Empirical Analysis of the Impact of Top Management Team Social Networks on the Homophily Effect of ESG Disclosure in Companies

Abstract

This study investigates the homophily effect in corporate information disclosure, specifically focusing on executive social networks. We analyze data from 385 privately listed companies in China’s Growth Enterprise Market between 2018 and 2021. An OLS regression model is employed to examine the presence of a homophily effect in ESG information disclosure by private enterprises, along with regional and industry variations. Additionally, we utilize a moderation effect model to assess the influence of executive social networks on the homophily effect of ESG information disclosure. We conduct robustness tests based on our findings. The results indicate a significant homophily effect in ESG information disclosure by private enterprises, with varying magnitudes across regions and industries. Furthermore, executive social networks positively moderate the homophily effect, suggesting that a more diverse social network among the executive team enhances the homophily effect of ESG information disclosure for private enterprises. These findings offer valuable insights for corporate low-carbon sustainable development.

1. Introduction

The “China Listed Companies ESG Development White Paper (2021)” reveals that as of June this year, over a thousand listed companies have disclosed ESG-related information, accounting for approximately one-fourth of all listed companies. This is a significant increase compared to only 317 companies in 2009. Notably, out of the constituents of the CSI 300 Index, 248 companies have published annual ESG reports, accounting for over 82%. The future focus is on strengthening the disclosure of ESG information. The “Guidelines for Corporate Governance of Listed Companies” in 2018 provided a significant description of its content framework. In 2021, the China Securities Regulatory Commission listed “enterprises should consciously provide relevant information to investors and fulfil corresponding obligations to protect their legitimate rights and interests from infringement” as one of the essential provisions.

ESG encompasses responsibilities in the areas of environment, society, and corporate governance. ESG information disclosure refers to the actions taken by companies to reduce market information asymmetry, enhance the company’s social reputation, improve social relationships, and meet the information demands of investors and other stakeholders. It involves the specific disclosure of information related to the company’s environment, society, and corporate governance. These are obligations and responsibilities that the company should fulfil throughout an accounting year through paper or online reports [1]. In the process of corporate management, the homophily effect refers to the phenomenon where a particular activity undertaken by one or more companies in a group leads to the simultaneous adoption of the same activity by other companies, resulting in imitation and convergence of behavior among companies. The concept of social networks refers to the relationship network established through communication and contact among individuals within a team. Social networks utilize interpersonal communication and contact to control social activities. In a company, senior executives can use social information to formulate the company’s development plans and corresponding business policies and increase public awareness [2].

In recent years, the changes in the behavior of companies within the same group, especially within the same industry, have attracted extensive attention due to the emergence of homophily theory in social psychology. This has provided new directions for research on corporate information disclosure. However, current research lacks an in-depth exploration of the impact of executive information acquisition channels on the homophily effect of ESG information disclosure by companies. This is primarily due to the difficulty in accurately measuring executive information acquisition channels. Social network analysis provides a new approach for this research, as information dissemination relies on diverse social networks, which broaden the executive’s information acquisition channels [3]. Therefore, the number of decision-maker social network channels and the company’s ability to acquire information are closely related, resulting in a strong degree of dependence on homophilic companies. Companies utilize social networks to acquire information and reflect it in their ESG information disclosure decisions. This study examines privately listed companies in the Growth Enterprise Market from 2018 to 2021 and conducts a comprehensive analysis of the social network information of relevant executive teams. It considers aspects such as the number of concurrent appointments, educational background, positions as political advisors or representatives, banking work experience, and association positions [2]. Subsequently, relevant models are established to explore whether there is a homophily effect in ESG information disclosure by sample companies, and regression methods are used to determine the relationship between the two.

This article’s innovations are as follows: First, existing literature mostly analyzes executive social networks or the homophily effect of information disclosure separately, with limited research on the influencing factors of the homophily effect of ESG information disclosure by companies. This article enriches the research on the homophily effect of corporate behavior by constructing a research model that examines the impact of mimetic behavior on the level of ESG information disclosure by companies. Second, evaluating the social network of executive teams based on five components expands the depth and breadth of research on corporate social networks. The research findings indicate that these five factors influence the social network, demonstrating the feasibility of this approach. Therefore, the conclusions of this study have practical implications for privately listed companies in terms of ESG information disclosure. Third, analyzing the homophily effect from the perspectives of regions and industries further expands the relevant research topics of the homophily effect.

2. Literature Review and Research Hypotheses

2.1. Mechanisms of Homophily in Information Disclosure Behavior

In a specific group, individuals are influenced by the behavior of other group members, leading to a change in their own behavior to align with others. Homophily studies the relationship between individual characteristics and the average characteristics of group members [4]. Similar phenomena, such as the “herding effect” [5,6], the “contagion effect” [7], and the “surge effect” [8], have been observed and explained. The presence of homophily within groups has been revealed through studies on interactions and mutual influences among individuals. Incorporating group influence into the interaction between individuals and the market expands the theoretical framework of classical economics and has important theoretical and practical implications. Traditional studies in corporate finance often overlook the influence of other companies on decision making and treat financial strategies as independent decisions. However, research on interactions and mutual influences among individuals has shown the presence of homophily within groups [9]. In real situations, the decisions of other companies can impact a company’s own decision making, leading to group effects in company decision making [10]. Previous research has demonstrated that the financial decisions of certain companies are influenced by the decisions of peer companies, with far-reaching effects on their financial decisions. As the economic and market environment changes, more companies choose to enhance their competitiveness and increase profits through cooperation with other companies. Leary and Roberts noted that companies consider information from industry peers when making financing decisions and incorporate it into their own decisions [11]. Kaustia and Rantala’s research indicated that companies make stock issuance decisions based on the behavioral performance of other companies in the same industry, suggesting the presence of group effects in stock issuance decisions. These effects arise due to information asymmetry among companies in the same industry and competition among industry firms [12]. Bratten et al. found that companies with a higher market value in an industry are considered “leaders”, while companies that imitate the behaviors of these leaders are called “followers”, implying that follower companies’ operational decisions are greatly influenced by leader companies in terms of earnings management decisions [13]. Kedia et al. argued that group effects influence corporate earnings management decisions. This study analyzes the impact and mechanisms of different types of companies in the stock issuance process by examining representative companies within the industry. In recent years, domestic scholars have also conducted in-depth studies on the daily decision making of companies and their roles and mechanisms within industries [14]. Based on previous literature, some scholars suggest that companies become aware of their peers’ merger and acquisition plans in advance and formulate favorable strategies with sufficient information support [15]. Kyissima et al. pointed out that companies refer to the decisions of their peer companies when selecting a capital structure to ensure the rationality and sustainability of their decisions [16,17]. Li proposed that industrial policies significantly influence investment homophily [18].

There is some research on regional homophily. For example, some researchers have found that companies within the same region tend to adopt similar financial strategies. If two companies are geographically close, it can stimulate sparks of technological innovation. Parson et al. pointed out that there is regional homophily in corporate financial illegitimacy tendencies [19]. Lu Rong and Chang Wei believed that corporate misconduct exhibits regional homophily, and violations in information disclosure are more pronounced [20]. Rashid et al. argue that excessive corporate debt will exhibit significant group effects within the same region, and these group effects will impact the operation of companies. Strong group effects will lead to excessive debt and overinvestment by companies, thereby weakening their debt repayment, profitability, and other issues. Starting from the causes of group effects, the authors analyze the issue of excessive corporate debt and provide strategies and recommendations [21]. Scholars have extended the study of group effects to government policy formulation, which goes beyond the corporate level. Deng and Zhao verified that local governments exhibit group effects when making major decisions, and the study shows that this kind of group effect among local authorities occurs in regions with similar characteristics in the same province and comparable levels of economic development [22]. Meyer explored the formation mechanisms and influencing factors of three different types of corporate information disclosure group effects in listed heavy-polluting industrial companies to verify whether there is a group effect in environmental information disclosure by listed companies in China [23]. Neo-institutionalism advocates simulating individuals’ decision making and behavior towards others in the same situation to achieve organizational stability and gain recognition of “legitimacy” [24]. In addition, DiMaggio and Powell proposed three mechanisms: coercion, imitation, and social norm mechanisms [25]. Subsequently, Haunschild and Miner identified three aspects that are mainly present when organizations imitate target selection: frequency-based imitation, feature-based imitation, and outcome-based imitation [26]. Among them, frequency-based imitation is the most common and representative. This article aims to explore the group effects of corporate ESG information disclosure decisions and deeply investigate their influencing mechanisms through frequency simulation studies.

Companies tend to imitate the structures and behaviors adopted by the majority of other companies, and this imitation is based on frequency [27]. This imitation between companies can be seen as a rationalization and legitimization of organizational structures [28]. Therefore, decision-makers in information disclosure only need to meet the minimum standards of mutual imitation, without striving to be the best. Otherwise, it may attract attention and bring hidden risks to future development. Based on this convention, companies make their own decisions based on the imitation of other companies. This view has been confirmed by numerous scholars and is recognized as truth. Some scholars also point out that with the widespread application of corporate social responsibility information disclosure in the entire industry, imitation by other companies has become more common, and the pressure for conformity among companies has increased [29]. Boateng’s research on the social responsibility disclosure of listed companies in China found that 85% of the sample companies based their disclosure intentions, timing, and level on China’s general standards [30]. Shen also found that the pressure for convergence not only affects the behavior of corporate information disclosure but also interferes with its quality [31].

Hypothesis 1A. 

Corporate ESG information disclosure is positively correlated with the average level of ESG information disclosure by other companies in the same region, indicating the presence of group effects.

Hypothesis 1B. 

Corporate ESG information disclosure is positively correlated with the average level of ESG information disclosure by other companies in the same industry, indicating the presence of group effects.

2.2. The Influence of Executive Social Networks on the Group Effects of Corporate ESG Information Disclosure

Both companies and individuals are individuals situated within various network relationships, and the external environment in which they operate has a profound impact on them. There are various connections between companies and the environment. As a company, it must constantly engage in resource exchange with the external world to obtain various resources and information necessary for sustainable development. Therefore, companies need to rely on external forces to maintain and promote their continuous growth and expansion. To gain advantages in fierce market competition, companies must rely not only on internal knowledge, human, material, and financial capital but also on the support of external social capital to ensure their competitive advantage. Companies have numerous interrelated social relationships with the outside world, forming a complex network structure known as the external social network. The various resources possessed by companies can be effectively utilized and accessed, which is referred to as external social capital, and social networks are an important form of social capital. Social capital can help companies obtain market opportunities in terms of new products, services, technologies, and other related economic benefits [32]. Therefore, establishing a broad corporate social network is a recommended measure to promote interaction and cooperation between companies and society.

Social networks are an important means for people to collect and disseminate hard-to-obtain information. With social networks at the core, you can gain more information from the Internet [33,34]. Portfolio managers hold a tighter grip and achieve higher returns on companies with social networks, indicating that social networks can convey important information in the stock market [9]. Establishing social network connections between corporate banks and bankers can effectively reduce loan costs and mitigate information asymmetry [35]. In the Internet era, a company’s management activities depend more on understanding market conditions, competitors, customers, and other factors, which are communicated and feedbacked through the Internet. Leveraging informational advantages, the Internet can deliver valuable information to companies, thereby improving their decision making [36]. Additionally, non-informational transmission methods such as political favors can also bring tangible benefits to companies [37]. The Internet has a significant impact on corporate decision making and is one of the important ways for companies to adjust their strategies and enhance their competitiveness. One of the advantages of the Internet is the phenomenon of knowledge spillover, where certain information in the network can be utilized by other members, thus forming a unique mechanism of information sharing [38]. Moreover, a company’s position can be seen as an effective means of substituting for explicit control arrangements. Therefore, the importance of networks depends on the relationships between network members and the degree of connection between the network and the external market. In the field of venture capital, network members can collaborate to suppress potential competitors, such as raising entry barriers and increasing economic rents for incumbent employees [39]. The information shared among network members is not only valuable in itself, but its credibility also allows the establishment of or damage to reputation capital [40]. Reputation capital is a form of social capital embedded in director networks. Companies associated with financial firms are more likely to adopt higher levels of corporate financing [41], and companies with well-developed networks can improve their performance through network resources [42]. Social network connections play a crucial role in the director labor market, as connections or involvement in other networks strongly incentivize directors to enter company boards [43].

Regarding the study of social networks, GAYE suggests that social networks promote the risk-taking ability of corporate decision making [44]. Wang points out that companies within clusters may vary their migration choices based on the diversity and core density of their networks [45]. Garst indicates that network financing effectively reduces costs for banks and companies through information transmission while also increasing a company’s turnover, which is beneficial for its long-term development [46]. Chan points out that having rich social network relationships can acquire social capital; however, excessive reliance on social network relationships may lead to biases within a company, thereby inhibiting the acquisition of entrepreneurial resources [47]. Rudd highlights that the key to a company’s financial activities depends on the control and informational advantages brought about by changes in the position of “structural holes” [48]. Liao suggests that the division of labor in the global agricultural value chain is influenced by the centrality, connection strength, and heterogeneity of the network [49]. Capelle-Blancard suggests that policy-based intermediaries emphasize information fairness and justice due to their economic independence. Local media has a significant advantage in obtaining information, challenging the market competition of private information transactions, whereas there is no apparent relationship between remote media and private information transactions [50]. Sierdovski argues that the concentration of human capital in social networks is significantly positively correlated with a company’s innovation ability among listed companies [51].

Based on the literature mentioned above, this study proposes the following hypothesis:

Hypothesis 2. 

Executive social networks have a positive moderating effect on the group effects of ESG information disclosure.

3. Research Design

3.1. Sample Selection

This study specifically focuses on privately listed companies in the Growth Enterprise Market (GEM) between 2018 and 2021. Companies with ST status and those with missing data were excluded, resulting in a sample of 385 companies and 1540 observations. The ESG information disclosure level of the companies was obtained from the WIND database, while other variables were sourced from the CSMAR database. The industry names were based on the new industry classification standards for listed companies published by the China Securities Regulatory Commission (CSRC). The sample companies were located in different regions, including North China, northeast China, East China, central China, South China, southwest China, and northwest China. Data processing was performed using SPSS 26.

3.2. Variable Definitions

The dependent variable is the ESG information disclosure level. The study utilizes the WIND database’s ESG composite score to measure the level of ESG information disclosure. The score, ranging from 1 to 10, combines the management practice score (70%) and the controversy event score (30%). It assesses companies across three dimensions, twenty-seven issues, and over three hundred indicators, providing a comprehensive evaluation of a company’s ESG management practices and its ability to address significant unexpected issues. The independent variable was the executive social network. The social network of the top management team, to some extent, represents the company’s overall social network. The richness of the executive team’s social network significantly influences the company’s information acquisition. This study considers the following five dimensions of the executive social network:

Directorships in other companies: if the ratio of executive team members holding positions in other companies exceeds the market average, the variable “Net1” is assigned a value of 1; otherwise, it is assigned a value of 0.

Educational background: if the ratio of executive team members with a master’s degree or above exceeds the market average, the variable “Net2” is assigned a value of 1; otherwise, it is assigned a value of 0.

Serving as deputies in CPPCC/NPC: if the ratio of executive team members who have been elected as deputies to the Chinese People’s Political Consultative Conference (CPPCC) or National People’s Congress (NPC) exceeds the market average, the variable “Net3” is assigned a value of 1; otherwise, it is assigned a value of 0.

Banking work experience: if the ratio of executive team members with previous work experience in banks exceeds the market average, the variable “Net4” is assigned a value of 1; otherwise, it is assigned a value of 0.

Association service history: if the ratio of executive team members with a history of serving in associations exceeds the market average, the variable “Net5” is assigned a value of 1; otherwise, it is assigned a value of 0.

The executive social network value is calculated using the above variables. It ranges from 0 to 5, with higher values indicating a richer social network and lower values indicating a less diverse network.

Net = Net1 + Net2 + Net3 + Net4 + Net5

Control variables (

Table 1. Definitions of Control Variables.

Control Variable Name	Variable Symbol	Variable Definition
Financial Leverage	Alr	End-of-year total liabilities divided by end-of-year total assets
Board Ownership	Num	Total number of directors holding shares in the company
Firm Size	Size	Natural logarithm of total assets at the end of the year
Profitability	ROE	Return on equity
Dual Roles	Dual	Whether the chairman and CEO positions are held by the same person (1 if yes, 0 if no)
Board Size	Board	Number of members on the board of directors
Ownership Concentration	Crio	Percentage of shares held by the largest shareholder
Proportion of Independent Directors	Out	Number of independent directors divided by the total number of directors
Company Growth	Growth	Asset growth rate: (end-of-period total assets—beginning-of-period total assets) divided by beginning-of-period total assets

): to mitigate the influence of other internal and external factors on ESG information disclosure and ensure the independence of social networks, the following variables are introduced as control variables.

3.3. Model Design

Based on Hypotheses 1A and 1B, this study constructs the following models:

$${\mathrm{ESG}}_{\mathrm{i},\mathrm{t}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1\mathrm{market}{1}_{\mathrm{i},\mathrm{t}-1}+{\mathsf{\beta }}_2{\mathrm{ESG}}_{\mathrm{i},\mathrm{t}-1}+{\displaystyle \sum }{\mathsf{\beta }}_{\mathrm{j}}{\mathrm{Controls}}_{\mathrm{j},\mathrm{i},\mathrm{t}}+{\in }_{{}_{\mathrm{i},\mathrm{t}}}$$

(1)

$${\mathrm{ESG}}_{\mathrm{i},\mathrm{t}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1\mathrm{market}{2}_{\mathrm{i},\mathrm{t}-1}+{\mathsf{\beta }}_2{\mathrm{ESG}}_{\mathrm{i},\mathrm{t}-1}+{\displaystyle \sum }{\mathsf{\beta }}_{\mathrm{j}}{\mathrm{Controls}}_{\mathrm{j},\mathrm{i},\mathrm{t}}+{\in }_{{}_{\mathrm{i},\mathrm{t}}}$$

(2)

In Models (1) and (2), ESG_i,t represents the level of ESG disclosure of company i in period t, $\mathrm{market}{1}_{\mathrm{i},\mathrm{t}-1}$ represents the average level of ESG disclosure of other companies in the same region as company i in the previous period t − 1, ${\mathrm{ESG}}_{\mathrm{i},\mathrm{t}-1}$ represents the level of ESG disclosure of company i in the previous period t − 1, and $\mathrm{market}{2}_{\mathrm{i},\mathrm{t}-1}$ represents the average level of ESG disclosure of other companies in the same industry as company i in the previous period t − 1. ${\mathrm{Controls}}_{\mathrm{j},\mathrm{i},\mathrm{t}}$ represents the control variables, with j indicating the variable index, and ε_i,t represents the error term.

For Hypotheses 2, the interaction term Net_i,t × $\mathrm{Market}{1}_{\mathrm{i},\mathrm{t}-1}$ is included to investigate the moderating effect of social networks. The models are as follows:

$${\mathrm{SG}}_{\mathrm{i},\mathrm{j}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1\mathrm{Market}{1}_{\mathrm{i}.\mathrm{j}-1}+{\mathsf{\beta }}_2{\mathrm{Net}}_{\mathrm{i},\mathrm{j}}+{\mathsf{\beta }}_3{\mathrm{Net}}_{\mathrm{i},\mathrm{j}}\times \mathrm{Market}{1}_{\mathrm{i},\mathrm{j}-1}+{\mathsf{\beta }}_4{\mathrm{ESG}}_{\mathrm{i},\mathrm{t}-1}+{\displaystyle \sum }{\mathsf{\beta }}_{\mathrm{j}}{\mathrm{Controls}}_{\mathrm{j},\mathrm{i},\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{j}}$$

(3)

$${\mathrm{ESG}}_{\mathrm{i},\mathrm{j}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1\mathrm{Market}{2}_{\mathrm{i}.\mathrm{j}-1}+{\mathsf{\beta }}_2{\mathrm{Net}}_{\mathrm{i},\mathrm{j}}+{\mathsf{\beta }}_3{\mathrm{Net}}_{\mathrm{i},\mathrm{j}}\times \mathrm{Market}{2}_{\mathrm{i},\mathrm{j}-1}+{\mathsf{\beta }}_4{\mathrm{ESG}}_{\mathrm{i},\mathrm{t}-1}+{\displaystyle \sum }{\mathsf{\beta }}_{\mathrm{j}}{\mathrm{Controls}}_{\mathrm{j},\mathrm{i},\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{j}}$$

(4)

In Models (3) and (4), the variables Net, Market1, and Market2 are centered to improve the fit of the moderation effect model.

4. Empirical Results Analysis

4.1. Descriptive Statistics and Correlation Analysis

Table 2. Descriptive Statistics of Variables.

Variable	Mean	Median	Standard Deviation	Minimum	Maximum
Net	1.8519	2	1.16206	0	5
Alr	0.3594	0.3517	0.17801	0.03	1.4
Num	3.22	3	1.587	0	8
Size	21.6387	21.5294	0.91988	19.53	26.45
ROE	0.0134	0.0685	0.66755	−19.67	5.32
Dual	0.49	0	0.5	0	1
Board	7.7	7	1.446	4	13
Crio	0.2706	0.2522	0.11814	0.03	0.75
Out	0.3883	0.4	0.05323	0.2	0.75
Growth	0.1591	0.1015	0.34133	−0.93	5.85

presents the descriptive statistics for each variable. The variable “ESG disclosure level” shows a relatively balanced distribution, with minimal difference between the median and mean values. The variable “top management team social networks” also exhibits a small difference between the mean and median values, indicating a certain degree of unbiasedness in its distribution. The number of directors holding shares in the company ranges from 0 to 8, with a small difference between the mean and median values, suggesting that 50% of the companies have a relatively low total number of directors holding shares. The company size ranges from 19.53 to 26.45, with small differences between the samples, indicating a relatively even distribution.

The net return on assets shows a wide range of values, with a maximum of 5.32 and a minimum of −19.67. The number of members on the board of directors ranges from 4 to 13, with an average of 7.7 and a median of 7, indicating that the majority of companies have a board size of 7 or more. The average shareholding proportion of the largest shareholder in most companies is 27.06%, and the asset growth rate is above 10.15%.

According to

Table 3. Correlation Analysis of Variables.

	ESG	Net	Crio	Dual	Market1	Market2	Board	Out	Num	Size	ROE	Alr	Growth
ESG	1
Net	−0.034 **	1
Crio	0.104 **	−0.013	1
Dual	−0.036 *	0.013	0.056 *	1
Market1	0.161 ***	−0.035	−0.103 **	−0.015	1
Market2	0.484 **	−0.042 *	−0.027	0.021	0.142 **	1
Board	0.034 *	0.080 **	0.022	−0.073 **	−0.024	−0.001	1
Out	0.002 *	0.006	−0.005	0.058 *	−0.001	0.019	−0.619 **	1
Num	0.067 **	−0.014	−0.050 *	−0.085 **	−0.078 **	0.016	0.441 **	−0.306 **	1
Size	0.008 *	0.264**	−0.044 *	−0.096 **	0.059 *	−0.032	0.138 **	−0.035	0.110 **	1
ROE	0.126 **	0.034	0.04	0.028	0.075 **	0.108 **	0.073 **	−0.024	0.086 **	−0.004	1
Alr	−0.192 **	0.088**	0.012	−0.035	−0.050 *	−0.204 **	0.041	−0.025	−0.051 *	0.322 **	−0.180 **	1
Growth	0.165 **	0.019	0.087 **	0.007	0.057 *	0.081 **	0.053 *	−0.016	0.113 **	0.160 **	0.178 **	−0.009	1

Note: ***, **, and * represent significance at the 1%, 5%, and 10% levels, respectively.

, the ESG disclosure level shows significant positive correlations with Market1 and Market2 and a significant negative correlation with the variable Net. Additionally, the ESG disclosure level exhibits significant positive or negative correlations with the control variables Board, Out, Dual, Alr, Size, Roe, Growth, Crio, and Num, suggesting the need to control for these variables in the analysis.

4.2. Same-Group Effect of Corporate ESG Disclosure

As presented in

Table 4. Regression Results of Same-group Effect of Corporate ESG Disclosure.

ESG Disclosure Level
	Model 1	Model 2
ESGi,t−1	0.796 ***	0.734
Market1i,t−1	0.042 ***
Market2i,t−1		0.165 ***
Alr	−0.039 **	−0.021 *
Crio	0.027 **	0.034 **
Dual	−0.013	−0.019
Board	0.004	0.002
Out	−0.025	−0.026
Num	0.004	0.003
Size	−0.037 **	−0.031 **
Roe	0.001	−0.002
Growth	0.031 **	0.029 **
Adjusted R2	0.668	0.688
F-value	282.669	309.867
Sample Size	1540	1540

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively.

, Market1_i,t−1 and Market2_i,t−1 show a significant positive correlation at the 1% level. This suggests that companies tend to imitate the market average of ESG disclosure levels of other companies in the same region and industry in the previous period, supporting Hypotheses 1A and 1B. The coefficients of ESG_i,t−1 in both models are significantly positive, indicating that the ESG disclosure level in the previous period has an impact on the current ESG disclosure level. Thus, we conclude that there is a same-group effect in ESG disclosure. Regarding the control variables, Alr and Size exhibit significant negative correlations, suggesting that the size and leverage of the company have an adverse impact on the ESG disclosure level. Conversely, Crio and Growth have significant positive coefficients, indicating that the concentration of ownership and the company’s growth promote the level of ESG disclosure.

4.3. The Moderating Effect of Top Management’s Social Network on the Same-Group Effect of Corporate ESG Disclosure

As indicated in

Table 5. Test of Regional Effects.

ESG Disclosure Level
	Model 3	Model 4
ESGi,t−1	0.428 ***	0.412 ***
Net	−2.615 ***	−2.529 ***
Neti,t × Market1i,t−1	2.652 ***
Neti,t × Market2i,t−1		2.566 ***
Market1i,t−1	0.019 *
Market2i,t−1		0.075 ***
Alr	−0.012	−0.004
Crio	0.022 **	0.025 **
Dual	−0.014	−0.017
Board	−0.015	−0.016
Out	−0.016	−0.017
Num	0.015	0.014
Size	−0.025 **	−0.023 **
ROE	0.031 ***	0.029 ***
Growth	0.01	0.01
Adjusted R2	0.827	0.831
F-value	565.475	581.477
Sample Size	1540	1540

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively.

, in Models (3) and (4), the standardized coefficient of Net is negative, suggesting a negative relationship between the richness of top management’s social network and the level of corporate ESG disclosure. This implies that within the same region or industry, a greater extent of top management’s social network is associated with a lower level of ESG disclosure. Net_i,t × Market1_i,t−1 is significantly positive at the 1% level, indicating that the same-group effect of ESG disclosure is positively moderated by the richness of top management’s social network. In other words, as the social network of top management becomes more extensive, the same-group effect of ESG disclosure is strengthened. This provides support for Hypothesis 2. Regarding the control variables, the standardized coefficients of Crio and Roe consistently show significant and positive effects in both models. This suggests that within the same region or industry, the concentration of ownership and profitability significantly enhance the level of ESG disclosure. Conversely, the standardized coefficient of Size consistently shows a significant and negative effect in both models, indicating that the firm’s size reduces the level of ESG disclosure.

4.4. Robustness Test

In order to ensure the reliability of the results of this study, five types of robustness tests were conducted: First, the per capita GDP was added as a control variable to control for macroeconomic conditions; second, the lagged effect of executive social networks was considered; third, the bootstrapping regression method was used to overcome the issue of small sample size; and fourth, a panel model with bidirectional fixed effects was employed based on the data characteristics and model applicability.

(1)

Divided into several groups and regression

To further determine whether the social networks of the executive team have an impact on the homophily effect of ESG disclosure when imitating companies expand their regional presence, we divide the regions of companies into eastern, central, and western regions according to the latest industry classification issued by the China Securities Regulatory Commission. We treat each region as a separate sample to further test the conclusions mentioned above. Models (5) and (6) are constructed as follows:

$${\mathrm{SG}}_{\mathrm{i},\mathrm{t}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1\mathrm{market}{3}_{\mathrm{i},\mathrm{t}-1}+{\mathsf{\beta }}_2{\mathrm{ESG}}_{\mathrm{i},\mathrm{t}-1}+{\displaystyle \sum }{\mathsf{\beta }}_{\mathrm{j}}{\mathrm{Controls}}_{\mathrm{j},\mathrm{i},\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$$

(5)

$${\mathrm{ESG}}_{\mathrm{i},\mathrm{t}}={\mathsf{\beta }}_0\times {\mathsf{\beta }}_1\mathrm{market}{3}_{\mathrm{i},\mathrm{t}-1}+{\mathsf{\beta }}_2{\mathrm{Net}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }}_3{\mathrm{Net}}_{\mathrm{i},\mathrm{t}}\times \mathrm{Market}{3}_{\mathrm{i},\mathrm{t}-1}+{\mathsf{\beta }}_4{\mathrm{ESG}}_{\mathrm{i},\mathrm{t}-1}+{\displaystyle \sum }{\mathsf{\beta }}_{\mathrm{j}}{\mathrm{Controls}}_{\mathrm{j},\mathrm{i},\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$$

(6)

Here, Market3_i,t−1 represents the average level of ESG disclosure of other companies in the same region as company i in the previous period t − 1.

As shown in

Table 6. Regression Results of Same-group Effect of ESG Disclosure by Region.

ESG Disclosure Level
	Eastern Region	Central Region	Western Region
Market3t−1	0.063 ***	0.112 **	0.299 ***
ESGi,t−1	0.786 ***	0.675 ***	0.645 ***
Alr	−0.027	−0.066	−0.01
Crio	0.02	0.106 **	0.079
Dual	−0.005	−0.059	−0.143 **
Board	0.016	0.005	−0.081
Out	−0.024	0.039	−0.016
Num	0	−0.007	−0.03
Size	−0.031 *	−0.125 **	0.052
ROE	−0.007	0.037	0.048
Growth	0.028	0.093 **	−0.061
Adjusted R2	0.659	0.654	0.782
F-value	224.162	31.972	29.328
Sample Size	1271	181	88

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively.

, the variables Market3_i,t−1 for all three regions are positively significant, indicating that companies imitate the average level of ESG disclosure of other companies in the same region in their ESG disclosure. ESG_i,t−1 is significant, indicating that ESG disclosure is influenced by the previous period. In

Table 7. Test of Moderating Effects.

ESG Disclosure Level
	Eastern Region	Central Region	Western Region
Market3t−1	0.023 *	0.018	0.205 ***
Net	−2.708 ***	−3.284 ***	−1.468 ***
Neti,t × Market3t−1	2.737 ***	3.365 ***	1.565 ***
ESGi,t−1	0.414 ***	0.346 ***	0.417 ***
Alr	−0.005	−0.016	0.009
Crio	0.014	0.056	0.03
Dual	−0.006	−0.042	−0.134 **
Board	−0.021	−0.009	−0.022
Out	−0.023	0.014	0.014
Num	0.007	0.045	0.019
Size	−0.028 **	−0.018	0.061
ROE	0.002	0.06 *	0.096 *
Growth	0.005	0.025	−0.008
Adjusted R2	0.826	0.818	0.854
F-value	465.868	63.373	40.173
Sample Size	1271	181	88

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively.

, the coefficients of Net and Net_i,t × Market3_i,t−1 are significant at a 1% level, indicating a positive and significant impact of the executive team’s social networks on ESG disclosure in large-scale regions such as the eastern, central, and western regions, consistent with the previous findings.

The same-group effect of executive social networks on corporate ESG may be driven by similar macroeconomic forces that encourage these companies to adopt ESG disclosure practices. To mitigate the potential influence of unobservable firm and CEO characteristics, we employ a fixed-effects model to control for firm-level fixed effects and time-fixed effects. We then re-estimate the model, and the results of our study provide support for the aforementioned hypothesis.

(2)

Control of macroeconomic environment

In order to overcome the potential impact of the macroeconomic environment on the ESG disclosure of firms, we introduced per capita GDP as a control variable to address this concern. The regression results, shown in

Table 8. Robustness Test.

Variables	(1)	(2)	(3)
Net	−2.507 ***	−2.589 ***	−2.589 ***
Control variables	Yes	Yes	Yes
Constant	1.773 **	1.771 **	2.225 **
FE of company	Yes	Yes	Yes
FE of year	Yes	Yes	Yes
R2	0.546	0.527	0.524
Sample of size	1271	1271	1271

Note: *** and ** indicate significance at the 1% and 5 levels, respectively.

Column (1), demonstrate that even after controlling for macroeconomic conditions, executive social networks still significantly contribute to the homophily effect in ESG disclosure, supporting the findings of this study.

(3)

Bootstrapping regression

Considering the relatively small sample size of panel data consisting of 385 firms over a period of 4 years, we employed the bootstrap method to obtain more robust regression results. Therefore, we conducted 1000 iterations of bootstrap sampling and performed regression analysis again. The results, presented in Table 8 Column (4), indicate that after 1000 iterations of bootstrap sampling, the regression coefficients and significance of executive social networks remained unchanged compared to the base regression, further confirming the robustness of our findings.

(4)

Alternative regression methods

In addition to the base regression, we also employed a panel model with bidirectional fixed effects for further analysis. The results, shown in Table 7 Column (4), consistently aligned with the base regression, demonstrating the robustness of our findings after the alternative regression method was applied.

5. Conclusions and Implications

5.1. Research Conclusions

This study examines the impact of the social networks of top management teams on the homophily effect of ESG disclosure in China’s privately-owned companies listed on the Growth Enterprise Market from 2018 to 2021. The conclusions are as follows: The current level of ESG disclosure in a company is positively correlated with the average level of ESG disclosure among other companies in the same region and industry in the previous period, indicating the presence of a homophily effect in ESG disclosure. Despite the negative influence of top management team social networks on ESG disclosure behavior, they have a positive moderating effect on the homophily effect of ESG disclosure. This suggests that in companies with rich top management team social networks, the homophily effect of ESG disclosure will be strengthened, while in companies with limited top management team social networks, the homophily effect of ESG disclosure will be weakened.

5.2. Research Implications

In summary, the social networks of top management teams in privately-owned listed companies have a positive impact on the homophily effect of ESG disclosure, indicating that companies with more extensive social networks among their top management teams exhibit a higher homophily effect in ESG disclosure. This study not only expands the literature on the influence of social networks on the homophily effect of ESG disclosure but also provides suggestions for companies and decision-makers for their future development. From the company’s perspective, building and enriching social networks can enhance ESG disclosure levels, gain more returns, and increase corporate value. From the decision-makers’ perspective, strengthening social network development not only improves a company’s competitiveness in a complex environment but also contributes positively to the rapid development of the overall socio-economic system. Therefore, for privately-owned enterprises, it is important to expand the social networks of top management teams, acquire external information, conduct proper analysis, make informed decisions, and promote high-level corporate development. The findings of this study have implications for the development of relevant theories and the formulation and implementation of policies.

5.3. Research Limitations

Due to constraints such as data availability, it is necessary to further validate the research findings of this study when extrapolating them from private enterprises to the overall context. The use of more detailed data would be required for this purpose. Additionally, exploring the specific impacts of regional and industry-specific public policies and providing more detailed policy recommendations for different regions and industries are all directions that warrant further investigation.