Environmental Regulation, Political Connections, and Corporate Green Investment

Abstract

Based on the implementation of China’s new Environmental Protection Law (the new EPL), using the difference-in-differences (DID) method and the PSM method, this paper examines the impact of changes in local environmental governance motivation on corporate environmental protection investments before and after the implementation of the new EPL. The results show that, before introducing the new EPL, the scale of green investment of politically connected enterprises was significantly lower than that of other enterprises; after the introduction of the new EPL, the increase in environmental protection investment by politically connected enterprises was significantly higher than that of other enterprises. This promotion effect is more potent for formally politically connected enterprises. Given this, we suggest that governments need to achieve fair law enforcement of environmental protection and avoid the distortion of ecological protection investment by political connections during economic transitions.

1. Introduction

Economic growth in Asian countries has been remarkable thanks to a booming domestic economy and increasing foreign investment. Nevertheless, in the past, high-polluting industries have sought partnerships with Asian countries to circumvent environmental regulations. The main reason Asian countries have become “pollution havens” is that they have not had the same stringent environmental regulations as North America and Europe in recent decades. As a result, Pacific nations suffer from escalating environmental degradation in addition to the effects of global climate change. Policy makers realize that economic development and environmental sustainability should go hand in hand. In the face of increasing global environmental issues, countries worldwide are actively exploring environmental policies to deal with this challenge [1,2]. China is facing the dual challenges of economic development and environmental protection. Therefore, the environmental performance of enterprises is receiving growing attention while pursuing rapid economic development. Green investments (or environmental protection investments) of enterprises are the means to control environmental pollution and a meaningful way to promote economic growth and assist in the harmonious development of the economy and the environment [3]. In recent years, most corporate environmental protection investment studies have shown that environmental regulation can promote green investment to some extent [3,4], but the environmental governance effects are not ideal [5].

We place the changes in local governments’ environmental governance motivation, corporate green investment, and political connections in the same research scenario to distinguish the root causes of the environmental governance dilemma. Specifically, we systematically analyze the impact of changes in local governments’ environmental governance motivations on corporate environmental protection investment decisions, assuming that polluting enterprises’ political connections remain unchanged. First, we need to find an exogenous policy shock that directly impacts local governments’ environmental governance motives. Next, we compare the differences on the scale of environmental protection investments between politically connected and nonpolitically connected enterprises before and after the policy shock. By doing this, we can achieve a proxy analysis of whether improving local governments’ environmental governance motivation can effectively alleviate the inhibition effect of political connections on corporate investment. On 1 January 2015, the Environmental Protection Law of China (hereinafter the new EPL) was introduced, which is known as the “strictest environmental protection law ever” [6]. The old Environmental Protection Law (hereinafter the old EPL) only focused on corporate environmental violations. The new EPL clarifies the legal responsibilities of environmental law enforcers and local officials, which has a strong policy impact on the motivation of local governments’ environmental governance. Therefore, implementing the new EPL is an exogenous policy shock that is in line with the concept of this paper.

Based on the exogenous policy shock of the new EPL, we use the data of heavy-polluting listed companies in Shanghai and Shenzhen A-shares from 2012 to 2019. We apply the DID method to test the impact of the change in the motivation of local governments’ environmental governance on the green investment of different enterprises. The results show that the scale of investment of politically connected companies was significantly lower than that of nonpolitically connected ones before the new EPL’s implementation. After the new EPL’s implementation, investments by politically connected enterprises increased significantly more than those of nonpolitically related companies. The green investment of enterprises with formal political connections has a more substantial effect, while the investment levels of enterprises with informal political connections do not change significantly. China’s environmental governance measures can shed light on those of other developing countries to help achieve the global goal of green development.

The rest of this paper is organized as follows. Section 2 describes the institutional background of the new EPL and the construction of a theoretical model. Section 3 presents the research design and methodology. Section 4 shows the results and discussion. Section 5 outlines the conclusions and policy implications.

2. Literature Review

Porter & Linde (1995) [7] proposed the “Porter Hypothesis”, which argues that enterprises’ environmental protection investment can prompt companies to increase innovation in and the use of clean technologies, offsetting environmental protection costs and achieving a win–win situation for both environmental governance and economic profitability. Research on corporate environmental governance behaviors such as corporate environmental disclosure [8], environmental performance [9,10], and technological innovation [11,12,13] has received increasing attention from scholars.

The reason why enterprises invest in environmental protection is mainly due to the pressure of environmental regulation. We define environmental regulations by reference to regulatory economics: environmental regulations are the general rules and specific actions enforced by administrative agencies to control pollution and manage natural resources with the purpose of protecting the environment, and internalizing externalities, including direct and indirect interventions. Taylor et al. (2005) [14], by studying the relationship between the R&D investment of power plants and government environmental regulations, found that power plants will increase environmental protection investment when they are pressured by the government to do so. Leiter et al. (2009) [15] used European industry data to verify that environmental regulation can encourage enterprises to increase environmental investment. Why is environmental regulation limited in its ability to promote corporate green investment to reduce pollution emissions? One view is that corporate political connections provide shelter for heavy-polluting firms to circumvent government environmental regulations [16,17]. Local governments impose relatively light penalties for polluting behavior by politically connected firms. Hence, compared to the higher cost of environmental governance, the cost of environmental violation penalties borne by enterprises is meager, resulting in weak environmental protection investment intentions. Another view is that a lack of motivation for local government is the root cause of the poor effect of ecological governance. Due to the need for local economic development, local governments may be shielding or hiding enterprises’ environmental pollution. In addition, they may purposefully tamper with the environmental quality data to meet the environmental assessment standards of their superiors [18].

Based on existing studies, this study makes two contributions.

(1)

Few studies have explored the relationship between political connections and corporate green investment from the perspective of the exogenous impact of environmental regulation. We reorganize the logic between environmental regulations, political connections, and corporate green investment decisions. The DID and PSM methods significantly alleviate the endogenous problem of mutual causality between local governments’ motivations and enterprises’ political connections.

(2)

We show that the profound reason for the difference in the scale of enterprise environmental protection investment is a lack of local government governance motivation rather than corporate political connections alone. Thus, we deepen our understanding of the relationship between government and enterprises during economic transformations and enrich the related literature.

3. Institutional Background and Theoretical Model

3.1. The Institutional Background of the New EPL

China’s first Environmental Protection Law (the old EPL) came into effect in September 1979 and has not been amended since then. The enforcement standards of the old EPL are relatively loose. For example, most penalties are less than 50,000 RMB. China updated its Environmental Protection Law (the new EPL) in 2015. This was the nation’s first major reform of its environmental policies in more than two decades. The updated EPL includes stricter punishments for polluters and also holds legal enforcers accountable. Specifically, the new EPL’s enforcers and local officials who violate the criminal law in environmental law enforcement must also bear criminal responsibility. The new EPL induced a policy shock for local governments and provides an opportunity for a quasi-natural experiment in our research design and implementation.

3.2. Construction of a Theoretical Model

To endogenize the choices of local governments and enterprises before and after the implementation of the new EPL, this paper refers to the model setting of Becker (1968) and Zhang (2019) [19,20]. It involves the probability of government punishment for polluting firms (p), the amount of emission reduction of firms (q), and the potential benefits that political connections may bring to the enterprise (t) and develops the analysis in two steps:

In the first step, before implementing the new EPL, the environmental standards are looser, and local governments and enterprises are not motivated to govern. Local governments tend to pursue economic growth at the cost of polluting the environment [21]. Local governments selectively supervise polluting companies and shelter companies with political connections, exposing them to a more lenient regulatory environment. While nonpolitically connected companies are forced to invest in emissions reduction, companies with political connections choose not to invest.

In the second step, after implementing the new EPL, the environmental protection standards are stricter, and the governance motives of local governments and enterprises are activated. Under stricter environmental governance requirements, local governments force both types of enterprises to invest in environmental protection and achieve emission reduction targets. At this time, the environmental protection investment of politically connected enterprises increases significantly compared to before the law’s implementation. In addition, such enterprises can cooperate with local governments in pollution control to obtain future financing, scarce resources, and government support. They are more willing to actively reduce emissions due to the potential benefits [22].

3.2.1. Model Setting

Three types of subjects are introduced. They are governments ($G$), politically connected firms ($E_P$), and nonpolitically connected firms ($E_{np}$). The model is solved using the backward projection method. To start, determine the government penalty probability, conceptualize firms’ strategic choices on whether to reduce emissions, and solve for the optimal government penalty strategy based on the company response function.

First, set up the enterprise decision model:

For enterprise $E_i$, $i\in \left\{p,np\right\}$, it is assumed that whether the enterprise conducts environmental pollution control only affects the costs of the enterprise. To generalize the model derivation, consider that the enterprise utility, $U_i{q}_i,F_i$, is only related to the emission reduction decision, $q_i$, and whether or not the firm is punished, $F_i$. Pollution control increases corporate governance costs, thereby reducing the investment in profitable projects and reducing business profits. Hence, the utility of enterprises decreases monotonically with increasing emission reductions. That is, $\left(\partial U\_i\right)/\left(\partial q\_i \right)<0$. The administrative punishment on the enterprise will lead to a decline in the enterprise’s profits. Therefore, the utility of the enterprise decreases monotonically with an increase in the penalty. That is, $\left(\partial U\_i\right)/\left(\partial F\_i \right)<0$. At the same time, it also simplifies the enterprise’s emission reduction, q, which takes the value Q (reduction) or 0 (no reduction).

(1)

If the company chooses not to reduce emissions:

If the emission reduction $q_i=0$, the probability of being punished by the government is $p_i$, the fine is F, and the likelihood of not being penalized by the government is $1-p_i$. The utility of the firm when it is penalized is $U_i(0,F)$, and the utility of the firm when it is not penalized is $U_i(0,0)$. According to $\left(\partial U\_i\right)/\left(\partial F\_i \right)<0$, $U_i\left(0,F\right)<U_i(0,0)$.

(2)

If the company chooses to reduce emissions:

The emission reduction $q_i=Q$, and the enterprise utility is $U_i(Q,0)$. We assume that enterprises have potential benefits, $t_i$. After increasing environmental protection investment, the potential benefits may come from two aspects: one is avoiding government penalties after the company conducts environmental pollution control, and the other comes from the government giving financial subsidies, tax incentives, etc. (Fan et al., 2007) [23]. Here, $t_p>t_{np}$. This means that the two types of companies have the same emission reduction Q, and the potential benefits of politically connected companies are greater than those of nonpolitically connected companies. Thus:

$$U_i(0,0)-U_i(Q,0)-t_i\le p_i\left[U_i(0,0)-U_i(0,F)\right]$$

where the left side represents the net cost of conducting environmental pollution control and the right side represents the benefits of conducting environmental pollution control (opportunity cost of pollution). It can then be shown that:

$$U_i(Q,0)+t_i\ge p_i{U}_i(0,F)+\left(1-p_i\right)U_i(0,0),i\in \left\{p,np\right\}$$

(1)

When Equation (1) is satisfied, the enterprise conducts environmental pollution control.

When the benefits of emission reduction are greater than the benefit of no emission reduction, enterprises will invest in environmental protection. The critical value is:

$$\widehat{P_{NP}}=\frac{U_{NP}\left(0,0\right)-U_{NP}\left(Q,0\right)-t_{NP}}{U_{NP}\left(0,0\right)-U_{NP}\left(0,F\right)}$$

$$\widehat{P_P}=\frac{U_{NP}\left(0,0\right)-U_{NP}\left(Q,0\right)-t_P}{U_{NP}\left(0,0\right)-U_{NP}\left(0,F\right)}$$

For both types of enterprises, when $p_i\ge \widehat{p}$, the enterprise invests in environmental protection and reduces emissions Q; when $p_i<\widehat{p}$, it does not invest in environmental protection and reduces emissions by 0. Since $t_{NP}<t_P$, $\widehat{P_{NP}}>\widehat{P_P}$. That is, the critical penalty probability of politically connected enterprises is lower than that of nonpolitical-related enterprises.

The above analysis shows that corporate environmental governance behavior is related to the critical penalty probability, while the government determines the penalty probability.

Second, the government decision-making model setting is:

Let $2Q>Q_2>Q>Q_1>0$. The maximum value is 2Q, which is the emission reduction amount when both types of enterprises carry out environmental pollution control.

(1)

Before the implementation of the new EPL, the environmental standards were looser (the total emission reduction should meet Q₁).

At this time, the government’s objective function is to minimize the cost of local firms in the case of total emission reductions up to Q₁. That is, the unit abatement cost c is minimized, and the government increasingly shelters politically connected enterprises ($\beta >1$). The government decision is as follows:

$$mi{n}_{\left\{qp,qnp\right\}}c{q}_{np}+\beta c{q}_p$$

(2)

$$s.t.q_{np}+q_p\ge Q_1$$

(2)

After the implementation of the new EPL, the environmental standards are stricter (the total emission reduction $\ge Q_2$).

At this time, the government objective function remains unchanged, but the total emission reduction increases, $\ge Q_2$, so the government decision is as follows:

$$mi{n}_{\left\{qp,qnp\right\}}c{q}_{np}+\beta c{q}_p$$

(3)

$$s.t.q_{np}+q_p\ge Q_2$$

3.2.2. Model Solution

(1)

Before the implementation of the new EPL:

For the government, Equation (2) requires $q_{np}+q_p\ge Q_1$, $Q_1<Q$, and one type of enterprise can be selected to reduce emissions, namely $q_p, q_{np}=\left(0, Q\right)$ or $q_p,q_{np}=(Q,0)$. When $q_p,q_{np}=\left(0, Q\right)$, the objective function is cQ. When $q_p,q_{np}=\left(Q,0\right)$, the objective function is βcQ. Since >1, $\beta cQ>cQ$, the optimal solution of (2) is $\left(q_p^{*},q_{np}^{*}\right)=(0,Q)$.

To urge both types of enterprises to achieve the above emission reduction targets, the government can design a biased environmental violation penalty probability $p_p^{*},p_{np}^{*}=(0,1)$ or any arbitrary $p_p^{*}\langle \widehat{P_P},p_{np}^{*}\rangle \widehat{P_{NP}}$.

The implication is that the environmental standards were looser before the implementation of the new EPL. The government can now choose one type of enterprise to reduce emissions. Since the government shelters enterprises with political connections, the government shows a bias toward the probability of environmental violation penalties $(p_{np}^{*}>p_p^{*})$. Hence, politically connected companies do not invest in emissions reductions, while nonpolitically connected companies invest in emissions reductions.

(2)

After the implementation of the new EPL:

For the government, Equation (3) requires $q_{np}+q_p\ge Q_2$, $2Q>Q_2>Q$. At this time, both types of enterprises must reduce emissions to meet the environmental standards. Therefore, the optimal solution of (3) is $q_p^{*},q_{np}^{*}=(Q,Q)$.

To urge the two types of enterprises to achieve the above emission reduction targets, the government can change the probability of incurring environmental violation penalties $p_p^{*},p_{np}^{*}$. At this time, $p_p^{*}>\widehat{P_P}$, $p_{np}^{*}>\widehat{P_{NP}}$.

The implication is that the environmental standards are stricter after implementing the new EPL. At this time, both types of enterprises need to reduce their emissions. The critical value of the government penalty probability is $\left(p_{np}^{*},p_p^{*}\right)=\left(\widehat{p_{np}},\widehat{p_p}\right)$. Compared with before, politically connected enterprises are more likely to be punished after the law’s implementation. However, at this time, the environmental protection investment by politically connected enterprises will gain preferential resources. Stricter environmental regulations and potential benefits will prompt a significant increase in green investment by politically connected enterprises.

3.2.3. Model Summary

To sum up, compared with companies without political connections, companies with political connections are less likely to be punished. Before the implementation of the new EPL, due to sheltering by the local government, the environmental violation penalty probability faced by politically connected enterprises is $p_p^{*}\in \left\{0,\widehat{P_P}\right\}$, and the environmental violation penalty probability faced by nonpolitical connection enterprises is $p_{np}^{*}\in \left\{\widehat{P_{np}},1\right\}$. That is, politically connected firms are less susceptible to penalties. Therefore, the rational choice for politically affiliated companies is not to invest in or to maintain a low level of green investment. After the implementation of the new EPL, the probabilities of environmental violation penalties faced with politically connected companies and nonpolitical connection companies are $p_i^{*}\in \left\{\widehat{P_P},1\right\}$ and $i\in \left\{p,np\right\}$, respectively. The new EPL stimulates the environmental governance motives of local officials. Only relying on certain enterprises to reduce emissions is not enough to achieve the environmental goal. The local government will increase punishments, supplying environmental protection investment motives to both types of enterprise. However, the increase in green investment of politically connected enterprises is more significant than that of others.

4. Methodology and Research Design

4.1. Methodology

The difference-in-differences (DID) method is a quasi-experimental approach that compares the changes in outcomes over time between a population enrolled in a program (the treatment group, i.e., companies with political connections) and a population that is not (the comparison group, i.e., companies without political connections). Difference-in-differences takes the before–after difference in the treatment group’s outcomes. This is the first difference. In comparing the same group to itself, the first difference controls for factors that are constant over time in that group. Then, to capture time-varying factors, difference-in-differences takes the before–after difference in the control group, which was exposed to the same set of environmental conditions as the treatment group. This is the second difference. Finally, difference-in-differences “cleans” all time-varying factors from the first difference by subtracting the second difference from it. This leaves us with the impact estimation. The detailed research methodology is shown in Table 1.

Table 1. The detailed research methodology.

Steps	Description
1. Research foundation	A brief introduction to the research foundation is presented as theoretical predictions in the model summary.
2. Methods	Difference-in-difference approach.
3. Restrictions of the method	Difference-in-differences relies on the equal trends assumption, which can be tested via placebo tests and other methods.
4. Criteria for method selection	Difference in differences requires data measured from a treatment group and a control group at two or more different time periods, specifically at least one time period before “treatment” and at least one time period after “treatment”.
5. Criteria for research population selection	Chinese A-share listed companies with heavy pollution. The scope of heavy pollution industries is based on “Guidelines for Environmental Information Disclosure” issued by the Ministry of Environmental Protection of China.

The research idea of DID is shown in Figure 1.

Table 1. The detailed research methodology.

Steps	Description
1. Research foundation	A brief introduction to the research foundation is presented as theoretical predictions in the model summary.
2. Methods	Difference-in-difference approach.
3. Restrictions of the method	Difference-in-differences relies on the equal trends assumption, which can be tested via placebo tests and other methods.
4. Criteria for method selection	Difference in differences requires data measured from a treatment group and a control group at two or more different time periods, specifically at least one time period before “treatment” and at least one time period after “treatment”.
5. Criteria for research population selection	Chinese A-share listed companies with heavy pollution. The scope of heavy pollution industries is based on “Guidelines for Environmental Information Disclosure” issued by the Ministry of Environmental Protection of China.

The research idea of DID is shown in Figure 1.

4.2. Sample Selection and Data Sources

This paper takes the listed companies in China’s A-share heavy-polluting industries from 2012 to 2019 as its research sample. Heavily polluting industries include chemical raw materials, nonferrous metal smelting, coal mining, etc. The scope of heavy pollution industries is based on the “Guidelines for Environmental Information Disclosure” issued by the Ministry of Environmental Protection of China, and the corresponding CSRC (China Securities Regulatory Commission) industry codes are B06, B07, B08, B09, B11, C17, C18, C19, C22, C25, C26, C28, C29, C31, and C32. Next, we match the above industry codes with the CSMAR database and obtain 1145 companies. The sample selection comprises (1) excluding ST and ST* sample companies, (2) excluding companies with an asset–liability ratio >1, and (3) excluding companies with missing data from 2012 to 2019. In addition, we excluded companies with changes in their political connectedness during the study period. Companies with formal and informal political ties are classified as formally politically connected. Under the above criteria, a total of 382 companies are obtained. Among them, 226 have political connections, including 87 formal political connections, 139 informal political connections, and 156 with no political connections. In general, 3056 observations are obtained for 382 enterprises over 8 years.

The corporate environmental protection investment data used in this paper come from the annual reports of listed companies and are collected manually. Political connections and political connection-type data are collected manually from the CSMAR database, Baidu Baike, Sina Finance, and other channels. Control variable data are obtained from the CSMAR database.

4.3. Variables Explanation

4.3.1. Enterprise Green Investment Scale (ln vest)

Green investments are investment activities that focus on projects or areas that are committed towards the preservation of the environment. We collected data from the “construction in progress” section of the annual report of listed companies—that is, projects directly related to environmental protection. The data included the purchasing of pollution control equipment, desulfurization, denitration, waste gas treatment, and wastewater treatment. We aggregate the project data yearly. The environmental protection investment is standardized by the total assets at the end of the year and multiplied by 100 to enhance the readability of the regression coefficient.

4.3.2. Political Connection (pc)

Referring to previous studies (Li et al., 2008; Gao et al., 2019) [24,25], if the chairman or general manager of a company has served in a government agency or has been a deputy of the National People’s Congress, a member of the CPPCC, or a party representative, the pc = 1. Otherwise, it is 0. Further, if the company’s chairman or general manager has served or is currently serving in a government agency, we define it as a formal political connection (pcgov), and pcgov = 1. If the company’s chairman or general manager has been a member of the NPC or CPPCC, this is defined as an informal political connection (pcre), and pcre = 1.

4.3.3. Implementation of the New EPL (time)

Before the law, time = 0, and after the law, time = 1.

4.3.4. Control Variables

We control the business characteristics, financial condition, and corporate governance. In addition, this paper controls for the year and industry effects.

Table 2. Definitions and descriptions of the variables.

Type	Variable	Description
Dependent variable	$lnvst$	Enterprise environmental protection investment scale/asset scale
Independent variables	$pc$	The value is 1 if the firm has political connections; otherwise, it is 0
	$pcgov$	Formal political connection is 1; otherwise, it is 0
	$pcre$	Informal political connection is 1; otherwise, it is 0
	$time$	If after 2015, the value is 1; otherwise, it is 0
Control variable	$asst$	Total assets of the firm (ln log)
	$age$	Firm listing years
	$soe$	If state-owned enterprise, the value is 1; otherwise, it is 0
	$lev$	Total liability/total assets
	$roa$	Net profit after tax/total assets
	$opportunity$	Tobin Q
	$cashflow$	Net cash flow from operations/total assets
	$bos$	Board of directors
	$top1$	If the controlling shareholder holds >50% shares, it is 1; otherwise, it is 0

lists the variable definitions and descriptions.

4.4. Empirical Model

First, build the model (1) to test the impact of political connection on corporate environmental protection investment before the new EPL.

$$lnvst=\alpha +{\beta }_1*pc\left(pcgov/pcre\right)+\eta *X+Year+Industry+\epsilon$$

(4)

where the dependent variable lnvst is the amount of corporate green investment, pc is a dummy variable that denotes whether the firm has a political connection, X is a series of corporate control variables, and Year and Industry are year’s fixed effect and industry fixed effect (the same below). The primary focus of this paper is the regression coefficient ${\beta }_1$. Secondly, to identify the impact of political connection types on corporate environmental protection investment, we distinguish between a formal political connection (pcgov) and informal political connection (pcre).

This paper references the previous literature (Gilje et al., 2016) [26]. It sets up the DID model (2) to compare the differences in environmental protection investment changes between politically connected and nonpolitical connected enterprises before and after the implementation of the new EPL. The coefficient ${\beta }_3$ of the pc*time represents the difference between the changes in the green investment amount of the two types of enterprises before and after the new EPL’s implementation. At the same time, it distinguishes between formal political connections and informal political connections.

$$lnvst=\alpha +{\beta }_1*pc\left(pcgov/pcre\right)+{\beta }_2*time+{\beta }_3*pc\left(pcgov/pcre\right)*time +\eta * X++Year+Industry+\epsilon$$

(5)

5. Results and Discussion

5.1. Summary Statistics

Taking 2015 as the time node,

Table 3. Descriptive Statistics.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
	Before the Law					After the Law
	N	Mean	Sd	Min	Max	N	Mean	Sd	Min	Max
$pc$	1146	0.592	0.492	0	1	1910	0.592	0.492	0	1
$pcgov$	1146	0.228	0.420	0	1	1910	0.228	0.419	0	1
$pcre$	1146	0.364	0.481	0	1	1910	0.364	0.481	0	1
$lnvst$	1146	0.798	0.038	0.652	0.904	1910	0.806	0.038	0.674	0.909
$time$	1146	0	0	0	0	1910	1	0	1	1
$asst$	1146	22.400	1.427	18.39	28.51	1910	22.85	1.406	18.84	28.64
$age$	1146	10.290	6.004	1	22	1910	14.29	6.113	4	27
$soe$	1146	0.487	0.500	0	1	1910	0.487	0.500	0	1
$roa$	1146	0.031	0.085	−0.691	2.163	1910	0.034	0.064	−0.746	0.379
$lev$	1146	0.483	0.215	0.014	1.000	1910	0.465	0.199	0.015	1.000
$opp~y$	1146	1.829	1.161	0.699	12.18	1910	2.002	1.705	0.691	28.34
$ca~w$	1146	0.054	0.076	−1.080	0.377	1910	0.064	0.065	−0.255	0.488
$bos$	1146	9.159	1.891	4	18	1910	8.894	1.871	0	20
$top1$	1146	0.248	0.432	0	1	1910	0.185	0.388	0	1

performs descriptive statistics on the variables by stage. The results show that the average value of corporate political connection during the sample period is 0.592, indicating that 59.2% of the samples in heavy-polluting industries have political connections. The average value of the formal political bond is 0.228, meaning that 22.8% of the sample in heavy-polluting industries have formal political connections. Moreover, the average value of informal political connections is 0.364, indicating that 36.4% of the sample have informal political connections. The variable standard deviation of environmental protection investment is relatively large, indicating that the scale of green investment varies greatly between different enterprises. After implementing the new EPL, the environmental protection investment of the sample enterprises increased.

5.2. Regression Results

5.2.1. The Impact of Political Connection on Corporate Green Investment

Table 4. The impact of political connections on corporate environmental protection investments.

	(1)	(2)	(3)	(4)	(5)	(6)
Variables	Before PSM			After PSM
	lnvst	lnvst	lnvst	lnvst	lnvst	lnvst
$pc$	−0.014 ***			−0.015 ***
	(−6.15)			(−4.72)
$pcgov$		−0.023 ***			−0.019 ***
		(−6.92)			(−4.11)
$pcre$			0.003			−0.005
			(1.18)			(−1.42)
$asst$	0.002 *	0.003 **	0.001	−0.001	−0.001	−0.002
	(1.77)	(2.41)	(1.05)	(−0.56)	(−0.40)	(−0.74)
$age$	0.001 **	0.001 **	0.001 ***	0.001	0.001	0.001
	(2.24)	(2.53)	(2.77)	(1.64)	(1.52)	(1.41)
$soe$	−0.007 ***	−0.005 *	−0.008 ***	−0.007 *	−0.005	−0.007 **
	(−2.84)	(−1.88)	(−2.94)	(−1.93)	(−1.40)	(−1.98)
$roa$	−0.021	−0.022	−0.019	−0.004	−0.007	0.001
	(−1.36)	(−1.55)	(−1.11)	(−0.10)	(−0.19)	(0.02)
$lev$	−0.020 ***	−0.018 **	−0.018 **	−0.028 ***	−0.027 ***	−0.027 **
	(−2.68)	(−2.38)	(−2.24)	(−2.67)	(−2.60)	(−2.54)
$opportunity$	−0.002	−0.001	−0.003	−0.008 ***	−0.007 ***	−0.008 ***
	(−1.08)	(−0.67)	(−1.22)	(−3.25)	(−3.18)	(−3.16)
$cashflow$	0.033 **	0.031 **	0.041 ***	−0.009	−0.016	−0.016
	(2.25)	(2.19)	(2.70)	(−0.34)	(−0.60)	(−0.56)
$bos$	0.001	0.001 *	0.001	0.000	0.000	0.000
	(1.56)	(1.81)	(1.40)	(0.24)	(0.49)	(0.13)
$top1$	−0.002	−0.002	−0.002	−0.002	−0.001	−0.002
	(−0.77)	(−0.89)	(−0.77)	(−0.37)	(−0.18)	(−0.43)
Constant	0.757 ***	0.731 ***	0.769 ***	0.864 ***	0.851 ***	0.870 ***
	(26.06)	(24.54)	(26.72)	(18.03)	(17.46)	(18.45)
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes
N	1146	729	885	567	368	450
R-squared	0.113	0.145	0.083	0.165	0.162	0.163

* p < 0.1, ** p < 0.05, and *** p < 0.01 and the value of t in parentheses.

reports the regression results. Columns (1)–(3) are the regression results before using PSM. In columns (1) and (2), the coefficients of pc and pcgov are significantly negative at the level of 1%, indicating that the political connection has a negative impact on green investment. Compared with other enterprises, the adverse effects of formal political connections on green investment are more significant. In addition, the negative effects of informal political connections on green investment are insignificant.

For the control variables, the enterprise scale (asst) and enterprise age (age) encourage enterprises to invest in environmental protection. The nature of enterprise property right (soe) has a negative effect on the environmental protection investment of enterprises; that is, the environmental protection investment of state-owned enterprises is less than that of private enterprises. Enterprise financial leverage (lev) has a negative effect on enterprise environmental protection investment. The cash flow (cashflow) of enterprises has a positive effect on their environmental protection investment.

To avoid potential self-selection bias, this paper uses the PSM method to match political connection enterprises with other enterprises one-to-one, and the regression results are shown in columns (4)–(6). In columns (4) and (5), the coefficients of pc and pcgov are significantly negative at the level of 1%, indicating that the political connection has a negative impact on investment. Compared with other enterprises, the negative effects of formal political connections are more significant. In contrast, the informal political association does not cause significant differences. These results are consistent with the regression results before PSM sample matching.

5.2.2. The Impact of Environmental Regulation and Political Connection on Corporate Green Investment

Table 5. The impact of environmental regulations and political connections on corporate green investments.

	(1)	(2)	(3)	(4)	(5)	(6)
Variables	Before PSM			After PSM
	lnvst	lnvst	lnvst	lnvst	lnvst	lnvst
$pc\ast time$	0.033 ***			0.032 ***
	(11.71)			(8.49)
$pc$	−0.015 ***			−0.017 ***
	(−6.56)			(−5.59)
$pcgov\ast time$		0.040 ***			0.045 ***
		(12.60)			(8.00)
$pcgov$		−0.025***			−0.025 ***
		(−9.48)			(−5.37)
$pcre\ast time$			0.003			0.006
			(1.22)			(1.48)
$pcre$			0.004 *			−0.001
			(1.69)			(−0.26)
$time$	−0.024 ***	−0.014 ***	−0.006 **	−0.017 ***	−0.010 **	−0.002
	(−7.02)	(−4.52)	(−1.96)	(−3.72)	(−2.36)	(−0.46)
$asst$	0.004 ***	0.005 ***	0.004 ***	−0.001	−0.001	−0.002
	(4.52)	(5.04)	(4.79)	(−0.56)	(−0.44)	(−1.24)
$age$	0.001 ***	0.001 ***	0.001 ***	0.001 ***	0.001 ***	0.001 ***
	(3.77)	(3.28)	(3.52)	(2.89)	(2.58)	(3.22)
$soe$	−0.004 **	−0.004 **	−0.003 *	−0.005 **	−0.005 **	−0.003
	(−2.45)	(−2.34)	(−1.90)	(−2.02)	(−2.29)	(−1.42)
$roa$	−0.027 **	−0.024 **	−0.028 **	−0.026 *	−0.023 *	−0.024
	(−2.30)	(−2.15)	(−2.30)	(−1.82)	(−1.69)	(−1.60)
$lev$	−0.026 ***	−0.024 ***	−0.024 ***	−0.018 ***	−0.018 ***	−0.017 ***
	(−5.83)	(−5.42)	(−5.27)	(−2.90)	(−2.82)	(−2.74)
$opportunity$	0.000	0.000	0.000	−0.001	−0.001	−0.001
	(0.09)	(0.40)	(0.19)	(−0.64)	(−0.47)	(−0.87)
$cashflow$	0.017 *	0.016	0.020 **	0.017	0.015	0.027 *
	(1.67)	(1.58)	(1.98)	(1.26)	(1.10)	(1.90)
$bos$	0.001 **	0.001 **	0.001 **	0.000	0.000	0.000
	(2.27)	(2.42)	(2.38)	(0.56)	(0.79)	(0.54)
$top1$	−0.002	−0.003	−0.003	−0.004	−0.005 **	−0.005 *
	(−1.30)	(−1.40)	(−1.41)	(−1.57)	(−1.99)	(−1.86)
Constant	0.721 ***	0.705 ***	0.704 ***	0.828 ***	0.823 ***	0.845 ***
	(37.26)	(35.98)	(36.06)	(25.57)	(25.14)	(25.83)
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes
N	3056	1944	2360	1572	1080	1252
R-squared	0.128	0.129	0.088	0.114	0.120	0.072

* p < 0.1, ** p < 0.05, and *** p < 0.01 and the value of t in parentheses.

shows the impact of implementing the new EPL on the relationship between political connections and corporate green investments. Columns (1)–(3) are the regression results before PSM matching. In columns (1) and (2), the coefficients of pc*time and pcgov*time are significantly positive at the level of 1%, which shows that the investment of politically connected enterprises increased more significantly compared with that of other enterprises after the implementation of the law. The green investment of formally politically connected enterprises has increased more significantly, while that of informally politically connected enterprises has not significantly changed. Columns (4)–(6) are the regression results after PSM. In columns (4) and (5), the coefficients of pc*time and pcgov*time are significantly positive at the level of 1%, which shows that the investment of politically connected enterprises increased more significantly compared with other enterprises after the implementation of the law. The environmental protection investment of formally politically connected enterprises has increased substantially, while that of informally politically connected enterprises has not changed significantly. The results are consistent with the regression results before matching.

5.3. Robustness Test

5.3.1. Parallel Trend Test

Bertrand (2004) [27] argued that the DID method is based on the premise that the treatment and control groups’ trends must be the same before the policy is implemented. Figure 2 shows the preliminary results. Before 2014, the investment amount of politically connected and nonpolitically connected enterprises showed a parallel downward trend, and the investment of politically connected enterprises was lower.

Since the end of 2014, the scale of green investment in these two types of enterprises has shown a clear trend of change. Under the impact of the new EPL, enterprises with political connections have significantly increased the scale of their investment, while the investment of nonpolitically connected enterprises shows a downward trend. Similarly, until 2014, the scale of environmental investments of firms with formal and informal political connections remained roughly the same. However, since the end of 2014, the investments in formally politically connected enterprises have increased significantly and exceed that of informally politically connected enterprises. Therefore, the DID model used in this paper conforms to the premise of the parallel trend hypothesis.

5.3.2. Placebo Test

To rule out the possibility of unobserved major national policies, this paper refers to Topalova (2010) [28] and conducts a placebo test using fictitious years of implementation of the new EPL. Specifically, we consider the situation where the new EPL implementation year was one year earlier or one year later. Suppose the coefficient of pc (pcgov/pcre)*time is insignificant. In that case, it means that the increase in green investments of politically connected enterprises results from the implementation the new EPL.

In columns (1)–(3) of

Table 6. Placebo test.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Advance a Year			Postpone a Year
	lnvst	lnvst	lnvst	lnvst	lnvst	lnvst
	−0.001 (−0.19)			0.006
				(0.93)
$pcgov\ast time$		−0.003			0.026
		(−0.77)			(0.04)
$pcre\ast time$			0.001			0.010
			(0.51)			(1.55)
$pc$	0.006 ***			0.005 ***
	(2.60)			(3.55)
$pcgov$		0.001			−0.009
		(0.55)			(−0.05)
$pcre$			0.005 **			0.006 ***
			(2.39)			(4.12)
$time$	−0.005	−0.005	−0.006 *	−0.007	−0.004	−0.006
	(−1.45)	(−1.53)	(−1.89)	(−1.08)	(−0.85)	(−1.21)
$asst$	0.004 ***	0.004 ***	0.004 ***	0.004 ***	0.004 ***	0.004 ***
	(4.32)	(4.67)	(4.80)	(4.34)	(4.67)	(4.79)
$age$	0.001 ***	0.001 ***	0.001 ***	0.001 ***	0.001 ***	0.001 ***
	(3.58)	(3.24)	(3.51)	(3.58)	(3.25)	(3.51)
$soe$	−0.004 **	−0.004 **	−0.003*	−0.004 **	−0.004 **	−0.003 *
	(−2.40)	(−2.27)	(−1.90)	(−2.38)	(−2.25)	(−1.91)
$roa$	−0.027 **	−0.027 **	−0.028 **	−0.027 **	−0.027 **	−0.027 **
	(−2.13)	(−2.17)	(−2.30)	(−2.15)	(−2.17)	(−2.25)
$lev$	−0.024 ***	−0.024 ***	−0.024 ***	−0.024 ***	−0.025 ***	−0.024 ***
	(−5.18)	(−5.30)	(−5.23)	(−5.21)	(−5.33)	(−5.26)
$opportunity$	0.000	0.000	0.000	0.069	0.019	0.022
	(0.06)	(0.17)	(0.20)	(0.06)	(0.17)	(0.21)
$cashflow$	0.022 **	0.021 **	0.021 **	0.022 **	0.020*	0.020 **
	(2.08)	(1.96)	(2.01)	(2.07)	(1.94)	(1.99)
$bos$	0.001 **	0.001 **	0.001 **	0.001 **	0.001 **	0.001 **
	(2.21)	(2.36)	(2.38)	(2.22)	(2.37)	(2.41)
$top1$	−0.003	−0.003	−0.003	−0.003	−0.003	−0.003
	(−1.39)	(−1.41)	(−1.42)	(−1.39)	(−1.43)	(−1.39)
Constant	0.710 ***	0.705 ***	0.703 ***	0.710 ***	0.706 ***	0.704 ***
	(36.12)	(35.47)	(36.24)	(36.28)	(35.43)	(36.32)
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes
N	3056	1944	2360	3056	1944	2360
R-squared	0.087	0.083	0.087	0.087	0.083	0.088

* p < 0.1, ** p < 0.05, and *** p < 0.01 and the value of t in parentheses.

, we can see the results if the implementation of the new EPL was advanced by one year to examine the impact of political connections, and columns (4)–(6) show the effect of postponing the implementation of the law by one year. We notice that whether the law is advanced or delayed, the coefficients of pc*time, pcgov*time, and pcre*time are insignificant. Hence, the results of this study are robust.

5.3.3. Redefinition of the Dependent Variable

The lnvst mentioned above is replaced by the amount of environmental protection investment/operating income (EIE) and the amount of environmental protection investment/net assets at the end of the year (EIN). Similarly, the variables are normalized and multiplied by 100 to regress to the DID model (2).

The results are shown in

Table 7. The estimation results after redefining the dependent variable.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	EIE	EIE	EIE	EIN	EIN	EIN
$pc\ast time$	0.055 ***			0.056 ***
	(11.26)			(11.69)
$pcgov\ast time$		0.071 ***			0.072 ***
		(12.38)			(13.28)
$pcre\ast time$			0.004			0.004
			(0.77)			(0.90)
$pc$	−0.028 ***			−0.024 ***
	(−7.14)			(−6.25)
$pcgov$		−0.042 ***			−0.043 ***
		(−9.15)			(−9.98)
$pcre$			0.002			0.007 **
			(0.61)			(1.98)
$time$	−0.040 ***	−0.023 ***	−0.008	−0.043 ***	−0.026 ***	−0.012 **
	(−6.40)	(−4.12)	(−1.40)	(−7.12)	(−4.76)	(−2.02)
$asst$	0.005 ***	0.005 ***	0.005 ***	0.002	0.003 **	0.003 *
	(2.86)	(3.19)	(2.99)	(1.60)	(2.14)	(1.91)
$age$	0.001 ***	0.001 ***	0.001 ***	0.002 ***	0.001 ***	0.001 ***
	(3.62)	(3.30)	(3.36)	(4.12)	(3.59)	(3.80)
$soe$	−0.005	−0.005 *	−0.004	−0.006 **	−0.006 **	−0.005 *
	(−1.64)	(−1.71)	(−1.35)	(−2.33)	(−2.31)	(−1.75)
$roa$	−0.061 ***	−0.055 ***	−0.061 ***	−0.047 **	−0.041 **	−0.048 **
	(−3.55)	(−3.32)	(−3.44)	(−2.43)	(−2.24)	(−2.42)
$lev$	−0.050 ***	−0.047 ***	−0.047 ***	0.014 *	0.017 **	0.017 **
	(−5.93)	(−5.50)	(−5.39)	(1.79)	(2.16)	(2.12)
$opportunity$	0.003	0.004 *	0.003	0.002	0.003	0.002
	(1.45)	(1.68)	(1.56)	(1.11)	(1.45)	(1.30)
$cashflow$	−0.021	−0.022	−0.015	0.020	0.018	0.026
	(−1.31)	(−1.38)	(−0.92)	(1.21)	(1.08)	(1.51)
$bos$	0.002 **	0.002 **	0.002 **	0.002 **	0.002 **	0.002 **
	(2.20)	(2.26)	(2.25)	(2.33)	(2.47)	(2.48)
$top1$	−0.007 **	−0.007 **	−0.007 **	−0.004	−0.004	−0.005
	(−2.15)	(−2.24)	(−2.22)	(−1.26)	(−1.36)	(−1.38)
Constant	0.478 ***	0.459 ***	0.455 ***	0.487 ***	0.462 ***	0.457 ***
	(13.83)	(13.22)	(13.03)	(14.54)	(13.68)	(13.40)
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes
N	3056	1944	2360	3056	1944	2360
R-squared	0.110	0.117	0.071	0.123	0.125	0.081

* p < 0.1, ** p < 0.05, and *** p < 0.01 and the value of t in parentheses.

, with columns (1)–(3) representing the results of the EIE and columns (4)–(6) representing the results of the EIN. From columns (1) to (6), the coefficients of pc*time and pcgov*time are significantly positive at 1%. The coefficient of pcre*time is not significant. Since implementing the new EPL, the green investment of politically connected enterprises has increased significantly, and the green investment of formally politically connected enterprises has increased considerably. In contrast, the environmental protection investment of informally politically connected enterprises has not increased significantly. In this way, the reliability of the research conclusions of this paper is verified.

5.3.4. Exclusion of the Influence of Regional Factors

The region fixed effect is added to the regression analysis to exclude the influence of regional factors, according to the province where the enterprise is registered. In columns (1) and (2) of

Table 8. Exclusion of the influence of regional factors.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Before the Law			After the Law
	lnvst	lnvst	lnvst	lnvst	lnvst	lnvst
$pc$	−0.015 ***			−0.015 ***
	(−5.98)			(−6.53)
$pcgov$		−0.023 ***			−0.023 ***
		(−7.69)			(−8.70)
$pcre$			0.001			0.002
			(0.49)			(1.06)
$pc\ast time$				0.033 ***
				(12.24)
$pcgov\ast time$					0.041 ***
					(13.21)
$pcre\ast time$						0.003
						(1.20)
$time$				−0.024 ***	−0.014 ***	−0.006 *
				(−7.12)	(−4.52)	(−1.81)
$asst$	0.004 ***	0.004 ***	0.003 **	0.005 ***	0.005 ***	0.005 ***
	(2.65)	(3.01)	(2.05)	(5.44)	(5.90)	(5.59)
$age$	0.001 **	0.001 **	0.001 ***	0.001 ***	0.001 ***	0.001 ***
	(2.14)	(2.49)	(2.64)	(3.60)	(3.09)	(3.27)
$soe$	−0.005 *	−0.002	−0.005 *	−0.002	−0.002	−0.001
	(−1.85)	(−0.90)	(−1.89)	(−1.16)	(−1.30)	(−0.77)
$roa$	−0.024	−0.025 *	−0.023	−0.026 **	−0.022 *	−0.026 **
	(−1.58)	(−1.71)	(−1.36)	(−2.13)	(−1.92)	(−2.09)
$lev$	−0.019 **	−0.016 **	−0.017 **	−0.022 ***	−0.021 ***	−0.020 ***
	(−2.55)	(−2.21)	(−2.19)	(−5.00)	(−4.59)	(−4.41)
$opportunity$	−0.001	−0.000	−0.001	0.000	0.001	0.001
	(−0.46)	(−0.16)	(−0.57)	(0.44)	(0.72)	(0.52)
$cashflow$	0.020	0.018	0.026 *	0.002	0.001	0.006
	(1.36)	(1.29)	(1.73)	(0.19)	(0.12)	(0.53)
$bos$	0.001 *	0.001 *	0.001	0.001 ***	0.001 ***	0.001 ***
	(1.68)	(1.79)	(1.47)	(2.86)	(2.97)	(2.94)
$top1$	−0.002	−0.001	−0.002	−0.000	−0.000	−0.000
	(−0.67)	(−0.55)	(−0.73)	(−0.13)	(−0.21)	(−0.21)
Constant	0.738 ***	0.718 ***	0.746 ***	0.711 ***	0.699 ***	0.698 ***
	(23.01)	(22.30)	(23.20)	(35.25)	(34.49)	(34.32)
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes
Province	Yes	Yes	Yes	Yes	Yes	Yes
N	1146	729	885	3056	1944	2360
R-squared	0.204	0.224	0.175	0.217	0.219	0.174

* p < 0.1, ** p < 0.05, and *** p < 0.01 and the value of t in parentheses.

, the coefficients of pc and pcgov are significantly negative at 1%, indicating that the political connection has a negative effect on environmental protection investment. Moreover, the negative effect of formal political connections on green investment is more significant, while the informal political connection is not significant. In columns (4) and (5), the coefficients of pc*time and pc*gov are significantly positive at 1%, which shows that the increase in the amount of environmental protection investment of politically connected enterprises has been more significant since the implementation of the new EPL. Nonetheless, the increase in green investment of formally politically related enterprises has been more prominent. In contrast, the representative type is not significant. Therefore, the results are consistent with those of previous sections.

5.3.5. Excluding the Impact of Overcapacity Industries

To exclude the interference of the de-capacity policy on the research findings, the sample of enterprises identified as overcapacity industries from 2012 to 2019 was excluded according to the policy issued by the Chinses State Council in 2013. The results are shown in

Table 9. Excluding the impact of overcapacity industries.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Before the Law			After the Law
	lnvst	lnvst	lnvst	lnvst	lnvst	lnvst
$pc$	−0.014 ***			−0.014 ***
	(−5.62)			(−6.00)
$pcgov$		−0.025 ***			−0.026 ***
		(−8.62)			(−9.36)
$pcre$			0.004			0.005 **
			(1.58)			(2.03)
$pc\ast time$				0.033 ***
				(11.49)
$pcgov\ast time$					0.041 ***
					(12.43)
$pcre\ast time$						0.004
						(1.28)
$time$				−0.025 ***	−0.014 ***	−0.006 *
				(−6.88)	(−4.38)	(−1.91)
$asst$	0.002 *	0.003 **	0.001	0.004 ***	0.005 ***	0.004 ***
	(1.78)	(2.47)	(1.10)	(4.42)	(5.05)	(4.79)
$age$	0.001 **	0.001 **	0.001 ***	0.001 ***	0.001 ***	0.001 ***
	(2.09)	(2.34)	(2.60)	(3.64)	(3.08)	(3.35)
$soe$	−0.008 ***	−0.005 **	−0.008 ***	−0.004 **	−0.004 **	−0.003*
	(−3.03)	(−2.00)	(−3.04)	(−2.46)	(−2.35)	(−1.82)
$roa$	−0.020	−0.022	−0.018	−0.027 **	−0.024 **	−0.028 **
	(−1.30)	(−1.53)	(−1.08)	(−2.28)	(−2.11)	(−2.31)
$lev$	−0.018 **	−0.016 **	−0.015 *	−0.026 ***	−0.024 ***	−0.024 ***
	(−2.38)	(−2.12)	(−1.89)	(−5.55)	(−5.24)	(−5.01)
$opportunity$	−0.002	−0.001	−0.003	0.000	0.000	0.000
	(−1.06)	(−0.63)	(−1.18)	(0.06)	(0.38)	(0.21)
$cashflow$	0.034 **	0.031 **	0.042 ***	0.017	0.015	0.019 *
	(2.24)	(2.14)	(2.68)	(1.58)	(1.45)	(1.83)
$bos$	0.001 *	0.001 *	0.001	0.001 ***	0.001 ***	0.001 ***
	(1.71)	(1.93)	(1.61)	(2.83)	(2.93)	(2.92)
$top1$	−0.002	−0.002	−0.001	−0.002	−0.003	−0.003
	(−0.57)	(−0.73)	(−0.55)	(−1.32)	(−1.46)	(−1.41)
Constant	0.759 ***	0.726 ***	0.771 ***	0.723 ***	0.705 ***	0.702 ***
	(25.96)	(24.12)	(26.56)	(37.05)	(35.42)	(35.74)
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes
N	1101	714	855	2936	1904	2280
R-squared	0.110	0.146	0.084	0.132	0.131	0.091

* p < 0.1, ** p < 0.05, and *** p < 0.01 and the value of t in parentheses.

. In columns (1) and (2), the coefficients of pc and pcgov are significantly negative at 1%, indicating that the political connection of enterprises has a negative effect on environmental protection investment. The negative effect of formal political connections on enterprise environmental protection investment is more significant. In contrast, the impact of informal political association is not significant. In columns (4) and (5), the coefficients of pc*time and pcgov*time are significantly positive at 1%, which shows that the green investment of politically connected enterprises has increased more significantly since the new EPL was implemented. The green investment of formally politically connected enterprises has increased more significantly. In contrast, representative-type political connections have not been significant, which verifies the reliability of the previous conclusions.

5.3.6. Excluding the Bidirectional Influence

To exclude the bidirectional influence between the dependent and control variables, this paper lags behind by one period of the continuous control variable. As is shown in

Table 10. Excluding the bidirectional influence.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Before the Law			After the Law
	lnvst	lnvst	lnvst	lnvst	lnvst	lnvst
$pc$	−0.015 ***			−0.016 ***
	(−5.26)			(−5.56)
$pcgov$		−0.025 ***			−0.026 ***
		(−7.19)			(−7.85)
$pcre$			0.002			0.003
			(0.79)			(1.23)
$pc\ast time$				0.033 ***
				(10.29)
$pcgov\ast time$					0.040 ***
					(10.96)
$pcre\ast time$						0.004
						(1.24)
$time$				−0.017 ***	−0.006 **	0.001
				(−4.73)	(−1.96)	(0.36)
$L.asst$	0.003 *	0.004 **	0.002	0.004 ***	0.005 ***	0.005 ***
	(1.86)	(2.31)	(1.30)	(4.44)	(4.82)	(4.94)
$L.age$	0.001 *	0.001 **	0.001 **	0.001 ***	0.001 ***	0.001 ***
	(1.82)	(2.05)	(2.22)	(3.37)	(2.83)	(2.99)
$soe$	−0.007 **	−0.004	−0.007 **	−0.003 **	−0.003 **	−0.002
	(−2.17)	(−1.38)	(−2.29)	(−1.97)	(−2.06)	(−1.38)
$L.roa$	−0.029 *	−0.030 *	−0.026	−0.028 ***	−0.023 **	−0.029 ***
	(−1.78)	(−1.94)	(−1.44)	(−2.67)	(−2.26)	(−2.65)
$L.lev$	−0.020 **	−0.018 **	−0.017 *	−0.025 ***	−0.024 ***	−0.023 ***
	(−2.27)	(−1.97)	(−1.84)	(−5.26)	(−4.97)	(−4.84)
$L.opportunity$	−0.002	−0.001	−0.002	0.000	0.000	0.000
	(−0.62)	(−0.24)	(−0.75)	(0.11)	(0.35)	(0.31)
$L.cashflow$	0.022	0.019	0.032 *	0.022 **	0.019 *	0.024 **
	(1.26)	(1.14)	(1.71)	(2.21)	(1.95)	(2.34)
$L.bos$	0.001	0.001	0.001	0.001 **	0.001 **	0.001 **
	(1.05)	(1.31)	(0.96)	(2.05)	(2.26)	(2.29)
$top1$	−0.003	−0.003	−0.003	−0.003	−0.003	−0.003 *
	(−1.00)	(−1.06)	(−1.02)	(−1.56)	(−1.59)	(−1.66)
Constant	0.734 ***	0.710 ***	0.746 ***	0.709 ***	0.695 ***	0.686 ***
	(20.53)	(19.39)	(21.03)	(34.10)	(33.00)	(32.61)
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes
N	764	486	590	2674	1701	2065
R-squared	0.107	0.138	0.073	0.137	0.131	0.096

* p < 0.1, ** p < 0.05, and *** p < 0.01 and the value of t in parentheses.

in columns (1) and (2), the coefficients of pc and pcgov are significantly negative at 1%, indicating that the political connection has a negative effect on green investment. In detail, the negative impact of the formal political connection is more significant, while the negative effect of the representative-type political connection is not significant. In columns (4) and (5), the coefficients of pc*time and pcgov*time are significantly positive at 1%, which shows that the investment of politically connected enterprises has increased more significantly since the implementation of the law. Meanwhile, the green investment of formally politically connected enterprises has increased significantly, while that of informally politically connected enterprises has not. The reliability of the conclusions of this paper is verified.

5.4. Discussion

The preliminary results show that the political connections of firms have a significant inhibitory effect on their environmental protection investments, which is consistent with the conclusions of the existing literature (Zhang et al., 2019) [20]. Further, we examine the impact of implementing the new EPL on the relationships between political connections and corporate investments. After the introduction of the new EPL, the increase in the environmental protection investment of politically connected enterprises is significantly higher than that of other enterprises. Among them, the green investments of formally politically connected enterprises are stronger. The results are consistent with the results of this paper’s model. Implementing the new EPL has strengthened the deterrent effect on local governments, forcing local governments to enhance the supervision of politically connected enterprises. This has increased the scale of corporate environmental protection investments. Moreover, when distinguishing between the types of political connections, it is also significant that the increase in green investments by formally politically connected enterprises is more pronounced. The conclusions drawn in this paper are consistent with the argument that environmental governance needs to rely on intensive environmental legislation (Zhou et al., 2017) [29].

We conducted strict robustness tests. First, Figure 2 shows that the DID method used in this article is advisable, and the trends of the treatment group and the control group before and after the law’s implementation show that the law has significantly increased the environmental investment of politically connected enterprises. In addition, the robustness of the results in this paper was verified by a placebo test, the substitution of dependent variables, and the exclusion of the influence of related factors.

6. Conclusions and Policy Implications

6.1. Conclusions

The main conclusions are as follows:

(1)

The empirical results support the “Porter Hypothesis”. Specifically, under a strict environmental regulation shock, the environmental governance motives of local governments are enhanced, which significantly promotes the scale of environmental protection investments. Our findings are consistent with Berman & Bui (2001) [30] and Leiter et al. (2009)’s [15] conclusions using US and European industry data.

(2)

The increase in green investments of politically connected companies has been significantly higher than that of enterprises without political connections. Among them, the investments of enterprises with formal political connections have increased significantly.

(3)

The second-order conclusion is that the crucial reason for environmental governance ineffectiveness in many developing countries is the local government’s lack of environmental governance motivation. The central government needs to make scientific and reasonable environmental policies to stimulate the environmental protection motivation of local governments.

6.2. Policy Implications

The conclusions of this paper have certain value for developing countries when dealing with the environmental protection dilemmas they may face.

(1)

Environmental regulations can institutionally establish a supervision and assessment mechanism for the effectiveness of local government environmental pollution control, clarify the environmental governance responsibilities of local governments, establish a lifelong responsibility system, and further improve the incentive structure of local governments that combine “positive incentives” and “negative incentives”.

(2)

During the economic transition, the government needs to achieve fair law enforcement with regards to environmental protection, block the interest conveyor belt between the government and enterprises, and avoid the distortion of ecological protection investments caused by political connections.

6.3. Future Directions

The problems and limitations of the study are as follows:

(1)

The availability of corporate environmental protection investment data is restricted. At present, environmental protection issues are getting more public attention, and listed companies are paying more attention to the disclosure of social responsibility. However, the green investment data of enterprises is not the mandatory disclosure content of corporate social responsibility (CSR) reports. Therefore, the data are not easy to obtain, and the unavailability of data shrinks the research samples in this paper.

(2)

The measurement range of political connection data is relatively narrow. In terms of political connections, we only considered the prominent characteristics when analyzing the relationship between enterprises and the government.

Future research needs to improve the availability of data and conduct in-depth research on environmental investments to obtain more complete conclusions. In addition, considering the social networks among the relatives and friends of government officials, future research should consider expanding the research on political connections from the perspective of the social network of corporate executives. Additionally, it could examine the influence of political connections in environmental investments based on alternative environmental regulations.

Moreover, inadequate environmental governance and governance corruption are common problems in the economic transformation of developing countries, and the new EPL’s long-term impact on enterprise environmental protection investment may be a topic for further exploration to examine the realization of the “Porter Hypothesis” in emerging countries.