0.009	−0.022
	(−1.178)	(−0.693)	(−1.164)	(0.225)	(−1.331)	(−0.433)	(−0.246)	(−0.645)
Lev	−3.060 ***	−2.958 ***	−2.994 ***	−4.628 ***	−3.009 ***	−3.365 ***	−3.388 ***	−3.046 ***
	(−6.176)	(−5.984)	(−6.055)	(−7.039)	(−6.075)	(−6.453)	(−6.501)	(−6.148)
Grow_Opp	−20.082 ***	−19.608 ***	−20.067 ***	−22.361 ***	−20.207 ***	−4.174 ***	−4.176 ***	−19.833 ***
	(−19.403)	(−18.975)	(−19.424)	(−16.991)	(−19.505)	(−4.362)	(−4.363)	(−19.144)
Constant	−404.312 ***	−398.582 ***	−404.852 ***	−456.842 ***	−408.771 ***	−401.362 ***	−401.830 ***	−405.602 ***
	(−27.366)	(−27.296)	(−27.908)	(−29.036)	(−27.296)	(−25.176)	(−27.296)	(−27.940)
R-squared	0.276	0.271	0.278	0.291	0.274	0.271	0.271	0.273
Years	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes

Note: *, **, *** show 10%, 5% and 1% significance levels respectively.

highlights the results of the fixed effect models. Specifically, Column 1 of Table 5 indicates that all the power sources of CEOs (OP, EP, and PP) are significantly and positively related to CSR except their structural power (SP) which shows a negative impact on CSR. Column 2 of Table 5 shows the results for board power and shareholder’s power, which indicate that an organization’s power variables (including BP, MP, and B_PP) positively influence CSR, but the shareholder’s power (Holders-P) has a negative effect on CSR. Overall, the regression models using fixed effects provide results consistent with the baseline analysis, confirming that our conclusions were not affected by the omitted-variable bias.

Second, we employed an instrumental variable approach known as two-stage least squares (2-SLS) as a standard remedy to address endogeneity. There is a probability that socially responsible or irresponsible organizations appoint and retain powerful managers at management. This study explored the probability of an issue of reverse causality by incorporating an endogenous choice of power variables. A 2-SLS estimation requires finding a valid instrumental variable which must be specifically correlated to an independent variable but should not affect the dependent variable [114,115]. Consistent with Bebchuk, Cremers, and Peyer [8], and Jiraporn and Chintrakarn [10], we utilized the industry average of power variables as it only corresponds to the power of a particular firm but has no relation with a firm’s CSR performance. In the first stage of 2-SLS, we regressed the power variables on the industry average of the manager’s power along with the control variables. In the first stage, we found that the coefficients of the industry average (power variables) were positive and significant. However, the results obtained from the first stage are not reported for the sake of parsimony. In the second stage, we used the fitted values of the first stage for power and considered them as independent variables in the second stage of regressions.

Table 6. Endogeneity test using the Two Stage Least square (2-SLS) model.

Column-1					Column-2
(CEO Power Dimensions)			Board Power Dimensions			Shareholder Power
Variables	SP	OP	EP	PP	BP	MP	B_PP	Holders_P
SP	−45.166 **
	(−7.470)
OP		3.373 ***
		(5.296)
EP			16.575 ***
			(3.136)
PP				10.926 ***
				(5.388)
BP					155.189 *
					(1.878)
MP						7.266 ***
						(2.958)
B_PP							1.861 ***
							(2.717)
Holders_P								−0.265 *
								(−1.676)
Age	0.003	0.009	−0.005	0.077	0.009	0.225 ***	0.237 ***	0.010
	(0.063)	(0.186)	(−0.105)	(1.269)	(0.184)	(4.424)	(4.706)	(0.215)
B_Own	2.131 *	−0.578	1.621	3.861 ***	1.560	−0.341	0.640	1.637
	(1.892)	(−0.480)	(1.452)	(3.092)	(1.394)	(−0.286)	(0.544)	(1.463)
F_Size	17.750 ***	16.599 ***	18.992 ***	19.995 ***	18.490 ***	3.697 ***	4.796 ***	18.503 ***
	(45.373)	(45.068)	(45.540)	(37.006)	(45.585)	(24.267)	(30.100)	(45.585)
Grow	−0.295 ***	−0.302 ***	−0.276 ***	−0.214 *	−0.297 ***	0.151	0.041	−0.311 ***
	(−3.297)	(−3.385)	(−3.094)	(−1.780)	(−3.322)	(1.597)	(0.437)	(−3.482)
ROA	3.097 ***	3.061 ***	3.082 ***	2.639 ***	3.066 ***	4.351 ***	4.163 ***	3.060 ***
	(4.560)	(4.513)	(4.544)	(3.716)	(4.515)	(6.002)	(5.880)	(4.505)
F_Age	0.033	0.059 *	0.042	0.121 ***	0.040	0.123 ***	0.006	0.047
	(0.977)	(1.764)	(1.262)	(2.819)	(1.209)	(3.464)	(0.180)	(1.411)
Lev	−2.988 ***	−2.931 ***	−2.939 ***	−4.645 ***	−2.955 ***	−2.755 ***	−3.375 ***	−3.006 ***
	(−6.030)	(−5.921)	(−5.938)	(−7.066)	(−5.962)	(−5.200)	(−6.472)	(−6.068)
Grow_Opp	−18.292 ***	−17.992 ***	−18.185 ***	−20.587 ***	−18.276 ***	2.896 ***	−3.899 ***	−18.083 ***
	(−18.182)	(−17.895)	(−18.108)	(−16.184)	(−18.172)	(3.254)	(−4.083)	(−17.954)
Constant	−319.525 ***	−506.636 ***	−423.502 ***	−434.492 ***	−447.171 ***	−393.162 ***	−437.983 ***	−406.906 ***
	(−14.401)	(−32.025)	(−24.280)	(−24.561)	(−21.096)	(−23.806)	(−26.581)	(−26.122)
R-squared	0.102	0.133	0.103	0.115	0.144	0.141	0.135	0.132
Ind and Years	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes

Note: *, **, *** show 10%, 5% and 1% significance levels respectively.

presents the results of the second stage of 2-SLS estimations. Consistent with the OLS and fixed effect models, Column 1 of Table 6 shows that CEOs with ownership, expert, and prestige power invest more in CSR, but CEOs who acquire structural power tend to reduce CSR investment. Column 2 of Table 6 indicates that the board’s supervision, monitoring, and political power motivates organizations to demonstrate better CSR performance, while shareholder’s power discourages organizations from pursuing CSR objectives. To sum up, the 2-SLS results lend support to our main analysis suggesting that the conclusion of this study appears to be robust to the possible endogeneity.

Finally, it is also probable that our main conclusion may be biased due to self-selection concerns. For instance, the characteristics of a more socially accountable and a less socially responsible organization might vary, which can lead to dissimilar conclusions for organizations to be socially responsible. In this scenario, organizations are referred as responsible due to organizational attributes, rather than because of managerial power dimensions. By considering this problem, our study adopted a dummy variable approach in which we assigned one if an organization acquired CSR scores greater than the industry median, and zero otherwise, and matched the organizations based on the control variables by employing a propensity score matching the (PSM) technique.

Table 7. Endogeneity test using the Propensity Score Matching model.

Column-1					Column-2
(CEO Power Dimensions)			Board Power Dimensions				Shareholder Power
Variables	SP	OP	EP	PP	BP	MP	B_PP	Holders_P
SP	−4.026 **
	(−3.401)
OP		2.291 ***
		(3.246)
EP			6.115 ***
			(4.036)
PP				1.016 ***
				(6.078)
BP					125.971 *
					(1.738)
MP						4.056 ***
						(3.358)
B_PP							1.004 ***
							(4.407)
Holders_P								−1.155 *
								(−1.806)
Age	1.158	5.224	1.008	0.007 ***	2.008	1.006	0.005 **	0.010
	(0.650)	(0.670)	(0.186)	(3.910)	(1.251)	(1.230)	(1.970)	(0.215)
B_Own	−0.006	0.007 ***	−0.013	−0.016 *	−0.025 *	0.005	−0.030 **	1.637
	(−0.630)	(3.700)	(−1.120)	(−1.730)	(−1.823)	(0.580)	(−2.500)	(1.463)
F_Size	−0.012 ***	0.006	−0.056	−0.064	−0.054	−0.067	0.012	0.006 ***
	(−2.910)	(0.740)	(−0.600)	(−0.980)	(0.586)	(−1.020)	(0.120)	(3.240)
Grow	0.272 ***	−0.074	5.543 ***	1.248 ***	5.475 ***	4.114 ***	4.603 ***	0.005
	(13.968)	(−1.140)	(8.690)	(6.200)	(8.600)	(10.640)	(6.310)	(0.590)
ROA	0.011	1.148 ***	−0.036	−0.017 **	−0.035	−0.034 *	0.061 **	−0.071
	(0.960)	(5.600)	(−1.370)	(−1.029)	(1.330)	(−1.840)	(1.980)	(−1.080)
F_Age	−0.956 ***	−0.006	0.010 *	−0.008 **	−0.010 *	−0.013 ***	0.009	4.117 ***
	(−6.820)	(−0.330)	(1.840)	(−2.160)	(−1.800)	(−2.990)	1.510	(10.660)
Lev	0.064 **	−0.012 ***	0.234 ***	0.208 ***	0.234 ***	0.246 ***	12.993 ***	−0.034 *
	(2.590)	(−2.870)	(7.370)	(12.851)	(7.380)	(12.400)	(13.140)	(−1.850)
Grow_Opp	0.397 ***	0.268 ***	0.025	0.011	0.019 ***	0.014	0.009	−0.013 ***
	(9.27)	(13.710)	(1.320)	(0.940)	(9.976)	(1.200)	(0.430)	(−2.950)
Constant	−1.121 **	−506.636 ***	−423.502 ***	−434.492 ***	−447.171 ***	−393.162 ***	−437.983 ***	−406.906 ***
	(−2.390)	(−32.025)	(−24.280)	(−24.561)	(−21.096)	(−23.806)	(−26.581)	(−26.122)
R-squared	0.102	0.133	0.103	0.115	0.144	0.141	0.135	0.132
Ind and Years	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes

Note: *, **, *** show 10%, 5% and 1% significance levels respectively.

reports the findings about the propensity score matching that indicates all the findings were consistent with the main evidence of this study. Moreover, the PSM results were similar to the alternative methodologies, which shows that our main conclusion was not driven by firm characteristics.

5. Implications

The findings of our study have some important implications in relation to theory and practice.

5.1. Theoretical Implication

The rationale for our study is found in the upper echelon theory. In line with UET, the findings of our study corroborate that the power characteristics of the upper echelons have a fundamental impact on corporate strategic choices, such as CSR investment. One of the significant implications, particularly for emerging economies, is that because emerging markets such as China are transitioning to a market-oriented economy, Chinese companies must deal with a high level of unpredictability in the institutional environment. Therefore, from the viewpoint of the upper echelon theory, in this situation, more powerful CEOs and boards of directors may aid corporations to cope better with environmental and societal uncertainties. In addition, as the upper management’s decision-making power grows, the probability of extreme performance grows as well. As a result, the greater the power of the CEO and the board of directors, the more flexible the firm’s performance will be. Consistent with this paradigm, our empirical evidence demonstrates that it is essential to consider the influence of not only the CEO’s power dimensions on CSR policies, but also the other top managers, such as the board of directors, etc.

5.2. Practical Implication

From a practical perspective, our study can help practitioners to understand which source of power is beneficial to direct a firm’s social behavior. In this vein, our findings corroborate that a higher structural power and controlling power of the largest shareholders may restrict management to make investments in CSR; thus, an increase in structural power and the concentration of shares in one of the shareholders are not advisable from a broader societal perspective. On the other hand, our study suggests that CEOs with higher ownership, longer tenures, and the highest educational backgrounds should be more socially responsible which may augment the social performance of firms. Similarly, this study stresses the need for large, independent, and politically connected boards that may function well as to effective monitoring and controlling and better supervision and surveillance to satisfy stakeholders’ interests. Overall, from a practical standpoint, this research represents an in-depth investigation of managerial power confirming which source of power is beneficial (or detrimental) in terms of societal welfare.

6. Conclusions

Prior studies on power have devoted much attention to explore the impact of the power sources of CEOs. The management in most of the organizations is viewed as a shared platform in which a set of top managers also hold the responsibility to devise and implement corporate policies. Thus, it is necessary to assess the societal impacts of power of the entire management team, which has been specifically underexplored. From this perspective, we examined the impact of managerial power heterogeneity (power accumulated by CEOs, and the board of director’s power bestowed by the organizations) and the largest shareholder’s power on CSR performance.

Using a panel sample from the largest emerging market, China, our study found that the structural power of CEOs lessens their tendency towards CSR activities, while other power dimensions such as ownership, expert, and prestige power prepares them to embrace social responsibilities. Besides, we found that board power dimensions, board supervision power, monitoring power, and political power improve CSR performance, whereas shareholder power tends to reduce a firm’s proclivity towards CSR. Our findings conclude that apart from the power of CEOs which ensures corporate outcomes, the other team members also significantly attain organizational outcomes through power.