The Effect of Local Government Environmental Concern on Corporate Environmental Investment: Evidence from China

Abstract

This paper uses machine learning tools to construct local government environmental concern indicators and empirically examines the impact of local government environmental concern on corporate environmental investment. From the resource endowment perspective, corporate resources’ moderating role is also verified. The major findings are as follows: (1) local government environmental concern has a significant positive effect on the environmental investment of corporations in their jurisdictions; (2) corporations with fewer financial and political resources will pay more attention to the local government’s intention when making environmental investment decisions, and the promotion effect of local government environmental concern on the environmental investment of such corporations is more prominent. Further analysis shows that this promotion effect is more significant in regions with a high intensity of environmental regulation and high levels of economic development, and is more effective for key regulated corporations. This paper verifies the effect of local government on micro-corporations in environmental governance from the perspective of environmental concern, broadens the boundary of research on the relationship between government and corporate environmental responsibility fulfillment, and enriches the study of factors influencing corporate environmental investment behavior. It also provides important empirical evidence for central and local governments to implement green development and build a government–business collaborative environmental governance system.

1. Introduction

Since the reform and opening up, China’s socialist economy has made world-renowned achievements, but the rapid economic development has brought a series of ecological problems, such as environmental pollution and resource depletion. As the considerable cost of environmental management has become a significant obstacle to economic development, China must transform its economic development mode and establish a green and coordinated development pattern. The Chinese government is placing increasing emphasis on high-quality economic development. The 18th National Congress of the Communist Party of China (CPC) included “building ecological civilization” in the party constitution. The 2022 Government Work Report of China also sets the promotion of comprehensive green transformation of China’s economy and society as the main direction of economic development. The above-mentioned programmatic documents indicate that building ecological civilization has been elevated to an unprecedented strategic level, and promoting the coordinated development of economic and environmental protection has become an essential task in China’s new development stage.

As micro-entities of economic operation, corporations’ production and operation activities are also the main sources of environmental pollution [1]. Integrating corporations into the environmental governance system and guiding them to fulfill their environmental responsibilities are crucial in ecological civilization construction. Under the pressure transmission pattern of “Central Government-Local Government-Corporation”, whether the central government’s environmental protection strategy can effectively promote corporations’ environmental management mainly depends on local governments’ policy implementation efforts. However, China’s government assessment mechanism has long-term-followed the “promotion tournament” model with GDP as the core indicator [2]. The distorted incentive mechanism induces short-sighted behavior of local governments and “bottom-up competition” in enforcing environmental regulations [3]. Some local governments even acquiesce to polluters’ excessive emissions to generate more tax contributions, ultimately leading to the failure of local environmental governance [4]. In order to raise the attention of local governments to environmental protection and to promote the effective implementation of environmental protection policies in various regions, the central government has introduced a series of binding measures for local governments, such as central environmental protection inspections, green performance appraisal, and leading cadres’ natural resource asset discharge auditing. These policies force and guide local governments to actively manage the environment through vertical supervision and incentive mechanism optimization and indirectly improve corporations’ environmental performance in their jurisdictions [5,6,7].

Environmental investments have the characteristic of seeking economic and social benefits simultaneously, while the benefits are slow to materialize. As profit-oriented organizations, corporations lack the initiative to make environmental investments [8,9]. Therefore, the regulatory pressure exerted by the government is an essential external driver for corporations to make environmental investments. A review of the literature reveals that different types of environmental regulations, such as environmental standards and regulations [4,10], environmental penalties [11], environmental courts [12], government enforcement strength [13], and environmental taxes [14,15], all have significant effects on corporate environmental investment. It has also been shown that indirect regulations enacted for local governments and officials, such as the central environmental protection inspectors [16] and the leading officials’ accountability audit of natural resources [17], also promote corporate environmental investment. However, since the environmental regulations faced by corporations are complex and diverse, the meaning of exploring the impact of a single regulation on corporate environmental investment is limited. According to the theory of planned behavior, environmental regulation is only a behavioral outcome driven by the government’s environmental beliefs. Conducting research from the perspective of the government’s environmental beliefs rather than from the perspective of a particular regulation can provide a more comprehensive assessment of the government’s impact on corporate environmental governance. Accordingly, this paper uses a machine learning method to analyze the local government work reports to construct environmental concern indicators and explore the impact of local government environmental concern on corporate environmental investments, thereby addressing the following questions. Does the local government environmental concern increase corporate environmental investment? Which characteristics of corporations are more likely to be influenced by the local government’s call for environmental protection? What economic and social externalities facilitate or inhibit this influence?

Compared with existing studies, the marginal contributions of this paper are mainly as follows: (1) This study verifies the influence of local government environmental concern on corporate environmental investment. Using empirical evidence from China, it reveals the effect of local government environmental governance motives on micro-corporation environmental performance at the awareness level, broadening the boundaries of research on the relationship between government and corporate environmental responsibility fulfillment while enriching the study of factors influencing corporate environmental investment behavior. (2) This study analyzes the influence mechanism of local government environmental concern on corporate environmental investment based on legitimacy and resource dependence theories. Starting from the corporate endogenous characteristics (resource endowment characteristics), it explores the differences in the sensitivity of corporations with different resource bases to the impact of local government environmental concern when making environmental investment decisions. Further, in the dimension of the external environment, the heterogeneity of the local government environmental concern’s effect on corporate environmental investment is explored. (3) Based on the traditional dictionary method, this paper introduces machine learning tools to analyze local government work reports. Using the “Seed Word Set + Word2Vec Similar Word Expansion” method, a scientific and objective text index is constructed, which can be applied to measure the local government environmental concern. It provides a new method for analyzing government work reports and guides future intersection research of econometrics and machine learning.

2. Theoretical Analysis and Research Hypothesis

2.1. Local Government Environmental Concern and Corporate Environmental Investment

Environmental concern can be defined as the perception of environmental problems and the willingness to engage in environmental activities [18,19]. Much of the literature on environmental concern has paid particular attention to its impact on corporate operations. Early studies have shown that environmentally conscious corporations have better financial performance than non-environmentally conscious corporations [20]. Johnson (1998) pointed out that the environmental preferences of multiple entities, such as suppliers and customers, influence corporate environmental decisions [21]. Accordingly, he constructed a matrix of green business indicators involving six entities. Many studies indicated that public environmental concern (PEC) influences corporate product line design decisions [22], reduces urban–rural pollution disparities of corporations [23], and significantly promotes corporate environmental investments [24]. Ahmed et al. (2003) found that corporate employees’ environmental concerns facilitate their environmental efforts [25]. The higher the senior managers’ environmental concerns, the more resources the corporation devotes to environmental activities. Cao and Chen (2019) believed that senior managers who are concerned about the environment have a better ability to perceive the pressures of environmental regulations and identify potential benefits and market opportunities that arise from environmental activities [26]. It has also been noted that government environmental concern is beneficial to developing environmental policies [27], and governments with high environmental concern tend to set high environmental tax rates for corporations [28]. However, the influence of government environmental concern on corporate environmental performance is still not explored.

Governmental concern for environmental issues is crucial to environmental governance. On the one hand, this is because the government can play an irreplaceable role in environmental governance by resource allocation, regulation formulation, and administrative intervention [29]. On the other hand, it is because the government shoulders a certain responsibility in the environmental field. According to the 2021 Chinese General Social Survey (CGSS), 51.33% of the surveyed residents believe the government should be responsible for mitigating China’s environmental problems [30]. In recent years, the Chinese central government has enacted several policies to incorporate environmental performance indicators into officials’ political performance assessments. These policies have changed the objective function of local officials during their tenure and elevated their environmental concern.

As micro-entities of environmental governance, corporations make environmental investments for two original motivations, yielding to compliance pressure [31,32] and seeking a competitive advantage [33]. Various factors influence corporate environmental investment behavior. However, from the perspective of the original motivation, the influence of external regulatory factors created by the government is more critical than internal factors such as corporate size [34,35], property rights nature [36], internal control [37], and industry attributes [38]. Promoting the fulfillment of corporate environmental responsibility in the jurisdiction is an essential task for local governments’ environmental governance. Whether local governments’ environmental concern can positively affect corporate environmental investment needs to be discussed from several perspectives.

From a legitimacy perspective, corporations have to make environmental investments to cope with environmental pressures. The central government’s green development strategy forces local governments to pay more attention to the environment while also transmitting environmental pressures to the corporations in their jurisdictions. This pressure is comprehensive and challenging to measure. It consists of both the institutional pressure generated through the implementation of environmental regulations and the non-institutional pressure generated through the promotion of public concern for the environment. According to organizational legitimacy theory, corporate behavior needs to comply with socially accepted norms, values, beliefs, and definitions [39]. In order to alleviate environmental pressure, corporations will reduce the negative externalities of their production and operation activities on the environment through environmental behavior, shoulder the corresponding social responsibility [40], and motivate employees to effectively communicate with the public and media to maintain sustainable growth [41], thus satisfying the environmental expectations of the local government and other social agents and improving environmental legitimacy to obtain space for survival and development. Under the assumption of law-abiding citizenship, corporations follow the local government’s environmental concern to make environmental investments, which can avoid the potential risk of suffering from local government penalties and legal sanctions. This behavior also shows the willingness to actively cooperate with the local government’s green development to stakeholders such as investors, consumers, and suppliers, thus creating an image of consciously fulfilling social responsibility and sending a positive signal that the corporation is in line with the social value system [42].

From a resource dependence perspective, corporations are not always in a passive position when facing compliance pressures. In order to access diversified resources such as capital, information, and technology, corporations tend to take actions to conform to the institutional environment proactively [43]. Especially in China’s socialist system, local governments hold a large number of resources and have the power to allocate resources [44], which makes the local government significantly impact corporate behavior [45,46]. Assisting local governments with environmental management tasks is beneficial for corporations to maintain a favorable interactive relationship with the local government. In order to gain the opportunity to obtain government resources or policy support from the government–business relationship, corporations have the incentive to actively follow the local government’s environmental concern to increase environmental protection investment [47]. At the same time, the green competitive advantage that the environmental investment behavior can bring to the corporation will also enhance the corporate initiative. From a cost–benefit perspective, environmental investment causes an increase in production costs in the short term [48]. However, the corporate environmental investment will eventually lead to the completion of green technological change, break the inherent production and business model, and then play a compensating effect to achieve the dual goals of profitability and environmental protection [49].

This positive link between local government environmental concern and corporate environmental investment is not definitive, due to the existence of an attitude–behavior gap [50]. Local governments may acquiesce to corporate pollution in their jurisdictions for economic development, generating opportunistic behaviors such as incomplete enforcement, environmental data falsification, and government–business collusion in the policy implementation process [3,49,51]. The local government environmental concern cannot be effectively translated into actual environmental management behavior in this scenario. However, considering the continuous improvement in the Chinese government’s assessment system and the increasingly strict central government environmental vertical management in recent years, the incentive and space for local governments to connive at pollution have become minimal. In addition, to avoid the environmental governance costs arising from the increase in local government environmental concern, corporations may respond in negative ways, such as reducing production, shutting down production, and shifting production and operation activities to regions with lower environmental pressure [52]. At this point, local government environmental concern cannot positively impact corporate environmental investment. Nevertheless, production cuts, shutdowns, and relocations also directly impact the daily operations of corporations, resulting in significant sunk costs. At the same time, the environmental investment by Chinese corporations is generally insufficient, and local government regulation still has an “incremental” impact on the scale of corporate environmental investment [10]. Based on rational economic man theory, most corporations at this stage are more inclined to choose profitable environmental investments rather than costly passive avoidance in the face of increased local government environmental concern.

Based on the above analysis, the following hypothesis is proposed.

Hypothesis 1 (H1).

Local government environmental concern is positively related to corporate environmental investment.

2.2. Moderating Role of Corporate Resource Endowment

Due to differences in their endogenous characteristics, micro-corporations will naturally differ in sensitivity to local government environmental concern. Since resource endowment affects both corporate legitimacy and resource dependence, we select it to represent corporate endogenous characteristics to explore its moderating role in the relationship between local government environmental concern and corporate environmental investment. When exploring the relationship between government and corporations, the priority factor that needs to be considered is political resources. Government subsidies, as non-reimbursable transfer payments provided by the government to micro-corporations, directly reflect political resources. Compared to corporations with more government subsidies, those with fewer subsidies are more eager for government resources to compensate for their competitive shortcomings. This situation makes them more inclined to cater to the standards set by the local government to obtain more government resources. Political connections, as an informal relationship between government and business, often bring many potential benefits to corporations [36], representing the political resources owned by corporations. Local governments may design biased penalties to shield politically connected corporations, making politically connected corporations generally subject to less stringent environmental controls. When non-compliance costs are lower than environmental investment, politically connected corporations will have little incentive to cooperate with the local government’s environmental actions. In contrast, non-politically connected corporations are more likely to be the target of selective regulation due to the lack of political protection [4], and their pressure and constraint from local government environmental concern are substantial. In exchange for the legitimacy needed to survive and thrive, non-politically connected businesses are more incentivized to follow the local government and make more environmental investments.

Similarly, corporations with insufficient financial resources are more sensitive to local government policy directions. It is true that when corporations face lower financing constraints, they face less cash flow pressure and are more capable of making investments [53]. However, the investment behavior of corporations with high financing constraints is more significantly influenced by the policy environment to obtain more financial resources [54]. Under the local government’s call for environmental protection, such corporations are more inclined to follow the policy direction to invest in environmental projects, which can help them achieve green transformation and broaden access to financing. Also, corporations with higher debt costs are more concerned about interacting with the local government in fulfilling environmental responsibilities. Corporations can show consistency with local government concern through environmental investment, conveying a good corporate image to creditors. Banks and other financial institutions will increase their willingness and confidence to lend to such companies, reduce the risk expectation of debt default, and decrease interest rates to lessen corporate borrowing costs.

In summary, from both legitimacy and resource dependence perspectives, corporations with fewer resources will respond more positively when perceiving local government environmental concern. On the one hand, corporations with fewer resources have less voice in their industries and regions and are more sensitive to policy directions. A lack of protection creates a sense of insecurity that enables them to cooperate with the local government in environmental management to seek their own legitimacy. On the other hand, corporations with fewer resources will also emphasize cooperation with the local government to seek resources and policy support that enable them to gain a competitive advantage.

Therefore, the following hypotheses are proposed.

Hypothesis 2a (H2a).

The fewer political resources corporations have, the stronger the promotion of local government environmental concern on corporate environmental investments.

Hypothesis 2b (H2b).

The fewer financial resources corporations have, the stronger the promotion of local government environmental concern on corporate environmental investments.

The logical framework of the theoretical part is shown in Figure 1.

3. Research Design

3.1. Sample Selection and Data Sources

We select the samples of Chinese listed corporations in Shanghai and Shenzhen A-shares that disclose the amount of environmental investment from 2008 to 2020. To further enhance the data representativeness, we treat the samples as follows: (1) exclude sample companies with the trading status of ST and PT; (2) exclude financial and insurance companies; (3) exclude samples with missing or abnormal variable data; and (4) winsorize all continuous variables at the upper and lower 1% level to eliminate the effect of extreme values. Finally, 1752 valid observations are obtained.

The data for this paper come from the following sources: (1) the data on corporate environmental investment are collected manually by the author from the social responsibility reports of listed corporations, such as the “Social Responsibility Report”, “Sustainable Development Report”, and “Environment, Society and Governance”; (2) the local government environmental concern data are obtained from the results of text analysis that is based on prefecture-level city government work reports; (3) the government subsidy data are obtained from the WIND database [55]; (4) the rest of the financial and corporate governance data are obtained from the CSMAR database [56].

3.2. Variable Definition and Measurements

(1)

Explained variable: Environmental investment (EI): We manually collect and sum up corporate environmental investment data from corporate social responsibility reports, including the “Social Responsibility Report”, “Sustainable Development Report”, and “Environment, Society and Governance” according to keywords such as “environmental investment”, “environmental investment”, “environmental management”, and “green investment”. On this basis, we take the natural logarithm of this data to measure the corporate environmental investment.

(2)

Explanatory variable: Local governmental environmental concern (ER20). The government work report is an annual summary and outlook of the government’s work—a visual representation of the government’s concern for the year—and the description of environmental protection reflects the government’s concern for environmental protection. In previous studies, scholars have used the dictionary method to calculate the frequency of environmental protection words or sentences to analyze the proportion of environmental protection content in local government work reports [57]. This paper improves the original dictionary method and adopts the “Seed Word Set + Word2Vec Similar Word Expansion” method to construct a local government environmental concern indicator. Compared with the word set of the dictionary method, the word set constructed by this method is more objective and comprehensive. The specific steps of the method are as follows. Firstly, we construct the seed word set (in Chinese) of environmental protection by reading the related literature and local government work reports. Second, the Word2Vec neural network similar word algorithm is used to expand the word set based on the seed word set. We collect each seed word’s first 20 similar words and remove the duplicate words to determine the environmental word set (in Chinese). Finally, we count the number of environmental protection words in the local government work report based on the environmental word set and calculate its proportion to the total number of words in the local government work report, then multiply by 100 as the local government environmental concern indicator (ER20). The seed words and some extended words of the environmental word set in this paper are shown in

Table 1. Environmental word set overview.

Environmental Word Set
Seed Word Set	Environmental protection, pollution control, dust reduction, dust removal, emissions, pollution management, waste classification, emission reduction, pollution, pollutants, solid waste, pesticides, carbon dioxide, consumption, conservation, circular economy, clean production, resourcefulness, energy, clean energy, coal to electricity, emissions, backward production capacity, heavy chemical industry, collaborative development, integrated watershed management, regional coordinated development, collaborative governance, joint governance, blue sky, wetlands, river chief system, stay green, livable, afforestation, forest fields, re-greening, harmonious coexistence, greening, ecology. (40 in total)
Expanded Word Set	Environmental protection, environmental governance, protection, environmental protection, building and raising, ecological construction, pollution prevention, protection of the environment, emission reduction, environmental monitoring, environmental pollution, ecological environment, energy saving, resource management, environmental protection regulations, civilization, environmental protection inspectors, dam water, background, environmental management, atmosphere, pollution control, haze, water, gas, coal pressure, coal control, car control, haze reduction, coal reduction, car control, dust reduction, pollution control, pollution management, mountain, dust suppression, dust control, etc. (614 in total)

.

Based on the above calculation process, we can obtain the values of local governmental environmental concern (ER20) for each prefecture-level city in China. We further calculate the mean value of local government environmental concern (ER20) for each city during the research period and draw a map reflecting the environmental concern strength of Chinese local governments, as shown in Figure 2. Missing values in a few regions are filled using the Kriging Interpolation method.

It can be seen from Figure 2 that the governments of northern cities around the Chinese capital Beijing are significantly more concerned about the environment than other regions. Meanwhile, the governments of coastal cities in eastern China, which have high economic development and legalization, also pay more attention to the environment. Conversely, governments in some cities of central China are less attentive to the environment, and the local governmental environmental concern is considerably lower in most northeastern and western cities.

(3)

Moderator variables: we select government subsidies and political connection as moderator variables that reflect corporate political resources while selecting financing constraint and debt cost as moderator variables that reflect financial resources. Government subsidies (GOV): the value of government subsidies received by the corporations and normalized with the operating income at the end of the period. Political connection (PC): referring to Fan et al. (2007) [58], we assign a value of 1 if the CEO or chairman is a former government official, a member of the National People’s Congress (NPC), or a member of the Chinese People’s Political Consultative Conference (CPPCC), and 0 otherwise. Financing constraint (SA): referring to Hadlock and Pierce (2010) [59], we take the absolute value of −0.737 × Size + 0.043 × Size2 − 0.04 × Age. Cost debt (CD): referring to Pittman and Fortin (2004) [60], we use the proportion of total corporate interest expense to total liabilities at the end of the period.

(4)

Control variables: With reference to relevant studies, we select corporate size (SIZE), board size (BDS), the dual role of the chairman (DUA), and cash flow (CASH) as control variables reflecting the basic conditions of the corporation, and debt-paying capacity (LEV), profit capacity (ROA), operating capacity (OC), and risk level (IL) as control variables reflecting the capabilities of the corporation.

The variables and definitions are shown in

Table 2. Research variables and definitions.

Variable Type	Variable Name	Variable Symbol	Variable Definition
Explained variable	Corporate Environmental Investment	EI	Natural logarithm of total corporate environmental investment
Explanatory variable	Local government environmental concern	ER20	(Word frequency of environmental keyword set/total number of words in the local government work report) × 100
Moderator variables	Government subsidies	GOV	Government subsidies received by corporations
	Political connection	PC	Assign a value of 1 if the CEO or chairman is a former government official, or NPC member, or CPPCC member, otherwise 0
	Financing constraints	SA	Calculate according to SA Index
	Cost debt	CD	Total corporate interest expenses/total liabilities
	Corporate size	SIZE	Natural logarithm of corporate total assets at the end of the period
	Board size	BDS	Number of the board of directors of listed companies
	The dual role of the chairman	DUA	Set as a dummy variable, the unity of the two positions of chairman and CEO is 1; otherwise, it is 0
	Cash Flow	CASH	(Net increase in cash and cash equivalents—Net cash flows from financing activities)/Total assets
	Debt-paying capacity	LEV	Total liabilities/total assets
	Operating capacity	OC	Operating income/total assets
	Profit capacity	ROA	Net income/total assets
	Risk level	IL	(Net income + income tax expense + finance costs + depreciation of fixed assets, depreciation of oil and gas assets, depreciation of productive biological assets + amortization of intangible assets + amortization of long-term amortization expenses)/(Net income + income tax expense)

.

3.3. Research Model

We construct model (1) to test H1: the effect of local government environmental concern on corporate environmental investment.

$${\mathrm{EI}}_{\mathrm{i},\mathrm{t}}=\mathsf{\alpha }+{\mathsf{\beta }}_0\mathrm{ER}{20}_{\mathrm{i},\mathrm{t}}+\mathsf{\lambda }{\displaystyle \sum }{\mathrm{Controls}}_{\mathrm{i},\mathrm{t}}+\mathrm{Year}+\mathrm{Industry}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$$

(1)

In model (1), ${\mathrm{EI}}_{\mathrm{i},\mathrm{t}}$ is the explained variable; $\mathrm{ER}{20}_{\mathrm{i},\mathrm{t}}$ is the explanatory variable; $\mathsf{\alpha }$ is the intercept term; ${\mathrm{Controls}}_{\mathrm{i},\mathrm{t}}$ is the control variable selected in this paper; Year is the year dummy variable to control the time-fixed effects; $\mathrm{Industry}$ is the industry dummy variable to control the industry-fixed effects; and ${\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$ is the random disturbance term. The regression coefficient ${\mathsf{\beta }}_0$ of the explanatory variable $\mathrm{ER}{20}_{\mathrm{i},\mathrm{t}}$ represents the effect of local government environmental concern on the corporate environmental investment, and its changes are the focus of the following tests. If ${\mathsf{\beta }}_0$ is significantly positive in the test results, then it indicates that local government environmental concern will promote corporate environmental protection investment in the jurisdiction. If ${\mathsf{\beta }}_0$ is significantly negative or insignificant, then it means that the local government environmental concern will inhibit the corporate environmental protection investment or has no effect on the corporate environmental protection investment.

We construct model (2) to test H2a and H2b: the moderating role of corporate resources.

$${\mathrm{EI}}_{\mathrm{i},\mathrm{t}}=\mathsf{\alpha }+{\mathsf{\beta }}_0\mathrm{ER}{20}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }}_1\mathrm{ER}{20}_{\mathrm{i},\mathrm{t}}\times {\mathrm{Moderator}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }}_2{\mathrm{Moderator}}_{\mathrm{i},\mathrm{t}}+\mathsf{\lambda }{\displaystyle \sum }{\mathrm{Controls}}_{\mathrm{i},\mathrm{t}}+\mathrm{Year}+\mathrm{Industry}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$$

(2)

Model (2) adds a moderator variable ${\mathrm{Moderator}}_{\mathrm{i},\mathrm{t}}$ and an interaction term $\mathrm{ER}{20}_{\mathrm{i},\mathrm{t}}\times {\mathrm{Moderator}}_{\mathrm{i},\mathrm{t}}$ to model (1). The regression coefficient ${\mathsf{\beta }}_1$ of the interaction term $\mathrm{ER}{20}_{\mathrm{i},\mathrm{t}}\times {\mathrm{Moderator}}_{\mathrm{i},\mathrm{I}}$ represents the moderating role of corporate resource endowment in the relationship between local government environmental concern and corporate environmental investment, and its changes are the focus of the moderating effects test. If ${\mathsf{\beta }}_1$ is significantly positive or negative in the test results, it indicates that corporate resources will promote or inhibit the impact of local environmental concern on corporate environmental investments.

4. Empirical Analysis

4.1. Descriptive Statistics

From the descriptive statistics in

Table 3. Descriptive statistics of main variables.

Variable Name	Mean	Standard Deviation	Minimum	Median	Maximum
EI	17.175	2.17	11.462	17.281	22.221
ER20	2.888	0.503	1.91	2.823	4.332
SIZE	23.246	1.536	20.384	23.195	27.703
BDS	9.186	1.925	5	9	15
DUA	0.197	0.398	0	0	1
CASH	−0.002	0.087	−0.27	0.004	0.21
LEV	0.472	0.189	0.077	0.482	0.864
OC	0.714	0.408	0.115	0.635	2.283
ROA	0.045	0.051	−0.118	0.038	0.204
IL	2.912	4.02	1.001	1.68	29.257

, it is clear that the mean and median of corporate environmental investment (EI) are 17.175 and 17.281, respectively. The median is close to the mean, indicating that the sample is roughly symmetrically distributed with low dispersion. Furthermore, the minimum value is 11.462 and the maximum value is 22.221. The difference between the maximum and minimum values is significant, and the standard deviation is 2.17, showing that there are obvious differences in environmental investment among different corporations. The mean and median of local government environmental concern (ER20) are 2.894 and 2.827, respectively, the mean is close to the median, and the standard deviation is small, illustrating a slight disparity in local governmental environmental concern across the regions. The sample tends to be normally distributed. The results of descriptive statistics for the remaining control variables are generally consistent with existing studies.

4.2. Benchmark Regression Results

Table 4. Benchmark regression results.

Variable	(1)	(2)	(3)
	EI	EI	EI
ER20	0.341 ***	0.262 ***	0.278 ***
	(2.99)	(2.70)	(2.88)
SIZE		0.812 ***	0.758 ***
		(24.72)	(18.94)
BDS		0.004	−0.000
		(0.19)	(−0.00)
DUA		−0.199 *	−0.196 *
		(−1.83)	(−1.81)
CASH		0.769	0.171
		(1.58)	(0.33)
LEV			0.536
			(1.60)
OC			0.464 ***
			(3.93)
ROA			2.965 **
			(2.36)
IL			0.039 ***
			(3.74)
Constant	16.264 ***	−2.468 ***	−2.089 **
	(49.02)	(−3.16)	(−2.47)
Industry FE	YES	YES	YES
Year FE	YES	YES	YES
Observations	1823	1791	1770
Adj R2	0.132	0.385	0.398

Note: The t-values are in parentheses. ***, **, * denote 1%, 5%, 10% significance levels, respectively.

shows the benchmark regression results. Column (1) is the regression result with no control variables added, column (2) is the regression result with the inclusion of control variables reflecting the basic conditions of the corporations, and column (3) is the regression result with the further inclusion of control variables reflecting the capabilities of the corporations. The results in columns (1) to (3) suggest that local government environmental concern (ER20) is positively correlated with corporate environmental investment (EI) (${\mathsf{\beta }}_0$ = 0.341, 0.262, and 0.278). It means that the stronger the environmental concern of the local government, the more environmental investment by the corporations in the jurisdiction, and H1 is verified. These results suggest that similar to public environmental concerns [24], local government environmental concern can also promote corporate environmental investment. Among the control variables, corporate size (SIZE) and the dual role of the chairman (DUA) are both negatively correlated with EI, indicating that the larger the corporate size and the more centralized the leadership power, the lower the corporate environmental investment. Operating capacity (OC) and profit capacity (ROA) are both positively correlated with EI, suggesting that the stronger the corporate operating capacity and profit capacity, the higher the corporate environmental investment.

4.3. Regression Results of Moderating Effect

As shown in

Table 5. Regression results of moderating effect.

Variable	(1)	(2)	(3)	(4)
	EI	EI	EI	EI
ER20	0.272 ***	0.393 ***	0.302 ***	0.309 ***
	(2.82)	(3.56)	(3.11)	(3.11)
GOV × ER20	−0.023 **
	(−2.25)
GOV	−0.001
	(−0.13)
PC × ER20		−0.377 **
		(−2.19)
PC		0.973 *
		(1.92)
SA × ER20			0.818 ***
			(3.61)
SA			−0.068
			(−0.41)
CD*ER20				15.333 **
				(1.98)
CD				4.356
				(0.92)
SIZE	0.758 ***	0.759 ***	0.771 ***	0.754 ***
	(18.88)	(18.88)	(17.95)	(18.77)
BDS	−0.001	−0.000	−0.003	0.001
	(−0.05)	(−0.00)	(−0.11)	(0.04)
DUA	−0.185 *	−0.163	−0.197 *	−0.171
	(−1.72)	(−1.49)	(−1.81)	(−1.56)
CASH	0.142	−0.078	0.161	0.060
	(0.28)	(−0.15)	(0.31)	(0.12)
LEV	0.550	0.597 *	0.526	0.522
	(1.64)	(1.77)	(1.59)	(1.54)
OC	0.456 ***	0.433 ***	0.478 ***	0.454 ***
	(3.86)	(3.63)	(4.02)	(3.80)
ROA	2.891 **	3.256 **	3.057 **	3.109 **
	(2.30)	(2.57)	(2.45)	(2.46)
IL	0.038 ***	0.037 ***	0.038 ***	0.039 ***
	(3.70)	(3.55)	(3.69)	(3.71)
Constant	−1.276	−2.416 ***	−1.556 *	−1.156
	(−1.51)	(−2.78)	(−1.71)	(−1.36)
Industry FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES
Observations	1770	1735	1770	1746
Adj R2	0.399	0.399	0.401	0.394

Note: The t-values are in parentheses. ***, **, * denote 1%, 5%, 10% significance levels, respectively.

, the results in column (1) show that the interaction term between the government subsidy (GOV) and the local government environmental concern (ER20) is negatively correlated with EI (β₁ = −0.023). The results in column (2) also indicate that the interaction term between the political connection (PC) and the local government environmental concern (ER20) is negatively correlated with EI (β₁ = −0.377). It means that government subsidies and political connection can weaken the promotion effect of local government environmental concern on corporate environmental investment. In other words, the fewer political resources a corporation has, the stronger the positive impact of local government environmental concern on corporate environmental investment. Hypothesis H2a is verified.

The results in column (3) show that the interaction term between the financing constraint (SA) and the local government environmental concern (ER20) is positively correlated with EI (β₁ = 0.818). The environmental investments by corporations with stronger financing constraints tend to be more influenced by local government environmental concern, which is consistent with the findings of previous studies [54]. The results in column (4) show that the interaction between the debt cost (CD) and the local government environmental concern (ER20) is positively correlated with the EI (β₁ = 15.333). This illustrates that the higher the debt cost of corporations, the more they are inclined to follow local government environmental concern to make environmental investments. In summary, the fewer financial resources the corporation has, the stronger the promotion effect of local government environmental concern on corporate environmental investment. Hypothesis H2b is verified.

The above empirical results and analysis demonstrate the moderating role of corporate resources in the relationship between local government environmental concern and corporate environmental investment. It also reveals an interesting phenomenon that the environmental investments of weaker corporations with fewer resources are more influenced by the local government, while resourceful corporations that have a stronger ability to make environmental investments are not sensitive to local government environmental concern. One possible explanation for this phenomenon is that the weaker corporations often have an urgent need for legitimacy and resources, and their environmental investment behavior has to follow the local government’s direction, which indirectly verifies that the influence of local government environmental concern on corporate environmental investment is transmitted through the legitimacy path and resource dependence path.

5. Robustness Tests

5.1. Replacing the Explained Variable

In the benchmark regression, we use the natural logarithm of total corporate environmental investment to measure the corporate environmental investment (EI) level. In the robustness test, to eliminate corporate scale effects, we use the ratio of corporate environmental investment to total corporate assets to measure the level of corporate environmental investment (EI). Column (1) of

Table 6. Robustness test results.

Variable	(1)	(2)	(3)
	EI	EI	EI
ER20	0.299 ***	0.190 **
	(3.55)	(2.20)
Time × Treat			0.741 **
			(2.50)
Time			−0.620
			(−1.27)
Treat			−1.158 ***
			(−4.38)
SIZE	−0.146 ***	0.498 ***	0.748 ***
	(−5.07)	(3.67)	(19.71)
BDS	0.008	0.018	−0.003
	(0.47)	(0.38)	(−0.12)
DUA	−0.200 ***	−0.028	−0.186 *
	(−2.63)	(−0.27)	(−1.82)
CASH	−0.290	−0.445	0.278
	(−0.66)	(−1.22)	(0.56)
LEV	0.374	−0.126	0.356
	(1.47)	(−0.22)	(1.14)
OC	0.214 *	0.062	0.490 ***
	(1.90)	(0.26)	(4.39)
ROA	2.927 ***	0.603	2.541 **
	(2.64)	(0.47)	(2.12)
IL	0.010	−0.015	0.038 ***
	(1.20)	(−1.47)	(3.94)
Constant	2.748 ***	4.678	0.136
	(4.55)	(1.55)	(0.12)
Industry FE	YES	NO	YES
Company FE	NO	YES	NO
Year FE	YES	YES	YES
Observations	1770	1771	1923
Adj R2	0.080	0.163	0.392

Note: The t-values are in parentheses. ***, **, * denote 1%, 5%, 10% significance levels, respectively.

shows the regression results after replacing the explanatory variables. Local government environmental concern (ER20) is still positively related to EI (β₀ = 0.299), again validating hypothesis H1, indicating that the regression analysis results are robust.

5.2. Replacing the Regression Model

We perform another robustness test by replacing the regression model. Specifically, the benchmark regression is validated again using two-way fixed effects models with individual and time effects. Column (2) of Table 6 shows the regression results after replacing the model. The local government environmental concern (ER20) is positively related to EI (β₀ = 0.190), verifying hypothesis H1 again and indicating that the regression results are robust.

5.3. Endogeneity Test

In order to alleviate the effect of endogeneity issues on the regression results, we again test the effect of local government environmental concern on corporate environmental investment by reconstructing the explanatory variables and using the difference-in-difference (DID) model. Taking the implementation of the new PRC Environmental Protection Law as the background, the policy dummy variables are constructed through the following steps. The PRC Environmental Protection Law was promulgated in April 2014 and has been in effect since 1 January 2015. Referring to the study of Liu et al. (2019) [10], we select 2015 as the policy time point, meaning that the time dummy variable Time is assigned to 1 for 2015 or after 2015 and 0 for before 2015. The degree of local government environmental concern is measured by whether local governments responded positively to the new Environmental Protection Law after its enforcement. Provinces that followed the new Environmental Protection Law by introducing local rules and regulations are the experimental group, and the policy dummy variable Treat is assigned to 1. Provinces and regions that did not respond are the control group, and Treat is assigned to 0.

The column (3) of Table 6 presents the regression results of DID, and it is clear that the coefficients of Time × Treat are significantly positive (β₀ = 0.741), indicating that local government environmental concern is positively correlated with corporate environmental investment, which indicates that the research results are robust.

6. Further Research

Based on the theoretical analysis in Section 2, the positive impact of local government environmental concern on corporate environmental investments is not certain. This promotion effect exists when local government environmental concern is effectively transformed into regional environmental pressure, and corporations in the jurisdiction respond positively. However, when local environmental concerns cannot be effectively translated into regional environmental pressure, due to local opportunistic behavior or when corporations in the jurisdiction respond negatively, this promotion effect is difficult to form. Based on the above analysis, this paper further empirically examines the relationship between local government environmental concern and corporate environmental investment in different external scenarios from the perspective of environmental regulation strength, economic development level, and environmental supervision strength.

6.1. Local Environmental Regulation Strength

Environmental regulation is the regulation of corporate pollution behavior, aiming to reduce the negative environmental externalities from corporate production and operation by converting the pollution costs into corporate internal costs. When local government environmental concern fails to bring high-intensity environmental regulation, corporate managers, as rational economic agents, are more willing to bear the low-cost environmental penalties rather than making slow-return, high-risk environmental investments [61]. Conversely, when the local government is concerned about environmental issues and implements high-intensity environmental regulations, the cost of environmental taxes and environmental penalties will increase significantly, making environmental investment a better choice [30]. With reference to the research of Ye et al. (2018) [62], we adopt the comprehensive index of pollution emission to measure the environmental regulation intensity of each city and divide the sample into a high-environmental-regulation-strength group and a low-environmental-regulation-strength group according to the annual median. The regression results are shown in

Table 7. Heterogeneity test results.

Variable	(1)	(2)	(3)	(4)	(5)	(6)
	EI High Environmental Regulation Strength	EI Low Environmental Regulation Strength	EI High Economic Development	EI Low Economic Development	EI High Environmental Supervision Strength	EI Low Environmental Supervision Strength
ER20	0.396 ***	−0.131	0.492 ***	−0.052	0.444 **	0.256 **
	(3.03)	(−0.85)	(4.06)	(−0.32)	(2.04)	(2.33)
SIZE	0.765 ***	0.733 ***	0.733 ***	0.775 ***	0.967 ***	0.708 ***
	(14.13)	(11.54)	(14.61)	(11.23)	(13.51)	(14.79)
BDS	−0.011	0.036	0.006	−0.000	−0.124 **	0.028
	(−0.33)	(1.06)	(0.20)	(−0.01)	(−2.26)	(1.04)
DUA	−0.361 **	0.030	−0.199	−0.205	−0.052	−0.257 *
	(−2.34)	(0.18)	(−1.44)	(−1.13)	(−0.29)	(−1.80)
CASH	−0.349	1.008	0.102	−0.211	2.073 **	−0.767
	(−0.50)	(1.25)	(0.17)	(−0.22)	(2.13)	(−1.20)
LEV	0.688	0.851	0.598	0.801	0.327	0.640
	(1.57)	(1.47)	(1.51)	(1.36)	(0.55)	(1.56)
OC	0.478 ***	0.267	0.497 ***	0.411 **	0.251	0.500 ***
	(2.71)	(1.62)	(3.47)	(2.03)	(1.07)	(3.51)
ROA	3.430 **	2.262	1.924	6.493 ***	6.564 **	2.267
	(2.00)	(1.09)	(1.25)	(2.88)	(2.50)	(1.54)
IL	0.034 **	0.030 *	0.034 **	0.041 **	0.072 ***	0.036 ***
	(2.41)	(1.95)	(2.47)	(2.52)	(2.92)	(3.26)
Constant	−2.295 **	−1.018	−2.102 *	−1.816	−6.091 ***	−1.221
	(−1.97)	(−0.78)	(−1.91)	(−1.32)	(−3.89)	(−1.21)
Industry FE	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES	YES
Observations	874	702	1111	657	369	1278
Adj R2	0.400	0.440	0.388	0.422	0.520	0.383

Note: The t-values are in parentheses. ***, **, * denote 1%, 5%, 10% significance levels, respectively.

. The results in columns (1) and (2) show that the coefficient of ER20 for the high-environmental-regulation-strength group is significantly positive (β₀ = 0.396, t = 3.03), which is consistent with the benchmark regression results. The coefficient of ER20 for the low-environmental-regulation-strength group is negative and insignificant (β₀ = −0.131, t = −0.85). This indicates that the local government environmental concern has a more significant role in promoting corporate environmental investment in regions with high environmental regulation strength.

6.2. Regional Economic Development Level

Regional economic development indicators are still an essential basis for assessing the performance of Chinese officials. According to public choice theory, in regions with sluggish economic growth, it is difficult for local governments to impose substantial environmental pressure on corporations in their jurisdictions, even if the local government has raised environmental concern. Some local governments may even engage in opportunistic behavior for economic development goals, such as introducing high-energy-consumption and high-pollution corporations that can boost the local economy [49]. Therefore, it is questionable whether the local government environmental concern can promote corporate environmental investment in regions where economic growth is urgently needed. Based on this, we adopt the GDP growth rate to measure the regional economic development and divide the sample into a high-economic-development group and a low-economic-development group according to the annual median. The regression results are shown in columns (3) and (4) of Table 7. The coefficient of ER20 for the high-economic-development group is significantly positive (β₀ = 0.492, t = 4.06), while the coefficient of ER20 for the low-economic-development group is positive but not significant (β₀ = −0.052, t = −0.32). This suggests that the faster the regional economic growth, the more significant the positive correlation between local government environmental concern and corporate environmental investment.

6.3. Environmental Supervision Strength

When a Chinese corporation is under key supervision for pollution issues, its environmental performance will become the focus of the local government’s attention, and its related information will be transmitted to the central environmental authorities for unified management. Clearly, such corporations hardly have the opportunity to avoid environmental pressure through negative methods such as production shifts. At the same time, since these corporations’ environmental pollution behaviors are more likely to be detected, the risk of facing penalties will be higher. In this situation, corporations have to pay more attention to changes in local government environmental concern and follow the local government more actively through environmental investments to ensure legitimacy [13] Accordingly, we divide the sample into a high-environmental-supervision-strength group and low-environmental-supervision-strength group based on whether corporations belong to the list of key emission units. As shown in columns (5) and (6) of Table 7, the coefficients of ER20 are significantly positive in both groups, which is consistent with the results of the benchmark regression. This indicates that under the influence of the local government environmental concern, corporations will follow the local government to make environmental investments regardless of whether they belong to the key supervision units. However, the regression coefficient of ER20 in column (5) (β₀ = 0.444) is much larger than that of ER20 in column (6) (β₀ = 0.256), suggesting that when corporations suffer from a high strength of environmental supervision, their environmental investment behavior is more influenced by local government environmental concern.

7. Conclusion and Implications

7.1. Research Conclusions

In this paper, we construct local government environmental concern indicators by analyzing the text of local government work reports using machine learning tools and examine the effect of local government environmental concern on corporate environmental investment using a sample of Chinese listed companies in Shanghai and Shenzhen A-shares that have disclosed their environmental investment from 2008 to 2020. Meanwhile, the moderating role of corporate resources is also explored. In further research, we empirically examine the relationship between local government environmental concern and corporate environmental investment in different external scenarios from the perspective of environmental regulation strength, economic development level, and environmental supervision strength. The main conclusions of the research are as follows. Compared to previous studies that only focused on the influence of a single government environmental regulation on corporate environmental investment [4,10,11,12,13,14,15,16,17], this research reveals the effect of government environmental governance motives on corporate environmental performance, which broadens the research boundary on the relationship between government and corporate environmental responsibility fulfillment and enriches the study of factors influencing corporate environmental investment behavior.

(1)

Local government environmental concern is positively related to corporate environmental investment, suggesting that local government environmental concern can facilitate corporate environmental investment in the jurisdiction.

(2)

The fewer political and financial resources corporations have, the stronger the promotion of local government environmental concern on corporate environmental investments. This suggests that the environmental investment behavior of corporations with few resources is more influenced by local government orientation due to their urgent need for legitimacy and resources, indirectly validating that the impact of local government environmental concern on corporate environmental investment is transmitted through the legitimacy path and resource dependence path.

(3)

The promotion effect is more significant in regions with a high strength of environmental regulation and high level of economic development, and is more effective among key regulated corporations. The finding proves that the ineffective transformation of local government environmental concern and passive corporate avoidance in some situations may lead to a decrease in the effectiveness of local government environmental concern in facilitating corporate environmental investment.

7.2. Policy Implications

The research conclusions provide valuable empirical evidence for central and local governments to implement green development and build a government–business collaborative environmental governance system. The main insights that can be obtained from this research are as follows.

(1)

The central government should continue to increase the proportion of green assessment in the officials’ assessment system and realize the one-vote rejection mechanism for environmental protection problems as early as possible, which is beneficial to enhance local governments’ concern for the environment and fully motivate them for environmental governance. It is necessary to strengthen environmental protection inspectors to monitor the implementation of environmental policies at the local level, to deter the opportunistic behavior of local governments. Special attention needs to be paid to the environmental governance of less developed regions. Policy and financial support should be given when necessary to prevent them from sacrificing the environment for economic development. Vertical management should be implemented for corporations with high levels of emissions, breaking up local protectionism and restricting corporations from moving polluting industries to regions with fewer environmental regulations.

(2)

Local governments should establish a view of governance with the concept of green development and convey the government environmental concern to corporations through multiple media, such as conferences, newspapers, and the Internet; lead the public, investors, consumers, and other entities to monitor the fulfillment of corporate environmental responsibility, enabling corporations to perceive the environmental pressure brought about by the local government environmental concern; and strengthen regional environmental regulations using methods including sewage charges, environmental taxes, and administrative environmental penalties, compelling corporations to consider environmental risks when making business decisions and increase their environmental investments following the local government’s will. Local governments should also use their resource allocation capabilities to strengthen corporations’ incentives for environmental investment and provide policy support in credit, taxation, and operating permits to corporations with outstanding environmental performance, adding additional benefits to corporate environmental investment.

(3)

Some resource-rich corporations may not be sensitive to preferential policies and environmental pressure, but their ability to make environmental investments should not be overlooked. The local government needs to develop a differentiated strategy and flexibly utilize investigations and visitations to emphasize the government environmental advocacy to these corporations, inducing them to take responsibility for environmental protection actively. For corporations with fewer resources, the local government should fully consider its ability to withstand environmental pressure and ensure that their production and operation activities are not significantly affected. At the same time, since they have limited resources to invest in environmental protection, the local government should provide them with more financial and policy support so that they have both the will and the ability to fulfill environmental responsibilities in response to the government’s call.

7.3. Limitations and Future Research Perspectives

Two major limitations of this study need to be addressed in future research. First, there are many ways for corporations to fulfill their environmental responsibilities, such as emission control, green innovation, green M&A, and environmental investment. In this paper, we only consider the impact of local government environmental concern on corporate environmental investment, while the impact of government environmental concern on other environmental performance indicators of corporations is not assessed. If possible, future research should pay attention to the impact of government environmental concern on corporate green innovation. This is because, in the long term, green innovation brings more benefits to corporations and the environment compared to environmental investments. Second, we analyze the text of the local government work report to construct a government environmental concern indicator using machine learning in the study, the essence of which is to measure the frequency of environmental words in the local government work report. However, this approach ignores the importance of sentence meaning. Future research can explore how to measure government environmental concern by calculating environmental sentence frequencies in government work reports to produce more accurate results.