How Does Environmental Protection Tax Affect Corporate Environmental Investment? Evidence from Chinese Listed Enterprises

Abstract

This study mainly investigates the nexus of environmental protection tax and corporate environmental investment using a quasi-natural experiment and the difference-in-difference method. On the basis of 422 Chinese listed enterprises from 2012 to 2020, empirical results show that environmental protection tax has significantly reduced corporate environmental investment in China. In terms of mechanism, environmental protection tax inhibits corporate environmental investment through the financing constraints and the substitution effect of innovation investment. In terms of heterogeneity, environmental protection tax has a significant inhibitory effect on environmental investment for large firms, state-owned firms, and firms located in cities with strict regulations and economically developed cities. Environmental protection tax has a negative effect on corporate environmental investment, mainly through direct effect, innovation substitution effect, and financing inhibition effect.

1. Introduction

As a series of world environmental issues such as global warming have emerged, and economic globalization has gradually deepened, sustainable development strategies have gradually reached a consensus in countries around the world. In 2015, the United Nations unanimously adopted a sustainable development agenda known as the Sustainable Development Goals (SDGs), which it hopes to achieve by 2030. One of the main targets of the SDGs is to ensure the lasting protection of the planet and its natural resources. This target shows that sustainable development is not only an economic goal but also a political one and that all countries should fulfill their responsibilities and contribute to the global sustainable development process [1]. The goal of sustainable development will drive enterprises to technological innovation, energy saving, and emission reduction, which will lead to stronger vitality and better corporate performance. As the largest developing country, China has taken a series of measures to promote environmentally friendly activities and urge firms to make environmental investments, such that China can fulfill its responsibility for environmental protection and lead other countries, especially developing countries, to perform environmental protection activities [2]. Environmental investment from enterprises is the main part of environmental investment, and it is a manifestation of a company’s environmental responsibility. By making corporate environmental investments, a company can establish a good image to the market as a company with good environmental awareness and enhance its corporate performance. Thus, we need to focus on the benefits of corporate environmental investment in sustainable development. In addition, according to statistics from Chinese environmental protection authorities, more than 80% of environmental pollution is caused by enterprises in their production and operation [3], especially those in heavily polluting industries. Hence, as a major source of pollution emissions [4], enterprises are expected to take the initiative to fulfill their social responsibility for environmental protection by making corresponding corporate environmental investments and increasing their efforts to promote environmental management. This study explores the influence of environmental tax implementation on corporate environmental investment and describes its intrinsic influence mechanism, filling a gap in the domestic environmental-tax-related literature.

Corporate environmental investment belongs to the category of corporate governance and capital investment. As a special type of investment, companies make environmental investments in pursuit of comprehensive benefits, including economic, environmental, and social benefits [5]. At the same time, by making environmental investments, companies can obtain benefits, such as good social reputation [6] and increased market value, which are conducive to increasing the long-term value of the company. However, given the strong externalities, long periods, high costs, and high risks associated with corporate environmental investments, not only do they fail to bring significant economic benefits in a short period, but the internal operating costs and operational risks are also increased. This phenomenon can lead to a decline in corporate financial performance and, in turn, make companies less willing to make corporate environmental investments. In addition, given the high opportunity costs associated with investing in environmental protection, companies may choose to avoid the opportunity costs, treat them negatively, or even avoid the environmental management work they should be doing. For firms, the motivation to actively invest in environmental protection is to cope with the institutional pressure from environmental regulations [7] and fulfill their social responsibility in environmental protection through investment and financing [5]. However, in view of the inadequate incentive mechanism, the environmental governance of all listed companies is irregular and unsound, thereby making corporate environmental investment a passive behavior, which adversely affects the enthusiasm of corporate environmental investment.

Environmental protection tax is a market-incentive type of environmental regulation tool, and it is a mandatory tool for the government to induce enterprises to conduct relevant environmental protection work. Some scholars at home and abroad have discussed the impact paths and policy effectiveness of environmental regulations on corporate environmental investment, but no concrete conclusion has been made from domestic and international empirical studies on the specific effects of environmental taxes on corporate environmental investment. The findings of positive correlation [8,9,10], negative correlation [11,12], “U-curve” relation [13], and “inverted U-curve” relation [14,15,16,17] have been confirmed under certain conditions, but their correctness and usefulness need to be further tested. In addition, although the effect of environmental regulation on corporate environmental investment has been explored by some scholars, a wide gap still exists in the research results at home and abroad in terms of the specific effect of environmental protection tax on corporate environmental investment and its influence mechanism and policy effectiveness, especially given that the environmental protection tax policy is relatively new and has been implemented in China for only two years now.

This study examines the effect of the implementation of environmental protection tax on firms’ environmental investment and its influence mechanism. This study initially examines the corporate environmental investment in 422 Chinese listed firms between 2012 and 2020. Moreover, firms in heavily polluting industries are treated as the treatment group, and those in non-heavily polluting industries are treated as the control group in a quasi-natural experiment. We find that the implementation of environmental protection tax policies has a negative effect on corporate environmental investment. Then, we investigate the firm-level and regional heterogeneity of the effect of environmental protection tax policy implementation on corporate environmental investment, finding that the disincentive effect of the environmental tax is more significant for large firms, state-owned enterprises (SOEs), firms located in cities with high implementation levels, and firms located in economically developed cities to make environmental investments. Finally, we explore and analyze the mechanism of the effect of environmental protection tax implementation on firms’ environmental protection investment (EPI), such that we can understand how the implementation of environmental protection tax affects corporate environmental investment, finding that the implementation of environmental protection tax inhibits corporate environmental investment through the substitution effect of firms’ technological innovation and enhancement of financing constraints.

This study makes the following contributions to the research in environmental protection. First, this study sheds new light on the relationship between environmental taxes and corporate investment in environmental protection. Although there exists a growing literature on the effect of environmental regulation on corporate environmental investment [13,14,15], few studies have investigated the effect of environmental protection taxes on corporate environmental investment decisions. Particularly, the environmental protection tax, a new policy only implemented in China in 2018, lacks relevant literature to support it. In comparison with traditional environmental regulation, studying the effects of environmental taxes on corporate environmental investments is important. Inspired by Huang et al. [14], this study examines the effect of the implementation of environmental tax on corporate environmental investment, finding that the implementation of environmental tax discourages firms from making EPIs. Second, this study contributes to the understanding of the substitution effect of technological innovation on corporate environmental investment and enriches the literature on environmental regulation. The existing literature has focused extensively on the role of environmental regulation in firms’ investment decisions [14], whereas few studies have considered the substitution effect from firm innovation. The Porter effect of environmental regulation is to increase environmental investment, which is a long-term effect, while our study is of short-term effect. The findings of this study enrich relevant literature on the relationship between environmental regulation and enterprise performance. In addition, the results of this study can help governments implement better policies to encourage firms to engage in innovative activities. We hold that the benefits of corporate innovation investment in environmental protection can replace corporate environmental investment.

The remainder of this paper is organized as follows. Section 2 provides a short review of the relevant literature and analyzes the system and presents the theoretical hypotheses. Section 3 presents the selected sample, data sources, and models. Section 4 reports and analyzes the empirical results. Section 5 reports the conclusions and provides relevant advice.

2. Literature Review, Institutional Background, and Research Hypothesis

2.1. Literature Review

Existing studies generally agree that different types of environmental regulatory instrument, such as environmental protection taxes, can have some effects on corporate environmental investments. The three main types of common environmental regulatory instrument are command-and-control, market based, and public participation [14,18]. Moreover, the four main types of finding on the relationship between environmental regulation and corporate environmental investment are positive and negative correlations and U-curve and inverted U-curve relationships. Huang et al. [14] used panel data of Chinese listed firms from 2008 to 2016 to conduct a regression analysis and applied the instrumental variables approach to address the endogeneity problem. They found that an inverted U-shape relationship exists between command-and-control environmental regulation and firms’ environmental investment. In addition, the market-based environmental regulation and public participation ones are positively related to corporate environmental investment. Wen et al. [19] analyzed financial panel data of non-financial sector firms for 2007–2016 from 11 Asian countries and applied EGLS and GMM methods to address heteroskedasticity and endogeneity. They found that carbon tax hurts corporate environmental investment, and the government green environmental strategy has a positive effect on corporate environmental investment. Moreover, the financial sector plays a moderating role in the above process. Lopez et al. [20] found that uncertainty in environmental regulation has a positive relationship with the decision to invest in corporate environmental protection. In addition, investment history is positively associated with a corporate environmental investment in a given year. Li et al. [15] used two-panel data of Chinese listed firms in the heavy pollution industry and found an inverted U-shape relationship between them.

The review and discussion above indicate that environmental regulation has a significant effect on firms’ environmental investment, but its effect is uncertain. The main reason for this is that their choice of environmental regulation proxy variables is not the same as the choice of measures they take for corporate environmental investment, which, in turn, leads to large differences in their research findings.

In addition, the effect of environmental regulation on enterprise environmental investments has to be analyzed in light of the actual situation in China. First is the effect at the regional level. Lei et al. [21] argued that environmental regulation under China’s environmental policy system has significant regional differences, with local governments being the main commanders and enforcers of implementing environmental regulation; thus, government-related behavior can influence the effectiveness of environmental regulation implementation. Wang et al. [22] also found that governments significantly reduce the enforcement standards of environmental regulations due to performance demands and pressure for local economic growth. Hence, we can speculate that the marketization process may significantly affect the implementation and effectiveness of environmental regulations. Second is the firm-level influence. Song et al. [23] found that, in terms of corporate ownership, SOEs are politically linked and have a greater advantage in market-based environmental regulation constraints compared with non-SOEs.

From the review and discussion above, the significant effect of regional and firm heterogeneity on this study can be easily found. Thus, this study will further investigate the effect of environmental protection tax reform on corporate environmental investments in different regions and firms in the background of China’s reality, such that the findings can be closer to it.

2.2. Analysis of the Institutional Background

Environmental protection tax, first proposed by the British welfare economist Pigou, also known as the “Pigou tax,” has been widely discussed in international taxation circles. The environmental protection tax consists of four major systems: emission, resource, energy, and transportation taxes, which internalize the social costs of environmental resource depletion and ecological damage into the internal operating costs of enterprises through market-based environmental regulation. The official implementation of the environmental protection tax on 1 January 2018 marked the completion of China’s transition from an emission tax to an environmental protection tax, and its policy implementation has been significantly enhanced compared with that in the past. The evolution of China’s environmental tax has been in the exploration and improvement stage and has experienced the following development history: the initial establishment and implementation stage (1978–2002), the total emission fee collection and adjustment stage (2003–2015), the fee to tax phase (2016–2018), and the environmental protection tax implementation phase (after 2018). Since the 18th Party Congress, the Party Central Committee and the government have raised the importance of ecological civilization to an unprecedented level, and the sustainable development concept of “green water and green hills are golden mountains” has taken root in people’s hearts. Moreover, the environmental protection tax has been used as an opportunity to practice the contemporary green development concept, which has been rapidly implemented nationwide. With the adoption of the Environmental Protection Tax Law of the People’s Republic of China in 2016 and the official implementation of the environmental protection tax in 2018, the cost of pollution must be internalized, the green transformation of taxation must be promoted, and energy companies must contribute to the green transformation and ecological development of the economy.

According to the Environmental Protection Tax Law of the PRC, the environmental tax is levied on enterprises, institutions and other production operators in China that directly discharge taxable pollutants into the environment, of which the taxable pollutants are air pollutants, water pollutants, solid waste, and noise, as stipulated in the environmental protection tax law. Additionally, taxation is based on pollutant discharge and is mainly the responsibility of ecological and environmental authorities and tax authorities.

The environmental protection tax policy is essentially changed from sewage charge to environmental tax. After the implementation of environmental protection tax, enterprises, including large enterprises, face higher cost constraints on pollution emissions [24]. Given its strong exogenous nature and progressive characteristics in time and space, environmental tax policy provides a good quasi-natural experiment to explore the effect of an environmental tax on the firm level. In this study, the environmental protection tax is selected as a quasi-natural experiment, and firms located in heavily polluting industries are considered the “treatment group”, whereas firms located in non-heavy polluting industries are considered the “control group” for the policy intervention. On the basis of the implementation of the environmental protection tax policy in 2018 and the differences in the responses of different corporations to the implementation of the policy, a quasi-natural experiment can be used to assess the effect of the environmental protection tax at the microlevel.

2.3. Research Hypothesis Development

As modern economic growth theory gradually focuses on green development, the effect of environmental regulation on corporate environmental investment has become a core issue in the field of environmental economics. On the one hand, given that pollution control activities bring large externalities to enterprises, the optimal level of emissions that maximizes the benefits of enterprises must be higher than the optimal level of emissions for society as a whole [25]. Therefore, the government can transform the social costs of pollution into the internal operating costs of the enterprises by establishing relevant environmental laws and regulations, thereby forcing them to reduce their polluting behavior [26]. Plenty of studies have focused on the effect of environmental regulation on firms’ production and business decisions. It is based on three theoretical hypotheses in environmental economics, namely costly regulation hypothesis [27], pollution haven hypothesis [28], and Porter hypothesis [29]. Furthermore, a large number of relevant theoretical and empirical studies have been conducted to prove its validity [30,31,32,33]. On the other hand, environmental taxes are a price-based environmental regulation instrument, which can incentivize firms to perform environmental protection accordingly through market-based instruments. However, the pressure of increasing internal operating costs and the intensity of enforcement of environmental regulations may induce firms to avoid investing in environmental protection through a range of means. On this basis, this study proposes the following theoretical hypothesis to be tested.

Theoretical Hypothesis 1 (H1):

The implementation of environmental protection tax has a disincentive effect on firms’ EPI.

Environmental regulations, such as environmental protection taxes, have been well-documented to force firms to engage in green technological innovation [34], and the implementation of environmental taxes is intended to encourage firms to make green transformation by engaging in green technological innovation. To avoid compliance costs and enhance their competitiveness, companies will choose to invest in green technology development instead of environmental protection. Moreover, investors are increasingly concerned about the environmental performance of companies in the context of increasing environmental awareness among the population, which has been documented in the literature [35]. If a firm does not implement environmental protection policies sufficiently, resulting in lower environmental performance, then investors may reduce their holdings significantly, thereby creating a strong financing constraint to force the firm to implement environmental protection policies. In addition, the shift from an emission fee to an environmental tax forces firms to pay a portion of their revenues to the appropriate authorities for environmental protection, which increases the pressure on firms to operate and affects their performance, thereby making them potentially subject to more financing constraints as a result and thus discouraging investment in environmental protection. The specific influence mechanism is shown in Figure 1. Accordingly, the following theoretical hypothesis is proposed.

Theoretical Hypothesis 2 (H2):

The implementation of environmental tax inhibits corporate environmental investment, mainly through the substitution effect of innovation investment and financing constraint effect, and the substitution effect of innovation investment is stronger than financing constraint effect.

After the implementation of the environmental protection tax, the financial and technological resources of enterprises of different sizes vary greatly, allowing large enterprises to avoid compliance costs by making green technology innovations instead of environmental investments. With the support of state capital and government policies, SOEs may have a stronger bargaining power than other types of enterprise in making environmental investments, and they can make green technology innovations more easily, thereby reducing their environmental investments to avoid compliance costs. Small- and medium-sized enterprises (SMEs) and other types of enterprise have to pay the corresponding compliance costs to maintain normal business operations due to the limitations of their capital and technological strengths. Therefore, environmental tax policies have heterogeneous intervention effects across different sizes and the nature of enterprises. In addition, according to the pollution haven hypothesis, heavy polluters may avoid compliance costs by shifting their capital from regions with stronger environmental regulations to regions with weaker environmental regulations, thereby gaining more profits. Areas with weaker economic development may choose to accept these heavy polluters for their own economic development, financial incentives, and performance aspirations to harbor their environmental violations and gain more economic benefits and local economic growth at the expense of ecological and environmental quality [36]. Relevant enterprises can reduce their environmental costs and gain higher profits, and relative companies can reduce their environmental protection costs to obtain higher profits. By contrast, areas with higher levels of economic development and higher enforcement capacity can implement environmental tax policies better due to better implementation mechanisms, policy enforcement, and stronger economic power, which enable enterprises in the area to better transform into green companies and replace environmental investments with R&D investments. Hence, there also exist heterogeneous intervention effects of environmental tax policies in regions with different levels of implementation and different levels of economic development. Accordingly, the following theoretical hypothesis is proposed.

Theoretical Hypothesis 3 (H3):

Environmental tax policies have heterogeneous intervention effects across firm sizes, firms with different ownership properties, and regions with different levels of enforcement and economic development.

3. Data and Methodology

3.1. Sample and Data

Referring to relative studies [14,19,37], this study selects Chinese industrial enterprises listed on the Shanghai and Shenzhen Stock Exchanges and sets the sample period for the annual data from 2012 to 2020, with heavy polluters as the treatment group and other types of industrial enterprise as the control group. The reason for using heavy polluting enterprises as the treated group is that heavy polluting enterprises are the main source of pollution emission [14], and they will pay more environmental protection tax than other enterprises, which can more effectively reflect the effect of environmental protection tax implementation on enterprise EPI. In 2012, China paid more attention to the construction of ecological civilization, and the government placed more emphasis on environmental protection and encouraged enterprises to invest in environmental protection. However, the micro data of enterprises’ EPI can only be obtained up to 2020; thus, the sample interval is selected as 2012–2020. To exclude the interference of extreme values, continuous variables are truncated at the 1% and 99% levels.

3.2. Variable Selection and Description

Dependent variable: EPI (lnEPI). The study chooses the natural logarithm of environmental investment [38,39] to measure corporate environmental investment, where environmental investment is the total corporate environmental investment.

Environmental protection tax (ET). The environmental protection tax was implemented and collected in 2018. Thus, this study treats the years in which the environmental protection tax was not implemented as 0 (i.e., ET = 1), and the years in which the environmental protection tax was implemented and thereafter were treated as 1 (i.e., ET = 1).

Heavy polluting enterprises (Treat). This study defines and classifies heavily polluted industries according to the “Guidelines for Disclosure of Environmental Information of Listed Companies (Draft for Public Comments)” issued by the Ministry of Environmental Protection in 2010 and the 2012 industry classification standards of the Securities and Futures Commission, where Treat = 1 if the sample is a heavy polluting enterprise, and Treat = 0 otherwise.

Referring to the relevant literature [2,14,26], the balance sheet ratio (Lev), opportunity cost (Choice), firm size (lnSize), growth (Growth), profitability (ROA), cash flow from operating activities (Flow), and equity concentration (Top1) are selected as control variables. The specific variables are defined in

Table 1. Variable description.

Type	Variable Name	Symbol	Definition
Dependent variable	Corporate environmental investment	lnEPI	ln (corporate environmental investment)
Core explanatory variable	Environmental protection tax	ET	Before 2018 (pre-implementation), ET = 0; in the year of implementation and thereafter, ET = 1
	Heavy polluting enterprises	Treat	Treat = 1 if the enterprise is in a heavily polluting industry, and Treat = 0 otherwise
Control variable	Gearing ratio	Lev	Total liabilities at end of period/total assets at end of the period
	Opportunity cost	Choice	Tobin Q
	Enterprise size	lnSize	ln (Total assets at end of period)
	Growth	Growth	Operating income growth rate
	Profitability	ROA	Net income/total assets at the end of the period
	Cash flow from operating activities	Flow	Net cash flow from operating activities/total assets at the end of the period
	Shareholding concentration	Top1	Percentage of shareholding of the largest shareholder

.

3.3. Empirical Model

The difference-in-difference (DID) method has an excellent performance in analyzing the effect of policy implementation, and it is an effective method to solve the problem of policy effect assessment, which can effectively alleviate the endogeneity issue due to correlation between core explanatory variable and disturbance term; thus, it has been widely used by scholars in environmental protection policy assessments [26,40,41,42]. The implementation of the environmental protection tax is an exogenous shock for each enterprise, and it will have a differentiated effect depending on the region, the feature of the industry, and other factors, which satisfy the basic assumptions of the DID method. Hence, the present study constructs a DID model to analyze the effect of the implementation of environmental protection tax on enterprises’ EPI. The specific model is shown as follows:

$$lnEP{I}_{it}={\beta }_0+{\beta }_{1}E{T}_t\times Trea{t}_i+{\sum }_{j=2}^n{\beta }_j\times X_{jit}+{\delta }_i+{\mu }_t+{\mathsf{\gamma }}_{\mathrm{city}}+{\epsilon }_{it}$$

(1)

where subscript i denotes the firm’s stock code, subscript t denotes the year in which it is located, lnEPI denotes the firm’s EPI, ET denotes whether the environmental protection tax policy started to be implemented in that year, Treat denotes whether the firm is a heavy polluter, ET × Treat denotes the interaction term of the variables ET and Treat, X denotes a series of control variables, δ denotes the firm fixed effect, μ denotes the time fixed effect, γ_city denotes the city feature, and ε denotes the random disturbance term. ${\beta }_j$ is the parameter of focus in this study, and if ${\beta }_1$ is significantly negative, then the implementation of environmental protection tax has a suppressive effect on enterprise environmental investment.

4. Empirical Analysis

4.1. Analysis of Descriptive Statistics

Table 2. Descriptive statistics for variables.

Variable	Obs	Mean	Std. Dev.	Min	Max
lnEPI	3788	10.5950	6.9430	0.0000	21.2140
ET	3788	0.3330	0.4710	0.0000	1.0000
Treat	3788	0.5770	0.4940	0.0000	1.0000
Lev	3788	0.4550	0.2000	0.0080	0.9930
Choice	3786	1.0760	1.3120	−0.8960	26.6350
Growth	3714	21.8962	60.2423	−68.1989	618.0289
ROA	3788	0.0300	0.0840	−1.6480	2.1630
lnSize	3788	22.5700	1.2730	18.3930	27.9660
Flow	3788	0.0860	0.4000	−17.1010	2.7370
Top1	3788	0.3570	0.1480	0.0390	0.9000

reports a descriptive statistical analysis of the main variables. The results show that the mean value of enterprise environmental investment is 10.595, with a standard deviation of 6.493, a minimum value of 0, and a maximum value of 21.214, indicating that a large variation exists in environmental investment occurring in different enterprises in different years. In addition, the mean value of the variable Treat is 0.577, indicating that there exist 2187 annual observations of enterprises belonging to the treatment group and 1601 belonging to the control group, which show even distribution.

4.2. Empirical Results of Baseline Regression

4.2.1. Baseline Regression

This study conducts a quasi-natural experiment, using DID for the treatment and control groups. The results of the baseline regression of environmental tax implementation on corporate EPI are shown in

Table 3. Effect of environmental protection tax policy on EPI.

Variable	lnEPI
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
ET × Treat	−1.069 ** (−2.11)	−1.039 ** (−2.05)	1.099 ** (−2.23)	−1.066 ** (−2.31)	−1.089 ** (−2.20)	−0.927 * (−1.86)	−0.927 * (−1.81)
Lev		−1.149 * (−1.76)		−0.444 (−0.31)		−0.532 (−0.37)	−0.532 (−0.44)
Choice		−0.226 * (−1.92)		−0.284 * (−1.85)		−0.0722 (−0.56)	−0.0722 (−0.78)
Growth		0.00212 *** (2.97)		0.0134 *** (7.49)		0.0135 *** (7.43)	0.0135 *** (7.30)
ROA		0.731 (0.53)		−0.423 (−0.26)		−2.078 * (−1.76)	−2.078 * (−1.93)
lnSize		0.961 *** (8.97)		0.634 * (1.72)		0.954 ** (2.31)	0.954 (1.63)
Flow		0.647 *** (2.88)		0.608 ** (2.19)		0.446 * (1.74)	0.446 * (1.73)
Top1		0.487 (0.64)		0.273 (0.11)		0.800 (0.32)	0.800 (0.31)
_cons	10.48 *** (53.77)	−10.56 *** (−4.77)	11.69 *** (145.15)	−2.117 (−0.26)	10.80 *** (41.76)	−10.20 (−1.15)	−10.20 (−0.82)
Firm FE	No	No	Yes	Yes	Yes	Yes	Yes
Year FE	No	No	No	No	Yes	Yes	Yes
City feature	No	No	No	No	No	No	Yes
N	3788	3712	3788	3712	3788	3712	3712
Adj. R2	0.067	0.098	0.069	0.073	0.366	0.370	0.370

Note: T statistics in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01. FE refers to fixed effects. City features refers to the economic growth rate and environmental regulation at the city level.

. The results show that the coefficient of ET × Treat is significantly negative regardless of the inclusion of control variables, indicating that the implementation of environmental protection tax has a suppressive effect on corporate EPI. The coefficients of ET × Treat are still significantly negative when we sequentially introduce firm fixed effects, year fixed effects, and city features in the model, consistently implying that the implementation of environmental protection tax will induce firms to reduce or shift the probability of EPI.

The coefficients of the control variables are generally in line with theoretical expectations, implying that the empirical results are relatively robust and reliable. The increase in gearing and opportunity cost will increase the difficulty of financing and the possibility of technological substitution. Thus, there exists a disincentive for enterprises to invest in environmental protection. Moreover, profitability, cash flow from operating activities, equity concentration, operating income growth rate, and enterprise size can motivate enterprises to invest in environmental protection, with operating income growth rate all significant at the 1% significance level, and operating cash flow all significant at the 10% significance level. The growth rate of operating income is significant at the 1% significance level, and operating cash flow is significant at the 10% significance level, implying that the growth of operating income can significantly motivate enterprises to invest in environmental protection.

4.2.2. Parallel Trend Test

The ability of the treatment and control groups to satisfy the parallel trend test is a prerequisite for valid estimation of the DID method (i.e., whether the time trends of the means of the explanatory variables in the treatment and control groups are consistent before the policy occurs). The most common methods of ex-ante parallel trend tests are mainly two methods. One is inscribing time trend graphs, and the other is regressions by calculating the mean values of corporate environmental investments in the treatment and control groups. According to the study of Jacobson et al. [43], the event study method is used to test the parallel trend hypothesis and analyze the dynamic effect of environmental protection tax implementation on corporate EPI, and the following model is constructed:

$$lnEP{I}_{it}={\beta }_0+{\sum }_{k=-3}^{k=2}{\beta }_{k}Trea{t}_{it}\times D_{ik}+{\sum }_{j=2}^n{\lambda }_j\times X_{jit}+{\delta }_i+{\mu }_t+{\epsilon }_{it}$$

(2)

where $D_{ik}$ indicates whether the sample year is a dummy variable for the kth year of enterprise i after the environmental protection tax reform, and a negative value of k indicates the (−k) year before the reform implementation, and the other variables are defined as in Model (1). Using the year before the implementation of environmental protection tax as the base year, ${\beta }_k$ denotes the difference between the corporate EPI in the treatment and control groups in the kth year after the implementation of environmental protection tax.

Figure 2 plots the estimation results of ${\beta }_k$at the 95% confidence interval (CI). The coefficient estimates of the environmental investment before the implementation of environmental protection tax are all insignificant at the 5% confidence level, indicating that no significant difference exists between the EPI of heavy polluters and other industrial enterprises before the implementation of environmental protection tax, which satisfies the parallel trend hypothesis. In addition, the coefficient estimates keep decreasing after the implementation of environmental protection tax, and they are significant at the 1% significance level in the second year after implementation, which indicates that the implementation of environmental protection tax is persistent, and its inhibitory effect on enterprises’ EPI increases over time. Although the environmental protection tax policy may have the expected effect on firm behavior before it is introduced, the result in Figure 2 shows that the expected effect is insignificant.

4.3. Heterogeneity Test

4.3.1. Enterprise Heterogeneity

The implementation of environmental protection tax forces firms to make ecological transformation and increase their investment in green technological innovation, thereby replacing corporate environmental investment. However, this situation will only allow SMEs to make corporate environmental investment due to the constraints of objective conditions, such as financing costs, market size, and the cost of following environmental regulation. To explore the heterogeneous impact of environmental protection tax reform on firms of different sizes when making corporate environmental investments, we divided the sample into small-, medium-, and large-scale firms based on the quartiles of firm size (lnSize). In addition, because SOEs receive financial support and policy support from the government, they may have stronger bargaining power and implementation power than non-SOEs to make ecological transformation after the implementation of environmental protection tax, thereby avoiding the cost of environmental regulation to reduce EPI. This study classifies the sample enterprises according to the nature of enterprise ownership (i.e., SOE or non-SOE). SOE = 1 if the firm is an SOE, and SOE = 0 if the firm is a non-SOE. The results of enterprise heterogeneity estimation are reported in

Table 4. Enterprise heterogeneity results of EPI of environmental protection tax policy.

Variable	Business Size			Ownership
	(1) Large	(2) Medium	(3) Small	(4) SOE	(5) Non-SOE
ET × Treat	−3.408 *** (−2.98)	−0.421 (−0.62)	0.119 (0.10)	−1.845 *** (−2.63)	0.147 (0.21)
Lev	−2.607 (−0.64)	0.123 (0.07)	4.853 * (1.74)	−0.626 (−0.28)	0.561 (0.30)
Choice	−0.312 (−0.60)	−0.507 ** (−2.45)	0.513 ** (2.50)	−0.0724 (−0.31)	−0.169 (−0.99)
Growth	0.0138 ** (2.52)	0.488 *** (6.83)	1.340 (0.84)	0.0149 *** (5.87)	0.592 * (1.66)
ROA	−11.22 (−1.37)	−3.517 (−1.38)	−0.750 (−0.59)	−6.844 ** (−2.08)	−1.062 (−0.89)
lnSize	3.024 * (1.95)	1.908 ** (2.21)	−0.583 (−0.50)	1.050 * (1.75)	0.805 (1.51)
Flow	−1.629 (−0.92)	1.096 * (1.79)	0.535 ** (2.28)	0.253 (0.35)	0.388 (1.53)
Top1	4.540 (0.73)	−1.334 (−0.33)	2.457 (0.55)	2.409 (0.73)	0.0709 (0.02)
_cons	−58.92 (−1.65)	−30.82 (−1.61)	19.34 (0.80)	−12.51 (−0.97)	−7.630 (−0.67)
Year fixed effects	Yes	Yes	Yes	Yes	Yes
Firm fixed effects	Yes	Yes	Yes	Yes	Yes
N	947	1893	945	1886	1899
Adj. R2	0.525	0.373	0.302	0.421	0.336

Note: T statistics in parentheses: * p < 0.1, ** p < 0.05, *** p < 0.01.

.

The effect of the implementation of environmental protection tax on corporate environmental investment is heterogeneous in terms of firm size and nature of ownership. The reform of environmental protection tax can significantly affect the corporate environmental investment of large-scale firms and SOEs; however, it is not significant for SMEs and non-SOEs. For large-scale firms, the coefficient of ET × Treat is significantly less than 0 at the 1% significance level, which indicates that the implementation of environmental protection tax can significantly affect the corporate EPI of large-scale firms. However, the coefficient of ET × Treat is not significant for the subsample of SMEs, which indicates that the implementation of environmental protection tax does not significantly affect the corporate environmental investment of SMEs. The specific mechanism may be that the limited assets and technology level of SMEs are not sufficient to increase the R&D investment and promote the technological transformation of enterprises as a substitute for corporate environmental investment. In addition, because SMEs have few financing channels and high financing costs to follow the cost of environmental regulations, the government introduced relevant policies such as environmental subsidies to support the environmental protection and financing of SMEs. Thus, the effect of the environmental protection tax implementation on the financing constraints of SMEs is not significant, which, in turn, leads to the effect on corporate EPI being insignificant. For SOEs, the coefficient of ET × Treat is significantly less than 0 at the 1% significance level, whereas non-SOEs are not significant. Thus, the degree of influence of green technology innovation on firms’ environmental investment substitution and financing constraints after the implementation of environmental protection tax is greater for SOEs than for non-SOEs, reflecting the stronger implementation of policies by SOEs compared with non-SOEs. These results indicate that large enterprises and state-owned enterprises are affected by the environmental protection tax policy and change the environmental protection investment. The results of enterprise heterogeneity also indicate that different types of firms have different responses to environmental protection tax.

4.3.2. Regional Heterogeneity

The purpose of China’s environmental protection tax reform is to strengthen the enforcement of environmental tax levies and prevent a portion of enterprises with strong bargaining power from reducing their payments related to environmental protection. By exploring the heterogeneous effect of the implementation of environmental protection tax policy in different regions, we can infer the potential effect of environmental protection tax implementation on EPI by enterprises in different regions. As the local government has strong bargaining power over SOEs in terms of environmental protection tax payment and EPI, they can avoid the increase in production costs due to changes in tax rates and can thus better replace EPI with technology development. In this study, the natural logarithm of the total value of regional state-owned assets is used to measure the relative enforcement of regional environmental protection policies, and regions with higher than the average value of state-owned assets are classified as high-enforcement regions, and vice versa as low-enforcement regions. In addition, the economic level, technology level, and development of a region have important effects on the implementation of environmental tax. Economically developed regions have relatively sound regulations and implementation tools, which can help them implement more efficient relevant environmental policies compared with other regions, making enterprises in the region more willing to obey them. We choose the natural logarithm of regional GDP to measure the economic development level of different regions, and regions are classified into developed regions (Megacities), medium-developed regions (Type I), and underdeveloped regions (Type II) according to the quartiles of regional GDP. The results of subsample estimation based on the execution level and economic development level are reported in

Table 5. Regional heterogeneity results of EPI of environmental protection tax policy.

Variables	Execution Level		City Size
	(1) High	(2) Low	(3) Megacities	(4) Type I	(5) Type II
ET × Treat	−1.971 *** (−3.11)	0.992 (1.18)	−2.279 ** (−2.33)	0.442 (0.57)	−0.315 (−0.25)
Lev	−0.931 (−0.47)	2.787 (1.30)	2.127 (0.83)	−1.329 (−0.55)	2.309 (0.69)
Choice	−0.128 (−0.72)	0.0134 (0.06)	−0.161 (−0.61)	0.0766 (0.49)	−0.508 (−1.26)
Growth	0.0117 *** (4.94)	0.0835 * (1.74)	0.0149 *** (3.95)	1.300 (0.65)	0.0273 (1.04)
ROA	−3.049 (−1.61)	−1.400 (−0.86)	−1.623 (−0.67)	−4.293 ** (−2.11)	3.817 (0.83)
lnSize	0.0292 (0.06)	2.158 *** (3.19)	−0.838 (−1.03)	1.672 *** (2.71)	1.091 (1.20)
Flow	0.958 * (1.75)	−0.721 (−1.09)	1.387 ** (2.47)	−0.0816 (−0.30)	0.759 (0.67)
Top1	0.818 (0.26)	0.519 (0.15)	4.860 (0.99)	1.472 (0.30)	−4.767 (−1.18)
Year fixed effects	Yes	Yes	Yes	Yes	Yes
Firm fixed effects	Yes	Yes	Yes	Yes	Yes
N	2168	1617	1247	1698	841
Adj. R2	0.408	0.335	0.438	0.046	0.036

Note: T statistics in parentheses: * p < 0.1, ** p < 0.05, *** p < 0.01. The area size order is as follows: Type II < Type I < Megacities.

.

As shown in Table 5, the disincentive effect of the implementation of environmental protection tax on corporate EPI is significant in regions with stronger policy implementation and economically developed regions. Thus, regions with higher levels of economic development and greater power of SOEs have higher action and efficiency in implementing environmental protection tax reform policy compared with other regions, forcing the companies located in this region to replace corporate EPI with green technological innovation. In addition, the substitution effect is more pronounced as the firms in the region are forced to replace corporate environmental investments by making corporate green technology innovations or to reduce corporate environmental investments by tightening the related financing constraints. Furthermore, the estimation results in Table 5 also indicate that the lagging effect of environmental protection tax implementation is more obvious in low-implementation regions, medium-developed regions, and underdeveloped regions. Thus, high-implementation and economically developed regions respond more thoroughly to the policy and pay more attention to the policy than other regions. The incentive effect on the relevant enterprises in the region is more obvious. However, the environmental protection tax has not been fully implemented in all regions of the country, and the reform still needs to be strengthened.

According to the pollution haven hypothesis, when a country’s environmental regulations are more stringent, polluting industries tend to move to countries with relatively weaker environmental regulations to avoid the “compliance costs”. Although the environmental protection tax policy is uniformly applied in China, the differences in enforcement and economic development levels across regions make the effect of environmental regulations become different, and heavy polluters are more likely to move their businesses to regions with weaker regulations rather than invest in environmental protection. The results in Table 5 also show the regional validity of the pollution haven hypothesis. The estimates in Table 5 also demonstrate the validity of the pollution haven hypothesis at the regional level.

4.4. Analysis of Mechanisms

4.4.1. Mechanism of Enterprise Technological Innovation

The implementation of China’s environmental protection tax reform is intended to make enterprises endure the necessary environmental treatment and environmental damage repair costs by replacing fees with taxes and guide them to raise their environmental awareness, increase their treatment efforts, and promote their green technological innovation. According to the Porter hypothesis, appropriate environmental regulations have an incentive effect on the promotion of corporate technological innovation [44]. The technological innovation of enterprises is an important step to solve the environmental pollution problem at the root and achieve the coexistence of environmental protection and economic development. On this basis, to investigate the influence mechanism of corporate technological innovation on corporate EPI after the implementation of environmental protection tax, this study selects the natural logarithm of R&D investment (lnRD) as the mediating variable and initially constructs a regression model of the interchange item ET × Treat on lnRD, and the specific econometric model is then established.

$$lnR{D}_{it}={\beta }_0+{\beta }_{1}E{T}_t\times Trea{t}_i+{\sum }_{j=2}^n{\beta }_j\times X_{jit}+{\delta }_i+{\mu }_t+{\epsilon }_{it}$$

(3)

The coefficients of the interchange item are then estimated to further investigate how environmental protection tax implementation affects corporate environmental investment through corporate technological innovation as a mediating channel.

Table 6. R&D investment as the dependent variable.

Variable	lnRD
	(1)	(2)	(3)
ET × Treat	0.672 * (1.71)	0.943 ** (2.40)	0.933 ** (2.36)
Lev	−4.376 *** (−6.39)	−3.533 *** (−2.92)	−3.451 *** (−2.87)
Choice	−0.231 ** (−2.39)	0.0251 (0.25)	0.00290 (0.03)
Growth	−0.0287 *** (−40.36)	−0.00389 *** (−2.73)	−0.00408 *** (−2.79)
ROA	−3.138 ** (−2.43)	−1.324 (−1.44)	−1.167 (−1.25)
lnSize	1.184 *** (11.60)	2.426 *** (8.23)	2.376 *** (7.10)
Flow	−0.544 (−1.23)	−0.237 (−0.89)	−0.235 (−0.88)
Top1	−1.054 (−1.32)	−0.361 (−0.20)	−0.210 (−0.11)
Year fixed effects	No	No	Yes
Firm fixed effects	No	Yes	Yes
N	3711	3711	3711
Adj. R2	0.065	0.142	0.142

Note: T statistics in parentheses: * p < 0.1, ** p < 0.05, *** p < 0.01.

displays the results of specific estimation.

As shown in Table 6, the coefficient of ET × Treat of Model (1) is significantly positive at the 10% significance level, and those of Models (2) and (3) are significantly positive at the 5% significance level, implying that the implementation of environmental protection tax has a significant positive driving effect on green technology innovation. In addition, the estimation results are still significantly positive after adding the firm fixed effects and time fixed effects, which consistently indicates that the implementation of environmental protection tax will have an incentive effect on enterprises’ green technology innovation. This finding is consistent with the purpose of policy implementation.

Given that firms have limited assets and need to avoid opportunity costs to maximize their interests, they may substitute corporate environmental investments by increasing their R&D investment in corporate green technology innovation to reduce opportunity costs and create economies of scale. In addition, companies can substitute corporate green technology innovation for corporate environmental investment to avoid the corresponding investment and financing costs and risks, which will also have a certain substitution effect on corporate environmental investment. To investigate the substitution effect of corporate technology innovation on corporate environmental investment, this study replaces the explanatory variable (EPI) in the previous test of corporate heterogeneity with the natural logarithm of research and development investment (lnRD) in a group regression of firms with different sizes and ownership properties. The specific estimation results are shown in

Table 7. Heterogeneity results of R&D investment of environmental protection tax policy.

Variable	lnRD
	Business Size			Ownership
	(1) Large	(2) Medium	(3) Small	(4) SOEs	(5) Non-SOEs
ET × Treat	0.633 (0.78)	1.381 ** (2.57)	−0.0564 (−0.09)	1.644 *** (2.86)	0.210 (0.42)
Lev	0.0563 (0.02)	−3.539 * (−1.92)	−3.239 (−1.49)	−3.353 ** (−2.02)	−2.958 * (−1.89)
Choice	−0.673 *** (−2.66)	0.0468 (0.32)	0.0255 (0.33)	0.182 (0.83)	−0.0625 (−0.55)
Growth	−0.00254 (−0.87)	0.0659 * (1.73)	1.257 (1.42)	−0.00263 * (−1.95)	0.279 * (1.80)
ROA	10.54 ** (2.07)	−1.059 (−0.73)	−1.185 (−1.20)	−2.702 (−1.02)	−1.388 (−1.46)
lnSize	3.428 ** (2.49)	3.009 *** (3.71)	1.434 ** (2.57)	1.031 *** (2.83)	3.097 *** (6.40)
Flow	−0.669 (−0.84)	−0.533 (−0.84)	−0.401 (−1.45)	−0.0328 (−0.03)	−0.284 (−1.30)
Top1	−1.013 (−0.29)	−1.216 (−0.32)	0.129 (0.09)	1.064 (0.62)	−2.044 (−0.75)
Year fixed effects	Yes	Yes	Yes	Yes	Yes
Firm fixed effects	Yes	Yes	Yes	Yes	Yes
N	947	1893	945	1886	1899
Adj. R2	0.164	0.064	0.034	0.091	0.180

Note: T statistics in parentheses: * p < 0.1, ** p < 0.05, *** p < 0.01.

.

As shown in Table 7, the coefficients for medium-sized enterprises are significantly positive at the 5% significance level, and those for SOEs are significantly positive at the 1% significance level, implying that the implementation of environmental protection tax can significantly increase the R&D investment of medium-sized enterprises and SOEs. Conversely, the substitution effect of corporate technological innovation is more pronounced for medium-sized enterprises and SOEs compared with other types of enterprises. The reason is that the opportunity cost, the corresponding investment, and financing risks for medium-sized enterprises and SOEs are greater than those for technological innovation. Moreover, they are more significant than those for other types of enterprise, prompting medium-sized enterprises and SOEs to substitute their investment in environmental protection through green technology research and development to obtain higher returns and form economies of scale. In addition, the implementation of the environmental protection tax is intended to motivate enterprises to achieve green transformation through technology innovation. Furthermore, SOEs can complete the green transformation better than non-state enterprises with the assistance of local governments, the support of state capital, and corresponding policy inclination. SOEs, with the promotion of state capital, should also do a good job of being a “role model” for other enterprises. Furthermore, SOEs that achieve the support of national capital should set a good example for other enterprises to promote the implementation of environmental protection tax.

4.4.2. Mechanism of Financing Constraints

Companies responding to the government’s call for pollution control requires corporate environmental investment, which is generally characterized by long period, low upfront returns, and high risk. Internal financing by the company’s strength alone may generate greater pressure of cash flow and operational risk. To avoid the pressure and risk, enterprises may prefer to invest in environmental protection through external financing; thus, the financing constraint has a certain effect on enterprises’ EPI. To investigate the influence mechanism of financing constraints on corporate EPI in the implementation of environmental protection tax, this study constructs the KZ index based on Kaplan et al. [45] by selecting five factors: cash flow, dividend distribution, cash holding, Tobin Q, and gearing ratio. The larger the KZ index is, the stronger the financing constraints on the firm will be. Then, the regression model of the interaction term ET × Treat on the KZ index (KZ) is established, and the model is shown as follows:

$$K{Z}_{it}={\beta }_0+{\beta }_{1}E{T}_t\times Trea{t}_i+\sum _{j=2}^n{\beta }_j\times X_{jit}+{\delta }_i+{\mu }_t+{\epsilon }_{it}$$

(4)

The coefficients are then estimated, and the specific results are shown in

Table 8. KZ index as the dependent variable.

Variable	KZ
	(1)	(2)	(3)
ET × Treat	0.254 *** (3.02)	0.219 *** (2.60)	0.181 ** (2.12)
Lev	−6.207 *** (−34.91)	−5.896 *** (−13.39)	−5.446 *** (−13.81)
Choice	0.639 *** (9.18)	0.584 *** (6.18)	0.494 *** (5.73)
Growth	0.0251 ** (2.05)	−0.0832 *** (−2.93)	−0.0777 *** (−2.79)
ROA	−9.430 *** (−8.55)	−8.741 *** (−5.89)	−8.097 *** (−6.08)
lnSize	−0.335 *** (−11.04)	−0.391 *** (−4.28)	−0.655 *** (−6.33)
Flow	−0.248 ** (−1.98)	0.00939 (0.07)	−0.0484 (−0.39)
Top1	−0.0552 (−0.48)	−0.692 ** (−1.97)	0.0966 (0.29)
Year fixed effects	No	No	Yes
Firm fixed effects	No	Yes	Yes
N	3355	3355	3355
Adj. R2	0.744	0.546	0.611

Note: T statistics in parentheses: ** p < 0.05, *** p < 0.01.

. As shown in Table 8, the coefficient of ET × Treat on KZ index in Model (1) is significantly positive at the 1% significance level, indicating that the implementation of environmental protection tax will significantly tighten the financing constraint. The coefficient of Model (2) is significantly positive at the 1% level, and one of the coefficients of Model (3) is positive at the 5% significance level after adding firm and time-fixed effects. This result consistently indicates that the implementation of environmental protection tax can increase the external financing constraints of firms and thus has a dampening effect on their EPI.

The level of response to the implementation of the environmental protection tax varies from city to city, which, in turn, leads to different financing constraints for firms to invest in environmental protection in different cities. Cities with higher levels of enforcement and developed economies can implement rules and regulations more effectively than other cities. They also have stronger policy enforcement and better supporting measures, which may result in greater financing constraints for firms in those cities compared with other regions. To avoid the higher financing constraints, firms may choose to transfer their assets to areas with lower financing constraints, thereby significantly reducing the amount of environmental investment in the city. To examine the inference above, this study replaces the dependent variable with the KZ index (KZ) based on the original regional heterogeneity and re-estimates the results, as shown in

Table 9. Regional heterogeneity results of KZ index of environmental protection tax policy.

Variable	KZ
	Executive Level		City Size
	(1) High	(2) Low	(3) Megacities	(4) Type I	(5) Type II
ET × Treat	0.189 * (1.76)	0.178 (1.29)	0.315 * (1.79)	0.179 (1.59)	0.450 *** (3.45)
Lev	−5.955 *** (−16.18)	−4.975 *** (−10.75)	−5.424 *** (−10.94)	−5.251 *** (−10.72)	−5.807 *** (−15.27)
Choice	0.436 *** (4.49)	0.602 *** (5.20)	0.862 *** (12.17)	0.399 *** (4.69)	0.853 *** (17.32)
Growth	0.0171 (0.30)	−0.0760 (−1.33)	−0.0159 (−0.20)	−0.00534 (−0.06)	−0.0569 (−1.04)
ROA	−9.080 *** (−11.72)	−8.780 *** (−5.85)	−9.591 *** (−8.55)	−8.810 *** (−9.87)	−12.30 *** (−13.16)
lnSize	−0.554 *** (−5.03)	−0.745 *** (−3.87)	−0.547 *** (−3.14)	−0.812 *** (−5.24)	−0.401 *** (−3.03)
Flow	0.0583 (0.44)	0.0569 (0.63)	0.144 (1.39)	0.262 (1.52)	−0.0173 (−0.15)
Top1	0.109 (0.32)	0.203 (0.29)	0.818 * (1.74)	−0.590 (−1.02)	0.666 (0.90)
Year fixed effects	Yes	Yes	Yes	Yes	Yes
Firm fixed effects	Yes	Yes	Yes	Yes	Yes
N	2010	1345	912	1640	803
Adj. R2	0.636	0.602	0.660	0.602	0.677

Note: T statistics in parentheses: * p < 0.1, *** p < 0.01. The area size order is as follows: Type II < Type I < Megacities.

.

The estimation results in Table 9 show that, in terms of enforcement, although the coefficients of financing constraints are not significantly positive for both cities located in high enforcement and low enforcement cities, the p-value of the coefficient for high enforcement cities is 0.103, which is basically significant, implying that the degree of financing constraints is more pronounced in high enforcement cities than in low enforcement cities after the implementation of environmental protection tax, which is consistent with the inference above. In addition, in terms of economic development level, the coefficient for economically developed regions is significantly positive at the 10% significance level. The coefficient for medium-developed regions is not significantly positive, but the p-value of its coefficient is 0.112, which is basically significant. The coefficient for underdeveloped regions is significantly positive at the 1% significance level, which implies that the financing constraint of low-developed regions is more significant than that of the other two types of region after the implementation of the environmental protection tax policy. The possible reason for this is that less-developed regions have fewer financing channels than better-developed regions, and attracting better companies for investment and financing is difficult for these regions due to their development level. Combined with the pollution haven hypothesis above, heavy polluters may use their negotiation power to achieve “government-enterprise collusion” with the local government after transferring their assets, thereby avoiding EPI to protect their interests. In addition, from the magnitude of the coefficients, we can infer that the regional financing constraints do not change significantly after the implementation of the environmental protection tax. This inference suggests that financing constraints are not the main mechanism influencing firms to invest in environmental protection.

4.5. Robustness Test

To ensure the robustness of the estimation results, the following regression analysis is conducted in this study. (1) The environmental protection tax reform that occurred in 2018 is treated as a quasi-natural experiment and faces the problem of policy self-selection. This study conducts propensity score matching (PSM) based on the control variables for the firms in the treatment and control groups to mitigate endogeneity while controlling for firm and time fixed effects. (2) Control variables are added to the regression model, which may be related to firms’ environmental investments, such as Tangibility, Agencost, and Separation. (3) To exclude the effect of some environmental policies’ implementation, such as the “de-capacitation” policy (implemented in 2013) on corporate environmental investment, the sample year interval is shortened, and the study starts from 2014 [46]. (4) The standard errors of the regressions are replaced with the standard errors of the clustering at the industrial level. The results of the robustness test are reported in

Table 10. Regression results of the robustness test.

Variables	lnEPI
	(1) PSM-DID	(2) Add Control	(3) Shorten Period	(4) Replace SE
ET × Treat	−0.893 * (−1.78)	−0.917 * (−1.83)	−1.001 ** (−2.07)	−1.354 ** (−2.10)
Separation		0.0392 (0.94)
Tangibility		−0.0506 (−0.85)
Agencost		−0.149 (−0.18)
Control	Yes	Yes	Yes	Yes
Year FEs	Yes	Yes	Yes	Yes
Firm FEs	Yes	Yes	Yes	Yes
N	3766	3785	2944	2698
Adj. R2	0.371	0.371	0.456	0.378

Note: T statistics in parentheses: * p < 0.1, ** p < 0.05.

.

The results in Table 10 report that the implementation of environmental protection tax still has a significant inhibitory effect on corporate investment in environmental protection, which is the same as in the previous findings, and the robustness test is passed. The sample obtained through the PSM method somehow moderates the potential threat of differences between the treatment and control groups, and the coefficient of ET × Treat is significantly negative at the 10% significance level, indicating that policy endogeneity does not interfere with the findings. After adding the control variables, replacing the original standard errors with robust standard errors for clustering at the industry level, and shortening the sample year interval, the results are still robust. Thus, the implementation of relevant environmental policies does not affect the previous conclusions.

5. Conclusions and Implications

On the basis of the quasi-natural experiment of the implementation of the Environmental Protection Tax Law, this study investigates the effects of exogenous interventions of environmental protection tax policies on corporate EPI by using DID based on firm-level EPI data and financial data during the period of 2012–2020. In addition, this study explores the potential influence of the implementation of environmental protection tax policies on firms’ EPI from the perspective of firm heterogeneity and regional heterogeneity. The empirical results show that ① the implementation of environmental protection tax has a significantly negative effect on firms’ investment in environmental protection. ② The environmental protection tax replaces investment in environmental protection mainly through corporate technological innovation, and the Porter hypothesis is supported. ③ The environmental protection tax also discourages investment in environmental protection by tightening external financing constraints. ④ There exists firm heterogeneity in the implementation effect of the environmental protection tax reform, and its disincentive effect on EPI is more significant in large firms and SOEs. In addition, there exists regional heterogeneity in the effect of environmental tax implementation on corporate environmental investment, which is more significant in high enforcement regions and economically developed regions. The above conclusions hold true through a series of robustness tests.

The findings of the study theoretically expand the research on the effects of environmental protection policies on environmental investment and the influence mechanism, practically providing important policy insights for the government to implement environmental protection tax policies more effectively and promote the sustainable development of enterprises. In response to people’s growing demand for green ecology, the environmental protection tax should be better used to promote green technology innovation, fulfill the responsibility of ecological protection, and comprehensively promote the green development of industrial enterprises. Specifically, first, the government should improve the implementation rules of the environmental protection tax system, strengthen the implementation of the policy, and ensure that the environmental protection tax system is implemented in an orderly manner. In addition, corporations should pay more attention to their environmental protection performance. Second, the government should optimize the environmental protection tax levy in terms of tax rate level, tax scope, and taxation standards, strengthening the push-back mechanism of environmental protection tax on enterprise green technology innovation to motivate enterprises to actively address pollution. Third, we should strengthen the monitoring of the effect of environmental protection tax on different types of enterprise and region, and investors should pay attention to the effect of the policy on different types of enterprise and different regions. The relevant departments should put emphasis on the “compliance costs” brought by the environmental tax reform to SMEs, reduce the investment and financing risks and pressures faced by SMEs in environmental management, and strengthen the monitoring of the implementation progress in the relevant regions to prevent the phenomenon of “collusion between government and enterprises” [47,48]. Fourth, the financing mechanism should be improved for the green transformation of enterprises, and related measures should be supported to reduce the burden of investment and financing for enterprises in environmental protection activities from the market perspective. In terms of mechanism, this study has only conducted a preliminary investigation of the influence mechanism, and room to explore the intrinsic influence mechanism remains. Future research should investigate its influence mechanism in depth.