Macro-Institutional Pressures and Firms’ Environmental Management Behavior: The Moderating Effect of Micro-Institutional Pressures

Abstract

How to promote firms’ environmental management behavior is a concern for academics. In order to explain the reasons why firms’ environmental management behavior exhibits heterogeneity from the institutional perspective, this paper seeks to investigate the relationship between macro-institutional pressures and firms’ environmental management behavior, with the moderating effects of micro-institutional pressures. Firms’ environmental management behavior is classified into substantive environmental management behavior and symbolic environmental management behavior. Macro-institutional pressures typically include environmental regulation, industry imitation, and media attention, while micro-institutional pressures include cognitive focus of executive and cognitive complexity of executive. A sample of 236 groups from 118 listed companies in China’s heavy-polluting industries is employed. It is found that environmental regulation is more likely to promote substantive environmental management behavior, while industry imitation and media attention are more likely to promote symbolic environmental management behavior. Moreover, cognitive focus of executive negatively moderates the relationship between environmental regulation and substantive environmental management behavior and the relationship between media attention and symbolic environmental management behavior. Cognitive complexity of executive positively moderates the relationship between macro-institutional pressures and firms’ environmental management behavior. The findings of this paper clarify the reasons for the heterogeneity of firms’ environmental management behavior from the institutional perspective, which contributes to improving the institutional environment, integrating executive cognition, and promoting firms’ environmental management behavior.

1. Introduction

Economic development has caused environmental pollution in recent years, such as air pollution, water pollution, and resource shortage [1,2]. The responsibility of firms for pollution control has raised attention from all walks of life. Environmental management behavior (EMB) (abbreviations: EMB = Environmental Management Behavior) of firms has triggered the research interest of scholars [3,4,5]. EMB refers to a series of measures taken by firms to reduce the negative impacts on the environment or improve their environmental images [6,7], including environmental innovation, recycling, emissions reduction, the pursuit of ISO 14001 certification, disclosing environmental information, establishing environmental management committees, and so on [1,8,9].

The characteristics and effects of different EMBs vary considerably, with some EMBs involving practical changes, while others are used as tools to manage perceptions of stakeholders. Therefore, referring to the classification by Ashforth and Gibbs (1990) [10], EMB is classified into substantive EMB and symbolic EMB. Substantive EMB is the behavior that firms take to reduce negative impacts on the environment, such as process improvement, product innovation, recycling, and emissions reduction, which involves in-depth and practical changes in the production and operation of firms [11]. Symbolic EMB is the behavior that firms use to convey legitimacy signals through publicity and commitment, including the pursuit of ISO 14001 certification, establishing environmental management committees, disclosing environmental information, releasing corporate social responsibility reports, and releasing environmental reports, in order to meet social expectations and manage images [7,12]. This paper is developed based on the two-dimensional classification of EMB.

Macro-institutional theory provides an explanation for the dynamics of EMB. This theory emphasizes macro-institutional pressures through formal or informal mechanism, tangible documents, and intangible awareness to control the sources of legitimacy of firms, adjusting their activities within the bounds of stakeholders’ permission, which can constrain and guide firms [13,14]. Government, suppliers, customers, competitors, industry associations, environmental NGOs, media, and the public all influence environmental decisions and behavior of firms [15,16,17,18]. DiMaggio and Powell (1983) [13] classify macro-institutional pressures into three types: coercive pressure, mimetic pressure, and normative pressure. According to this classification, scholars have explored the effects of macro-institutional pressures on EMB. Chen et al. (2018) [1] and Qi et al. (2021) [19] investigate the impact of coercive pressure on firms’ environmental technology innovation. Cormier et al. (2005) [20] analyze the influence of mimetic pressure on environmental information disclosure. Castka and Prajogo (2013) [21] find that normative pressure can drive ISO 14001 certification. Combining the existing research, it can be found that different types of macro-institutional pressures can trigger heterogeneous EMB of firms. However, most of the existing studies focus on one specific EMB, resulting in a lack of in-depth discussion and comparison of different institutional pressures. Consequently, the effects of the three types of macro-institutional pressures on substantive EMB and symbolic EMB will be compared. The fine-grained study will help to clarify the reasons for the heterogeneity in firms’ EMB.

Theoretically, macro-institutional pressures make firms behave in a similar way, that is, institutional isomorphism [13]. However, in reality, firms even in the same region and industry that face the same macro-institutional pressures show differences in their decisions and behavior. Therefore, it is not comprehensive to explain heterogeneous EMB of firms only from the macro-institutional perspective. The micro-institutional factors are also worth considering. Micro-institutional theory emphasizes the roles of individuals’ experiences, perceptions, and explanations, focusing on the analysis of the impact of executive cognition on firms’ decisions and behavior, which provides an explanation for firms’ deviant behavior in response to macro-institutional pressures [22]. The theory can be complementary to macro-institutional theory and fill the above-mentioned gaps. Accordingly, two types of micro-institutional pressures based on micro-institutional theory are introduced, namely, cognitive focus of executive and cognitive complexity of executive, in order to investigate the interactive effects of macro-institutional pressures and micro-institutional pressures on heterogeneous EMB.

Listed companies in the heavy-polluting industries in China’s Shanghai and Shenzhen A-shares are selected as the sample. The reasons are as follows. China, as the second largest economy in the world, suffers from serious environmental pollution along with its economic development [23]. The Chinese government insists on the goal of coordinated development of economy, society, and environment. Heavily polluting industries should be held primarily responsible for environmental pollution. They are the focus of government regulation and media supervision [24,25]. It is of great theoretical and practical significance to take China’s heavily polluting industries as the sample for improving environmental governance.

The potential innovations of this paper are as follows. (1) Heterogeneous EMB is classified into two dimensions according to its characteristics, that is, substantive EMB and symbolic EMB, which is helpful to enrich the research on heterogeneous EMB. (2) The main effects of different types of macro-institutional pressures on heterogeneous EMB are examined. We identify three main sources of macro-institutional pressures, that is, environmental regulation, industry imitation, and media attention, in order to compare their effects on firms’ heterogeneous EMB. (3) The research on micro-institutional pressures on heterogeneous EMB is extended. We explore the moderating effects of cognitive focus of executive and cognitive complexity of executive in the relationship between macro-institutional pressures and heterogeneous EMB so as to provide accurate explanations for the causes of firms’ heterogeneous EMB. In conclusion, this paper explains the reasons why firms’ EMB exhibits heterogeneity from the institutional perspective and achieves a combination of micro-level and macro-level research to improve the institutional environment and promote firms’ EMB.

Basic theories and research hypotheses are presented in Section 2. Research design is given in Section 3. Results of research are shown in Section 4. Finally, Section 5 shows the conclusions and implications.

2. Basic Theories and Research Hypotheses

2.1. Basic Theories

2.1.1. Macro-Institutional Theory

Macro-institutional theory argues that the necessary prerequisite for firms’ development is legitimacy. Complying with institutional requirements is important for firms to gain legitimacy, which can facilitate firms gaining recognition from stakeholders [26,27]. DiMaggio and Powell (1983) [13] classify macro-institutional pressures into three types: coercive pressure, mimetic pressure, and normative pressure. Coercive pressure, characterized by rigidity and deterrence, refers to the coercive force exerted by organizations such as governments on firms through laws, regulations, and public policies. Compared to developed economies, China’s economic development is more susceptible to environmental regulation of governments [14]. Mimetic pressure, characterized by universality and propagation, originates from peers or other firms in the same industry, who provide templates of practice that guide the optimization of firms’ behavior [16]. In an uncertain environment, the demonstration effect of industry imitation provides firms with viable solutions [19,28]. Normative pressure comes from the dissemination of institutional norms by professionalized organizations. In particular, media attention, as an important source of normative pressure, directly affects the image and reputation of firms, exerting constraints on their behavior [29]. In summary, environmental regulation, industry imitation, and media attention are selected as the three main sources of macro-institutional pressures to explore their effects on firms’ heterogeneous EMB.

2.1.2. Micro-Institutional Theory

Micro-institutional theory suggests that the background, experience, and preferences of individuals influence their cognition. The decisions of firms based on individuals’ cognition are different, which makes some firms comply with external pressures, while others resist them [22]. Thus, executive cognition constitutes the micro-institutional environment of the firm [22]. As a result of differences in executive cognition, firms respond heterogeneously to the same macro-institutional pressures, thus discouraging institutional isomorphism [14]. Nadkarni and Barr (2008) [30] further classify executive cognition into cognitive focus of executive and cognitive complexity of executive. In particular, cognitive focus of executive refers to the number of key concepts in the structure of executive cognition, with the well-differentiated hierarchy among concepts, that is, some concepts dominate executive cognition more than others. Cognitive complexity of executive refers to the diversity of concepts and inter-concept relationships in the cognitive structure of the executive, with equal hierarchy between concepts [31]. Both affect the strategic decisions and behavior of firms [31].

2.2. Research Hypotheses

2.2.1. Environmental Regulation and EMB

Environmental regulation aims to reduce environmental pollution and restrain firms’ environmental violations by means of environmental legislation, taxation, and fines. Such initiatives sanction excess emissions by increasing the cost of non-compliance for firms [23,32]. At the same time, environmental regulation provides financial supports to firms through such incentives as government subsidies and tax breaks to promote technological improvements and product innovations in the production [32], which can promote substantive EMB of firms. According to the Porter hypothesis, the mechanism of innovation compensation for environmental regulation can compensate for environmental expenditures and promote technological innovation [33,34]. Thus, environmental regulation can promote environment-oriented activities in products and technologies, thereby reducing negative impacts on the environment. In contrast, it is difficult for firms to meet the requirements of environmental regulation through symbolic responses. Therefore, the following hypothesis is proposed:

H1. 

Compared with symbolic EMB, environmental regulation is more likely to promote firms’ substantive EMB.

2.2.2. Industry Imitation and EMB

Industry imitation provides the templates of practice that are available and accessible to firms. The demonstration effect allows for the sharing of industry norms and standard templates, leading firms to imitate peers and optimize their own EMB, which increases the chances of gaining legitimacy [35,36]. It has been shown that industry imitation can drive such symbolic EMB as environmental information disclosure and the certification of environmental management system [37,38]. Firms have discretion in making environmental information disclosures, which allows them to imitate the practices of peers [39]. The pursuit of ISO 14001 certification by firms is also driven by peers [40]. In contrast, such substantive EMB as process improvement and product innovation requires sufficient capital and human resources, as well as specific knowledge and technology based on their own R&D team, which is difficult to obtain through industry imitation [41,42]. Therefore, the following hypothesis is posited:

H2. 

Compared with substantive EMB, industry imitation is more likely to promote firms’ symbolic EMB.

2.2.3. Media Attention and EMB

Media attention, as an essential community force, is an effective complement to government regulatory and irreplaceable governance mechanism [24,43]. The media disseminates information to stakeholders through news reports, addressing information asymmetry and opacity [44]. Compared to other macro-institutional pressures, media attention is time-sensitive. When negative information is reported, firms need to respond quickly to minimize crisis losses. The crude oil spill of BP in the Gulf of Mexico in 2010 caused irreversible environmental pollution and attracted the attention of media. After the incident, BP quickly launched crisis communications and made environmental commitments to reestablish public trust [45]. In contrast, substantive responses such as product innovation require long time cycles [19], which are contrary to the time-sensitive principle of media attention. Therefore, media attention is more powerful in leading firms to adopt symbolic EMB to respond to pressures and shape their images. The following hypothesis is formulated:

H3. 

Compared with firms’ substantive EMB, media attention is more likely to promote symbolic EMB.

2.2.4. The Moderating Role of Micro-Institutional Pressures

To further understand the heterogeneity of EMB, it is critical to consider the micro-institutional pressures. Executive cognition is an essential factor in the micro-institutional domain of firms [14]. Due to differences in background, experience, knowledge, and preferences, macro-institutional pressures are filtered and interpreted through the framework of executive cognitive, leading to firms’ heterogeneous behavior [31,46]. Executive cognition includes two dimensions: cognitive focus of executive and cognitive complexity of executive [30]. Cognitive focus dominates executives’ interpretation of environmental information through a few key concepts, while cognitive complexity disperses executives’ thinking through diverse concepts and inter-concept relationships [31,47,48].

Cognitive focus of executive can maintain the stability of firms’ strategies and behavior, which plays an important role in shaping firms’ values [30]. However, cognitive focus of executive is also prone to generating an isolation mechanism for the external environment [31]. The small number of key concepts in the cognitive framework can lead executives to ignore novel information in the external environment, tend to deny and reject change, constrain formulation of adaptive solutions, and continue to invest resources into established activities to maintain the status quo, thus limiting the response of firms to macro-institutional pressures. Compared to production management, marketing management, and quality management, environmental management emerged late [2,49]. Therefore, environmental management is a relatively new and groundbreaking management initiative for executives, both substantive and symbolic, requiring a shift in their cognition. Cognitive focus of executive confines executive attention to traditional management and ignores the importance of environmental management, thus resisting the impetus of macro-institutional pressures on EMB. The following hypothesis is proposed:

H4a. 

Cognitive focus of executive plays a negative moderating role in the relationship between environmental regulation and substantive EMB.

H4b. 

Cognitive focus of executive plays a negative moderating role in the relationship between industry imitation and symbolic EMB.

H4c. 

Cognitive focus of executive plays a negative moderating role in the relationship between media attention and symbolic EMB.

Cognitive complexity of executive facilitates the information processing and responses to external pressures, which is conducive to introducing new management practices to create competitive advantages [50,51]. The strategic decisions of firms are influenced by the number of key concepts possessed by executives. The more key concepts and the more complex of inter-concept relationships, the more out-of-the-box the executive think and the stronger their ability of information collection and processing [31]. Higher cognitive complexity allows executives to perceive macro-institutional pressures more clearly. Executives tend to introduce new management initiatives in response to macro-institutional pressures, rather than being tied to a specific behavior. Manral (2011) [52] states that cognitive complexity can facilitate executives’ deep understanding of information, improve problem-solving skills, and stimulate innovative behavior in the firm. Executives with high cognitive complexity are able to perceive the intention of macro-institutional pressures to drive EMB. Therefore, they tend to conduct EMB to respond to macro-institutional pressures, gain legitimacy, and improve relationships with stakeholders. The following hypothesis is posited:

H5a. 

Cognitive complexity of executive plays a positive moderating role in the relationship between environmental regulation and substantive EMB.

H5b. 

Cognitive complexity of executive plays a positive moderating role in the relationship between industry imitation and symbolic EMB.

H5c. 

Cognitive complexity of executive plays a positive moderating role in the relationship between media attention and symbolic EMB.

The theoretical framework is shown in Figure 1.

3. Research Design

3.1. Research Sample

Listed companies in heavily polluting industries in China’s Shanghai and Shenzhen A-shares are selected as the sample. The sample is processed as follows. (1) Companies with less than 10 years of establishment are excluded. According to the research of Nadkarni and Narayanan (2007) [31], the short establishment time of firms cannot ensure the integrity of the formation of executive cognition. Therefore, firms that have been established for more than 10 years are selected. (2) The heavily polluting industries are divided into 12 sub-industries, such as thermal power, iron and steel, cement, coal, and so on, according to the Classified Management Directory of Environmental Protection of Listed Companies issued by China Ministry of Environmental Protection. (3) Companies marked as ST and *ST are excluded. (4) Companies with missing data are excluded. Considering the delay of EMB, the observation intervals of independent variables, moderating variables, and control variables are taken from 2019 to 2020. The observation intervals of dependent variables are taken from 2020 to 2021. The final sample of 236 groups of observations consisting of 118 listed companies with two-year data is obtained. To eliminate the effect of outliers, winsorize processing is taken for the (1%, 99%) quantile of the continuous variables.

3.2. Research Variables

3.2.1. Dependent Variables

Substantive EMB (Sub). The natural logarithm of the number of environmental patents plus one is adopted [11].

Symbolic EMB (Sym). According to the definition of symbolic EMB, three items are considered, that is, whether to disclose environmental vision, whether to release corporate social responsibility report, and whether to release environmental report. If the firm takes the initiative, the value of 1 is assigned, otherwise 0. Then, the summation is performed. Symbolic EMB can take values of 0, 1, 2, and 3 [11].

3.2.2. Independent Variables

Environmental regulation (Reg). The regional GDP of the place the firm is located in is divided by the industrial smoke (dust) emissions, industrial wastewater emissions, industrial sulfur dioxide emissions, and industrial nitrogen oxide emissions, respectively. Principal component analysis is employed to synthesize them into one indicator to measure environmental regulation. Generally, the greater regional GDP per unit of emission, the higher the environmental regulation [53].

Industry imitation (Imi). The measurement of industry imitation should be differentiated for different EMB. Industry imitation of substantive EMB is obtained by calculating the industry average of environmental patents, which is denoted by Imi1. Industry imitation of symbolic EMB is obtained by calculating the industry average of the three items, that is, whether to disclose environmental vision, whether to release corporate social responsibility report, and whether to release environmental report, which is expressed as Imi2 [11,19].

Media attention (Media). The number of online news reports about the firm is considered. The natural logarithm is obtained by adding one to the value [24].

3.2.3. Moderating Variables

Referring to the research of Nadkarni and Narayanan (2007) [31], text analysis method is applied to measure two dimensions of executive cognition. Specifically, the section of Discussion and Analysis of Operations in the annual report of the firm is analyzed. The steps are as follows. The first step is to find the statements in the annual report describing the cause–effect relationships between environment and strategy, such the statements with key words as “because, due to, by means of, through, so, therefore, thus”, and so on. The second step is to develop the one-to-one, one-to-many, many-to-one, or many-to-many cause–effect relationships with specific concepts in statements according to the code scheme [31]. The final step is to draw a cause–effect logic map with UCINET software based on the above cause–effect relationships to obtain the relevant data.

In order to ensure the reliability and validity of the coding, the following measures are taken. First, the coding is carried out by two postgraduate students who are trained uniformly. Before the formal coding, they are asked to read the code scheme carefully to define and scope each concept. In case of disagreement between coders, full discussion and consistent interpretation are required. In the formal coding, coders read and analyze the same firm’s annual report, judge the cause–effect relationships, and draw the cause–effect logic map. If the inconsistency is less than 20%, a unified result is reached by two coders after sufficient discussion. If the inconsistency is higher than 20%, it should be judged by the expert. If no consensus is reached after full discussion, the sample should be removed.

Cognitive focus of executive (Focus). The eigenvector centrality method is employed. The centrality of each concept is measured. The largest centrality in the cause–effect logic map is selected to measure cognitive focus of executives [31]. The calculation formula is as follows.

$$Focus=Max\left\{\right[\alpha \left]\right[\sum A_{ij}{C}_j\left]\right\}$$

(1)

where α indicates the parameter that is obtained by using UCINET software to process the centrality of each concept in the cause–effect logic map of executive cognition. A_ij represents the relationship matrix between concept i and other concepts. C_j represents the centrality of concepts related to concept i.

Cognitive complexity of executive (Complexity). The cause–effect relationship between concepts is measured [31]. The calculation formula is as follows.

$$Complexity=\raisebox{1ex}{$N_R$}\!\left/ \!\raisebox{-1ex}{$N_C$}\right.$$

(2)

where N_R represents the total number of inter-concept relationships in the executive cognition. N_C represents the total number of concepts in the executive cognition.

3.2.4. Control Variables

Firm size (Size). Firm size is related to EMB of firms [54]. The natural logarithm of the number of employees in the firm is taken [11].

Firm age (Age). Firm age usually affects executives’ cognition and firms’ EMB [30,55]. The natural logarithm of the number of years the firm’s establishment.

Growth (Growth). Growth affects the environmental activities of firms [56]. The growth rate of operating revenue is taken.

Financial performance (FP). Good financial performance contributes to EMB [11]. Return on assets is adopted.

Concentration of equity (Con). It has been shown that large shareholders influence the motivation of firms to implement EMB [57,58]. The shareholding ratio of the largest shareholder is employed.

Board size (Board). In general, large board size can drive EMB [59]. The natural logarithm of the number of board members is taken [60].

Industry type (Industry). The heavily polluting industries are divided into 12 sub-industries, which are transformed into dummy variables in the empirical analysis [2].

The detailed variable abbreviations, variable measurements, and data sources are shown in

Table 1. List of variables.

Variable	Abbreviation	Measurement	Data Source
Substantive EMB	Sub	Ln(the number of environmental patents +1)	State Intellectual Property Office of the People’s Republic of China
Symbolic EMB	Sym	The scores of environmental vision, corporate social responsibility report, and environmental report	China Stock Market and Accounting Research Database, Annual Report
Environmental regulation	Reg	Principal component analysis	China City Statistical Yearbook
Industry imitation (Substantive EMB)	Imi1	Industry average	State Intellectual Property Office of the People’s Republic of China
Industry imitation (Symbolic EMB)	Imi2	Industry average	China Stock Market and Accounting Research Database, Annual Report
Media attention	Media	Ln(the number of online news reports +1)	China Listed Companies Financial News Database
Cognitive focus of executive	Focus	Cause-effect logic map	Annual Report
Cognitive complexity of executive	Complexity	Cause-effect logic map	Annual Report
Firm size	Size	Ln(the number of employees)	China Stock Market and Accounting Research Database
Firm age	Age	Ln(the number of years the firm’s establishment)	Same as above
Growth	Growth	The growth rate of operating revenue	Same as above
Financial performance	FP	Return on assets	Same as above
Concentration of equity	Con	The shareholding ratio of the largest shareholder	Same as above
Board size	Board	Ln(the number of board members)	Same as above
Industry type	Industry	Dummy variable	Same as above

.

3.3. Research Model

The following models are used for the regression.

$$\begin{gathered} {Sub}_{it}=\alpha +{\beta }_1{Reg}_{it}+{\beta }_2{Imi1}_{it}+{\beta }_3{Media}_{it}+{\beta }_4{Focus}_{it}+{\beta }_5{Complexity}_{it}+{\beta }_6{Reg}_{it}\times {Focus}_{it} \\ +{\beta }_7{Reg}_{it}\times {Complexity}_{it}+{\beta }_8{Imi1}_{it}\times {Focus}_{it}+{\beta }_9{Imi1}_{it}\times {Complexity}_{it} \\ +{\beta }_{10}{Media}_{it}\times {Focus}_{it}+{\beta }_{11}{Media}_{it}\times {Complexity}_{it}+\sum {\beta }_j{Control}_{j-it}+{\epsilon }_{it} \end{gathered}$$

(3)

$$\begin{gathered} {Sym}_{it}=\alpha +{\beta }_1{Reg}_{it}+{\beta }_2{Imi2}_{it}+{\beta }_3{Media}_{it}+{\beta }_4{Focus}_{it}+{\beta }_5{Complexity}_{it}+{\beta }_6{Reg}_{it}\times {Focus}_{it} \\ +{\beta }_7{Reg}_{it}\times {Complexity}_{it}+{\beta }_8{Imi2}_{it}\times {Focus}_{it}+{\beta }_9{Imi2}_{it}\times {Complexity}_{it} \\ +{\beta }_{10}{Media}_{it}\times {Focus}_{it}+{\beta }_{11}{Media}_{it}\times {Complexity}_{it}+\sum {\beta }_j{Control}_{j-it}+{\epsilon }_{it} \end{gathered}$$

(4)

The meanings of the variables’ abbreviations are given in Table 1. Subscript it denotes the data of the i-th firm in year t. α is the intercept term of the regression model. β is the coefficient of the variable. Subscript j denotes the j-th control variable. ε is the random disturbance term.

Regression model usually includes the mixed regression model, the random effects regression model, and the fixed effects regression model. In order to select the appropriate regression model, we compare the three regression models as follows. The Wald F test is used to compare the mixed regression model with the fixed effects regression model. The results show that the F-value is 22.53 and p < 0.01, rejecting the original hypothesis, which indicates that the fixed effects regression model is appropriate. The Lagrange Mutiplicator test is used to compare the mixed regression model with the random effects regression model. The results show that Chi2 is 157.22 and p < 0.01, rejecting the original hypothesis, which indicates that the random effects regression model is appropriate. The Hausman test is applied to compare the fixed effects regression model with the random effects regression model. The results show that Chi2 is 22.86 and p < 0.01, rejecting the original hypothesis, which indicates that the fixed effects regression model is appropriate. After the above tests, the fixed effects model is ultimately appropriate for regression.

4. Results and Discussion

4.1. Descriptive Statistics and Correlation

Table 2. Descriptive statistics.

Variable	N	Mean	S.D.	Min	Max
Sub	236	1.174	1.236	0	4.745
Sym	236	1.131	0.965	0	3
Reg	236	0.851	1.903	−1.308	5.688
Imi1	236	1.124	0.51	0.299	2.137
Imi2	236	1.086	0.292	0.429	1.667
Media	236	5.283	1.208	3.178	7.53
Focus	236	0.593	0.092	0.413	0.836
Complexity	236	0.574	0.147	0.313	0.833
Size	236	8.426	1.259	5.398	11.09
Age	236	2.671	0.508	1.386	3.296
Growth	236	0.099	0.271	−0.608	0.963
FP	236	0.042	0.060	−0.154	0.216
Con	236	37.73	16.31	8.880	74.570
Board	236	2.164	0.202	1.609	2.708

shows descriptive statistics of variables. The mean and standard deviation of substantive EMB are 1.174 and 1.236, respectively. The coefficient of variation of substantive EMB is 1.053. The mean and standard deviation of symbolic EMB are 1.131 and 0.965, respectively. The coefficient of variation of symbolic EMB is 0.853. This indicates that there is more variation in the level of substantive EMB among firms than symbolic EMB.

Table 3. Results of correlation analysis.

Variable	Sub	Sym	Reg	Imi1	Imi2	Media	Focus
Sub	1
Sym	0.326 ***	1
Reg	0.510 ***	0.182 ***	1
Imi1	0.351 ***	0.079	0.123 *	1
Imi2	0.265 ***	0.203 ***	0.100	0.303 ***	1
Media	0.320 ***	0.310 ***	0.170 ***	0.079	0.154 **	1
Focus	0.143 **	−0.015	−0.140 **	0.077	0.055	−0.083	1
Complexity	0.326 ***	0.252 ***	0.179 ***	0.198 ***	0.027	0.319 ***	−0.232 ***
Size	0.476 ***	0.512 ***	0.261 ***	0.242 ***	0.205 ***	0.596 ***	0.011
Age	−0.011	0.073	−0.065	0.132**	0.211 ***	0.051	0.060
Growth	−0.105	−0.183 ***	−0.005	0.004	−0.175 ***	−0.165 **	−0.083
FP	0.016	0.279 ***	−0.125 *	0.063	0.103	0.246 ***	−0.016
Con	0.223 ***	0.143 **	0.026	0.230***	0.212 ***	0.220 ***	−0.023
Board	0.246 ***	0.244 ***	0.144 **	0.105	0.071	0.168 ***	0.063
	Complexity	Size	Age	Growth	FP	Con	Board
Complexity	1
Size	0.241 ***	1
Age	0.101	0.125 *	1
Growth	−0.039	−0.118 *	−0.049	1
FP	0.044	0.166 **	0.049	−0.019	1
Con	0.134 **	0.382 ***	0.132 **	−0.121 *	0.153 **	1
Board	0.220 ***	0.278 ***	0.226 ***	0.035	0.033	0.048	1

Note: *** p < 0.01, ** p < 0.05, * p < 0.1.

reports the results of correlation analysis. Table 3 indicates that environmental regulation is positively related to substantive EMB (β = 0.510, p < 0.01). Industry imitation is positively related to symbolic EMB (β = 0.203, p < 0.01). Media attention is positively related to symbolic EMB (β = 0.310, p < 0.01). The results are consistent with our preliminary speculation. The maximum value of variance inflation factor (VIF) is 1.84, indicating that there is no multicollinearity problem, which satisfies the basic requirements for subsequent regression analysis.

4.2. Results of Regression

The results of regression are shown in

Table 4. Results of regression.

Variable	Sub			Sym
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Reg	0.261 ***			0.020
	(−2.74)			(−0.32)
Imi1		0.310 **
		(−1.76)
Imi2					0.410 **
					(−1.92)
Media			−0.047			0.497 ***
			(−0.25)			(−3.85)
Control	Yes	Yes	Yes	Yes	Yes	Yes
Cons	14.987 ***	15.059 ***	19.162 ***	7.469 **	4.467	6.418 **
	(−3.72)	(−3.13)	(−4.66)	(−2.2)	(−1.19)	(−2.1)
N	236	236	236	236	236	236
R2	0.263	0.235	0.204	0.048	0.079	0.171
Adjusted R2	0.241	0.211	0.180	0.018	0.051	0.145
F	6.631 ***	5.907 ***	6.057 ***	1.071 ***	1.94 ***	3.688 ***

Note: *** p < 0.01, ** p < 0.05.

. The dependent variable from model 1 to model 3 is substantive EMB, while the dependent variable from model 4 to model 6 is symbolic EMB. Model 1 shows that environmental regulation has a significant positive impact on substantive EMB (β = 0.261, p < 0.01), while model 4 shows that environmental regulation has no impact on symbolic EMB (β = 0.020), indicating that environmental regulation is more likely to promote substantive EMB. Hypothesis 1 is supported, that is, compared with symbolic EMB, environmental regulation is more likely to promote firms’ substantive EMB. Model 2 and model 5 show that industry imitation has significant positive impacts on substantive EMB (β = 0.310, p < 0.05) and on symbolic EMB (β = 0.410, p < 0.05). To further to test the strength of the effect, the Bootstrap method is used to compare the coefficient difference (symbolic EMB minus substantive EMB) and the sampling times is set to 5000. If the confidence interval is to the right of 0, it indicates that industry imitation is more likely to promote symbolic EMB. The result shows that 95% confidence interval falls on [0.105, 0.773]. Hypothesis 2 is supported, that is, compared with substantive EMB, industry imitation is more likely to promote firms’ symbolic EMB. Model 3 shows that media attention has no impact on substantive EMB (β = −0.047), while model 6 shows that media attention has a significant positive impact on symbolic EMB (β = 0.497, p < 0.01), indicating that media attention is more likely to promote symbolic EMB. Hypothesis 3 is supported, that is, compared with firms’ substantive EMB, media attention is more likely to promote symbolic EMB.

Models 7 and model 8 in

Table 5. Regression results of the moderating effects.

Variable	Sub		Sym
	Model 7	Model 8	Model 9	Model 10	Model 11	Model 12
Reg	0.253 **	0.174
	(−2.29)	(−1.53)
Imi2			0.409 *	0.415 **
			(−1.9)	(−2.22)
Media					0.495 ***	0.517 ***
					(−3.81)	(−4.02)
Focus	0.670		0.277		0.089
	(−1.55)		(−0.39)		(−0.16)
Complexity		0.674		0.023		−0.468
		(−0.92)		(−0.05)		(−1.00)
Reg × Focus	−0.147 **
	(−1.73)
Reg × Complexity		0.831 **
		(−2.27)
Imi2 × Focus			−2.515
			(−1.20)
Imi2 × Complexity				3.043 ***
				(−2.67)
Media × Focus					−0.326 *
					(−1.80)
Media × Complexity						0.685**
						(−2.19)
Control	Yes	Yes	Yes	Yes	Yes	Yes
Cons	14.509 ***	12.779 ***	4.278	4.002	6.340 **	6.293 **
	(−3.7)	(−3)	(−1.15)	(−1.22)	(−2.13)	(−2.03)
N	236	236	236	236	236	236
R2	0.338	0.309	0.088	0.126	0.183	0.202
Adjusted R2	0.312	0.281	0.052	0.092	0.150	0.170
F	5.923 ***	6.344 ***	1.729 ***	3.241 ***	3.283 ***	4.319 ***

Note: *** p < 0.01, ** p < 0.05, * p < 0.1.

test the moderating role of executive cognition in the relationship between environmental regulation and substantive EMB. The coefficient of interaction term (Reg × Focus) in model 7 is significantly negative (β = −0.147, p < 0.05), indicating that cognitive focus of executive has a moderating effect on environmental regulation and substantive EMB. In order to visualize the moderating effect, the figure is further drawn. Figure 2 shows that when cognitive focus of executive is low, environmental regulation can promote substantive EMB. However, when cognitive focus of executive increases, the promoting effect of environmental regulation on substantive EMB is weakened. It can be inferred that cognitive focus of executive plays a negative moderating role in the relationship between environmental regulation and substantive EMB, that is, Hypothesis 4a holds. The coefficient of interaction term (Reg × Complexity) in model 8 is significantly positive (β = 0.831, p < 0.05), indicating that cognitive complexity of executive has a moderating effect in the relationship between environmental regulation and substantive EMB. It is further shown by Figure 3 that the slope of the positive effect of environmental regulation on substantive EMB is greater when cognitive complexity of executive is higher, indicating that cognitive complexity of executive plays a positive moderating role in the relationship between environmental regulation and substantive EMB, that is, Hypothesis 5a holds.

Model 9 and model 10 examine the moderating effect of executive cognition in industry imitation and symbolic EMB. The coefficient of interaction term (Imi2 × Focus) in model 9 is insignificant (β = −2.515). Hypothesis 4b is not supported, that is, cognitive focus of executive fails to play a negative moderating role in the relationship between industry imitation and symbolic EMB. The coefficient of interaction term (Imi2 × Complexity) in model 10 is significantly positive (β = 3.043, p < 0.01), indicating a moderating effect of cognitive complexity of executive in industry imitation and symbolic EMB. As shown by Figure 4, the promoting effect of industry imitation on symbolic EMB is stronger when cognitive complexity of executive is higher, indicating that cognitive complexity of executive plays a positive moderating role in the relationship between industry imitation and symbolic EMB, that is, Hypothesis 5b holds.

Similarly, model 11 and Figure 5 suggest that cognitive focus of executive plays a negative moderating role in the relationship between media attention and symbolic EMB, that is, Hypothesis 4c holds. Model 12 and Figure 6 suggest that cognitive complexity of executive plays a positive moderating role in the relationship between media attention and symbolic EMB, that is, Hypothesis 5c holds.

4.3. Robustness Tests

Firstly, the robustness test is conducted by using the measurement of variables. Substantive EMB is measured by whether the firm actually carries out special activity on environmental protection that involve investment of human and material resources, that is, 1 for adopting this activity and 0 otherwise. The pursuit of ISO 14001 certification by firms is a typical symbolic EMB [11]. Therefore, symbolic EMB is measured by whether the firm obtains ISO 14001 certification, that is, 1 for obtaining the certification and 0 otherwise [11]. Meanwhile, return on equity is used to measure financial performance instead of return on assets [61]. Regression analysis is performed using probit model, since the dependent variables are the binary discrete variables.

Table 6. Regression results of robustness test (changing the measurement of variables).

Variable	Sub			Sym
	Model 13	Model 14	Model 15	Model 16	Model 17	Model 18
Reg	0.326***			0.037
	(−5.18)			(−0.74)
Imi1		0.154
		(−0.89)
Imi2					0.922 ***
					(−2.75)
Media			0.062			0.716 ***
			(−0.71)			(−6.48)
Control	Yes	Yes	Yes	Yes	Yes	Yes
Cons	−2.277 **	−2.759 ***	−2.817 ***	−6.666 ***	−7.348 ***	−8.095 ***
	(−2.10)	(−2.73)	(−2.78)	(−5.33)	(−5.78)	(−5.86)
N	236	236	236	236	236	236
LR chi2	51.61	17.38	17.10	59.91	67.08	108.78
Pseudo R2	0.1580	0.0532	0.0523	0.1866	0.2090	0.3389

Note: *** p < 0.01, ** p < 0.05.

and

Table 7. Regression results of the moderating effects of robustness test (changing the measurement of variables).

Variable	Sub		Sym
	Model 19	Model 20	Model 21	Model 22	Model 23	Model 24
Reg	0.320 ***	0.463 ***
	(−4.61)	(−4.53)
Imi2			1.048 ***	1.278 ***
			(−3)	(−3.46)
Media					0.679 ***	0.718 ***
					(−5.94)	(−5.63)
Focus	2.628 **		−1.564		−0.258
	(−2.2)		(−1.50)		(−0.24)
Complexity		1.081		3.258 ***		2.063 ***
		(−1.43)		(−4.57)		(−2.64)
Reg × Focus	−1.371 ***
	(−3.43)
Reg × Complexity		2.454 ***
		(−4)
Imi2 × Focus			5.485
			(−1.64)
Imi2 × Complexity				4.905 *
				(−1.8)
Media × Focus					−1.847 *
					(−1.72)
Media × Complexity						2.568 ***
						(−3.14)
Control	Yes	Yes	Yes	Yes	Yes	Yes
Cons	−2.907 **	−2.646 **	−6.720 ***	−8.323 ***	−7.896 ***	−9.167 ***
	(−2.20)	(−2.27)	(−5.02)	(−6.09)	(−5.28)	(−6.03)
N	236	236	236	236	236	236
LR chi2	83.37	72.68	71.45	92.27	111.78	129.81
Pseudo R2	0.2552	0.2224	0.2226	0.2874	0.3482	0.4044

Note: *** p < 0.01, ** p < 0.05, * p < 0.1.

show the consistent results, indicating that the findings are robust.

In addition, there may be sample selection bias since the sample consists of only 118 firms. Thus, the Heckman two-stage method is applied as a robustness test. The probit model is applied in the first stage to test whether the firm implements EMB, that is, 1 for implementation and 0 otherwise. The inverse Mills ratio (IMR) generated in the first stage is placed in the regression model as a control variable in the second stage.

Table 8. Regression results of robustness test (Heckman two-stage method).

Variable	Sub			Sym
	Model 25	Model 26	Model 27	Model 28	Model 29	Model 30
Reg	0.240 ***			0.018
	(−5.45)			(−0.53)
Imi1		0.843
		(−1.95)
Imi2					0.510 **
					(−2.14)
Media			0.025			0.598 ***
			(−0.13)			(−15.44)
IMR	−0.777	−3.109 **	−2.673 **	−0.436	−0.937 **	−0.220
	(−1.19)	(−1.85)	(−1.86)	(−0.84)	(−1.78)	(−0.73)
Control	Yes	Yes	Yes	Yes	Yes	Yes
Cons	−3.565 ***	−4.799 *	−4.686 *	−3.011 ***	−3.413 ***	−3.382 ***
	(−4.33)	(−1.97)	(−2.24)	(−4.66)	(−4.56)	(−7.97)
N	236	236	236	236	236	236

Note: *** p < 0.01, ** p < 0.05, * p < 0.1.

and

Table 9. Regression results of the moderating effects of robustness test (Heckman two-stage method).

Variable	Sub		Sym
	Model 31	Model 32	Model 33	Model 34	Model 35	Model 36
Reg	0.230 ***	0.168 ***
	(−5.36)	(−3.66)
Imi2			0.483 *	0.448 *
			(−2.28)	(−2.21)
Media					0.588 ***	0.588 ***
					(−14.73)	(−14.42)
Focus	1.758 *		−0.695		0.226
	(−2.4)		(−1.14)		(−0.55)
Complexity		1.191 *		0.654		−0.384
		(−2.22)		(−1.42)		(−1.22)
Reg × Focus	−1.653 **
	(−3.13)
Reg × Complexity		1.151 ***
		(−4.48)
Imi2 × Focus			1.068
			(−2.64)
Imi2 × Complexity				1.287 **
				(−1.03)
Media × Focus					−0.496 **
					(−1.42)
Media × Complexity						0.479 **
						(−2.13)
IMR	−0.265	−0.298	−0.820 ***	−0.394	−0.182	−0.308
	(−0.43)	(−0.41)	(−1.74)	(−0.69)	(−0.60)	(−0.85)
Control	Yes	Yes	Yes	Yes	Yes	Yes
Cons	−4.125 ***	−2.940 ***	−3.048 ***	−3.369 ***	−3.502 ***	−3.369 ***
	(−4.89)	(−3.91)	(−4.16)	(−5.39)	(−7.36)	(−7.93)
N	236	236	236	236	236	236

Note: *** p < 0.01, ** p < 0.05, * p < 0.1.

show the same results, demonstrating the robustness of the findings.

4.4. Discussion

Consistent with the previous studies, environmental regulation can promote EMB, such as green innovation [62,63] and emission reduction [64,65]. This paper further compares the strength of the effect of environmental regulation on substantive EMB and symbolic EMB. The results show that compared with symbolic EMB, environmental regulation is more likely to promote firms’ substantive EMB. Environmental regulation is of a rigid and deterrent character. It can provide policy guidance and financial subsidies to firms for product innovation and technological improvement, which involves substantive changes.

Previous studies have shown that industry imitation can promote EMB, such as environmental information disclosure [66], environmental reporting [67], and environmental innovation [68], which is consistent with our findings. This paper further shows that compared with substantive EMB, industry imitation is more likely to promote firms’ symbolic EMB. This is because symbolic EMB involves simple and external behavior such as disclosure and publicity. The behavior of other firms in the same industry can provide a template for the target firm to optimize its own EMB.

The findings of this paper indicate that compared with firms’ substantive EMB, media attention is more likely to promote symbolic EMB, which is in line with previous research into the idea that media attention can promote EMB [69]. Media attention, through coverage of either positive or negative environmental news about firms, is closely related to symbolic EMB. These findings confirm that the effects of different types of macro-institutional pressures on firms’ heterogeneous EMB are also different.

The effects of micro-institutional pressures in the relationship between macro-institutional pressures and firms’ heterogeneous EMB are also explored in this paper. Extensive previous research has been conducted on the relationship between executive cognition and EMB. Executive cognition affects EMB, such as low-carbon management [70] and eco-innovation [71]. This paper further incorporates executive cognition into an integrated research framework of macro-institutional pressures and firms’ heterogeneous EMB. The findings suggest that cognitive focus of executive plays negative moderating roles in the relationship between environmental regulation and substantive EMB, as well as in the relationship between media attention and symbolic EMB, implying that cognitive focus of executive inhibits macro-institutional pressures from driving EMB. In addition, cognitive complexity of executive positively moderates the relationship between macro-institutional pressures and EMB, that is, in the environmental regulation-substantive EMB, industry imitation-symbolic EMB, and media attention-symbolic EMB relationship, indicating that cognitive complexity of executive enables macro-institutional pressures to effectively drive EMB, whether substantive or symbolic. The findings on micro-institutional pressures further explain the reason for the heterogeneity of firms’ EMB.

In contrast to the expected Hypothesis 4b, cognitive focus of executive does not play a negative moderating role in the relationship between industry imitation and symbolic EMB. This may be because industry imitation is widespread and accessible, targeting not only executives, but also board members, employees, and other stakeholders who have access to the templates and imitate them [19]. Furthermore, symbolic EMB does not involve the actual change in production and operation, which is relatively easy to implement [9]. Thus, the resistant effect of cognitive focus of executive in the promoting path of industry imitation on symbolic EMB is not obvious, that is, there is no negative moderating effect of cognitive focus of executive in the relationship between industry imitation and symbolic EMB.

5. Conclusions and Implications

5.1. Conclusions

Based on macro-institutional theory and micro-institutional theory, the impacts of three types of macro-institutional pressures and the moderating roles of micro-institutional pressures on firms’ heterogeneous EMB are explored by using the sample of Shanghai and Shenzhen A-share 118 listed companies in the heavy-polluting industries from 2019 to 2020. The results show that (1) environmental regulation is more likely to promote substantive EMB, while industry imitation and media attention are more likely to drive symbolic EMB. (2) Cognitive focus of executive plays negative moderating roles in the relationship between some macro-institutional pressures and EMB. Specifically, the negative moderating roles of cognitive focus of executive exist in the relationship between environmental regulation and substantive EMB and the relationship between media attention and symbolic EMB. Cognitive complexity of executive plays positive moderating roles in the relationship between macro-institutional pressures and EMB. Specifically, the positive moderating roles of cognitive complexity of executive exist in the relationship between environmental regulation and substantive EMB, the relationship between industry imitation and symbolic EMB, and the relationship between media attention and symbolic EMB.

5.2. Implications

5.2.1. Implications for Academia

Firstly, it enriches the relevant research in the field of environmental management. Most of the previous research on firms’ EMB has tended to take EMB as a whole or only focus on a specific environmental measure. We can learn that firms’ EMB is diverse and different. This paper provides a dimensional division of heterogeneous EMB. Based on the characteristics and effects of EMB, it is classified as substantive EMB and symbolic EMB. The findings of study based on this dimensional division are accurate.

Secondly, a fine-grained comparative study of the relationship between macro-institutional pressures and firms’ heterogeneous EMB is conducted. This paper refines and identifies the three main sources of macro-institutional pressures, namely, environmental regulation, industry imitation, and media attention. The strengths of the effects of three types of macro-institutional pressures on EMB are compared. The findings found that environmental regulation is more likely to promote substantive EMB, while industry imitation and media attention are more likely to drive symbolic EMB, confirming the conjecture that the effects of different types of macro-institutional pressures on EMB are different, which helps to precisely identify the motivation of heterogeneous EMB.

Finally, research in the micro-institutional domain has been deepened and enriched. A combined research framework of macro-domain and micro-domain has been developed. Most of the previous studies are based on macro-institutional theory to explore the effects of macro-institutional pressures on firms’ EMB, while neglecting the micro-institutional domain. Micro-institutional pressures are an important component of the institutional environment. This theory focuses on the role of decision makers in the process of dealing with external pressures in the firm, highlighting the effect of individual cognition in the firm’s decisions and behavior. In discussing the effects of different types of macro-institutional pressures on heterogeneous EMB, micro-institutional pressures are integrated as moderating variables through a combination of macro-domain and micro-domain studies to provide a precise explanation for firms’ heterogeneity EMB.

5.2.2. Implications for Regulators

It is found that environmental regulation can effectively promote substantive EMB. The findings suggest that specific external conditions are necessary for substantive behavior. Therefore, it is crucial to explore the environment in which firms are more willing to engage in substantive EMB. In the construction of the institutional environment, regulators should make full use of laws, regulations, and financial means to promote substantive EMB, thereby improving the overall environmental management of the industry. In addition, the findings suggest that industry imitation and media attention promote symbolic EMB. Regulators promote the optimization of symbolic EMB such as disclosure and publicity by regulating the practice templates disseminated within the industries and monitoring of media disclosure mechanisms, in order to reduce information asymmetries and improve environmental governance.

5.2.3. Implications for Executives

Conducting EMB is necessary for improving environmental quality and achieving sustainable development. Executives should accurately identify substantive EMB and symbolic EMB, clarify the differences between the two in terms of objectives and measures, and develop different environmental strategies and schemes. Substantive EMB and symbolic EMB have different characteristics and effects. Specifically, substantive EMB improves the technology and management of firms through environmental protection behavior, such as technological upgrade, product innovation, recycle, emission reduction, and so on, to gain competitive advantages. Symbolic EMB can green the external image and gain good reputation through publicity, disclosure, and commitment. Executives can integrate both EMBs to achieve sustainable development of firms.

Executives are encouraged to take the necessary initiatives to deal with institutional pressures. With regard to environmental regulation, executives should keep abreast of the relevant policies issued by government departments, such as environmental laws, regulations, and guidelines, to clarify policy direction and obtain cutting-edge information of the industry. With regard to industry imitation, executives ought to be open-minded and engage in timely communication with other firms to learn about excellent environmental management practices. With regard to media attention, it is advisable for executives to actively promote the firm’s internal environmental policy, objectives, and philosophy to present an eco-friendly image to the media and the public.

Executive cognition is a moderator for firms to respond to external pressures. A high level of cognitive complexity helps executives to receive novel information and updated knowledge from the environment. Executives conduct EMB to address macro-institutional pressures by integrating information, measuring the risk-benefit of projects, and raising awareness of environmental management. Consequently, it is essential to increase cognitive complexity of executive in order to promote environmental governance. There are several ways to increase cognitive complexity of executive. Firstly, establishing external communication mechanism by executives is crucial, such as connecting with professional organizations and other firms, collaborating with suppliers, perceiving the needs of customers, which helps broaden external information channels and enrich executive cognition. It is desirable to take measures such as organizing forums, attending lectures, and distributing questionnaires. Moreover, optimizing internal communication mechanism should also be taken seriously by executives. Promoting cooperation between the top and bottom level to ensure the accurate release of policies and timely feedback on issues is necessary. Communication skills training for staff should also be carried out. Executive cognition is in line with the actual complex situation of the firm, so that they can make targeted environmental decisions.

5.3. Limitations and Future Research

Due to the unavailability of some data, only a sample of 118 companies in the heavily polluting industries was selected. We hope to expand the sample size and industry in future. Meanwhile, only the effects of environmental regulation, industry imitation, and media attention on heterogeneous EMB are verified. Future research can explore other components of macro-institutional pressures to provide more explanations for firms’ heterogeneous EMB.