The Influence of Environmental Protection Tax Law on Urban Land Green Use Efficiency in China: The Nonlinear Moderating Effect of Tax Rate Increase

Abstract

Due to the basic carrier function of land, the economic and ecological effects of Environmental Protection Tax Law (EPTL) will be reflected in the land use. Therefore, this article investigates the effect of EPTL on land green use efficiency (LGUE). To be specific, based on the panel data of 278 prefecture-level cities in China from 2012 to 2020, LGUE is evaluated through a global super efficiency epsilon-based measure (EBM) with unexpected output. Then, the reform of “sewage fee-to-tax” is regarded as a natural experiment to accurately evaluate the effect of EPTL on LGUE. The result that the implementation of EPTL significantly drives LGUE is confirmed. The mechanism tests show that the implementation of EPTL enhances the intensity of green innovation, promotes the optimization of industrial structure, and thereby improves LGUE. Moreover, we find that the moderating effect of tax rate increase is nonlinear and exhibits an inverted U-shape. That is, below a certain value, the tax rate increase will strengthen the EPTL’s ability to improve LGUE. However, after exceeding the value, the tax rate increase will weaken the EPTL’s ability to improve LGUE. Targeted suggestions are proposed for improving the environmental protection tax system and LGUE.

1. Introduction

With the acceleration of urbanization in China, non-agricultural land use is in rapid expansion. According to the National Bureau of Statistics of China, as of 2021, the urban built-up area in China has reached 62,420 square kilometers, an increase of over 22,000 square kilometers compared to 2010. However, unreasonable land use has caused problems such as high pollutant emissions and high carbon emissions [1,2]. The above issues not only exacerbate urban greenhouse gas emissions, but also reduce the carbon sequestration capacity of land [3]. The Outline of the 14th Five-Year Plan (2021–2025) for National Economic and Social Development and Vision 2035 of the People’s Republic of China pointed out that we should achieve new progress in ecological civilization construction, accelerate green transformation and development, and comprehensively improve the efficiency of resource utilization. Land is the core element input in the process of socio-economic development. Its utilization should focus on improving land green use efficiency (LGUE), that is, not limited to pursuing economic benefits, but including the environmental benefits in the evaluation index system [4]. The LGUE, reflecting the green and high-quality transformation of land use under the carbon emission reduction goal, is the key to solving structural contradictions in production, achieving carbon peak and carbon neutrality, and promoting high-quality development. From the perspective of factors affecting LGUE, in addition to direct factors such as inputs and outputs, exogenous policy shocks can also have an impact, especially related policies in environmental regulation. Therefore, taking China’s first unitary taxation specifically reflecting the green tax system [5], Environmental Protection Tax Law (EPTL), as the starting point, this paper tries to clarify the impact of switching a pollutant discharge fee to an environmental tax on urban LGUE.

The original pollution discharge fee system under administrative regulations presents various drawbacks such as low collection rate, local administrative intervention, and insufficient enforcement rigidity, making it difficult to adapt to the high-quality development [6]. The key way to control pollution and improve resource utilization efficiency is not to accelerate economic growth, but to require more strict and effective environmental regulations. On 1 January 2018, China’s EPTL was officially implemented, which solved various drawbacks of the original pollution charge system by the “sewage fee-to-tax”. As of 2023, the environmental protection tax has been levied for 5 years. According to relevant research, compared to cities that did not increase tax rates, cities that increased tax rates experienced a higher pollution reduction and lower land expansion [7,8]. So, from the perspective of green and high-quality transformation of land use, can the implementation of EPTL promote the urban LGUE? If the promoting effect holds, what are the potential impact mechanisms? Moreover, the EPTL grants provinces the right to independently adjust the specific tax rates for pollutants within the prescribed scope, with 45% of provinces increasing the specific tax rates [9]. So, do different increases in tax rates lead to different policy effects? The scientific evaluation and quantification of the above issues are of great significance to better leverage the positive role of taxation in environmental protection, improve the level of land resource utilization, and achieve high-quality development in China.

The research on environmental tax started from Pigou’s research on the externality of production in the 1920s, flourished in the 1990s when OECD countries carried out environmental tax reform, and matured in the early 21st century. From the historical context of literature, relevant research followed the basic logic of “discovering problems—constructing theories—guiding practice—rediscovering problems—deepening theories—guiding practice” [10]. Pigou (1932) was the first to systematically study the theoretical issues of environment and taxation, but did not provide a very detailed explanation of the theoretical mechanism of environmental tax [11]. Tullock (1967) and Pearce (1991) pointed out that environmental pollution has additional costs for social welfare, and environmental tax, linking the polluter’s payment with the controller’s cost compensation, is an ideal tool for internalizing environmental externality [12,13]. Since the 1990s when OECD countries launched a wave of environmental tax reforms, research on the effects of environmental tax became increasingly mature and provided useful decision-making guidance. Early research on environmental tax often believed that environmental tax naturally contributes to improving environmental quality [14,15]. Therefore, subsequent research often focuses on analyzing the impact of environmental tax on the economy and society, usually within the framework of “double dividend”, in an attempt to explore the existence and conditions of the “double dividend”. The so-called “double dividend” refers to the dual benefits that environmental tax reform will generate: on the one hand, reducing pollution emissions, improving environmental quality, and obtaining a “green dividend”; on the other hand, improving economic efficiency, promoting Pareto improvement, and obtaining a “blue dividend” [13]. Based on the intensity of the double dividend, Goulder (1995) divides it into strong and weak forms [16]. In fact, the weak form emphasizes the comparison of effects in different tax return methods; the strong form emphasizes the impact of environmental tax reform on economic efficiency. With the deepening of research, many scholars found that environmental tax reform can drive economic growth, increase social employment, and even promote social welfare and social equity on the basis of reducing pollution emissions [17]. Wesseh and Lin (2019) established a Computable General Equilibrium (CGE) model to simulate the policy effects of implementing carbon tax reform in Libya, finding that both unified and local environmental tax reforms would lead to a decrease in energy consumption and improve environmental quality, while alleviating unemployment and improving social welfare [18]. While the research on environmental tax attracted attention from governments around the world, some scholars question the double dividend hypothesis [19,20]. However, some studies not only raised doubts and criticisms, but also further explored the specific conditions under which the double dividend hypothesis exists, usually involving the issue of optimal tax rates. Hu et al. (2020) conducted a policy effect analysis on the implementation of EPTL in China and found that due to the low tax rate, environmental tax does not even have a basic environmental dividend, thus negating the double dividend effect of environmental tax [21]. Fullerton et al. (2008) analyzed several possible scenarios and conditions for double dividend [22]. If the environmental tax rate is lower than the optimal level, increasing the environmental tax rate may result in a double dividend; if the environmental tax rate is already at the optimal level, further increasing the tax rate can only bring an environmental dividend; if pollution has been controlled through administrative measures, raising the environmental tax rate may not even achieve the basic environmental dividend; if the environmental tax is higher than the optimal level, further increasing the tax rate will only bring about a reduction in social welfare.

After the transformation of environmental tax from a theory to policy practice, relevant research closely aligns with the development strategies and public policy needs, which has been fully reflected in the recent literature. Given that Japan established its greenhouse gas emission reduction target in 2015, Lee et al. (2018) analyzed the possibility of Japan achieving its 2030 emission reduction target through carbon tax under different assumptions [23]. Research showed that even without nuclear energy, carbon tax policy can reach the emission reduction target while ensuring GDP growth. Aydin and Esen (2018) evaluated the impact of environmental tax on carbon emissions in EU countries and found that environmental tax has a threshold effect on emission reduction [24]. Only when the tax rate exceeds a threshold can emissions be effectively reduced, but there is no clear evidence to prove the existence of the double dividend. Drawing on the experience of developed countries, China officially proposed a pollution discharge fee system in 1978 and officially piloted it in 1979. After nearly 40 years of operation, the EPTL was implemented on 1 January 2018. Therefore, existing research mainly focuses on the discussion of the pollution discharge fee system, which can be summarized into three aspects: first, the emission reduction effect. A large number of scholars found that imposing pollution discharge fees can significantly suppress pollution emissions [25,26]. However, some scholars questioned the emission reduction effect of the pollution fee system [27]. Second, the economic effects. Some scholars found that imposing pollution discharge fees has a significant inhibitory effect on economic development [28]. However, some scholars found that imposing pollution discharge fees can help promote economic growth, such as improving corporate performance [29], generating a technical progress effect [30], and promoting green technology innovation [31]. Third, the combined effects. Scholars began to integrate the environmental and economic effects of the pollution fee system into a theoretical framework to study the dual benefits in order to achieve a win-win situation between environmental protection and economic development. Shi et al. (2019) found that a pollution fee can not only improve the environmental quality of Shandong Province, but also promote economic growth [32]. With the proposition of environmental protection tax in 2015, scholars attempted to use CGE models to simulate the effects of environmental tax reform. Xiao et al. (2015) used the CGE model to explore the impact of environmental tax on the Chinese economy and found that environmental tax is beneficial for environmental improvement, but has a negative impact on economic variables [33]. As of 2023, the EPTL has been levied for 5 years. Some scholars began to pay attention to the policy effects of EPTL, but focused on the pollution reduction effect [34,35,36]. Only a few studies examined the economic effects of EPTL, such as green technology innovation [31], corporate performance [37,38], and water resource utilization efficiency [39].

Due to the basic carrier function of land, the economic and ecological effects of EPTL should be reflected in the land use. However, the existing literature on LGUE mainly focuses on measurement [40], spatial-temporal pattern [41], regional differences [42], influencing factors [43], etc. With regard to the analyses of influencing factors, land resource mismatch [44], new urbanization [45], land finance [46], and urban agglomeration [47] are taken into account. “Porter effect” is the concept that well-designed environmental regulations (economic tools such as environmental tax or tradable permit) can not only offset compliance costs, but also stimulate enterprises to improve efficiency, thereby getting a win-win solution of “pollution reduction” and “efficiency enhancement” [48]. However, there are limited studies on the impacts of environmental regulations on urban LGUE. A few studies explored the effect of environmental regulation on LGUE at the provincial level [49]. Some research evaluated the impact and internal mechanisms of low-carbon city pilot policy on LGUE at the urban level [2].

In summary, scholars conducted extensive discussions on the environmental, economic, and combined effects of the pollution fee system, but the effectiveness of the pollution fee system has always been controversial. The low effectiveness of the pollution discharge fee system is closely related to drawbacks of low collection rate, local administrative intervention, and insufficient enforcement rigidity. Therefore, the key to improving the effectiveness is to overcome the aforementioned shortcomings. Therefore, the EPTL has been officially implemented, marking the formal transformation from the pollution discharge fee system to the environmental protection tax system. Whether the change can improve the effectiveness of the pollutant discharge fee and thus achieve a win-win situation for environmental protection and economic development still remains to be explored. However, the existing literature mostly discusses the environmental or economic effects separately, and rarely analyzes combined effect. Furthermore, the controversy over the dual dividend of environmental tax is often related to the issue of optimal tax rate, and the existing literature also lacks discussion. Thus, taking the reform of “sewage fee-to-tax” as a natural experiment, the global super efficiency EBM model with unexpected output and the difference-in-differences model are used to investigate the impact of EPTL on LGUE empirically. We conclude that the implementation of EPTL significantly improves LGUE. Then, the impact mechanisms are explored. We find that EPTL improves LGUE by enhancing the intensity of green innovation and promoting the upgrading of industrial structure. Moreover, we investigate the moderating effect of tax rate increase and find it is nonlinear and exhibits an inverted U-shape. Specifically, below a certain value, the tax rate increase will strengthen the EPTL’s ability to improve LGUE. However, if the tax rate increase exceeds the value, the implementation of EPTL will weaken the EPTL’s ability to improve LGUE.

The possible contributions are as follows: First, from the perspective of green and high-quality transformation of land use under the carbon emission reduction goal, we examined the policy effect of EPTL. The reform of the pollution discharge fee system to the environmental protection tax system will have a significant impact on the green transformation of land use, but existing research is relatively lacking. The green transformation of land use, as an important link between environmental protection and economic development, is an essential point to investigate the combined effect of EPTL. Therefore, this paper focuses on the green transformation of land use, which enriches the research on policy effects of EPTL. Second, we theoretically explore the impact mechanisms of EPTL on LGUE. The implementation of EPTL can fundamentally overcome the defects of the pollution discharge fee system, thus promoting the green transformation of land use. Therefore, the article clarifies the impact mechanism of EPTL on green transformation of land use, which helps to deeply understand the internal logic. Third, from the perspective of optimal tax rates, this paper studies the moderating effect of tax rate increase, attempting to explain the controversy over the double dividend of environmental tax. The controversy over whether EPTL can achieve a double dividend is closely related to the rights granted by EPTL to provinces to independently adjust the tax rates. Therefore, analyzing the nonlinear impact of environmental tax can help provide a theoretical basis and decision-making reference for achieving better policy effectiveness.

2. Policy Background and Research Hypotheses

2.1. Policy Background

Drawing on the experience of developed countries, the Chinese government promulgated the Environmental Protection Law (for trial implementation) in 1979, establishing the pollution discharge fee system and piloting it in some regions. To strengthen the management of pollutant discharge fee collection, the Regulation on the Administration of Collection and Use of Pollutant Discharge Fees was officially implemented in July 2003, which clarified the pollution discharge fee system on the basis of type and quantity of pollutants. In order to achieve the emission reduction targets of China’s eleventh Five-Year Plan and twelfth Five-Year Plan, since 2003, provinces gradually adjusted the collection standards for pollutant discharge fees, increasing the discharge fee of air and water pollutants. Despite the positive emission reduction effects of the pollution discharge fee system, due to the low collection rate, narrow collection scope, lack of compulsion and standardization, local administrative intervention, and insufficient enforcement rigidity, it was difficult for the pollution discharge fee to achieve the ideal effect [50]. In December 2016, the EPTL, China’s first tax law embodying the “green tax system”, was adopted at the 25th meeting of the Standing Committee of the 12th National People’s Congress and officially implemented on 1 January 2018.

Based on the pollution discharge fee, the implementation of EPTL adopts a gradual reform model to achieve a smooth transition from fee to tax. The main characteristics are as follows. First, high compulsion and standardization. The basis of taxation shifted from an administrative setting to statutory taxation, making it more mandatory; the legal foundation shifted from scattered to unified, and the institutional design shifted from extensive to refined, making it more standardized [51]. Second, sufficient enforcement rigidity. The environmental protection tax adopts a levy model of “enterprise declaring, tax department levying, environmental department collaborating, and information sharing”, which can ensure that the nominal amount is consistent with the actual amount [52]. Third, less local administrative intervention. Pollution discharge fees are set by the government, which are administrative fees. In pursuit of economic growth, local governments may intervene in the collection procedure of pollution fees. However, environmental protection tax is a statutory tax that is disconnected from departmental interests, making it difficult for local governments to intervene. Fourth, a positive emission reduction incentive mechanism. The EPTL provides tax incentives to emission reduction enterprises, implementing the policy of “paying more taxes for more emissions, and paying less taxes for fewer emissions”. Fifth, high proportion of local income sharing. The income from pollution discharge fees is divided in a ratio of 1:9 between the central and local governments [53]. After the implementation of EPTL, the State Council issued Notice on the Attribution of Environmental Protection Tax Revenue, which clearly states that “all environmental protection taxes are treated as local income”. Sixth, a dynamic tax rate adjustment mechanism. The EPTL adopts a tax rate adjustment mechanism of “the state sets a bottom line, and local governments can float up”. Local governments have the right to choose the levy standards for pollutants based on the local situation. The environmental protection tax in some regions still follows the original standards for collecting pollution discharge fees, but some regions have raised the tax rate.

2.2. Theoretical Analysis and Research Hypotheses

2.2.1. Impact of EPTL on Urban LGUE

Environmental protection tax (also called Pigouvian tax) can internalize the social cost of pollution into the production cost, further controlling the negative externality of environmental pollution and allocating environmental resources through the market mechanism. A reasonable design of an environmental protection tax system helps to identify the price signal of pollutant emissions. For one thing, it strengthens the control of pollution emission; for another, it drives resources to flow to greener fields. After the implementation of EPTL, some regions raised the tax rates. Moreover, the characteristics of environmental protection tax such as less local administrative intervention, high compulsion and standardization, and sufficient enforcement rigidity mean that the levy intensity increased. The dual improvements of levy standard and intensity led to enhanced environmental regulations. According to the Porter hypothesis, levying environmental taxes can force enterprises to innovate, thereby achieving “emission reduction” and “efficiency enhancement” [48]. Moreover, the double-dividend hypothesis suggests that environmental taxes can not only effectively suppress pollution emissions and achieve a “green dividend”, but local governments can also make use of environmental taxes to diminish the distorted effects of existing tax systems, thereby achieving a “blue dividend” [13]. Therefore, after the implementation of EPTL, the economic and environmental benefits in urban land use improve, while the emissions of pollutants reduce, leading to an improvement in urban LGUE.

On the basis of the above analysis, we propose the following hypothesis:

Hypothesis 1.

The implementation of EPTL can improve urban LGUE.

2.2.2. Impact Mechanism of EPTL on Urban LGUE

The Porter hypothesis indicates that targeted and flexible environmental regulations can offset compliance costs and reduce uncertainty about environmental investment, thus stimulating green technology innovation [48]. First, the strengthening of environmental regulations brought by EPTL makes enterprises concerned about the insufficient utilization of resources and potential environmental innovation opportunities, thereby generating enterprises’ external pressure to promote green technology innovation. Under the enterprise competitiveness theory, pressure from external sources can reduce innovation inertia and promote innovative thinking, which is conducive to urban innovation intensity [54]. Second, EPTL not only inherits the national mandatory force and universal binding force, but also demonstrates that the government determines to practice environmental protection in the long term, thereby reducing the uncertainty about green investment and driving green technology innovation. Third, tax incentives provide incentives for polluting enterprises to innovate in source control and end-of-pipe control. Finally, a high proportion of local income sharing is conducive to increasing government support for green technology innovation. For one thing, it can improve enterprise productivity; for another, it can reduce pollutant emissions. Thus, urban LGUE is improved.

On the basis of the above analysis, we propose the following hypothesis:

Hypothesis 2.

The implementation of EPTL can improve urban LGUE by increasing the intensity of urban green technology innovation.

On the basis of profit maximization, enterprises will make discretionary decisions based on their own conditions, showing heterogeneous behaviors, thus promoting the upgrading of industrial structure. The differences in tax rate increases among different provinces offer space for polluting enterprises to evade environmental regulations through inter-regional shift. On the one hand, high tax rates will have a “crowding-out effect” on pollution-intensive industries in the jurisdiction. On the other hand, tax incentives will form incentives for the clean industry in other regions, thereby promoting the immigration of the clean industry. Moreover, under the dual pressure of high tax rates and emigration costs, some pollution-intensive enterprises will transform into a clean sector or an exit market. As the market share of clean industry increases, the industry structure is optimized, leading to an improvement in urban LGUE.

On the basis of the above analysis, we propose the following hypothesis:

Hypothesis 3.

The implementation of EPTL can improve urban LGUE by the optimization of industry structure.

2.2.3. Moderating Effects of Tax Rate Increases

Due to differences in regional environmental carrying capacity and pollutant emission status among provinces, there are differences in tax rate increases. In the stage where the tax rate increase is relatively low, the negative impact of environmental tax is relatively small due to the lower cost of compliance. When the increase in amount increases, according to the Porter hypothesis and the pollution haven hypothesis [55], increasing the intensity of environmental regulation will make enterprises act actively instead of passively. On the one hand, by increasing R&D investment, the level of green innovation will be improved, thus improving enterprise productivity; on the other hand, pollution-intensive enterprises will conduct inter-regional shift, intra-regional shift, or exit the market. At this point, the tax rate increase will increase the EPTL’s ability to improve LGUE. However, When the tax rate increase is too large, excessive environmental regulation intensity will generate significant compliance cost, squeeze out profit space and innovative resources, increase production costs, and diminish innovation ability and enterprises’ competitiveness, thereby leading to an adverse effect on urban LGUE.

On the basis of the above analysis, we propose the following hypothesis:

Hypothesis 4.

The moderating effect of tax rate increase is nonlinear. Specifically, below a certain value, the tax rate increase will increase the EPTL’s ability to improve LGUE. However, if the tax rate increase exceeds the value, the implementation of EPTL will weaken the EPTL’s ability to improve LGUE.

3. Research Design

3.1. Measurement Model

3.1.1. Measurement Model of Urban LGUE

To rank Decision-Making Units (DMU) and compare LGUE in different periods, we select the global super efficiency EBM model with unexpected output to measure urban LGUE. Taking prefecture-level cities as DMUs, for ${DMU}_i$, $m$ inputs $x_i=\left(x_{i1},x_{i2},\dots ,x_{im}\right)$ can produce s expected outputs $y_i=\left(y_{i1},y_{i2},\dots ,y_{is}\right)$ and $p$ unexpected outputs $b_i=(b_{i1},b_{i2},\dots ,b_{ip})$. The production possibility sets based on global technology is:

$$P=\left\{\begin{array}{c}\left(\overline{x},\overline{y},\overline{b}\right)|{\sum }_{t=1}^T{\sum }_{j=1,j\ne 0}^n{x}_j^t{\gamma }_j^t\le \overline{x^t};\\ {\sum }_{t=1}^T{\sum }_{j=1,j\ne 0}^n{b}_j^t{\gamma }_j^t\le \overline{b^t};\end{array}\right.\left.\begin{array}{c}{\sum }_{t=1}^T{\sum }_{j=1,j\ne 0}^n{y}_j^t{\gamma }_j^t\ge \overline{y^t}\\ {\sum }_{t=1}^T{\sum }_{j=1,j\ne 0}^n{\gamma }_j^t=1;\gamma \ge 0\end{array}\right\}$$

(1)

$\left(\overline{x},\overline{y},\overline{b}\right)$ is the optimal solution of the model; $\gamma$ is the weight variable. The global super efficiency EBM model with unexpected output is:

$$K^{*}=\underset{\theta ,\phi ,\gamma ,s^{-},s^{+}}{\mathit{min}}\frac{\theta +{\epsilon }_x{\sum }_{i=1}^m\frac{w_i^{-}{s}_i^{-}}{x_{io}}}{\phi -{\epsilon }_y{\sum }_{r=1}^s\frac{w_r^{+}{s}_r^{+}}{y_{ro}}-{\epsilon }_b{\sum }_{q=1}^p\frac{w_q^{b-}{s}_q^{b-}}{b_{qo}}}$$

(2)

$$\left\{\begin{array}{c}{\sum }_{t=1}^T{\sum }_{j=1,j\ne 0}^n{x}_{ij}^t{\gamma }_j^t-s_i^{-}\le \theta x_{io},i=1,\dots ,m\\ {\sum }_{t=1}^T{\sum }_{j=1,j\ne 0}^n{y}_{rj}^t{\gamma }_j^t+s_r^{+}\ge \phi y_{ro},r=1,\dots ,s\\ \begin{array}{c}{\sum }_{t=1}^T{\sum }_{j=1,j\ne 0}^n{b}_{qj}^t{\gamma }_j^t-s_q^{b-}\le b_{qo},q=1,\dots ,p\\ {\sum }_{t=1}^T{\sum }_{j=1,j\ne 0}^n{\gamma }_j^t=1\\ \gamma \ge 0,s_i^{-}\ge 0,s_r^{+}\ge 0,s_q^{b-}\ge 0\end{array}\end{array}\right.$$

(3)

$K^{*}$ indicates the optimal efficiency of the EBM model. $s_i^{-}$, $s_r^{+}$, and $s_q^{b-}$, respectively, represent non-negative relaxation of the input factor, expected output factor, and unexpected output factor. $w_i^{-}$, $w_r^{+}$, and $w_q^{-}$, respectively, represent the weight of input factor, expected output factor, and unexpected output factor. $\theta$ and $\phi$ are the planning parameters for the radial part. With a value range of [0, 1], $\epsilon$ is the key parameter that represents the combination degree of radial and non-radial relaxation.

3.1.2. Econometrics Model

Compared to traditional OLS estimation, the difference-in-differences model can largely alleviate the problem of endogeneity. In addition, using fixed effects estimation also alleviates the problem of missing variable bias. The central government stipulated the upper and lower limits of tax rate. According to local conditions, local governments can independently set tax rates. The environmental protection tax in some regions still follows the previous standards for collecting pollution discharge fees, but Beijing, Tianjin, Chongqing, Hebei, Henan, Jiangsu, Shandong, Hunan, Sichuan, Guizhou, Hainan, Guangdong, Guangxi, and Shanxi raised the tax rates. Whether regions raised tax rates provides a practical basis for distinguishing between the treatment and control groups. To effectively evaluate the effect of EPTL on urban LGUE, this paper takes cities raising tax rates as the treatment group and other cities as the control group, and constructs the following difference-in-differences model:

$${ULGUE}_{it}={\alpha }_0+{\alpha }_1{Treat}_i\times {Post}_t+\sum _{j=2}^J{\alpha }_j{Control}_{ijt}+{\gamma }_t+{\mu }_i+{\epsilon }_{it}$$

(4)

$i$ and $t$ are city and year, respectively. ${ULGUE}_{it}$ is the level of urban LGUE of city $i$ in year $t$. ${Treat}_i$ is a dummy variable. If the city raised the tax rate, the value is 1; otherwise it is 0. ${Post}_t$ is a dummy variable. If $t\ge 2018$, the value is 1; otherwise it is 0. ${\alpha }_1$ is the focus of this paper. ${Control}_{ijt}$ are control variables that influence urban LGUE. ${\epsilon }_{it}$ is random error term. ${\mu }_i$ is individual fixed effect. ${\gamma }_t$ is the time fixed effect.

3.2. Variable Selection

3.2.1. Dependent Variable

The Urban LGUE ($ULGUE$) can be interpreted as the ratio of the outputs of economy and ecology to the inputs of land and non-land factors under certain technical conditions. Its connotation is to maximize the environmental and economic outputs with the minimum input of land factor and the minimum environmental loss. Based on traditional land use efficiency measurement, the measurement of LGUE considers the environmental loss during the land use process and incorporates unexpected outputs into the model. In order to comprehensively reflect the urban LGUE, with reference to Chen et al. (2022) [56], this paper selects nine indicators to construct the evaluation index system of urban LGUE (

Table 1. Evaluation index system of urban LGUE.

Layer of Criteria	Layer of Factors	Layer of Indicators	Unit
inputs	labor	number of employees at the end of the year	km2
	land	area of built districts	104 persons
	capital	total investment in fixed assets of the whole society	108 CNY
outputs	economic benefits	GDP	108 CNY
	environmental benefits	green coverage rate of built districts	%
	unexpected outputs	carbon emissions	ton
		industrial so2 emissions	ton
		industrial soot emissions	ton
		industrial waste water	104 tons

).

3.2.2. Independent Variables

Treatment dummy variable ($Treat$) and time dummy variable ($Post$). When a city raised the tax rate, the value of $Treat$ is 1; otherwise it is 0. If $t\ge 2018$, the value of Post is 1; otherwise it is 0.

3.2.3. Control Variables

Referring to Liu et al. (2022) [2] and You et al. (2022) [57], regional economic development ($Lngdp$) is determined by the natural logarithm of per capita GDP of a prefecture-level city. Government intervention ($Gover$) is determined by the ratio of general public budgeting expenditure to urban GDP. Education ($Education$) is determined by the ratio of the quantity of students in higher education institutions to the total population. Foreign direct investment ($FDI$) is determined by the ratio of the quantity of foreign-invested enterprises to industrial enterprises above the designated size. The industrial structure ($Industry$) is determined by the GDP proportion of the secondary and tertiary industry.

3.3. Data Description

The empirical test sample of this paper is panel data of 278 cities in China from 2012 to 2020. Data of urban LGUE indicators come from the China City Statistical Yearbook (2013–2021) and China Stock Market and Accounting Research Database. The missing values are filled by linear interpolation. This article conducts Winsorize treatment at the 1% and 99% levels to avoid the impact of extreme values on regression results and deflates all nominal values to real values based on 2012. The descriptive statistics are displayed in

Table 2. Descriptive statistics.

Variable	N	Mean	S.D.	Minimum	Maximum
$ULGUE$	2502	0.656	0.158	0.254	1.125
$Lngdp$	2502	10.745	0.556	9.007	13.056
$Gover$	2502	0.206	0.099	0.081	0.625
$FDI$	2502	0.039	0.043	0.003	0.232
$Education$	2502	0.035	0.017	0.013	0.111
$Industry$	2502	0.879	0.075	0.618	0.994

.

4. Empirical Results

4.1. Baseline Regression Results

Columns (1) and (2) of

Table 3. Baseline regression results.

	$\mathit{U}\mathit{L}\mathit{G}\mathit{U}\mathit{E}$
	(1)	(2)
$Treat\times Post$	0.024 * (0.006)	0.014 * (0.005)
control variables	no	yes
constant term	yes	yes
time fixed effect	yes	yes
individual fixed effect	yes	yes
sample size	2502	2502

* represent the coefficients are significant at 1% confidence levels. The standard error reported in parentheses is clustering to provinces.

show the regression results. The coefficient of $Treat\times Post$ is significantly positive at a confidence level of 1% without control variables in Column (1). Column (2) suggests that after adding control variables, the coefficient of $Treat\times Post$ remains positive and significant at the 1% confidence level. Thus, the implementation of EPTL can improve urban LGUE. After the implementation of EPTL, some regions raised the tax rates. Moreover, the characteristics of environmental protection tax such as less local administrative intervention, high compulsion and standardization, and sufficient enforcement rigidity mean that the levy intensity increased. The implementation of EPTL brings about dual improvements in levy standard and intensity, resulting in an increase in the intensity of environmental regulations. For one thing, it strengthens the control of pollution emission; for another, it drives resources to flow to greener fields. Ultimately, it drives urban LGUE. Hypothesis 1 is verified.

The impact mechanism of the implementation of EPTL on LGUE in Hypothesis 2 and Hypothesis 3 needs to be further tested. Moreover, although the overall promotion effect of EPTL has been verified, it remains to be verified whether its promotion effect still holds as the tax rate increases. The following is intended to explore the impact and moderating mechanisms. The specific verifications are in Section 4.3 and Section 4.4.

4.2. Robustness Tests

4.2.1. Parallel Trends Test and Dynamic Effects Analysis

The parallel trend assumption is a prerequisite for using the difference-in-differences model, and the treatment and control groups should meet the parallel trend assumption before the policy occurs (ex ante). On the contrary, if there is a certain difference between the treatment and control groups, then the results of the difference-in-differences model can no longer represent the net effect of the policy, and there is a high possibility that other factors may affect the changes in the explanatory variables. This paper uses the event study methodology to verify whether the LGUE of treatment and control groups is significantly different before the official implementation of EPTL. The model is:

$${ULGUE}_{it}={\alpha }_c+\sum _{k=-5}^2{\alpha }_k{Policy}_{it}^k+\sum _{j=5}^J{\alpha }_j{Control}_{jt}+{\gamma }_t+{\mu }_i+{\epsilon }_{it}$$

(5)

Among them, $k$ represents each year minus 2018, and ${Policy}_{it}^k$ represents the implementation of policy in each year. The base year is 2012. Figure 1 shows the results, and the parallel trends test result indicates that the estimated coefficient did not pass the significance tests before 2018. This proves that before the official implementation of EPTL, the changes in LGUE between the treatment and control groups are not significantly different. Thus, the parallel trend assumption is satisfied.

From the perspective of dynamic effects, after the official implementation of EPTL, the difference in efficiency between cities raising tax rates and other cities gradually increases, suggesting that the effect of EPTL on LGUE is positive. However, the LGUE only shows a significant upward trend after the second year (after 2019). Thus, the impact of EPTL on LGUE has a delay of about one year.

4.2.2. Placebo Tests

To avoid the trend changes in the treatment and control groups after the policy intervention time point are influenced by other policy or random factors, we further conducted placebo tests. Referring to Liu and Lu (2015) [58], due to the fact that the model in this paper is a single-time-point difference-in-differences model estimation, it is only necessary to randomly select cities as the treatment group based on the same number as the real treatment group. This article randomly selects 119 cities from the sample cities as the false treatment group cities, and the other cities as the false control group cities. The estimated coefficient of the impact of implementing placebo policy on LGUE can be obtained. We repeated the above process 500 times to obtain 500 regression coefficients and their corresponding p-values. Finally, the coefficient kernel density and scatter diagram of coefficient and p-value are drawn. The coefficient kernel density estimations are shown in Figure 2a. Figure 2b shows the scatter diagram of coefficient and p-value. The real coefficient value of benchmark regression is the vertical solid line, and the average value of the coefficient estimated 500 times is the vertical dotted line. As shown in Figure 2a, the mean value of coefficients is near 0, and the coefficients have an approximately normal distribution. The vast majority of p-values in Figure 2b are greater than 0.1. The above results suggest that EPTL does not significantly affect LGUE under random sampling. Thus, the placebo tests passed.

4.2.3. Propensity Score Matching and Difference-in-Differences (PSM-DID)

The benchmark regression results indicate that the EPTL will improve urban LGUE, but sample selection bias may lead to estimation errors. Propensity score matching is adopted to control the difference between the treatment and control groups. The steps are as follows. First, select the feature variable, and perform 1:1 nearest neighbor matching to obtain the propensity score. Then, calculate the standard deviation of the propensity score to calculate the caliper width. Finally, based on the caliper width, match by using the kernel matching method. Figure 3 shows the sample features before and after matching, and it can be observed that after matching, the difference between the treatment and control groups is significantly reduced. Finally, the difference-in-differences model is used to estimate the matched samples, and

Table 4. PSM-DID regression results.

	$\mathit{U}\mathit{L}\mathit{G}\mathit{U}\mathit{E}$
	(1)	(2)
$Treat\times Post$	0.024 * (0.006)	0.015 * (0.005)
control variables	no	yes
constant term	yes	yes
time fixed effect	yes	yes
individual fixed effect	yes	yes
sample size	2452	2452

* represent the coefficients are significant at 1% confidence levels. The standard error reported in parentheses is clustering to provinces.

displays the results.

4.3. Impact Mechanism Analysis

After the implementation of EPTL, some cities raised the tax rates. Moreover, the characteristics of environmental protection tax such as less local administrative intervention, high compulsion and standardization, and sufficient enforcement rigidity mean that the levy intensity increased. The dual improvements of levy standard and intensity led to the strengthening of environmental regulations. The Porter hypothesis indicates that well-designed environmental regulations may offset compliance costs, reduce uncertainty about the value of environmental investment, and stimulate green technology innovation. Also, tax incentives provide incentives for polluting enterprises to innovate in source control and end-of-pipe control. High proportion of local income sharing is conducive to increasing government support for green technology innovation. Moreover, the strengthening of environmental regulations led to the relocation, upgrading, and exit of polluting enterprises within the jurisdiction, thus optimizing the industry structure.

Drawing on the approach of Wang et al. (2023) [59], the urban green innovation intensity ($GTIA$) is measured by green patent applications. The (1) column of

Table 5. Impact mechanism tests.

	$\mathit{G}\mathit{T}\mathit{I}\mathit{A}$	$\mathit{S}\mathit{T}\mathit{R}1$	$\mathit{S}\mathit{T}\mathit{R}2$	$\mathit{S}\mathit{T}\mathit{R}3$
	(1)	(2)	(3)	(4)
$Treat\times Post$	0.145 * (0.027)	0.001 (0.001)	−0.012 * (0.003)	0.011 * (0.002)
control variables	yes	yes	yes	yes
constant term	yes	yes	yes	yes
time fixed effect	yes	yes	yes	yes
individual fixed effect	yes	yes	yes	yes
sample size	2502	2502	2502	2502

* represent the coefficients are significant at 1% confidence levels. The standard error reported in parentheses is clustering to provinces.

shows the mechanism test result of urban green innovation intensity. The result shows that after adding control variables, $Treat\times Post$ is significantly positive at the 1% confidence level, suggesting that EPTL significantly increases the intensity of urban green innovation, which is consistent with the empirical analysis result of Huang et al. (2022) at the enterprise level [60]. Hypothesis 2 is verified.

Drawing on the approach of Niu et al. (2023) [61], the GDP proportion of three industries is used as the dependent variable, and $Treat\times Post$ is treated as the independent variable. The (2), (3), and (4) columns of Table 5 are the mechanism test results of the GDP proportion in the primary industry ($STR1$), the GDP proportion in the secondary industry ($STR2$), and the GDP proportion in the tertiary industry ($STR3$), respectively. The results show that the estimated coefficient of $Treat\times Post$ corresponding to $STR1$ is not significant, the estimated coefficient of $Treat\times Post$ corresponding to $STR2$ is significantly negative at the 1% confidence level, and the estimated coefficient of $Treat\times Post$ corresponding to $STR3$ is significantly positive at the 1% confidence level, indicating that the implementation of EPTL significantly promotes the optimization of urban industry structure. Hypothesis 3 is verified.

4.4. Moderating Effect of Tax Rate Increase

To effectively identify the differences in the impact of EPTL on urban LGUE under different tax rate increases, this paper further constructs the following model based on the difference-in-differences model:

$$\begin{array}{c}{ULGUE}_{it}={\alpha }_0+{\alpha }_1{Treat}_i\times {Post}_t\times {ATI}_i+{Treat}_i\times {Post}_t\times {{ATI}_i}^2\\ +{\displaystyle \sum _{j=2}^J}{\alpha }_j{Control}_{ijt}+{\gamma }_t+{\mu }_i+{\epsilon }_{it}\end{array}$$

(6)

${ATI}_i$ is the increase in tax rate of city $i$.

Table 6. Moderating effect of tax rate increase.

	$\mathit{U}\mathit{L}\mathit{G}\mathit{U}\mathit{E}$
	(1)	(2)
$Treat\times Post\times ATI$	0.026 * (0.010)	0.032 * (0.010)
$Treat\times Post\times {ATI}^2$	−0.019 * (0.004)	−0.020 * (0.004)
control variables	no	yes
constant term	yes	yes
time fixed effect	yes	yes
individual fixed effect	yes	yes
sample size	2502	2502

* represent the coefficients are significant at 1% confidence levels. The standard error reported in parentheses is clustering to provinces.

shows the results, where columns (1) and (2), respectively, report the estimated moderating effects of tax rate increases without and with control variables. Column (1) shows that, without control variables, the coefficient of $Treat\times Post\times {ATI}^2$ is significantly negative at the 1% confidence level. From column (2), it can be seen that with control variables, the coefficient of $Treat\times Post\times {ATI}^2$ is still negative and significant at the 1% confidence level. Therefore, when the tax rate increases by a certain threshold, the EPTL can drive urban LGUE. When it exceeds this threshold, the reform of EPTL will suppress urban LGUE.

The reason is that in the stage where the tax rate increase is relatively low, the negative impact of environmental tax is relatively small due to the lower cost of compliance. When the increase in amount increases, increasing the intensity of environmental regulation will make enterprises act actively instead of passively. At this point, the tax rate increase will increase the EPTL’s ability to improve LGUE. However, when the tax rate increase is too large, excessive environmental regulation intensity will generate significant compliance cost, squeeze out profit space and innovative resources, increase production costs, and diminish innovation ability and enterprises’ competitiveness, thereby leading to an adverse effect in urban LGUE. Hypothesis 4 is verified.

5. Conclusions and Policy Implications

Whether abroad or domestically, empirical research showed that imposing an environmental tax can achieve an environmental dividend, but there has always been controversy over whether it can achieve an economic dividend. The urban LGUE reflects the green and high-quality transformation of land use. As an important link for the organic coordination of environmental protection and economic development, it is an important starting point to investigate the combined effect of environmental tax. Thus, we examined the policy effect of EPTL on LGUE. Moreover, disputes over the policy effects of environmental tax often involve the issue of optimal environmental tax rates. Therefore, we also included an analysis of the nonlinear moderating effect of tax rate increase. Specifically, based on the global super efficiency EBM with unexpected output and the difference-in-differences model, China’s reform of “sewage fee-to-tax” is regarded as a natural experiment in our research to investigate the effect of EPTL on LGUE empirically. We conclude that the implementation of EPTL significantly improves LGUE. Second, the impact mechanisms are explored. We find that EPTL improves LGUE by promoting the intensity of green innovation and the upgrading of industrial structure. Moreover, we find that the moderating effect of tax rate increase is nonlinear and exhibits an inverted U-shape. Specifically, below a certain value, the tax rate increase will strengthen the EPTL’s ability to improve LGUE. However, if the tax rate increase exceeds the value, the implementation of EPTL will weaken the EPTL’s ability to improve LGUE.

The conclusions of this paper reveal the driving effect of EPTL on urban LGUE. However, a higher tax rate increase does not necessarily mean a better policy effect, and it should be maintained within a reasonable range to achieve a better policy effect. In theory, this article expands the research on the driving factors of green and high-quality transformation in land use. In practice, it provides important implications for the government to improve the environmental protection tax system and promote sustainable land use. The details are as follows.

First, improve the implementation rules of EPTL. At present, there are some problems in the implementation of EPTL, such as unclear division of rights and responsibilities between the environmental protection department and the tax department, and insufficient attention from the tax department, which will weaken the policy effect of EPTL. Therefore, the government should continue to improve the implementation rules and effectively connect the pollution discharge fee system with the environmental protection tax system. Specifically, strengthen departmental collaboration; clarify the main body, specific procedure, and rights and obligations of various links in tax collection collaboration; and improve the efficiency of trans-departmental collaborative collection. Also, the scope of environmental tax collection is too narrow. Local governments can appropriately expand the scope of taxation based on the special needs of local pollutant reduction, such as incorporating carbon dioxide, volatile organic compounds, etc., into the scope of environmental protection tax.

Second, ensure the enforcement rigidity of EPTL. In addition to the rationality of the system, the mechanism of EPTL to promote the green transformation of land use also lies in the enforcement rigidity. If the enforcement rigidity is insufficient, it will greatly reduce the policy effect of EPTL. Therefore, the government should effectively ensure the effective implementation of EPTL and strengthen the rigidity of law enforcement. Specifically, we should further improve China’s pollution source discharge declaration system. To enhance the enforcement rigidity and effectiveness of environmental protection departments, we can consider comprehensively utilizing methods such as irregular inspections and spot checks by environmental law enforcement departments, online monitoring of key pollution sources, and public opinion supervision to reduce the phenomenon of enterprises’ under-reporting.

Third, fully utilize environmental protection tax revenue and establish innovative compensation and tax incentive mechanisms. The increase in production costs caused by environmental protection tax may squeeze R&D expenditures, which is not conducive to green technology innovation and the improvement of LGUE. Moreover, the crowding out effect of environmental protection tax on polluting enterprises has adverse impacts on local economies. The State Council made it clear that all environmental protection taxes are treated as local income, and local governments should fully utilize environmental protection tax revenue and establish innovative compensation and tax incentive mechanisms. The innovation compensation mechanism can solve the dilemma of green R&D in enterprises, promote green technology innovation, and drive sustainable land use. The tax incentive mechanism can introduce clean enterprises to the local community, alleviate the adverse impacts of environmental protection tax on the economy, achieve coordinated development of environmental protection tax reform and industrial upgrading, and improve urban LGUE.

Fourth, establish a dynamic adjustment mechanism and gradually increase tax rates in stages. The environmental tax rates in most regions of China are relatively low. Considering that at the beginning of the introduction of environmental protection tax, the smooth implementation of the law should be ensured first, so it is not advisable to excessively increase it too quickly. However, in the long run, in order to truly constrain enterprise pollution discharge through tax, various regions should further increase the tax rates of air and water pollutants in the later stage based on local economic development level, industrial structure, and other factors. Also, due to the game between enterprises and the government, if the tax rate difference between regions is too large, it will inevitably cause inter-regional transfer of polluting enterprises. However, if the tax rate is too low, the policy effect cannot be realized. Therefore, the country should timely raise the minimum limit of environmental tax to meet the cost of pollution control while reducing tax rate differences between regions.

Fifth, maintain the tax rates for pollutants within the prescribed tax adjustment range. Affected by environmental performance assessments, local governments increased their emphasis on pollution control. However, significantly increasing the tax rates for pollutants in a short period of time does not reduce pollutant emissions, but rather increases production costs for enterprises. Therefore, when formulating the tax rates for environmental protection tax, local governments should fully consider the actual situation of local economic development, and the increase in tax rates for pollutants should not be too high. For regions where the increases in tax rate are too high, local governments should promptly assess the adverse effects of the environmental protection tax when increasing policy support for the clean sector. Innovative compensation mechanisms and policy incentives should be used to help transform the polluting industry and encourage the entry of the clean industry, alleviating the inhibitory effect of high regulatory costs on urban LGUE.

In short, increasing the tax rates of pollutants through environmental protection tax can generally drive the LGUE, but specific implementation standards need to consider local realities in order to make policies play a greater role. It is necessary to establish a good communication and cooperation mechanism between central and local governments, as well as between local governments, in order to promote sustainable economic development.

Despite the above meaningful findings in this article, there are still the following limitations.

First, the spatial effects of EPTL were not considered. The difference in regional environmental protection tax rates may lead to the inter-regional transfer of pollution, known as the pollution heaven hypothesis. A systematic study of this issue helps to formulate regional coordinated environmental tax policies and prevent the inefficiency of individual actions. This is also our plan for further in-depth research.

Second, the specific optimal environmental tax rate value was not explored. Systematic research on this issue will provide practical reference for tax rate formulation. However, this issue is actually quite complex. Not only do we need to consider the tax rate standard of each city, but we also need to consider the tax rate standards for different pollutants. Therefore, in the next step, we will explore the specific optimal environmental tax rate based on spatial effects.