Does Environmental Decentralization Affect the Supply of Urban Construction Land? Evidence from China

Against the background of Chinese decentralization, the preferences and choices of local governments significantly affect the scale and structure of urban construction land supply. Due to the shortage of financial funds and the political performance pursuit of local governments, environmental decentralization gives local governments greater autonomy in environmental management, and increases the possibility for local governments relying on land transfer income to make up for the financial gap and provide public goods and services. This paper analyses the impact of environmental decentralization on the construction land supply scale of local government based on the panel data of 30 provinces in China from 2003 to 2015. The results indicate that: (1) environmental decentralization has a positive effect on the increase in urban construction land supply scale; (2) environmental decentralization affects urban construction land supply by strengthening land financial dependence and distorting land resources misallocation; (3) there are regional disparities in the effect of environmental decentralization on urban construction land supply. The impact is greater in regions with high financial pressure, high economic growth pressure, and low environmental protection pressure. In summary, some policy suggestions are put forward to reasonably supply urban construction land against the background of Chinese decentralization.


Introduction
Land resources are extremely important for ensuring food security, facilitating economic development, and protecting the ecological environment [1,2]. Urbanization has accelerated the conversion of agricultural land around cities to non-agricultural construction uses, which is a common problem faced by the whole world [3,4]. This phenomenon is particularly acute in developing countries, such as India and China [5]. Over the years, the important role of land in China's rapid economic development cannot be ignored. A great deal of construction land has been provided by the Chinese government to support urbanization and industrialization [6]. According to data from the "China urban construction statistical yearbook", the area of urban construction land in China increased from 6720 km 2 to 52,761.3 km 2 , with an average annual growth of 6.06%, from 1981 to 2016. In contrast, this rate of urban population growth over the same period was only 2.98%. However, economic growth that heavily relies on land not only causes many social problems such as land finance, illegal land use, inefficiency of industrial land, and soaring housing prices [7], but also leads to environmental problems, such as declining biodiversity and environmental pollution [8]. Governments are working to curb urban sprawl through environmental regulation [9][10][11][12][13].
In response to environmental problems, including those caused by urban sprawl and land resources misallocation, the Chinese government has adopted a decentralized Land 2023, 12, 601 3 of 18 in the following aspects: (1) environmental decentralization is incorporated into the Chinese decentralization framework, which expands the Chinese decentralization system. (2) Based on the Chinese decentralization framework, the urban construction land supply behaviors of local governments are analyzed. The impact of environmental decentralization on urban construction land supply is emphatically analyzed, which provides a new concept for studying the influencing factors of local government land supply. (3) Combined with the incentive mechanism of urban construction land supply under Chinese decentralization, the impact of environmental decentralization on land supply in regions with different financial pressure, economic growth pressure, and environmental protection pressure is considered, which provides policy directions for coordinating economic development and environmental protection.

Institutional Background
China has a unique institutional background of fiscal decentralization and political centralization, which is different from the fiscal decentralization and federal political system of Western countries [36]. In terms of finance, China implements a tax-sharing system, and the fiscal budgets of governments at all levels are relatively independent. This means that local governments have certain economic decision-making power and can govern according to their own preferences [37]. In terms of politics and bureaucracy, China is in power with one party and implements vertical management with upward responsibility, which forms a multi-departmental M-shaped hierarchy [38,39]. The superior (or central) government has the right to appraise the performance of local government officials and decide on their promotion [40].
Corresponding to the delegated powers, China implements a decentralized environmental management system [41]. That is, environmental decentralization. Environmental decentralization originates from environmental federalism, which studies the environmental management functions of all levels of government [42]. In the "pyramid" system, where the central government is located above and the local governments are located below, local governments are authorized to participate in environmental management affairs [43]. It is generally believed that environmental centralization can effectively control cross-regional pollution, and that environmental decentralization is more flexible and targeted [44][45][46][47]. Since the reform and opening up in 1978, the evolution of environmental decentralization in China can be divided into three stages [48]. The first stage was from 1978 to 1993. During this stage, financial and administrative powers were highly decentralized, and the degree of environmental decentralization was high. The central government lacked supervision over local environmental affairs. The second stage was from 1994 to 2007. During this stage, with the increase in central government's fiscal revenue, the environmental management powers of central government improved, and the degree of environmental decentralization decreased. The third stage is from 2008 to the present. With the adjustment of the government environmental protection institutions and the increased emphasis on environmental protection, the central government decentralized environmental administrative power and enhanced environmental supervision power and environmental governance incentives. Several instances in the literature have confirmed the beneficial effects of environmental decentralization in China, including promoting green development and pollution control [49,50]. However, some scholars believe that the institutional background of fiscal decentralization and political centralization in China makes the incentives and constraints of local government environmental protection mismatched. When decentralizing environmental rights, local governments will relax environmental control, which forms a race to the bottom and aggravates environmental pollution [51,52].
The fiscal decentralization, political centralization, and environmental decentralization implemented in China can be summarized as Chinese decentralization.

Theoretical Hypothesis
The institutional arrangement of Chinese decentralization affects the construction land supply by local government in three aspects: fiscal decentralization, political centralization, and environmental decentralization. In terms of fiscal decentralization, the 1994 tax-sharing reform led to a misalignment of the executive and fiscal powers of the government [53]. Fiscal revenues flow upwards to the central government, and the responsible affairs remain in the local government, resulting in a huge fiscal gap [54,55]. Decentralization has led local governments to provide land for construction to ease fiscal pressures and cover fiscal deficits. This is manifested in the expansion of fiscal revenue through "land rent" and "land tax" [56]. The "land rent" refers to the one-time income obtained by transferring land. Selling commercial and residential land at a high price can obtain more "land rent" [57,58]. The "land tax" refers to attracting investment by selling industrial land at low prices, and developing industries to obtain continuous tax revenue [59]. When it comes to political centralization, Chinese officials are usually accountable to the top. Local government officials are promoted through excellent performance appraisals [60]. Economic growth is the core of performance measurement, so there is fierce competition around GDP growth [61][62][63]. In order to win the GDP championship, local governments often sell industrial land at reserve prices or free of charge to attract investment and promote economic growth, which increases the supply of construction land [28,29]. The importance of environmental indicators in performance appraisals has increased in recent years, which limits the expansion of construction land [36]. Environmental decentralization affects the supply of construction land by relaxing or tightening environmental constraints after weighing the importance of economic growth and environmental protection. After incorporating environmental decentralization into the Chinese decentralization framework, the incentives and constraints of local government construction land supply are shown in Figure 1. The fiscal decentralization, political centralization, and environmental decentralization implemented in China can be summarized as Chinese decentralization.

Theoretical Hypothesis
The institutional arrangement of Chinese decentralization affects the construction land supply by local government in three aspects: fiscal decentralization, political centralization, and environmental decentralization. In terms of fiscal decentralization, the 1994 tax-sharing reform led to a misalignment of the executive and fiscal powers of the government [53]. Fiscal revenues flow upwards to the central government, and the responsible affairs remain in the local government, resulting in a huge fiscal gap [54,55]. Decentralization has led local governments to provide land for construction to ease fiscal pressures and cover fiscal deficits. This is manifested in the expansion of fiscal revenue through "land rent" and "land tax" [56]. The "land rent" refers to the one-time income obtained by transferring land. Selling commercial and residential land at a high price can obtain more "land rent" [57,58]. The "land tax" refers to attracting investment by selling industrial land at low prices, and developing industries to obtain continuous tax revenue [59]. When it comes to political centralization, Chinese officials are usually accountable to the top. Local government officials are promoted through excellent performance appraisals [60]. Economic growth is the core of performance measurement, so there is fierce competition around GDP growth [61][62][63]. In order to win the GDP championship, local governments often sell industrial land at reserve prices or free of charge to attract investment and promote economic growth, which increases the supply of construction land [28,29]. The importance of environmental indicators in performance appraisals has increased in recent years, which limits the expansion of construction land [36]. Environmental decentralization affects the supply of construction land by relaxing or tightening environmental constraints after weighing the importance of economic growth and environmental protection. After incorporating environmental decentralization into the Chinese decentralization framework, the incentives and constraints of local government construction land supply are shown in Figure 1. Under the specific institutional background of Chinese decentralization, environmental decentralization gives local governments greater autonomy in choosing economic development goals and environmental protection goals [64]. Local governments' economic incentives and political incentives do not match the environmental protection constraints, which causes local governments to prioritize economic growth over environmental welfare. Local governments ignore environmental issues and supply a great deal of construction land to generate revenue and boost the economy.
Hence, the first hypothesis is obtained: environmental decentralization will increase the urban construction land supply by local government.
When environmental management power is decentralized, the local government will compare the importance of economic incentives, political incentives, and environmental Under the specific institutional background of Chinese decentralization, environmental decentralization gives local governments greater autonomy in choosing economic development goals and environmental protection goals [64]. Local governments' economic incentives and political incentives do not match the environmental protection constraints, which causes local governments to prioritize economic growth over environmental welfare. Local governments ignore environmental issues and supply a great deal of construction land to generate revenue and boost the economy.
Hence, the first hypothesis is obtained: environmental decentralization will increase the urban construction land supply by local government.
When environmental management power is decentralized, the local government will compare the importance of economic incentives, political incentives, and environmental constraints. Against the background of the shortage of local government's fiscal revenue and the pursuit of political performance, environmental decentralization increases the motivation of local governments to pursue "land rent", "land tax", and "attracting investment from land", so that local governments reduce environmental protection requirements. This leads to the intensification of the local government's land financial dependence and land resources misallocation. Thus, the scale of construction land supply is increased.
Hence, the second hypothesis is obtained: environmental decentralization increases the urban construction land supply scale of local governments by increasing land financial dependence and land resources misallocation.
Environmental decentralization affects construction land supply by changing the economic incentives, political incentives, and environmental constraints of local governments. Therefore, in regions with different financial pressures, economic growth pressures, and environmental protection pressures, the incentives and constraints of land supply are also distinct. In regions with high financial pressure, local governments have stronger economic incentives. When environmental power is decentralized, there is a stronger incentive to provide a great deal of land at the expense of the environment [65]. In regions with high economic growth pressure, local governments have a stronger motivation to attract investment by land [66,67]. Environmental decentralization will increase the supply of industrial land, which expands the scale of construction land supply. In regions with high environmental protection pressures, local governments have stricter constraints on the preservation of the environment. When environmental power is devolved, local governments will optimize local environmental management and limit the scale of land supply [68].
Hence, the third hypothesis is obtained: there are regional disparities in the effects of environmental decentralization on urban construction land supply. In the regions with high financial pressure, high economic growth pressure, and low environmental protection pressure, the impact of environmental decentralization on construction land supply is more significant.

Model Settings
In the context of Chinese decentralization, environmental, fiscal decentralization, and political centralization jointly affect the urban construction land supply by local governments. Thus, the following fixed-effect model is established [50].
where LS it is the explained variable, representing construction land supply; ED it is the explanatory variable, representing environmental decentralization; FD it is fiscal decentralization; FAI it and IPC it are fixed asset investment and environmental protection investment, respectively, representing political centralization; X it are other control variables, including per capita GDP, population, urbanization rate of land, the proportion of secondary and tertiary industries, financial pressure; i and t represent the provinces and periods under consideration, respectively; β 0 is the constant term; β 1 -β i are coefficients; θ i and µ t represent the province and time effects, respectively; ε it is the random error term. In order to make the data smoother and reduce heteroscedasticity, the non-proportional data are dealt with as logarithms.

Variable Selection
(1) Explained variable. The construction land supply scale is used to describe the construction land supply behavior of local government. The data sources of urban construction land supply in the existing literature mainly include the "China Land and Resources Statistical Yearbook" [33], the land survey results sharing application service platform, and the network of "landchina" [69,70]. Among them, the time span of the data in the "China Land and Resources Statistical Yearbook" is relatively long. Therefore, this paper uses the construction land transfer area in the "China Land and Resources Statistical Yearbook" to express the scale of land supply.
(2) Explanatory variable. Environmental decentralization reflects the functional division of central and local governments in environmental management affairs. The higher the Land 2023, 12, 601 6 of 18 degree of environmental decentralization, the greater the authority of local governments in environmental affairs. Referring to the existing literature, the relative personnel number of environmental protection agencies at all levels is used to calculate environmental decentralization [71,72]. The calculation formula for environmental decentralization is as follows.
where ED it is environmental decentralization; LEPP it , POP it , GDP it are the personnel number of environmental protection agencies, population, and gross domestic product of province i in year t; LEPP t , POP t , GDP t are the number of personnel in environmental protection agencies, population, and gross domestic product of whole country in year t; (1 − GDP it /GDP t ) is the economic scaling factor for reducing endogenous interference [73].
(3) Control variables. Under the framework of Chinese decentralization, the impact of fiscal decentralization and political centralization on land supply cannot be ignored. Therefore, three special control variables were selected, including fiscal decentralization, fixed asset investment per unit of GDP, and industrial pollution control completed investment. Fiscal decentralization typically includes revenue decentralization and expenditure decentralization [74]. Fiscal revenue incentives can increase local government construction land supply. Thus, this paper uses fiscal revenue decentralization. Fixed assets investment is used to represent the promotion incentives of political centralization [75], and industrial pollution control completed investment is used to represent the environmental constraints of political centralization [76]. For the control of other social and economic factors, referring to the research of Wang (2015), Zhou (2019), and Li (2021) [32,33,50], six control variables are selected from the five aspects: economic development, population size, urban expansion, industrial structure, and financial pressure. Specific indicators include the per capita GDP and its square, population, land urbanization rate, the proportion of secondary and tertiary industries, and financial pressure.
The contents of the variables are shown in Table 1.

Data Sources
Since the personnel number of environmental protection agencies at provincial level has only been updated to 2015, the panel data of 30 provinces (Hong Kong, Macao, Tai Table 2. The spatiotemporal evolution of urban construction land supply and environmental decentralization is plotted in Figure 2.

Basic Estimation Results
For fear of unrealistic regression results, it is necessary to check that the data is stable before regression. Since this paper uses short panel data, the LLC unit root test is used. The test results indicate that the data used for regression are all stationary. The basic   Figure 2(a1,a2) shows that there is more construction land supply in the east than in the west, and the land supply in the Yangtze River Delta and Pearl River Delta is the largest. From 2003 to 2015, land supply in some provinces in the central and western regions of China increased. Figure 2(b1,b2) shows that the provinces closer to the capital (Beijing) have a higher degree of environmental decentralization. From 2003 to 2015, environmental decentralization in the central region decreased, while it marginally increased in the western region.

Basic Estimation Results
For fear of unrealistic regression results, it is necessary to check that the data is stable before regression. Since this paper uses short panel data, the LLC unit root test is used. The test results indicate that the data used for regression are all stationary. The basic estimation results are shown in Table 3. FE1 is the estimation result of adding no control variables, and FE2 and FE3 are the estimation results of gradually adding special control variables and other control variables. Note: values in parentheses are t statistics; **, and *** represent that the coefficients are significant at the levels of 5%, and 1%, respectively.
The regression results in Table 3 show that after adding all control variables, the coefficient of environmental decentralization is significant, at the level of 5%, which indicates that environmental decentralization increases the scale of construction land supply. The effect of fiscal decentralization is positive, at a 1% significance level, suggesting that the greater financial autonomy of local governments, the stronger the motivation to provide a great deal of construction land. The coefficient of fixed asset investment per unit of GDP is positive, at a 1% significance level, suggesting that the scale of construction land supply is expanding due to the competition of local governments to attract investment. The effect of industrial pollution control investment is negative, at a 5% significance level, indicating that environmental protection constraints limit the expansion of construction land supply. The coefficient of square per GDP is markedly negative, at the level of 1%, indicating that there is an inverted U-shaped Kuznets curve relationship between economic development and construction land supply. The scale of construction land supply first increases and then decreases with economic growth. The proportion of secondary and tertiary industries is significantly negative, at the level of 1%, indicating that industrial structure upgrading can reduce construction land supply. The effect of local government financial pressure on land supply is markedly positive, at the level of 5%, indicating that the greater financial pressure, the more construction land supply. Population size and land urbanization have a positive influence on construction land supply, but they are not significant enough.

Robustness Test
To make the regression results more robust and reliable, the explained and explanatory variables are replaced, and the estimation method is changed. The robustness test results are shown in Table 4. In Table 4, FE4 is the regression result of replacing the explained variable with land transfer income (LTI). FE5 is the regression result of replacing the explanatory variable by environmental decentralization with a lag of one period. In addition, the model may suffer from heteroscedasticity and autocorrelation, and the feasible generalized least squares (FGLS) method is used to mitigate these problems [37,77]. The third column in Table 4 are the regression results with FGLS. The main work of environmental management includes administration, supervision, and monitoring. Therefore, subdivided environmental decentralization can be used for robustness testing. FE6, FE7, and FE8 are the estimated results of environmental administrative decentralization, environmental supervision decentralization, and environmental monitoring decentralization, respectively.   Note: values in parentheses are t statistics; *, **, and *** represent that the coefficients are significant at the levels of 10%, 5%, and 1%, respectively. Table 4 indicates that environmental decentralization has a remarkable positive influence on land transfer income. The impacts of fiscal decentralization and competition for attracting investment on land transfer income are also significantly positive. The impact of environmental decentralization lagging one period on land supply is significant, at a level of 1%, indicating that the current land supply is affected by environmental decentralization in the previous period. The regression results with FGLS are not much different from the basic estimation. From the perspective of subdivided environmental decentralization, the influences of environmental administrative and monitoring decentralization are not significant, while the impact of environmental supervision decentralization is significant, at a level of 5%. The reason for this may be that administrative and monitoring decentralization increases the flexibility and positiveness of environmental management and helps to improve environmental performance. Supervision decentralization weakens the importance of environmental constraints and reduces environmental performance. Robustness test results are consistent with basic estimation results. It can be considered that the impact of environmental decentralization on urban construction land supply is stable.

Endogenetic Test
There are many factors that affect urban construction land supply. Although some variables affecting land supply are controlled in this paper, there may still be some factors that are not considered by the model. In addition, the bidirectional causality relationship between explanatory variables and explained variables may also produce endogeneity. In this paper, the first-order and second-order lags of environmental decentralization are used as instrumental variables, and two-stage least square (2SLS) regression is used to Land 2023, 12, 601 10 of 18 eliminate endogenous effects [78]. The generalized method of moments (GMM) can deal with endogenous issues [79]. In this paper, the system generalized method of moments (sys-GMM) is used to further avoid the endogenous and weak instrumental variables [50]. The endogeneity test results are shown in Table 5.

2SLS
Sys-GMM 1st Step (ED) 2nd Step (  Note: values in parentheses are t statistics; *, **, and *** represent that the coefficients are significant at the levels of 10%, 5%, and 1%, respectively; the p values of Underidentification test, Hansen J statistic, AR (2) and Sargan test are in parentheses; the critical value of 10% significance is shown in the brackets of Weak identification test. Table 5 shows that the impact of environmental decentralization on urban construction land supply is significantly positive, at a 10% level, whether estimated using the 2SLS model or the sys-GMM model. After accounting for endogeneity, the impact of environmental decentralization on urban construction land supply is still significantly positive.

Intermediary Mechanism
Theoretical analysis has shown that land finance dependence and land resources misallocation increase the scale of local government construction land supply. Land financial dependence can be measured by indicators such as the proportion of land transfer income to GDP [80], the proportion of land transfer income to local government fiscal revenue [81,82], and per capita land transfer income [83]. In this paper, the proportion of land transfer revenue in the general public budget revenue of the local government is used to represent the land financial dependence of local government. Land resources misallocation in this paper refers to the unreasonable allocation between different uses of construction land. This is because the government supplies a great deal of industrial land at low prices, resulting in a high proportion of industrial land. Scholars believe that agreement transfers go hand in hand with the supply of industrial land at low prices. Thus, the proportion of land transferred by agreement to the total land transfer area can be used to measure the degree of land resource misallocation [16]. The estimated results of the intermediary effect are shown in Table 6. Table 6 shows that the impact of environmental decentralization on land financial dependence and land resource misallocation is significantly positive, at a level of 1%, which indicates that environmental decentralization exacerbates land financial dependence and land resource misallocation. Comparing the columns (3)-(5) in Table 6, it can be found that the impact of environmental decentralization on land supply is significantly reduced after controlling land financial dependence and land resource misallocation. It shows that the two intermediary variables have a remarkable influence on urban construction land supply [84]. Note: values in parentheses are t statistics; **, and *** represent that the coefficients are significant at the levels of 5%, and 1%, respectively.

Heterogeneity Analysis
After environmental decentralization, urban construction land supply by local governments should be balanced and chosen among fiscal revenue, economic growth, and environmental protection. Local governments with high pressure on fiscal revenue and economic growth may loosen environmental control and transfer a great deal of construction land for obtaining more land transfer income and better political performance. Local governments with high environmental protection pressures are more inclined to protect the environment and will not supply a great deal of construction land. In order to contrast the different influences in the regions with different financial pressure, economic growth pressure, and environmental protection pressure, this paper uses the median of pressure to divide the provinces into two types of regions. Referring to Li (2015), the calculation of financial pressure is attained by the proportion of the disparity between fiscal expenditure and fiscal revenue to GDP [50]. The provinces are divided into high and low financial pressure regions based on the median of the average financial pressure in each province from 2003 to 2015. Providing more land can boost economic growth [85]. The regions with high economic growth pressure have a stronger willingness to supply more construction land. Referring to Yang (2016), regional economic growth pressure is represented by the divergence between the local GDP growth rate of the current year and the last year [86]. Similarly, the regional division standard is the median of the average economic growth pressure. Chemical oxygen demand (COD) and SO 2 are often used to indicate pollution levels [87,88]. COD is mainly produced by industrial enterprises, and is often used as an important indicator to measure environmental pollution. Therefore, COD emissions per unit of GDP are used to measure environmental protection pressure. The regional division of environmental pressure is in accordance with financial pressure and economic growth pressure. The estimated results for regions with distinct financial pressures, economic growth pressure, and environmental protection pressures are shown in Table 7.
From the perspective of financial pressure, the effect of environmental decentralization on urban construction land supply is significantly positive, at a level of 10%, in the regions with high financial pressure, but it is not significant in the regions with low financial pressure. Moreover, the regression coefficient of environmental decentralization in regions with high financial pressure is larger (0.609 > 0.396). The results show that environmental decentralization prompts local governments with high financial pressure to increase construction land supply to ease financial pressure.
From the perspective of economic growth pressure, the effect of environmental decentralization on urban construction land supply is markedly positive, at a level of 10%, in regions with high economic growth pressure. The influence is negative in regions with low economic growth pressure, and it is not significant. The regression coefficient of environmental decentralization in regions with high economic growth pressure is larger Land 2023, 12, 601 12 of 18 (0.845 > −0.007). It shows that the motivation of local governments, which relies on construction land supply to boost economy, distorts the impact of environmental decentralization.  Note: values in parentheses are t statistics; *, **, and *** represent that the coefficients are significant at the levels of 10%, 5%, and 1%, respectively.
From the perspective of environmental protection pressure, the impact of environmental decentralization on urban construction land supply is significant, at a level of 5%, in the regions with low environmental protection pressure, and it is not significant in the regions with high environmental protection pressure. Moreover, the regression coefficient of environmental decentralization in the regions with low environmental protection pressure is larger (0.760 > 0.245), which indicates that environmental decentralization has a greater influence on construction land supply in regions with low environmental protection pressure. The cause for such a phenomenon is that the environmental constraints in regions with low environmental protection pressure are relatively weak. When the degree of environmental decentralization increases, local governments tend to supply more construction land to improve economic and political performance.

Policy Suggestions
The research of this paper provides some policy ideas for local governments to reasonably supply urban construction land against the background of Chinese decentralization.
(1) The environmental management power and responsibilities of the central and local governments should be reasonably divided, and the incentive and restraint mechanisms for construction land supply should be adjusted. The original intention of environmental decentralization is to realize the localization of environmental management and improve the efficiency of environmental management. However, in practice, local governments are encouraged by fiscal revenue and political performance to loosen environmental control and increase construction land supply in the background of environmental decentralization. By decentralizing the environmental administrative power and centralizing the environmental supervision power, local government's target choice and land supply behavior can be efficaciously restrained. In addition, strengthening the central environmental protection constraints can increase the environmental preferences of local governments and encourage them to reasonably supply construction land. Some measures should be used, such as strictly implementing the environmental protection supervision system and improving the importance of environmental protection indicators in performance appraisal.
(2) Differentiated environmental decentralization should be implemented to form differentiated construction land supply incentives. The strong economic and political incentives brought by urban construction land to local governments make environmental decentralization ineffective in regions with high financial pressure, high economic growth pressure, and low environmental protection pressure. The central government ought to strengthen environmental centralization in these three types of regions. While empowering local governments with environmental administrative power, the central government's regulatory powers ought to be enhanced to restrict local governments' land supply behavior that damages the environment. In regions with low financial pressure, low economic growth pressure, and high environmental protection pressure, the central government should increase the degree of environmental decentralization to improve environmental management efficiency.
(3) The land financial dependence of local governments ought to be weakened. Looking for new sources of income and changing financing methods can alleviate the financial pressure of local governments and weaken their land financial dependence. Specific measures can be implemented from two aspects: open source and throttling. Firstly, the fiscal and taxation systems ought to be innovated to increase the tax revenue of local governments, such as levying property tax and inheritance tax, and so on [89]. Secondly, BOT (build-operate-transfer) and other financing modes could be adopted in infrastructure construction to alleviate the financial pressure of the government [90]. Moreover, the land lease can be changed to annual lease, and the land transfer income can be changed from one-time income to continuous income. While ensuring the government's stable income, the urban construction land supply for obtaining short-term economic profits can be controlled.
In addition, urban planning is an important means to control the expansion of construction land scale. The government's environmental policy objectives should be fully considered when formulating land use plans [91]. In the use of land, it is necessary to strengthen the guiding and standardizing roles of planning, and realize the scientific and rational allocation of land for different purposes, such as urban construction land and agricultural land.

Discussion
In recent years, China's environmental problems have gradually come to light. Under the environmental decentralization system, the environmental management power of the central government has gradually been strengthened [92]. This shows that the central government is trying to find a balance of environmental decentralization to coordinate the economic growth goals and the environmental protection goals of local government. We integrated environmental decentralization into the Chinese decentralization framework. Then, under the framework of Chinese decentralization, the incentives and constraints of environmental decentralization, fiscal decentralization, and political centralization on the supply of construction land by local governments were studied.
Our research enriches the Chinese decentralization framework. Previous studies have only emphasized the incentives of fiscal decentralization and political centralization on local government land supply [64,85,93]. Tang (2019) focused on the impact of environmental politically binding indicators on local governments' land violations [36]. However, there has been no in-depth analysis of the impact of environmental management systems on local government land supply. More importantly, the expansion of urban construction land against the background of resource constraints and environmental protection has been more strictly regulated [94]. Clarifying the incentives and constraints of government land supply is conducive to achieving the rational allocation of urban land.
Our research results suggest that environmental decentralization reduces the performance of land environmental management in China. This differs from Laskowski (2005) and Blundell (2021), in that environmental decentralization improves performance by improving environmental policy adaptability [44,45]. As Ulph (1998) argued, information asymmetry between local and central governments in the context of Chinese decentralization distorts the performance of environmental decentralization [46]. Among the three environmental management affairs of administration, supervision, and monitoring, supervision decentralization has the most obvious promotion of the expansion of construction land. Our research also shows that the impact of environmental decentralization on construction land supply regionally varies. Where economic, political, and environmental protection constraints differ, the performance of environmental decentralization is also different. Fredriksson's study of 110 countries also concluded that there were regional differences in environmental decentralization performance [95].
Although our study is based on the specific situation in China, it is also meaningful to analyze government actions in the supply of urban construction land in other countries. The impact of local government incentives and constraints on urban construction land supply also exists in Western countries. Götze (2021) found that municipalities in Germany and the Netherlands also provide a loose supply of urban land due to financial incentives [96]. Perrin (2018) and Kurnia (2021) studied the impact of decentralization on the expansion of urban construction land in France, Italy, and Indonesia [17,18].

Conclusions
The economical and intensive use of land resources is an important measure to push ecological civilization construction. Practice in western countries has proved that decentralization can reduce information asymmetry and principal-agent risks, and improve policy efficiency. However, decentralization may also allow local governments to loosen environmental regulations to attract investment, which could boost economic development. Under the Chinese decentralization framework, fiscal decentralization and political centralization provide incentives for local governments to supply a great deal of urban construction land to make up for fiscal gaps and support economic growth. Environmental decentralization increases the possibility of local governments to loosen environmental regulation, which promotes the expansion of construction land supply. The efficiency of environmental decentralization is further reduced. Based on the theoretical analysis of the effect of environmental decentralization on construction land supply, the interprovincial panel data from 2003 to 2015 were used for analysis. The main conclusions are as follows: (1) Environmental decentralization promotes the expansion of urban construction land supply scale. Environmental decentralization intensifies the positive incentives of economic and political incentives to the supply of construction land, and weakens environmental constraints. Due to the important role of construction land in local economic development, local governments usually choose to relax environmental control and supply a great deal of construction land.
(2) Environmental decentralization promotes the expansion of urban construction land supply by strengthening land financial dependence and distorting land resources misallocation. Environmental decentralization increases the possibility of "land rent", "land tax", and "attract investment by land", while weakening environmental constraints. This raises land financial dependence and the misallocation of land resources, which accelerates the swell of construction land supply scale.
(3) Environmental decentralization has diverse incentives and constraints in regions with distinct financial pressure, economic growth pressure, and environmental protection pressure, and has different impacts on construction land supply. In regions with high financial and economic growth pressure, local governments, driven by economic and political incentives, choose to relax environmental constraints and supply a great deal of construction land. In regions with low environmental protection pressure, local governments have lower environmental constraints. The impact of environmental decentralization on urban construction land supply is greater than that in regions with high environmental protection pressure.

Limitations and Future Research
There are some deficiencies in this study. First, this paper only pays attention to the effect of environmental decentralization on the urban construction land supply scale of local government. However, the construction land supply structure of different uses and industries is also affected by the preferences and behaviors of local government, which is a content worthy of study. Second, due to data limitations, this paper only focuses on the impact of decentralization and centralization between the central government and provincial governments on urban construction land supply. However, Chinese governments at or above the county level are empowered to supply land. The impact of decentralization and centralization between the upper and lower governments below the provincial level on land supply is also worth studying. Third, this paper only studies the impact of environmental decentralization on construction land supply in the context of China's specific system. However, there may be differences in the performance of environmental decentralization in different institutional contexts. In the future, comparative analysis with other countries could also be carried out.