Research on the Impact and Spillover Effect of Green Agricultural Reform Policy Pilot on Governmental Environmental Protection Behaviors Based on Quasi-Natural Experiments of China’s Two Provinces from 2012 to 2020

Abstract

The green agricultural reform policy pilot embraces the development trend of the times. The green agricultural policy pilot implemented in China’s Zhejiang Province is an attempt to find a balance between environmental protection and economic development in order to achieve the national “dual carbon” goals. Have the goals been achieved? Zhejiang Province is the national pilot zone of green agricultural development. Jiangsu Province is notably homogeneous with Zhejiang Province but has not been included in the pilot policy. Based on the comparative studies of the two provinces with quasi-natural experiment logic, this paper analyzes the mechanism of the national green agricultural pilot zone policy acting on the environmental protection behavior of local governments and the actual effect of such policy using the data of 22 cities in both provinces from 2012 and 2020 as the panel data and relying on the DID model. According to the research findings, the national green agricultural development pilot zone policy has significantly increased the investment of local governments in environmental protection in Zhejiang Province where the policy has been fully implemented, and the negative spillover effect has reduced the economic governance efficiency of local governments in the short term. However, the investment of local governments in Jiangsu Province where the policy has not been fully implemented has not increased significantly and the economic governance efficiency has not reduced significantly. Double robustness tests based on the parallel trend test and DID-PSM (Differences-In-Differences and Propensity Score Matching) have verified the reliability of the research findings. Both Zhejiang and Jiangsu are provinces with developed industrial economies. Agriculture only accounts for a small proportion of their GDP. In addition, there is a natural conflict between green agriculture and industrial manufacturing in these two provinces. However, for those provinces dominated by agriculture or tourism, the effect of such green agricultural development policy may be different, which requires follow-up deeper research in an effort to thoroughly learn about the impact of agricultural policy pilot on local environmental protection behaviors, especially economic performance.

1. Introduction

In the 21st century, we are confronted with more extreme and complex climate and environmental changes, which will cause a significant effect to the life and production of human beings. The exploration of a sustainable agricultural development model under the current climate and environmental changes with the initiative of the United Nations Sustainable Development Goals is a positive response of human beings to the relationship between man and nature. Environmental protection follows the development trends of the times, echoes with people’s increasing pursuit of a better ecological environment and directly affects people’s life quality. As the principal of social governance, the government shoulders the responsibility of environmental protection and serves society as the facilitator and the implementation body of green agricultural development. High local environmental quality is a prerequisite of green agricultural development. In order to upgrade agriculture into a green development model, relevant policies are promulgated and implemented to guide local governments, farmers and related stakeholders to build up the consideration of green agricultural development, and finally, realize sustainable development. The national green agricultural pilot zone policy falls into the category of environmental regulations. China unveiled its “No.1 central document” for 2022, giving top priority to the issues relating to “agriculture, rural areas and farmers” for 19 consecutive years. It also called for “promoting the green development of agriculture and rural areas, building the national green agricultural development pilot zone and implementing green agricultural development evaluation”.

Boosting the green transition of agriculture is the fundamental requirement for promoting the high-quality development of agriculture. Steady implementation of the national green agricultural development zone and agricultural modernization demonstration zone is an important task for building China into a beautiful country. Both zones are the regional comprehensive environmental regulation policies promulgated by the government to boost rural development by focusing on the relevant industries. They are of great significance for promoting the transition from extensive agriculture to green agriculture. Relevant academic research focuses on the following two aspects:

First, conducting research on green agricultural development. Green agricultural development has steadily become the goal and trend of China’s agricultural development. Many scholars have conducted research on the concept, path and countermeasures of green agricultural development. For example, according to the research findings of Wei, Q. et al. [1], green agricultural development refers to the high-quality and high-efficiency green development model of resource conservation, environmental friendliness and ecological conservation in the environment of agricultural producing areas, production processes and full life cycle of agricultural products. Yin, C.B. et al. [2] believed that the promotion of green agricultural production should adapt to actual local conditions and implement targeted policies. Xu, X. and Song, W. [3] proposed to strengthen the quality safety supervision of agricultural products and promote the green transition of agriculture. In the opinion of Yu, F.W. et al. [4], the environmental quality of agricultural production is the first guarantee for green agricultural development. According to the research findings of Ren, X.G. et al. [5], the promulgation of policies on green agricultural development can boost the sustainable development ability and ecological environmental governance. As implied by Guo. H. et al. [6], digital inclusive finance can effectively promote green agricultural development. However, agricultural insurance has an inhibitory effect on green agricultural development [7]. Additionally, Zhang, F. et al. [8] analyzed the spillover effect of agricultural science and technology innovations in green agricultural development using a spatial econometric model and proved that the level of green agricultural development has distinct spatial characteristics. The eastern and northwestern regions showed a fluctuating downward trend and the southwestern region showed a fluctuating upward trend. Moreover, the green agricultural development of southwestern coastal regions is faster than that of other regions [9].

Second, conducting research on the environmental protection behavior of the government. From the relationship of environmental protection behavior of government with citizens and enterprises, scholars analyze the game of government in economic development and environmental protection and discuss the methods, effects and influencing factors of environmental protection behavior of local governments. For example, Mao, Y.M. [10] pointed out that the environmental protection inspection system can significantly improve urban air quality in the short term, and the factors such as city location, air pollution level and government scale will significantly adjust the long-term effects of environmental protection inspection policy. According to the research of Qi, Y. and Cai, Q. [11], local governments should implement systematic and targeted governance innovations in order to achieve environmental protection goals. In recent years, popular research on the relationship between environmental protection behavior of government and citizens has emerged in the academic circle. In the opinion of Wang, F. [12], the public will not consciously participate in environmental protection and enterprises will not stop pollution if the government does not involve itself in environmental protection. On the contrary, government rewards and punishments are directly related to the number and cost of citizens participating in environmental protection, thus affecting the selection of corporate environmental protection behavior. As implied by Men, D. and Xiong, R.C. [13], the government investment in ecological environmental protection and public ecological civilization education can affect the public ecological behaviors indirectly by influencing the public sense of environment and environmental protection. Shi, S.X. and Gan, C.Y. [14] pointed out that the public satisfaction with environmental protection of local governments is positively correlated with environmental protection behavior. Some scholars have discussed the impact brought by the environmental protection behavior of government on corporate microeconomics. For example, Yu, Z.M. [15] proved that the implementation of regulatory talk on environmental protection not only produces an impact on the environmental governance behavior of local governments but also leads to some innovative corporate environmental strategies. According to Liu, L. et al. [16], a “U-shaped” relationship exists between environmental policy and corporate environmental investment in 2022. Incorporating official achievement assessment into environmental indicators helps strengthen the emphasis of local governments on corporate green innovation [17]. After the implementation of environmental tax, the corporate environmental investment has increased significantly and environmental economic policies and environmental administrative policies have effectively driven listed companies to increase their total environmental investment [18]. Some scholars have considered the duration and spatial spillover of China’s industrial pollution emissions according to the environmental governance effect with local fiscal expenditures. The research findings prove that local environmental protection expenditure can effectively reduce urban industrial pollution emissions [19]. Based on provincial panel data, Zheng, J. et al. [20] pointed out that the fiscal decentralization of governments in different economic development stages causes different impacts to environmental governance. Hu, C. et al. [21] proved the negative relationship between agricultural policy and agricultural carbon emission. Driven by agricultural policies, environmental protection will weaken the positive relationship between agricultural mechanization and agricultural carbon emissions. According to the findings of Guan, H. and Liu K. [22], low-carbon city policy improves the urban green productivity and reduces the urban carbon emission intensity. In light of the relevant research on policy implementation and governmental environmental protection behavior, the academic circle is shifting focus to the potential relationship between green agricultural development goals and governmental environmental protection behaviors. The implementation of green agricultural reform policy is one of the possible potential factors promoting governmental environmental protection behaviors.

In summary, government investment in environmental protection and green agricultural development are gaining an increase in popularity among various circles of society. The academic circle starts research on the environmental protection behavior of government and the concept, principles, problems, path and countermeasures of green agricultural development. However, the current research on the relationship between policy reform and governmental environmental governance behavior is dominated by theoretical analysis and superficial quantitative review. Moreover, the current research focuses on such pilot policies as innovative cities and low-carbon cities and seldom discusses the effect of green agricultural development pilot policy. China has attached great importance to the green agricultural development and promulgated relevant policy documents in recent years. “The Plan for the Construction of National Sustainable Agricultural Development Pilot and Demonstration Zones”, was issued in August 2016, initiating the first round of the green agriculture reform pilot. Under such circumstances, how has the green reform pilot policy evolved by 2022? Does the green agriculture reform pilot affect the environmental protection behavior of local governments? What is the environmental protection effect of this pilot policy since it was implemented seven years ago? Has the pilot policy effectively promoted the environmental protection behavior of local governments in Zhejiang as one of the first pilot provinces? Will negative spillover effects be caused by the economic governance performance? Jiangsu Province, which is highly homogeneous with Zhejiang in politics, economy and culture, was not included in the first batch of pilot provinces. Are there significant differences between Jiangsu as a non-pilot province and Zhejiang as a pilot province? All the above questions have not been answered and studied in the current research. With reference to these questions, this paper intends to conduct research on the national green agricultural development pilot zone policy and environmental protection behavior of local governments from the macro perspective. The possible marginal contributions include the following aspects: (1) In light of the research perspective, this paper explores the impact of the national green agricultural development pilot zone policy on the environmental protection behavior of government with the pilot policy as the starting point. (2) As for research methods, this paper extends the heterogeneous policy model to enrich the relevant research on the policy effect evaluation.

The Differences-in-Differences model (hereinafter referred to as the DID model) is a method based on natural experiments to evaluate pilot policies through panel data regression. The DID model is applicable to the evaluation of the effect of the green agricultural development pilot zone policy. In comparison with traditional regression analysis methods, the DID model can effectively eliminate the impact of time effects and test the robustness of results through parallel trend testing to avoid endogenous problems. Based on the quasi-natural experiment logic, this paper analyzes the mechanism of the national green agricultural pilot zone policy acting on the environmental protection behavior of local governments, and the actual effect of such policy, using the data of 22 cities in two provinces from 2012 and 2020 as the panel data and relying on the DID model. Zhejiang Province is the national pilot zone of green agricultural development. Jiangsu Province is notably homogeneous with Zhejiang Province but has not been fully included in the pilot policy. The research has eliminated the uncertainty between green agricultural development regulation and governmental environmental protection behavior. Guided by green agricultural development, the government is not sure about its choice between economic development and environmental protection. Furthermore, the government is not sure about the consequence of its responses to the green agricultural development goals. The national green agricultural development pilot zone is a pathfinder for green agricultural development. This paper has reviewed the relationship and spillover effect between the pilot policy and local governmental environmental protection behavior and analyzed the feasibility, stability and effectiveness of the pilot policy, hoping to provide empirical and theoretical support for future research and government decisions.

2. Evolution of the National Green Agricultural Development Pilot Zone Policy

Green agricultural development is a key component in implementing the rural revitalization strategy, while the national green agricultural development pilot zone is a comprehensive pilot and demonstration platform to promote green agricultural development. A plan was issued in August 2016, initiating the green agriculture reform pilot program. A notice was issued in June 2017, initiating the evaluation and validation of the first batch of pilot zones. In September 2017, the General Office of the State Council issued a notice, calling for promoting the setting of the national sustainable agricultural development pilot and demonstration zones and also building them into the pilot zones of green agricultural development.

“The Notice on Initiating the Construction of the First Batch of National Sustainable Agricultural Development Pilot and Demonstration Zones and Implementing the Pilot Program of Green Agricultural Development” issued in 2017 confirmed 40 green agricultural development pilot zones in the first batch. Zhejiang became the first province fully covered by the pilot policy. The other 39 pilot zones are cities, county-level cities or districts. “The Notice on Publishing the List of the Second Batch of National Green Agricultural Development Pilot Zones” issued in 2019 has included Hainan into the pilot program, making it become the only province fully covered by the pilot program in the second batch. All the other 49 pilot zones are still cities, county-level cities or districts. According to the “Notice on Publishing the List of the Third Batch of National Green Agricultural Development Pilot Zones” issued in 2022, all 49 pilot zones were cities, county-level cities or districts without any provincial pilot zone. A total of two provincial pilot zones and 128 urban pilot zones have been established in three batches as of 2022.

By analyzing and summarizing the evolution of this policy, we can find that the national green agricultural development pilot zones have the following two characteristics: (1) In the spatial layout, they integrate points with areas for overall promotion. The green agricultural development pilot zones involve provincial zones and zones of cities, county-level cities and districts. At first, a batch of cities was confirmed as pilot zones to work out experience and disseminate it nationwide. Meanwhile, the promotion combines points with areas to reduce the cycle of the “pilot program”. (2) Local agriculture is endowed with advantageous resources and regional characteristics. Therefore, these zones play a role in radiating and leading the green agricultural development of other regions. See

Table 1. Evolution of policies in the National Agricultural Green Development Pilot Zone.

BATCH	First	Second	Third
Year	2017	2019	2022
Policy paper	Notice on launching the first batch of national pilot demonstration zones for sustainable agricultural development and carrying out pilot work on green agricultural development	Notice on the release of the second batch of national agricultural green development pilot zone list	Notice on the release of the third batch of national agricultural green development pilot zone list
Policy Number	Ministry of Agriculture and Rural Affairs, China [2017] No. 121	Ministry of Agriculture and Rural Affairs, China [2019] No. 23	Ministry of Agriculture and Rural Affairs, China [2022] No. 22
Provincial pilot	Zhejiang Province	Hainan Province	None
Other pilot cities, county-level cities and districts	Thirty-nine	Forty	Forty-nine

for details.

3. Materials and Methods

3.1. Theoretical Basis

Starting with the experimental theory of science of public policy and the trade-off theory of economics, this research explores and analyzes the intrinsic relationship between the pilot policy and the efficiency of local governments to help the policy bring more benefits and provide effective feedback for the transformation of China’s green agricultural development.

In light of the experimental theory in the science of public policy, the formulation of a public policy is not a game and faces a high uncertainty due to government decisions. Therefore, prudent social experiments or small-scale trials should be launched before a policy is formulated and promulgated so as to reduce ignorance in the decision-making process and defects in the implementation process. Xi, J. P., General Secretary of CPC Central Committee, pointed out that pilot programs are launched to explore the path and pattern of reform, and then provide replicable and pervasive experience and practices for large-scale reforms. The trade-off theory in economics usually assumes that current resources are limited and fixed and local governments must abort other policies or cancel other goals while adopting a certain policy or trying to achieve a certain goal.

Based on actual situations, we can easily find that environmental protection must be accompanied with the loss of economic benefits at the current level of science, technology and productivity and a contradiction exists between local economic development and environmental protection. However, high environmental quality is the premise of green agricultural development. The national green agricultural development pilot zone is a policy experiment launched by China to explore the path of transition to green agricultural development. The policy experiment aims to seek experience in the process of sustainable and high-quality agricultural development, helping the country successfully achieve the goal of agricultural and rural modernization in the new era. Do local governments need to make a trade-off between economic development and environmental protection while promoting the transition to green agriculture? In the process of trade-off, how do local governments seek a balance between economic development and environmental protection? This research intends to answer the above questions.

3.2. Research Hypothesis

3.2.1. Effect of the Policy in Driving Local Governments to Implement Environmental Protections

Hypothesis 1 (H1):

The national green agricultural development pilot zone policy has increased the investment of local governments in environmental protection.

Livelihood calls for a larger amount of investment in environmental governance, which is also an important way for local governments to improve environmental quality. Any candidate for the pilot policy must perform better in the prevention and control of non-point agricultural pollution. Agricultural and rural pollution prevention and control are a weak link in agricultural environmental governance. It is necessary for local governments to allocate a large amount of investment in order to modernize the ecological environment governance system and governance ability, ensure the smooth pollution prevention and control of environmental infrastructure construction, implement key environmental infrastructure projects and build ecological infrastructures [23]. Meanwhile, it needs to improve the green financial policy system and help agricultural enterprises out of financing difficulties and high-cost financing [24]. Moreover, the implementation of land transfer [25], farmland protection [26] and other related agricultural policies requires human resources, financial resources and material resources. What does the national green agricultural development pilot zone mean to local governments? Are local governments willing to allocate limited resources to agriculture with a long return cycle? Will the setting of such a pilot zone have an impact on the proportion of local government revenue and fiscal expenditure? Based on the actual situation in the “13th Five-Year Plan” period, the expenditure of the central budget on ecological and environmental protection was CNY 349.9 billion in 2016, CNY 356.3 billion in 2017, CNY 390.5 billion in 2018 and CNY 411.7 billion in 2019 (the data are the final budget), showing a state of continuous growth. Based on the empirical research of China’s provincial panel, Jin, D.C. et al. [27] proved that the increase of government investment in environmental protection can significantly reduce the level of local pollution emissions and plays a notably effective role in regional environmental governance.

3.2.2. Effect of the Green Agricultural Development Pilot Zone in Negatively Driving the Economic Development of Local Governments

Hypothesis 2 (H2):

The pilot policy has increased the investment of local governments in environmental protection in the short term. However, it causes a significant negative spillover effect because it has reduced the economic governance efficiency of local governments.

Environmental protection is negatively correlated with economic development. Policy regulation will increase the corporate production and operation costs in the short term, thus inhibiting the production capacity of enterprises, which is not conducive to the rapid economic growth of enterprises in the short term [28,29]. Protecting the ecological environment needs to slow down industrial development. However, taking into account environmental protection will increase the production cost in the process of industrial development, thereby narrowing the profit space and causing the contradiction between profit and environmental protection. The resources owned by the government are limited. Agriculture is the foundation of the national economy, and industry is also indispensable for economic development. If the government increases agricultural investment, it will inevitably slow down the economic development to a certain extent. According to the research findings of Liu, L. et al. [30], the degree of water and soil loss control will rise by 0.18% to 0.40% for every 1% increase of the agricultural sector in the local GDP, while the degree of water and soil loss control will decline by 0.18% to 0.37% for every 1% increase of the industrial sector in the local GDP. As implied by Yang, Y.D. et al. [31], environmental regulation will result in not only increased investment of industrial enterprises in environmental protection, but also a heavier financial burden. Local governments will provide enterprises with excessive environmental incentives, thereby reducing corporate earnings. Therefore, the government should make a trade-off between economic development goals and ecological environmental governance. High environmental quality, which is known as the prerequisite for the green agricultural transition and upgrade, may force the government to give priority to environmental governance and improve the local environmental quality through policy support and investment inclination. However, the limited resources held by the government will slow down the local economic development, thereby reducing the economic governance efficiency of local governments. It is noticeable that only short-term results are involved in the research because the national green agricultural development pilot zone policy was implemented not long ago. The government promotes agricultural transformation and upgrading through the allocation of social resources. However, the contribution of agriculture to economic growth has a lag characteristic, which is reflected in the fact that agricultural transformation and upgrading needs a long period of time to feed the improvement of local environmental quality and economic development. In addition, agricultural development and economic growth should be mutually reinforcing. In the long term, the green agricultural transition and upgrade will help improve the local environmental quality and substantially promote the local economic development and people’s living standards, thus achieving the sustainable development goal of harmonious coexistence between man and nature.

3.3. Sampling

The balanced panel data of Zhejiang and Jiangsu from 2012 to 2020 were involved in the research. The two reasons for selecting these two provinces to conduct quasi-natural experiments are as follows.

First, both the provinces are located in eastern China and are homogeneous and comparable in terms of natural resources, economic development level, history, culture, local conditions and customs. Geographically, the two provinces are located in the Yangtze River Delta region and border each other. From the perspective of natural resources, both provinces are famous as “the land of milk and honey” and belong to the Yangtze River Delta urban cluster. In terms of economic development, Zhejiang has almost the same amount of per capita GDP and public budget expenditure with Jiangsu. Taking 2021 as an example, the GDPs of Zhejiang and Jiangsu were CNY 7351.6 billion and CNY 11,636.4 billion, respectively. At the end of 2021, Zhejiang had a resident population of 654 million and Jiangsu had a resident population of 840.54 million. The public budget expenditure was CNY 1101.7 billion for Zhejiang and CNY 1458.6 billion for Jiangsu (http://tjj.zj.gov.cn/art/2022/2/24/art_1229129205_4883213.html (accessed on 24 February 2022); http://tj.jiangsu.gov.cn/art/2022/3/3/art_85275_10487745.html (accessed on 3 March 2022)). Furthermore, the two provinces have similar history and culture, and both of them are the birthplace of ancient Chinese culture.

Second, the whole province of Zhejiang was included in the pilot policy in 2017 while only two cities of Jiangsu Province were listed as pilot cities. Xuzhou City and Taizhou City of Jiangsu Province were included in the pilot policy as the first batch of pilot cities in 2017. Therefore, the data of these two cities were excluded from the research process and would not be involved in follow-up tests. In summary, the statistics cover 22 prefecture-level cities, including all prefecture-level cities of Zhejiang Province and all prefecture-level cities of Jiangsu Province (excluding Xuzhou City and Taizhou City). The experimental group consists of 11 prefecture-level cities of Zhejiang while the control group consists of 11 prefecture-level cities of Jiangsu. The data sources include China Economic and Social Big Data Research Platform, China Carbon Emission Accounts and Datasets and statistical yearbooks of all related prefecture-level cities, in which some missing values are calculated and supplemented by linear interpolation. The descriptive statistical results of these variables are shown in

Table 2. Descriptive statistics of variables.

Type of Variable	Variable	Sample	Mean	Variance	Minimum	Maximum
Explained variable	$Policycity$	198	7.874	4.800	2.800	13.110
	$LnPerEP$	198	5.554	0.373	3.680	6.940
Explanatory variable	$Policycity$	198	0.500	0.251	0.000	1.000
Control variables	$LnPergdp$	198	11.323	0.194	8.670	12.100
	$LnAgri$	198	14.892	3.498	13.540	23.450
	$LnPerWW$	198	2.984	0.392	1.360	4.240
	$LnPerS{O}_2$	198	1.482	0.996	−1.100	3.110
	$LnPerC{O}_2$	198	2.069	0.364	0.430	3.230

.

3.4. Model Description and Variable Selection

3.4.1. Description of the DID Model

In order to study the impact of the pilot program on the governance efficiency of local governments, we break down the governance efficiency of local governments into two indicators, including economic efficiency and environmental protection behavior. In order to verify the above hypothesis, the variation of economic governance efficiency and environmental protection behavior of government is estimated using the two-way fixed effects model based on a DID panel with dummy variables. The model is presented as follows:

$$Y_{it}={\beta }_0+{\beta }_1{\mathit{DID}}_{it}+{\mathit{\beta X}}_{it}+{\mu }_{city}+{\gamma }_{year}+{\epsilon }_{it}$$

(1)

Subscripts $i$ and $t$ in Equation (1) represent the $i$th region and the $t$th year, respectively. Where $Y_{it}$ is the governance efficiency of government as an explained variable, $X_{it}$ is the control variable group, ${\mu }_{city}$ represents individual fixed effect, ${\gamma }_{year}$ represents time fixed effect and ${\epsilon }_{it}$ is the error term. The core explanatory variable estimation parameter β₁ of the DID model measures the impact of the pilot policy on the governance efficiency of local governments. When the pilot policy does affect the governance efficiency of local governments, the value of coefficient β₁ should be significantly positive if it improves the efficiency of government governance and the value should be significantly negative if it reduces the efficiency of government governance.

3.4.2. Variable Selection

(1) Explained variable: Economic governance efficiency and government environmental behavior. As for the economic governance efficiency ($Goveffcy$), we refer to the research of Zheng, K.F. et al. [32], in which it is measured by the ratio of actual GDP to regional fiscal expenditure. By combining the research of Zhang, X.M. et al. [33] with Geng, C.X. et al. [34], the environmental protection investment indicator is thus obtained. After considering the availability of data, the government fiscal expenditure on energy conservation and environmental protection is cited as the core variable. In order to effectively control the differences brought about by population factors in various regions, the follow-up research takes the natural logarithm ($LnPerEP$) of fiscal expenditure on energy conservation and environmental protection per capita as the government investment in environmental protection, and the two variables serve as the explained variable together.

(2) Explanatory variable: The national green agricultural development pilot zone policy. The core explanatory variable is the pilot dummy variable ($Policycity$). The pilot program was implemented in two batches (the first batch in 2017 and the second batch in 2019) and has just started in the second batch pilot cities. Therefore, only the first batch of cities was selected for analysis in the research process. Considering the heterogeneous influence caused by different levels of urban development, the prefecture-level cities of Zhejiang and Jiangsu with similar levels of urban development were selected as the experimental group and the control group for analysis. The setting of core explanatory variable depends on whether the city is listed among the first batch of pilot cities. The core explanatory variable of $Policycity$ consists of two dummy variables. In detail, the policy pilot is the first dummy variable, pilot cities constitute the experimental group and are defined as 1. Non-pilot cities constitute the control group and are defined as 0. Policy time is the second dummy variable. Only the pilot cities in the first batch were selected for analysis. Therefore, the period of 2017 and later is defined as 1 and the period before 2017 is defined as 0.

(3) Control variables. We select the following control variables: (1) Level of economic development: The level of economic development is closely related to the local agricultural development foundation [35] and governmental environmental protection behavior [36]. Most of the regions with developed economies often have a high-quality agricultural development foundation and local governments attach great importance to environmental protection. In order to control the impact caused by the level of economic development, measurements are performed with the natural logarithm of ($LnPergdp$) per capita GDP of the actual resident population. (2) Level of agricultural development: In research with the agricultural pilot policy as the core explanatory variable, it is necessary to control the level of local agricultural development so as to better estimate the effect of the agricultural pilot policy on governmental environmental protection behaviors with the natural logarithm ($LnAgri$) of added value of the primary industry. (3) Level of industrial development: The level of industrial development plays an important role in local economic development, and the energy consumption caused by industrial development will also increase the pressure of local environmental protections. In order to accurately control the impact of industrial development on the governmental environmental protection behavior, two variables, including industrial wastewater discharge and sulfur dioxide discharge, are introduced to the control; wherein, the industrial wastewater discharge per capita is measured with the natural logarithm ($LnPerWW$) of industrial wastewater discharge per capita of resident population, while the sulfur dioxide discharge is measured with the natural logarithm ($LnPerS{O}_2$) of sulfur dioxide discharge per capita of resident population. (4) Objective environmental quality: The deterioration of local environmental quality will force the government to increase the investment in environmental protection. In order to accurately estimate the impact of agricultural pilot policy on the governmental investment in environmental protection, it needs to control for the local objective environmental quality. With reference to the research of Leng, C.Y. and Liu, J.P. [37], the objective environmental quality is measured using the natural logarithm ($LnPerC{O}_2$) of carbon dioxide emissions per capita of resident population.

4. Results

4.1. Benchmark Model Estimation

Model 1 and model 3 are used for the estimation without considering the impact of the pilot policy as the control variable on the economic governance efficiency and environmental protection investment of local governments. Model 2 and model 4 are used for the estimation after involving the relevant control variables, while model 5 is used for the estimation after including the government investment in environmental protection into the benchmark model of economic governance efficiency. See

Table 3. Baseline model estimation results.

Variable	Model 1	Model 2	Model 3	Model 4	Model 5
	$\mathit{G}\mathit{o}\mathit{v}\mathit{e}\mathit{f}\mathit{f}\mathit{c}\mathit{y}$	$\mathit{G}\mathit{o}\mathit{v}\mathit{e}\mathit{f}\mathit{f}\mathit{c}\mathit{y}$	$\mathit{G}\mathit{o}\mathit{v}\mathit{e}\mathit{f}\mathit{f}\mathit{c}\mathit{y}$	$\mathit{L}\mathit{n}\mathit{P}\mathit{e}\mathit{r}\mathit{E}\mathit{P}$	$\mathit{G}\mathit{o}\mathit{v}\mathit{e}\mathit{f}\mathit{f}\mathit{c}\mathit{y}$
$Policycity$	−1.460 *** (0.182)	−1.264 *** (0.174)	0.326 *** (0.061)	0.319 *** (0.064)	−0.906 *** (0.171)
$LnPerEP$					−1.121 *** (0.197)
$LnPergdp$		0.548 ** (0.262)		−0.004 (0.096)	0.543 ** (0.240)
$LnAgri$		2.857 *** (0.642)		0.378 (0.234)	3.281 *** (0.593)
$LnPerWW$		0.012 (0.293)		0.231 ** (0.107)	0.271 (0.272)
$LnPerS{O}_2$		0.148 (0.156)		−0.017 (0.057)	0.129 (0.143)
$LnPerC{O}_2$		0.504 (0.384)		−0.188 (0.140)	0.293 (0.354)
City Effect	YES	YES	YES	YES	YES
Year Effect	YES	YES	YES	YES	YES
$Cons$	8.198 *** (0.061)	−41.900 *** (9.654)	5.471 *** (0.020)	−0.383 (3.523)	−42.330 *** (8.838)
N	198	198	198	198	198
R-squared	0.929	0.942	0.896	0.901	0.952

Note: *, ** and *** are significant at the level of 10%, 5% and 1%, respectively. The values in brackets are standard error.

for details.

Before including the control variables, the regression coefficient of government economic governance efficiency in model 1 is −1.460 and passes the significance test at the level of 1%, while the regression coefficient of government environmental protection investment in model 3 is 0.326 and passes the significance test at the level of 1%. After including the control variables, the regression coefficient of government economic governance efficiency in model 2 is −1.264 and passes the significance test at the level of 1% and the regression coefficient of government environmental protection investment in model 4 is 0.319 and passes the significance test at the level of 1%. In comparison with the result before including the control variables, the result after including the control variables is more plausible and accurate.

In summary, the regression coefficient of the core explanatory variable $Policycity$ to government environmental protection investment is significantly positive regardless of the inclusion of control variables. After a city is approved as a pilot city, its government will place more emphasis on environmental protection, thus making greater efforts in environmental protection. In this view, the hypothesis H₁ (the national green agricultural development pilot zone program has increased the investment of local governments in environmental protection) is proven. In addition, the regression coefficient of the core explanatory variable $Policycity$ to the economic governance efficiency of government is significantly positive regardless of the inclusion of control variables. In other words, the pilot program has reduced the economic governance efficiency of local governments.

The coefficient of $LnPerC{O}_2$ is not significant in all models, which may be correlated with the proportion of agricultural carbon emissions to total carbon emissions. According to the Annual Report 2018 of China Carbon Emissions, the total agricultural carbon emissions are 801.61 million tons of CO2e (carbon dioxide equivalent), only accounting for 6.85% of China’s total carbon emissions. Among all sources of agricultural carbon emissions, the carbon emissions from animal husbandry production processes, such as intestinal fermentation of cattle and sheep, livestock feces control, residual manure of pasture and livestock manure return to cropland, account for 41.8% of the total agricultural carbon emissions [38]. Jiangsu and Zhejiang are located to the east of China’s 400 mm isohyet. The agricultural development there is dominated by planting, forestry and fishery. Carbon emissions come from different social production sectors. All carbon emissions from energy and industrial sectors account for about 80% of the total carbon emissions. For this reason, the coefficient of $LnPerC{O}_2$ is not significant in the regression model with the policy pilot as the core independent variable. It is also noticeable that the coefficient of $Policycity$ is significantly negative in the regression model of economic governance efficiency of government, but significantly positive in the regression model of government investment in environmental protection. This fact indicates the trade-off of local governments between economic governance efficiency and environmental protection investment. Has the setting of such pilot zones reduced the economic governance efficiency of local governments while increasing the investment in the environmental protection of local governments?

In order to verify the conjecture, we have referred to the research of Ding, H. F et al. [39] and included $LnPerEP$ into the benchmark regression based on model 2 by keeping the control variables and the effects of regions and years unchanged. Then the intermediary test model 5 on the environmental protection investment and economic governance efficiency of local governments was established. See Table 3 for the results. After including the government environmental protection investment indicator $LnPerEP$ into the benchmark regression as a control variable, the regression coefficient of $LnPerEP$ is −1.121 and passes the significance test at the level of 1%. This fact indicates that such pilot zones have reduced the economic governance efficiency of local governments while increasing the environmental protection investment, thereby proving the hypothesis H₂. Therefore, local governments should still consider how to accurately discover the balance between economic development and environmental protection.

In order to show the impact of the pilot policy on the governance efficiency of local governments more intuitively, the line chart concerning the variation of government efficiency between the pilot province (Zhejiang) and the non-pilot province (Jiangsu) is obtained using a three-dimensional line chart. See Figure 1 for details. “Economic efficiency” shows that the economic governance efficiency of all prefecture-level governments in Zhejiang Province has decreased by 1.26 units after the implementation of the pilot policy in comparison with non-pilot cities. “Environmental protection expenditure” shows that the environmental protection level of all prefecture-level governments in Zhejiang Province has increased by 0.47 units after the implementation of the pilot policy in comparison with non-pilot cities.

4.2. Robustness Test

4.2.1. Robustness Test Based on Parallel Trend

A common change trend in the temporal aspect of both the experimental group and the control group before the implementation of the policy is an important prerequisite for the test using the DID model. In other words, the parallel trend hypothesis must be met. In particular, both the experimental group and the control group should have the same change trend in the government economic efficiency and environmental protection investment before the implementation of the pilot policy, allowing for no significant difference. Model (2) is built on the basis of an event study to conduct parallel trend testing:

$$Y_{it}=\alpha +{{\displaystyle \mathrm{\sum }}}_{t=-5}^{t=3} {\beta }_i{D}_{it}+\gamma X_{it}+{\epsilon }_{it}$$

(2)

where $D_{it}$ is a series of dummy variables and $t$ indicates the $t$th year of the pilot policy implementation. If $t$ is 1, it indicates the first year after the pilot policy implementation. If $t$ is −1, it indicates the first year before the pilot policy implementation. The parallel trend test focuses on the preceding coefficient ${\beta }_t$ of $D_{it}$. ${\beta }_t$ represents the degree of difference between the experimental group and the control group in the $t$th year since the implementation of the pilot policy (also the difference between economic efficiency and environmental protection investment). When $t$ is less than 0, it indicates that there is no significant difference between the experimental group and the control group before the implementation of the pilot policy if ${\beta }_t$ is not significantly more than 0, thus meeting the parallel trend hypothesis. When $t$ is less than 0, it indicates that there has been a significant difference between the experimental group and the control group before the implementation of the pilot policy if ${\beta }_t$ is significantly different from 0, thus failing to meet the parallel trend hypothesis.

Figure 2 and Figure 3 show the results of parallel trend test for economic efficiency and environmental protection investment of government. Before the implementation of the pilot policy, the regression coefficients of the above two indicators are not significantly different from 0. This fact indicates that there is no significant difference between the experimental group and the control group before the implementation of the pilot policy, thus meeting the parallel trend hypothesis. The regression coefficients are significantly different from 0 in the starting year of the policy implementation and after it. This fact indicates that significant changes have occurred to the economic efficiency and environmental protection investment of government after the implementation of the pilot policy. The coefficient value of government economic efficiency and government environmental protection investment for the experimental group has risen significantly after the implementation of the policy implementation. This result coincides with the benchmark estimation result and proves the result robustness (see Supplementary Material).

4.2.2. Robustness Test Based on PSM-DID Estimation

In order to ensure the reliability of estimation results, the robustness test for the impact brought by the pilot policy on the environmental governance of local governments is further implemented using the Propensity Score Matching and Differences-In-Differences method (hereinafter referred to as the PSM-DID model). The process is as follows: (1) Use Logit regression model to match the covariant in the benchmark model to estimate the score and find out the control group with no significant difference from the experimental group using the neighborhood matching method. (2) Perform balance tests to the matched experimental group and control group. If no significant difference is found among the covariant after matching, it indicates that the matching has succeeded. (3) Perform the DID method to the successfully matched experimental group and control group and then check whether the result is consistent with the estimation result of benchmark models. The results are shown in

Table 4. Sample feature comparison before and after matching.

Variable	Sample	Mean		T	p
		Experimental	Control
$LnPergdp$	Before	11.227	14.419	−3.144 ***	0.002 ***
	After	11.227	11.264	−0.584	0.560
$LnAgri$	Before	15.079	14.705	1.412	0.161
	After	15.079	14.867	0.803	0.424
$LnPerWW$	Before	3.017	2.951	0.746	0.457
	After	3.017	2.847	1.935 *	0.055 *
$LnPerS{O}_2$	Before	1.426	1.539	−0.798	0.426
	After	1.426	1.356	0.488	0.626
$LnPerC{O}_2$	Before	2.002	2.136	−1.570	0.118
	After	2.002	1.898	1.148	0.252

Note: *, ** and *** are significant at the level of 10%, 5% and 1%, respectively.

. After performing propensity score matching to the sample data based on the radius matching method, the results indicate that the standard deviation of all control variables, excluding $LnPerWW$, are effectively reduced. Among them, the value p of the variable $LnPerWW$ is still insignificant at the level of 5%. Therefore, no significant difference exists between the experimental group and the control group after matching. The effect of the propensity matching score is favorable, which is suitable for the subsequent DID test.

See

Table 5. Regression analysis of the PSM-DID method.

Variable	Model 6	Model 7	Model 8	Model 9	Model 10
	$\mathit{G}\mathit{o}\mathit{v}\mathit{e}\mathit{f}\mathit{f}\mathit{c}\mathit{y}$	$\mathit{G}\mathit{o}\mathit{v}\mathit{e}\mathit{f}\mathit{f}\mathit{c}\mathit{y}$	$\mathit{L}\mathit{n}\mathit{P}\mathit{e}\mathit{r}\mathit{E}\mathit{P}$	$\mathit{L}\mathit{n}\mathit{P}\mathit{e}\mathit{r}\mathit{E}\mathit{P}$	$\mathit{G}\mathit{o}\mathit{v}\mathit{e}\mathit{f}\mathit{f}\mathit{c}\mathit{y}$
$Policycity$	−1.502 *** (0.190)	−1.347 *** (0.182)	0.310 *** (0.068)	0.266 *** (0.069)	−1.058 *** (0.176)
$LnPerEP$					−1.086 *** (0.206)
Control variables	NO	YES	NO	YES	YES
City Effect	YES	YES	YES	YES	YES
Year Effect	YES	YES	YES	YES	YES
$Cons$	8.102 *** (0.068)	−33.120 *** (9.878)	5.427 *** (0.024)	−4.224 (3.753)	−37.710 *** (9.073)
N	170	170	170	170	170
R-squared	0.931	0.944	0.902	0.912	0.954

Note: *, ** and *** are significant at the level of 10%, 5% and 1%, respectively. The values in brackets are standard error.

for the detailed results of the PSM-DID test. Model 7 and model 9 show the results after including the control variable while model 6 and model 8 show the results before including the control variable. Model 10 shows the robust test estimation of model 5. After comparing with the results of estimation using the benchmark model and the PSM-DID method, both the direction and significance of the coefficients remain unchanged, proving the robustness of regression results.

5. Discussions and Conclusions

5.1. Discussions

The national green agricultural development pilot zone is a policy experiment conducted by China to realize green agricultural development and agricultural ecological transformation and also an effective response to China’s current extensive agriculture and traditional agriculture. The setting of the national green agricultural development pilot zone is an important action to vigorously explore the path of China’s green agricultural development in the new era. This research focuses on the relationship and spillover effect between the pilot policy and government environmental behavior and has proven that the steady implementation of the pilot policy can effectively improve the environmental protection behavior of local governments. These empirical conclusions have the following implications for the formulation of policies.

First, the pilot cities should vigorously implement the pilot policy. Meanwhile, the state can further expand the scope of pilot zones to achieve high-quality agricultural development and build China into a beautiful country. The pilot policy can significantly improve the environmental protection behavior of local governments. Therefore, it is recommended to make full use of the positive impact brought by the policy on the environmental protection investment of local governments, thus maximizing the policy effects. In order to strengthen the environmental protection investment mechanism of the pilot cities, policy makers are recommended to sum up the experience of the cities where the pilot policy has been better implemented, encourage different regions to learn from the advanced experience and reasonably integrate and design different policies.

Second, when choosing pilot cities in batches, the characteristics, such as urban administrative level, resource endowment and economic level, should be taken into account, thus developing differentiated ways for green agricultural transformations. The research proves that the implementation of the pilot policy will negatively reduce economic performance to a certain extent. Targeted rather than rigidly uniform goals and plans should be formulated according to different regions and economic development levels, thus allowing for flexibility. Policies should be implemented according to the specific site and time and green agriculture goals should be set dynamically. For example, allow developed cities to achieve the goals ahead of schedule and undeveloped cities to improve economic development and then achieve the goals. In the process of realizing environmental protection goals, efforts should be made to avoid causing a negative impact to the overall development of economically backward regions. In particular, efforts should be made to coordinate the pilot policy with other supporting compensation policies relating to economic development while formulating and improving low-carbon pilot policies and make better use of the green economic development effect of pilot cities through reasonable policy combinations.

5.2. Conclusions

The pilot policy of the national green agricultural development pilot zone is a key issue in the context of rural revitalization strategy in the new era and a type of government environmental regulation. It is also a hot topic that has attracted much public attention. 11 prefecture-level cities of Zhejiang Province, as the pilot province of the national green agricultural development pilot zone policy, were selected to constitute the experimental group according to the quasi-natural experiment logic. In addition, 11 prefecture-level cities of Jiangsu Province (excluding Xuzhou City and Taizhou City), which is highly homogeneous with Zhejiang Province in terms of geographical location and economic development, were selected as the non-pilot cities to constitute the control group. With the data of these 22 prefecture-level cities from 2012 to 2020, this paper explores the impact of the pilot policy on the environmental protection behavior of local governments using the DID model. The robustness of the benchmark model has been verified through the parallel trend test and PSM-DID estimation. This paper focuses on the relationship between the pilot policy of the national green agricultural development pilot zone, economic governance efficiency and environmental protection investment of local governments and is effective feedback relating to the pilot policy. By analyzing the deep logic between the pilot policy and the governance efficiency of local governments, this paper helps the pilot policy play a better role. In conclusion, the research has very important practical value. The main research conclusions are as follows.

First, the promotion of the national green agricultural development pilot zone has increased the investment of local governments of these pilot cities in environmental protection. According to the result of data analysis, it was found that the regression coefficients of the pilot policy to the environmental protection investment of the pilot cities in Zhejiang Province are significantly positive regardless of the inclusion of control variables. This fact indicates that the local governments have specially scaled up the environmental protection investment in fiscal expenditures. Such positive impact is more significant in the pilot cities. The close correlation between the level of green agricultural development and local environmental quality may be the possible cause of the above fact. The setting of these pilot cities drives local governments to pay more attention to environmental governance for the green transition and high-quality development of agriculture. The promulgation of policies and measures relating to environmental protection is an effective way to improve local environmental governance, compelling governments to increase the local public finance budget for environmental protection and allocate more resources to environmental protection. The environmental protection investment of local governments is increased accordingly. Environmental protection behaviors of governments to boost the agricultural transition and upgrade are different because of different economic development levels and agricultural development foundations for various regions. The level of economic development is closely related to the local solid foundation for industrial and agricultural development. High-quality agricultural development is often based on a beautiful ecological environment, whereas high-speed industrial development often causes environmental pollution. For those regions with developed economies, the local governments should give higher priority to the balance between industry and agriculture. In addition to financial support and policy inclination to boost environmental protection, local governments should shut down or rectify some traditional industrial enterprises causing severe pollution and high energy consumption, thereby combining prevention with control. For those regions with a lower level of economic development, priority should be given to “prevention”. It is recommended that local governments should strictly maintain the lower limit of environmental quality, implement environmental protection standards throughout the entire process of investment promotion and economic development and boost agricultural development with a high environmental quality. Furthermore, local governments should attach importance to scientific and technological innovation while striving for the transition to green agriculture, thus helping improve local environmental quality while promoting green agricultural development. Government support for environmental protection can help the pilot cities efficiently explore different green agricultural development models, play an exemplary role in promoting the green agricultural development of other similar regions and finally increase the comprehensive benefits of agriculture in boosting local economic development. Various actions relating to the pilot policy will provide different cities featuring advantaged agricultural resources and solid agricultural development foundation with the path of transition to green agriculture. Through the trials in different pilot cities, pervasive pilot experience is obtained to bring the exemplary and driving role of these pilot cities into full play for other cities nationwide.

Moreover, the settings of such pilot zones have increased the investment of local governments in environmental protection. However, the reduced economic governance efficiency of local governments in these pilot cities is a consequent significant negative spillover effect. The regression coefficients of the pilot policy to the economic governance efficiency of government are significantly negative regardless of the inclusion of control variables. A possible explanation lies in that the investment in local economic development will inevitably decrease if the government invests more in environmental protection. This explanation also coincides with the contradictory characteristics between environmental protection and economic development. Social resources held by the government are limited in the same period. When increasing the investment in agriculture or environmental protection, the government must reduce the investment in other aspects. If expecting to follow the path of high-quality and environment-friendly development, the government should necessarily lose economic benefits and slow down economic development over a short period of time. Furthermore, in order to increase environmental benefits and improve environmental quality, local governments should shut down some enterprises that cause severe environmental pollution, which further dilutes their economic performance. Xi, J. P., General Secretary of CPC Central Committee, said that nature is the true treasure (http://news.cntv.cn/special/xddtsthb/ (accessed on 15 August 2005). In the short term, realizing the green transition of agriculture will lose economic benefits to a certain extent. However, in the long run the sustainable economic development model based on the green transition of agriculture will bring greater development opportunities, thus effectively boosting local economic benefits and achieving the high-efficiency coordinated development of all social production sectors. In the current stage, all levels of government should closely follow and solve the following problems: how to find a balance between economic development and environmental protection, and then accelerate a push for the goal of building a great agricultural power and making steady progress in revitalizing the rural industry, talent, culture, ecology and organization mentioned in the Report to the 20th National Congress of the Communist Party of China.

The setting of the national green agricultural development pilot zones is in line with the objective laws of China’s agricultural development in the current stage. It also plays a critical role in steering China’s agricultural development, promoting the green transition of agriculture and improving the level of sustainable agricultural development. In fact, benefits contributed by the policy are different at different time nodes. The setting of such pilot zone drives local governments to make a trade-off between economic development and the green transition of agriculture in the short term. However, in light of the long-term feedback of the policy implementation, the loss of economic benefits in the short term will surely help local economies achieve orderly and healthy development more effectively and the green transition of agriculture will facilitate all industries to realize high-quality development. The setting of the national green agricultural development pilot zone is an important action to vigorously explore the path of China’s green agricultural development in the new era.

The two provinces selected in this paper are homogeneous and comparable in terms of natural resources, economic development level, history, culture, local conditions and customs. Therefore, only Zhejiang and Jiangsu have been selected in this research. Hainan was selected as the pilot province in the second batch in 2019. However, the data relating to the effect of the policy implementation are not fully available. With the lapse of time, the follow-up data will be gradually unveiled and the data of pilot cities in the third batch will be constantly enriched and improved. At that time, the follow-up research will not only analyze the provincial data but also focus on the data of pilot cities nationwide so as to assist in proving the necessity of promoting the national green agricultural development pilot zones.