Coupling Coordination of China’s Agricultural Environment and Economy under the New Economic Background

Abstract

On the basis of the panel data of 31 provinces in China from 2011 to 2020, this paper first constructs an index system through the Pressure-State-Response (PSR) model and conducts a comprehensive evaluation of China’s agricultural environment according to the entropy weight TOPSIS model. Second, a coupling coordination degree model is established to calculate the degree of coupling coordination between the agricultural economy and the environment in each province. Finally, a spatial Durbin model is established to analyze the influencing factors of China’s agricultural economy. Results show that: ① the overall environment in the eastern region has little change, and the overall level is relatively backward; the agricultural environment in the central region is uneven; the agricultural environment in the western region is quite different from north to south. ② The regions with a high level of coupling coordination are mainly concentrated in the central and southern regions, and the performance is relatively intensive. The agricultural economy and the environment in the western region are extremely uncoordinated, and as is the overall coupling coordination between the agricultural economy and the environment in the eastern region in general. Further improvement is also needed. ③ Fixed asset investment, total power of agricultural machinery, rural electricity consumption, rural population, and rural per capita disposable income all have important influences on China’s agricultural economy. ④ The rural population size has a positive and the largest effect on the agricultural economy, whereas rural per capita disposable income has a negative effect on the agricultural economy. Moreover, improving farmers’ enthusiasm for farming is one of the key issues to be solved urgently.

1. Introduction

The construction of a new economic development pattern with the domestic economic cycle as the main and domestic and international dual cycle is an important strategic layout of China during the 14th Five-year Plan period, and its influence on China’s society and economy is comprehensive and fundamental. With the development of agricultural modernization, China’s agriculture has eliminated the old development model of the past and entered a new stage of development. Efforts to develop an agro-ecological economy have become an inevitable trend of future agricultural development. The development of the agricultural ecological economy is inseparable from the agricultural ecological environment. However, due to the reduction in arable land and the substantial decline in land quality in recent years, along with the unreasonable structural adjustment of land, the sown area of grain in China has been greatly reduced, and its food security has been threatened. At the same time, inevitable environmental pollution problems often occur in the human activities of farming, which are contrary to the current concept of the new pattern of economic development. At present, research on the agricultural economy and environment has mainly focused on the construction of relevant index systems, comprehensive evaluations, and exploration of influencing factors. In the process of promoting the development of the agricultural economy and improving its development level in a comprehensive way, the status of the agricultural ecological environment should be given attention. A good agricultural ecological environment can create favorable conditions for the normal growth of crops, provide more support for the development of the agricultural economy, and meet the needs for increased benefits and scientific development.

Zeng and Han [1] constructed panel vector autoregression for a generalized method of moments to discuss the effects of environmental regulation and technological innovation on agricultural economy. They used the impulse response function and variance decomposition method to analyze the regional differences in the effects of environmental regulation and scientific and technological innovation on the agricultural economy between the main and non-main grain-producing areas. The authors concluded that environmental regulation has a significant positive effect on the agricultural economy, and that regional differences exist in the contribution of environmental regulation and technological innovation to agricultural economic changes. Moreover, the environmental regulation contribution of the main grain-producing areas is greater than that of the non-main grain-producing areas.

Ye and Hui [2] reported that the agricultural economic growth accounting method does not consider pollution factors and ignores the loss caused by agricultural production pollution. As a result, agricultural total factor productivity is overestimated by nearly 100%. Redundant input and excessive pollution are the main causes of inefficiency in agricultural production, and the latter has a more obvious effect on it. China’s agricultural productivity growth mainly comes from technological progress, and technological efficiency has a limited role in promoting it. Considering whether the cost of pollution will have a greater influence on the performance evaluation of economic growth in various regions and ignoring pollution factors may lead to policy bias.

Yao and Chen [3] indicated that in the short term, ecological agricultural technological innovation will slowly show its role in promoting agricultural economic growth in the periods of Lags 2 and 3. in the long run, the environmentally friendly agricultural technological innovation and technology The promotion effect of innovation promotion degree on economic growth is slow and effective in the long term, whereas agricultural economic growth is the continuous driving force of ecological agricultural technological innovation. This innovation plays a role in promoting the promotion of technological innovation initially and then weakly.

Although studies have been conducted on the coordinated development of the agricultural economy and the environment, the autocorrelation effects between regions have not been included [4]. This paper innovatively selects a spatial econometric model to explore the spatial spillover effect of agricultural economic development between regions and combines the coupling degree and coordination degree models to explore the coordinated development of agricultural economy and the environment, which is more meaningful for research.

2. Research Methods and Data Sources

2.1. Research Methods

2.1.1. Measuring Method of GTFP of Grain

First, the proportion of the $j$th sample under the $i$th index is calculated, which is regarded as the probability in the relative entropy calculation. The calculation formula is as follows:

$$\begin{gathered} P_{ij}=\frac{Z_{ij}}{{{\displaystyle \sum }}_{i=1}^n{Z}_{ij}} \\ {{\displaystyle \sum }}_{i=1}^n{P}_{ij}=1. \end{gathered}$$

(1)

Then, the information entropy of each indicator is calculated, the information utility value is computed, and the entropy weight of each indicator is obtained by normalization. For the $j$th indicator, the calculation formula of its information entropy is

$$e_j=-\frac{1}{\mathit{ln}n}{{\displaystyle \sum }}_{i=1}^n{P}_{ij}\mathit{ln}(p_{ij}).$$

(2)

The larger $e_{ij}$ is, the greater the information entropy of the $j$th indicator will be, indicating that the information of the $j$th indicator is less, and the information utility value is

$$d_j=1-e_j.$$

(3)

By normalizing the information utility value, the entropy weight of each indicator is obtained as

$$w_j=\frac{d_j}{{{\displaystyle \sum }}_{j=1}^m{d}_j}.$$

(4)

The maximum value $Z^{+}$ and the minimum value $Z^{-}$ are, respectively, calculated as follows:

$$\begin{array}{ll}{Z}^{+}& =\left(Z_1^{+},Z_2^{+},\dots ,Z_m^{+}\right)\\ & =\left(\max \left\{Z_{11,}{Z}_{21},\dots ,Z_{n1}\right\},\max \left\{Z_{12,}{Z}_{22},\dots ,Z_{n2}\right\},\dots ,\max \left\{Z_{1m,}{Z}_{2m},\dots ,Z_{nm}\right\}\right)\end{array}$$

(5)

$$\begin{array}{ll}{Z}^{-}& =\left(Z_1^{-},Z_2^{-},\dots ,Z_m^{-}\right)\\ & =\left(\min \left\{Z_{11,}{Z}_{21},\dots ,Z_{n1}\right\},\min \left\{Z_{12,}{Z}_{22},\dots ,Z_{n2}\right\},\dots ,\min \left\{Z_{1m,}{Z}_{2m},\dots ,Z_{nm}\right\}\right)\end{array}$$

(6)

Then, the distance $D_i^{+}$ of the maximum value of the ith evaluation object and the distance $D_i^{-}$ of the minimum value of the $i$th evaluation object are, respectively, defined as follows:

$$\begin{gathered} D_i^{+}=\sqrt{{\displaystyle \sum }_{j=1}^m{(Z_j^{+}-Z_{ij})}^2} \\ {D}_i^{-}=\sqrt{{{\displaystyle \sum }}_{j=1}^m{(Z_j^{-}-Z_{ij})}^2}. \end{gathered}$$

(7)

Finally, the unnormalized score $E_i$ of the i-th evaluation object is calculated as follows:

$$E_i=\frac{D_i^{-}}{D_i^{+}+D_i^{-}}.$$

(8)

Moreover, the larger $E_i$ is, the higher the comprehensive level of the sustainable development of the evaluation object will be.

2.1.2. Coupling and Coordination Model

First, the pressure index $v_i$, state index $u_i$, and response index $q_i$ of each province are calculated, which can be calculated as follows:

$$u_i=\frac{\sqrt{{{\displaystyle \sum }}_{j=7}^9{w}_j{(Z_j^{-}-Z_{ij})}^2}}{\sqrt{{{\displaystyle \sum }}_{j=7}^9{w}_j{(Z_j^{-}-Z_{ij})}^2}+\sqrt{{{\displaystyle \sum }}_{j=7}^9{w}_j{(Z_j^{+}-Z_{ij})}^2}},$$

(9)

$$q_i=\frac{\sqrt{{{\displaystyle \sum }}_{j=10}^{20}{w}_j{(Z_j^{-}-Z_{ij})}^2}}{\sqrt{{{\displaystyle \sum }}_{j=10}^{20}{w}_j{(Z_j^{-}-Z_{ij})}^2}+\sqrt{{{\displaystyle \sum }}_{j=10}^{20}{w}_j{(Z_j^{+}-Z_{ij})}^2}},$$

(10)

where $w_j$ is the weight calculated by the entropy method used above, and $Z_{ij}$ is the normalized data above.

Then, the coupling degree $C_i$ between the three subsystems in each province is calculated according to the coupling degree model. The coupling degree model is

$$C_i={\left[\frac{v_i{u}_i{q}_i}{{(v_i+u_i+q_i)}^k}\right]}^{\frac{1}{k}},$$

(11)

where $k$ is the adjustment coefficient. Given that three subsystems of pressure, state, and response layers are involved, $k$ is taken as 3. The value range of $C_i$ is on $\left[0,1\right]$, and the larger $C_i$ is, the stronger the interaction among the three systems will be.

Finally, the coordination degree $D_i$ among the three subsystems in each province is calculated according to the coordination degree model. The coordination degree model is as follows:

$$\begin{gathered} T_i=\alpha v_i+\beta u_i+\gamma q_i \\ {D}_i=\sqrt{C_i{T}_i}, \end{gathered}$$

(12)

where $T_i$ is the comprehensive coordination index of the three subsystems in each province; and $\alpha$, $\beta$, and $\gamma$ are undetermined coefficients. This paper considers that the pressure, state, and response layers of each province are equally important. $D_i$ is the coordination degree among the three subsystems in each province, and its value range falls on $\left[0,1\right]$. The larger $D_i$ is, the better the coordination degree among the three systems will be.

2.1.3. Exploratory Spatial Data Analysis (ESDA) Method

ESDA takes the spatial correlation measure as the core, finds agglomerations and singular observations through spatial distribution, and reveals the regional structure of spatial variables. It is an important indicator associated with the attribute values of its adjacent spatial points. A positive correlation indicates that the change of the attribute value of a unit has the same trend as its adjacent spatial units. Thus, the spatial phenomenon has agglomeration, and a negative correlation is the opposite. ESDA includes two categories: global and local space autocorrelation.

(1) Global spatial autocorrelation. Global spatial autocorrelation is a description of the spatial characteristics of a geographical phenomenon or an attribute in the entire region. It summarizes the degree of spatial dependence of geographical phenomenon or attribute values in the spatial range, and it assesses whether there exist aggregation characteristics. The most commonly used correlation index is Moran’s I. The index is calculated as follows:

$$I\left(d\right)=n\frac{{{\displaystyle \sum }}_{i=1}^n{{\displaystyle \sum }}_{j=1}^n{w}_{ij}\left(X_i-\overline{X}\right)\left(X_j-\overline{X}\right)}{{{\displaystyle \sum }}_{i=1}^n\left(X_i-\overline{X}\right)}{{\displaystyle \sum }}_{i=1}^n{{\displaystyle \sum }}_{j=1}^n{w}_{ij},$$

(13)

where $n$ is the number of research objects; $X_i$ is the observed value; $\overline{X}$ is the average value of $X_i$; and $w_{ij}$ is the spatial connection matrix between research objects $i$ and $j$, indicating the potential interaction between spatial units. Moran’s I value is between $\left[0,1\right]$. When the value is greater than zero, a positive spatial autocorrelation exists, and the spatial entity exhibits an aggregated distribution. When the value is less than zero, a negative spatial correlation exists, and the spatial entity exhibits discrete distribution. When the value is equal to zero, spatial entities are randomly distributed.

(2) Local spatial autocorrelation. Global spatial autocorrelation reveals the degree of spatial dependence of objects as a whole but cannot reflect local spatial heterogeneity. Local spatial autocorrelation can be used to describe the similarity of attributes between a spatial unit and its adjacent areas. It can express the degree to which each local unit obeys the global general trend and spatial heterogeneity, and it shows how the degree of spatial dependence varies with location. Its commonly used reflection indicator is the local Moran’s I index. Moreover, combined with local indicators of spatial association (LISA) aggregation graph to study the local spatial distribution law, its essence is to decompose the global Moran’s I index into each regional unit, and the calculation formula is as follows:

$$I_i=Z_i{{\displaystyle \sum }}_{j=1}^n{W}_{ij}{Z}_j,$$

(14)

where $Z_i$ and $Z_j$ are the normalized values of the observed values of spatial units $i$ and $j$, respectively; and $W_{ij}$ is the spatial weight. At a given level of significance, when its value is greater than zero, there exists a positive local spatial autocorrelation, and the similarity value is clustered. When its value is less than zero, there exists a negative local spatial autocorrelation, and the similarity value is scattered. When its value is equal to zero, a local spatial autocorrelation is irrelevant. LISA agglomeration map is used to identify the hot and cold spots of local spatial agglomeration and reveal spatial singular values.

2.1.4. Spatial Panel Measurement Method

According to the different ways of expressing “space” when the model is set, the spatial econometric model is mainly divided into a spatial lag model and a spatial error model. The spatial lag model reflects that the influencing factors of the dependent variable will be applied to other regions through the spatial transmission mechanism, whereas the spatial error model reflects that the regional spillover is the result of random outflow. The spatial Durbin model (SDM) is an extended form of the spatial lag model and spatial error term model, which can be established by adding corresponding constraints to the two models. The SDM is a SAR model (spatial lag model) enhanced by adding spatial lag variables, which can be expressed as follows:

$$y=\rho W_{1}y+X{\beta }_1+W_2\overline{X}{\beta }_2+\epsilon ,$$

(15)

where $W_1$ is the spatial correlation of the dependent variable, $W_2$ is the spatial correlation of the dependent variable, and the two can be set to the same or different matrices; ${\beta }_2$ is the spatial autocorrelation coefficient of the exogenous variable; and $\epsilon$ is the normal independent and identical distribution. Given the random disturbance term of (14), the model can be simplified as:

$$y={(1-\rho W_1)}^{-1}-\left(X{\beta }_1+W_2\overline{X}{\beta }_2+\epsilon \right),$$

(16)

$$\epsilon ~N\left(0,{\sigma }^2-1\right),$$

(17)

where ${\beta }_2$ is a parameter vector of $\left(Q-1\right)\times 1$, which is used to measure the marginal influence of the explanatory variables in the adjacent regions on the dependent variable y. $\overline{X}$ is multiplied by and W to obtain a spatially lagged explanatory variable reflecting the average observed value of the adjacent area.

To analyze all the effects of the explanatory variables in detail, they are divided into direct and indirect effects according to the source. Particularly, the direct effect can be divided into two types, one is the direct effect of the explanatory variable on the explained variable, and the other is the feedback effect caused by the independent variable affecting the dependent variable in the adjacent area. The indirect effect is the spatial spillover effect of the explanatory variable, which can also be divided into two types. One is the influence of the independent variable of the adjacent area on the dependent variable of the local area, and the other is the influence of the independent variable of the adjacent area to change the dependent variable, which, in turn, affects the dependent variable of the area.

2.2. Data Sources

The research data in this paper include statistical panel data and appropriate spatial data. To explore the coordinated development of China’s agricultural economy and the environment, this paper collects panel data from 31 provinces in China with a time series from 2011 to 2020. Variables include the water quality compliance rate of centralized drinking water sources, total grain output (10,000 tons), per capita agricultural, forestry, animal husbandry, and fishery output value (yuan). The data are mainly from the China Statistical Yearbook and the Statistical Yearbook of Various Localities. For a small part of missing data, this paper adopts the process of taking the mean and taking the mode. The spatial data used are from the 1:1.5 million vector data provided by the National Basic Geographic Information Data Center. The data mainly include spatial relationships and related attribute information.

3. Time Series Evolution of China’s Agricultural Environment

3.1. Construction of China’s Agricultural Environmental Index System

On the basis of the PSR model established by the OECD, this paper selects indicators from pressure, state, and response layers. The environmental indicators chosen by most scholars are not the same [5], but in general, they all include seven major directions of forestry, farming, atmosphere, water quality, output value, pollution control, and resource utilization, from which nine indicators are selected.

To evaluate the agricultural environment of each province comprehensively, several indicators are used; however, due to the different types of indicators, they can be divided into benefit, cost, intermediate, and interval indicators, which all need to be transformed uniformly. For the benefit index see ref. [6]. In this paper, pesticide use intensity, NO₂ concentration, PM10 concentration, SO₂ emission intensity, and COD emission intensity are cost-type indicators, and the rest are benefit-type indicators, which need to be reversed to convert them into benefit-type indicators. The formula is

$$\frac{X_{\max }-X}{X_{\max }-X_{\min }}.$$

(18)

On the basis of the PSR model, following the principle of comprehensiveness and systematicness, all the indicators given are taken as the criteria of the comprehensive evaluation system. Many existing methods have been used to determine the weights among the indicators, such as the analytic hierarchy process, Delphi method, entropy method, gray prediction method, and coefficient of variation method. To reduce the subjectivity of the weighting, avoid the large deviation of the evaluation results caused by the extreme value of the index, and objectively show the relative importance of each index, this paper uses the entropy method to determine the weight between the indicators [7] and finally obtain a variable system. The final weights of the variable system obtained are shown in

Table 1. Index system weight.

First Level	Second Level	Third Level	Weights	Unit	Direction
Pressure	Forest	Forest Cover Rate	23.87%	%	+
	Farming	Pesticide Use Intensity	13.09%	10,000 tons/ha	−
	Atmosphere	NO2 Concentration	8.30%	μg/m3	−
		PM10 Concentration	4.12%	μg/m3	−
State	Water Quality	Centralized Drinking Water Source Water Quality Compliance Rate	3.76%	%	+
	Output Value	Gross Agricultural Output	31.74%	billion	+
Response	Pollution Control	SO2 Emissions	5.14%	tons	−
		COD Emissions	7.72%	tons	−
	Resource Utilization	Agricultural Water Use Rate	7.95%	%	+

.

Observing the weight of each indicator, the total agricultural output value has the highest weight, reaching 31.74%. The agricultural output value plays the greatest role in the evaluation of the agricultural environment, and the forest coverage rate is as high as 23.87%. The weights of pesticide use intensity, NO₂ concentration, COD emissions, and agricultural water use rate are not much different and all are above 10%. Therefore, the above three indicators also play an important role in environmental evaluation [8].

3.2. Comprehensive Assessment of China’s Agricultural Environment

The comprehensive score $E_i$ of the sustainable development level of each province is obtained, and the score is located in the interval of $\left[0,1\right]$. The agricultural environment level is divided into five grades with an interval of 0.2, from low to high: I, II, III, IV, and V, corresponding to the level of sustainable development rated bad, poor, medium, good, and excellent [9]. The agricultural environment level of each province is sorted and analyzed. According to the TOPSIS method, the comprehensive scores of agricultural environments from 2011 to 2020 are calculated based on the weights of relevant indicators in each province.

Table 2. Scores for China’s agricultural environment.

Province	2011		2020		Province	2011		2020
	Score	Rank	Score	Rank		Score	Rank	Score	Rank
Anhui	0.345	II	0.426	III	Guangxi	0.494	III	0.642	IV
Beijing	0.336	II	0.382	II	Guizhou	0.373	II	0.546	III
Fujian	0.491	III	0.554	III	Hainan	0.449	III	0.486	III
Gansu	0.264	II	0.391	II	Hebei	0.394	III	0.493	III
Guangdong	0.463	III	0.635	IV	Henan	0.454	III	0.64	IV
Jiangxi	0.461	III	0.519	III	Liaoning	0.375	III	0.449	III
Inner Mongolia	0.312	III	0.371	III	Ningxia	0.286	II	0.292	II
Tibet	0.29	II	0.288	III	Heilongjiang	0.443	III	0.64	IV
Xinjiang	0.312	II	0.41	III	Hubei	0.428	III	0.551	III
Yunnan	0.447	III	0.603	IV	Hunan	0.466	III	0.592	III
Zhejiang	0.466	III	0.508	III	Jilin	0.384	II	0.422	III
Chongqing	0.368	II	0.444	III	Jiangsu	0.353	II	0.497	III
Qinghai	0.288	II	0.29	II	Shandong	0.426	III	0.559	III
Shanxi	0.29	II	0.312	II	Shaanxi	0.422	III	0.543	III
Shanghai	0.247	II	0.274	II	Sichuan	0.444	III	0.66	IV
Tianjin	0.258	II	0.27	II	-	-	-	-	-

shows the scores of the Chinese agricultural environment.

From the results obtained, the comprehensive level of China’s agricultural environment is mostly concentrated at the medium level. Particularly, the agricultural environment in Guangdong, Yunnan, Guangxi, Henan, Heilongjiang, and Sichuan will reach a good level in 2020, but the overall environmental level has not reached an excellent level [10]. To better explore the regional differences in the level of environmental development, it is divided into western, central, and eastern regions to explore separately. The visualization results are shown in Figure 1.

From Figure 1, in the eastern region, Zhejiang, Fujian, Guangdong, and Hainan show a significant lead in the agricultural environment, whereas the agricultural environment in Tianjin and Shanghai is not ideal [11]. The overall environmental change in the eastern region is not large. The overall level is also relatively backward compared with the central and western regions, which shows that the eastern region does not pay considerable attention to the development of agriculture [12]. Environmental construction has reached a good level. Particularly, Henan developed rapidly between 2015 and 2020, and the environmental conditions of each region have developed well between 2011 and 2020 and have made great progress. In the western region, the north and south sides also show great differences. The environmental conditions in Tibet, Gansu, Qinghai, Ningxia, and Xinjiang are all at a relatively low level, and the environmental conditions have not changed significantly in 10 years [13], whereas the levels in southwestern regions, such as Sichuan and Guizhou, are at a relatively low level. Moderate and large improvements are observed over 10 years.

4. Coupling Coordination Degree of China’s Agricultural Environment and Economy

The coupled coordination degree model is used to analyze the coordinated development level of subjects. Coupling degree refers to the interaction between two or more systems to achieve a dynamic relationship of coordinated development, which can reflect the degree of interdependence and mutual restriction between systems. The degree of coordination refers to the degree of benign coupling in the coupling interaction relationship, which can reflect the quality of coordination. The agricultural economic growth index is represented by the per capita output value of agriculture, forestry, animal husbandry, and fishery (total output value of agriculture, forestry, animal husbandry, and fishery/persons engaged in the primary industry), considering population changes. At the same time, to consider the influence of inflation on agricultural economic growth, the gross output value is calculated using the constant price in 2011 as the base period [14]. The coupling degree between the comprehensive agricultural environment index and agricultural economic growth index in each province from 2011 to 2020 is calculated [15]. The coupling and coordination status of China’s agricultural environment and economy in 2020 is shown in

Table 3. 2020 agricultural economy and environment coupling coordination degree.

Province	Coupling	Coordination	Rank	Province	Coupling	Coordination	Rank
Beijing	0.511	0.178	Mild Disorder	Hunan	0.998	0.779	Well-coordinated
Tianjin	0.985	0.055	Moderately Disordered	Inner Mongolia	0.999	0.321	Barely Coordinated
Hebei	0.999	0.624	Intermediate Coordinator	Guangxi	0.969	0.761	Well-coordinated
Shanghai	0.882	0.050	Moderately Disordered	Chongqing	0.986	0.408	Primary Coordination
Jiangsu	0.992	0.688	Well-coordinated	Sichuan	0.999	0.943	Quality Coordination
Zhejiang	0.914	0.448	Primary Coordination	Guizhou	0.966	0.571	Intermediate Coordinator
Fujian	0.976	0.607	Intermediate Coordinator	Yunnan	0.981	0.715	Well-coordinated
Shandong	0.990	0.870	Quality Coordination	Tibet	0.752	0.065	Moderately Disordered
Guangdong	0.995	0.849	Quality Coordination	Shaanxi	0.958	0.553	Intermediate Coordinator
Hainan	0.847	0.377	Barely Coordinated	Gansu	0.998	0.192	Verge of Disordered
Shanxi	0.998	0.176	Verge of Disordered	Qinghai	0.919	0.080	Moderately Disordered
Anhui	0.993	0.493	Primary Coordination	Ningxia	0.964	0.092	Moderately Disordered
Jiangxi	0.961	0.513	Intermediate Coordinator	Xinjiang	1	0.408	Primary Coordination
Henan	1	0.955	Quality Coordination	Liaoning	0.999	0.467	Primary Coordination
Hubei	1	0.720	Well-coordinated	Jilin	0.982	0.357	Barely Coordinated
Heilongjiang	0.979	0.784	Well-coordinated	-	-	-	-

.

In 2020, the coupling and coordination status of most regions can reach coordination. Particularly, Sichuan, Shandong, Guangdong, and Henan are ahead of most other regions in terms of coupling and coordination and have reached a high-quality coordination level in the final coupling and coordination status evaluation. The coupling degree of almost all regions is above 0.9, which shows that the degree of interaction between the agricultural environment and the economy is extremely high. The coordination degree of Sichuan, Shandong, and Guangdong is also extremely high, and the benign role in the interaction is dominant. By contrast, the coordination degree of Beijing, Tianjin, Shanghai, Ningxia, Qinghai, and Tibet is extremely low, all below 0.1. There exist more malignant interactions behind the high coupling degree of these regions. To better show the spatiotemporal changes in the coupling coordination status, the coupling coordination degree in 2011 and 2020 is visualized, and the coupling coordination heat map is shown in Figure 2.

The performance of the heat map shows that compared with the coupling coordination in 2011 and 2020, great progress, in general, has been observed, and the coupling coordination level in each region has basically improved considerably, especially in the central and eastern regions [16]. The progress is highly significant, and some areas have reached the level of high-quality coordination in the level of coupling coordination. The regions with a higher level of coupling coordination are more concentrated in the central and southern regions, and the performance is relatively dense, so the influence of spatial effects can be considered. The western region generally shows extreme incoordination between the agricultural economy and the environment. The eastern region still shows a certain degree of coupling and coordination, but it still needs further improvement compared with the central region [17].

5. Spatial Correlation Test and Spatial Econometric Analysis

5.1. Spatial Autocorrelation Test

According to the first law of geography, everything is related to everything else, but near objects are more related than distant ones. To explore whether there exists a spatial autocorrelation effect in China’s agricultural economy [18] and whether the development of the agricultural economy is affected by the spatial correlation effect, according to Moran’s I formula, Geoda software is used to calculate the Moran’s I of per capita agricultural, forestry, animal husbandry, and fishery output value of 31 provinces in China. The calculation results are shown in

Table 4. Moran’s I of China’s Agricultural Economy.

Year	I	E(I)	sd(I)	z	p-Value
2011	0.389	−0.033	0.109	2.044	0.020
2012	0.371	−0.033	0.109	1.873	0.031
2013	0.374	−0.033	0.109	1.910	0.028
2014	0.374	−0.033	0.109	1.912	0.028
2015	0.367	−0.033	0.109	1.841	0.033
2016	0.361	−0.033	0.109	1.774	0.038
2017	0.359	−0.033	0.110	1.751	0.040
2018	0.349	−0.033	0.110	1.660	0.048
2019	0.353	−0.033	0.110	1.698	0.045
2020	0.339	−0.033	0.110	1.563	0.049

.

Overall, Moran’s I was positive from 2011 to 2020, and its p-values were all below 0.05, with a high level of significance. Therefore, a high degree of positive correlation exists among the 31 provinces in China [19]. From the numerical point of view, the degree of spatial autocorrelation generally shows a state of decline year by year, but it has always remained above 0.3. Particularly, between 2011 and 2014, the spatial autocorrelation effect decreased slightly. However, the spatial correlation effect between 2015 and 2020 began to decline rapidly but remained at a certain level. Thus, in 2011, a high positive correlation existed between the per capita agricultural, forestry, animal husbandry, and fishery output value between regions. That is, the agricultural economy of developed areas and the developed regions are adjacent to each other, whereas underdeveloped regions are adjacent to one another [20]. This finding has the same characteristics as the coupling coordination degree of agricultural economy and environment in China, that is, the agglomeration of environmental development and the agglomeration of agricultural economic development are related to each other [21].

To explore the spatial autocorrelation of local areas, Stata software is used to calculate the LISA index of each area, and the LISA cluster diagram is shown in Figure 3.

Overall, the high–high (H–H) agglomeration areas are mainly concentrated in the central region and are also involved in the eastern and western regions [22]. The specific provinces are Hubei, Hunan, Jiangxi, Jiangsu, Inner Mongolia, Qinghai, Sichuan, Liaoning, Ningxia, and Henan. Low–low L–L agglomeration areas are mainly scattered in China’s border provinces, including Guangdong, Yunnan, Chongqing, Zhejiang, Anhui, Shandong, Beijing, Heilongjiang, and Xinjiang. From the perspective of changing trends, the H–H agglomeration of China’s agricultural economy is showing a trend of spreading from the central region to the surrounding areas. The L–L agglomeration in the central and western regions is gradually decreasing. Thus, under the new economic development pattern, the agricultural economy is developing rapidly, the overall agricultural economy in the central region is developing to a high level, and it is constantly driving the development of the surrounding areas [23].

5.2. Theoretical Framework Analysis of Influencing Factors

At present, research on agricultural economics has mainly focused on the construction of relevant index systems, comprehensive evaluation, and exploration of influencing factors [24]. In the process of promoting the development of the agricultural economy and improving its development level in an all-around way, the status of the agricultural ecological environment should be given attention. A good agricultural ecological environment can create favorable conditions for the normal growth of crops, provide more support for the development of the agricultural economy, and meet the needs for increased benefits and scientific development [25].

Spatial Geographical Factors. Through the theoretical analysis of spatial econometrics, a certain attribute or phenomenon of an area and the attribute value of its adjacent area are related to the previous phenomenon. Therefore, when making suggestions, the influence of spatial correlation should be included, such that the proposed recommendations are more scientific and reliable. This study expects that the spatial and geographic factors positively affect China’s agricultural economy.

Production Scale Factor. From a traditional point of view, the production scale of a region is directly related to the agricultural economic output value. When the production scale of a region is larger, the corresponding crop yield will increase, and its production scale is mainly reflected in the sown area, fixed asset investment, and crop yield. In addition, the increase in the production scale of the region may be due to a large amount of resource input from the local government. Thus, this study expects that each factor in the production scale will have a positive effect on China’s agricultural economy [26].

Technical Factors. According to previous seeding technology, in most cases, the stacking of the number of chemical fertilizers used should be considered to increase crop yield. However, the stacking of chemical fertilizers not only increases yields but also brings environmental and soil fertility damage. Therefore, the effect of the use of chemical fertilizers on the agricultural economy needs further analysis. Conversely, the total power of agricultural machinery represents the equipment power of the current agricultural farming. The larger the total power is, the higher the agricultural production efficiency will be. Therefore, the influence of agricultural production efficiency on the agricultural economy is positive. Rural electricity consumption can also reflect the agricultural production conditions. However, there may be some areas where the electricity consumption efficiency is not high, so the influence of this variable on the agricultural economy needs to be further explored [27].

Social Factors. With the continuous improvement of urban construction, an increasing rural population has been attracted to cities, resulting in a sharp decrease in the rural population, and the rural population represents the local productivity. Thus, there exists a positive influence. Conversely, the disposable income of rural per capita income can reflect the income status of farmers, and it can involve farmers’ enthusiasm for production. When the income situation is not good, farmers may actively cultivate to improve their quality of life. Therefore, this factor is highly important for the agricultural economy. The effect should be negative.

5.3. Spatial Measurement Analysis

5.3.1. Spatial Panel Measurement Method

With the development of geographic information system technology, the use of spatial econometric models to explore agricultural research has gradually increased. The geographic weighted model is an extension of the ordinary linear regression model. It embeds the spatial position of the data into the regression equation, which is an efficient method for observing spatial non-stationarity and spatial dependence of data. To analyze the influencing factors of the agricultural economy and then explore how the agricultural economy and the environment can develop together, relevant indicators are selected from the three aspects of economy, ecology, and society. Data are searched through various statistical yearbooks, and the total grain output (10,000 tons), crop sown area (1000 hectares), and total investment in fixed assets (100 million yuan) are selected as production indicators. For technical indicators, the number of chemical fertilizers (10,000 tons), the total power of agricultural machinery (kWh), and rural electricity consumption (100 million kWh) are used. The per capita output value of agriculture, forestry, animal husbandry, and fishery is used as the explained variable, and the other indicators are used as the explanatory variables. The above data are all from the China Statistical Yearbook and the statistical yearbooks of various provinces. The final index system is shown in

Table 5. China’s agricultural economic indicator system.

First Level	Second Level	Definition	N	Average Value
Production	Tot	Total grain output (10,000 tons)	310	1992.7227
	Cro	Crop sown area (thousand hectares)	310	5722.3176
	Inv	Total investment in fixed assets (100 million yuan)	310	18,786.8500
Technology	Fer	Fertilizer usage (tons)	310	195.6563
	Pow	Total power of agricultural machinery (kWh)	310	3261.3195
	Ele	Rural electricity consumption (100 million kWh)	310	285.2590
Society	Pop	Rural population (10,000 people)	310	1875.9000
	Inc	Rural per capita disposable income (yuan)	310	12,561.3323

.

Before conducting the quantitative regression, the variance inflation factor (VIF) method is applied to the indicators in the above indicator system to eliminate the multicollinearity interference between the independent variables. The results are shown in

Table 6. VIF.

Index	VIF	1/VIF
lnTot	26.54	0.038
lnCro	23.19	0.043
lnFer	14.12	0.071
lnPop	9.60	0.104
lnPow	9.30	0.108
lnInv	8.14	0.123
lnEle	4.32	0.031
lnInc	4.14	0.242

.

As shown in Table 6, the VIF value of Tol, Cro, and Fer is 26.54, 23.19, and 14.12, respectively, which are far greater than the empirical value of 10. Therefore, to ensure the accuracy of subsequent spatial econometric regression, the above three indicators should be deleted. The index system is shown in

Table 7. Influencing factors of China’s agricultural economy.

Index	Definition	N	Average Value
Inv	Total investment in fixed assets (100 million yuan)	310	18,786.8500
Pow	Total power of agricultural machinery (kWh)	310	3261.3195
Ele	Rural electricity consumption (100 million kWh)	310	285.2590
Pop	Rural population (10,000 people)	310	1875.9000
Inc	Rural per capita disposable income (yuan)	310	12,561.3323
Out	Per capita output value of agriculture, forestry, animal husbandry, and fishery (yuan)	310	3384.5452

.

5.3.2. Spatial Panel Model Testing and Model Selection

To effectively avoid pseudo-logarithmic regression in advance, heteroscedasticity should be eliminated, and stationary data should be obtained without changing the basic properties and time correlation of each time variable series. We take a logarithm for each variable series. When selecting a spatial econometric model, the LM test is required to explore whether there a correlation exists between the residual series of the model [28]. The calculation results obtained using Stata software are shown in

Table 8. LM checklist.

Test		Statistic	p-Value
Spatial error effects	Moran’s I	3.745	0.000
	Lagrange multiplier	0.705	0.013
	Robust Lagrange multiplier	1.336	0.002
Spatial lag effects	Lagrange multiplier	2.284	0.012
	Robust Lagrange multiplier	2.951	0.048

.

As shown in Table 7, when the per capita output value of agriculture, forestry, animal husbandry, and fishery is used as the dependent variable, the statistic of Moran’s I index is 3.745, and the p-value is 0.000, which further indicates that there exists a spatial effect in the output value of agriculture, forestry, animal husbandry, and fishery. The difference between the spatial error effect and the spatial lag effect is as follows. The test p-values are all less than 0.05, which is highly significant. Thus, the model has spatial error effects and spatial lag effects [29]. Therefore, the mixed OLS model is rejected, and the spatial panel model is used for further analysis of the problem. To choose a reasonable spatial econometric model, a Hausman test is also required to determine whether to choose a fixed-effect model or a random-effects model. On this basis, an appropriate spatial panel model (spatial panel autoregressive model or spatial panel error model) is selected. The null hypothesis is that the p-value of the model with per capita agriculture, forestry, animal husbandry, and fishery as the dependent variable is 0.0024 when the individual effects in the random effect model are not correlated with the explanatory variables [30].

5.3.3. Spatial Panel Model Estimation Results

On the basis of the spatial Doberman regression selected above, a spatial panel estimation is performed on the influencing factors of China’s agricultural economy. The processed data are imported into Stata software and combined with Geoda to output the spatial adjacency weight matrix to calculate the estimation results of the spatial panel Dubin regression. The output results are shown in

Table 9. Spatial measurement estimation result.

Index	Model I		Model II		Model III		Model IV
	No Fixed Effects		Time Fixed Effects		Spatial Fixed Effects		Two-Way Fixed Effects
	Coef	p > z	Coef	p > z	Coef	p > z	Coef	p > z
lnInv	0.090 ***	0.000	0.236 ***	0.001	0.090 **	0.020	0.053 ***	0.004
lnPow	0.133 ***	0.000	0.147 ***	0.003	0.133 **	0.037	0.147 ***	0.000
lnPop	−1.021 ***	0.000	0.553 ***	0.000	−1.021	0.155	−1.121 ***	0.000
lnInc	0.212 **	0.047	−0.612 ***	0.000	0.212	0.107	1.876 ***	0.000
lnEle	0.041	0.196	0.126 ***	0.003	0.041 **	0.032	0.011	0.687
ρ	0.248 ***		0.215 ***		0.248 **		−0.000
R2	0.1220		0.8028		0.1220		0.1360
LogL	334.9123		−141.9636		334.9123		383.4904

Note: **, and *** indicate that they passed the significance test at the 10%, 5%, and 1% levels, respectively.

. Models I, II, III, and IV in the table correspond to the no-fixed effect, space fixed effect, time fixed effect, and time–space two-way fixed effect of the spatial panel of China’s agricultural economy, respectively [31].

The regression coefficients of each model indicate that the regression coefficients of Model II all passed the significance test, and the goodness of fit is considerably greater than that of Models I, III, and IV. Thus, Model II is better than Models I, III, and IV. By contrast, Model I does not consider the spatial effect before each province, and it assumes that the regions have the same level of agricultural economic development. Model III considers the effect of space [32], but does not affect the effect of time. It is considered in the model, so different degrees of deviation will exist in the fitting of the model. Model IV takes the regional differences and time effects of China’s agricultural economy into the model and theoretically avoids the time and regional differences. However, from the perspective of the goodness of fit [33], the fitting effect is better under the time fixed effect model. The reason may be that the spatial fixed effect not only affects the local agricultural economic conditions but also has specific radiation effects on other regions. At the same time, under the time difference, the time fixed effect model is better than the two-way fixed effect model.

5.3.4. Robustness Test of the Estimation Results of the Spatial Panel Model

Before further empirical analysis, it is necessary to test the robustness of the model. In the selection of dependent variables, the fitting effect of only using the agricultural output value (Agr) as the dependent variable may not be ideal, but they will still be approximate. The per capita output value for agriculture, forestry, animal husbandry, and fishery (Out) is used as the indicator of agricultural economic growth, and population changes are taken into consideration [34]; at the same time, in order to consider the impact of inflation on agricultural economic growth, the total output value of agriculture, forestry, animal husbandry, and fishery is taken as the constant price in 2011. Calculated in the base period, using the total agricultural output value as a robustness test. The fitting results are shown in

Table 10. Robustness test result.

	LnOut			LnAgr
	SDM	SAR	SEM	SDM	SAR	SEM
LnInv	0.236 ***	0.247 ***	0.261 ***	0.155 ***	0.154 ***	0.159 ***
	(0.001)	(0.000)	(0.000)	(0.021)	(0.014)	(0.006)
LnPow	0.147 ***	0.172 ***	0.215 ***	0.200 ***	0.202 ***	0.241 ***
	(0.003)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
LnPop	0.553 ***	0.555 ***	0.545 ***	0.570 ***	0.562 ***	0.575 ***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
lnInc	−0.612 ***	−0.512 ***	−0.438 ***	−0.717 ***	−0.689 ***	−0.612 ***
	(0.000)	(0.000)	(0.003)	(0.000)	(0.000)	(0.000)
lnEle	0.126 ***	0.105 ***	0.091 ***	0.172 ***	0.177 ***	0.159 ***
	(0.003)	(0.002)	(0.011)	(0.000)	(0.000)	(0.000)
ρ	0.215 ***	0.086 ***	-	0.203 ***	0.095 ***	-
	(0.000)	(0.023)	-	(0.000)	(0.008)	-
R2	0.8028	0.7415	0.7313	0.7864	0.7088	0.7232

Note: *** indicate that they passed the significance test at the 10%, 5%, and 1% levels, respectively.

.

Observing the fitting results, the spatial autoregressive coefficients all pass the significance test, and the autoregression coefficients are all positive. Thus, the improvement of the agricultural economy in the surrounding areas will lead to the improvement of the agricultural economy in the region, which will further lead to the regional H–H characteristics of aggregation. From the results of the two dependent variables, the SDM has the best fitting effect [35]. When the observed autoregression coefficient is positive, the spatial autoregression coefficient of the SDM is also the highest. From the results of the control variables, the estimated results of the two dependent variables are basically the same, including the regression coefficient results. The LR test is further used to explore whether the SDM can be degenerated into the SEM and the SAR model. The software is used to compare the SDM with SEM and SAR and finally compare the significance of the results. In both pairs of comparisons, the p-value is 0.0000, which is extremely significant; thus, the model passes the LR test, and the model is considered to pass the robustness test [36].

5.3.5. Analysis of Estimation Results of Spatial Panel Model

The SDM (fixed effects model) of China’s agricultural economy under time fixed effects, namely, the fixed effects regression model, is a time panel variable data model analysis method. That is, an experimental result only wants to compare the length difference between the specific different categories or other categories of each different independent variable and the various interactions between the specific different categories or other categories of each different independent variable. This model is designed to conduct related experiments that infer the same independent variable, another specific class, or a specific class to include in the future. The fixed-motion effect comprehensive regression model is a panel variable analysis method based on the spatial science panel model data, in which the individual intercept changes with time but cannot change with the individual time length. The time fixed effect regression model refers to a regression model with different intercepts for individuals in different time segments (specific time) [37]. If the intercept of a model has a certain significant length difference between different partial sections, but the intercepts are the same for different partial time individual series (different individuals), then a time fixed effect regression model should be re-established.

On the basis of the fitting results of Model II in Table 8, this paper explores how China’s agricultural economy is affected by its influencing factors. The regression coefficient of the total investment in fixed assets is 0.236, and its p-value is 0.001. The significance test shows that for every percentage point increase in the total investment in fixed assets, China’s agricultural economy will also increase by 0.236%. The growth of the agricultural economy has played a significant role in promotion. The regression coefficient of the total power of agricultural machinery is 0.147, and the p-value is also lower than 0.05, which is highly significant. For every percentage point increase in the total power of agricultural machinery, China’s agricultural economy will also increase by 0.147. From this point of view, the role of production equipment in improving agricultural production efficiency is still extremely large, and it plays a vital role in improving productivity. The p-value of the rural population is 0.000. Through the significance test, from the regression coefficient, the agricultural economy will increase by 0.553% if the rural population increases by one percentage point. In comparison with other variables, this variable has the greatest positive effect on the agricultural economy. The population is one of the most important factors affecting the development of the agricultural economy. To further promote the development of the local agricultural economy, the government should introduce more agricultural policies to attract people back to the countryside and leave the local people in the countryside. It plays a significant role in promoting the development of the agricultural economy. The influence of rural per capita disposable income on the agricultural economy is negative, and its regression coefficient is −0.612, which is significant at the 1% significance level. If it increases by one percentage point, then China’s agricultural economy will drop by 0.612 percentage points. The current income of rural people is still relatively low. When rural people obtain higher disposable income, they may leave the countryside to pursue more. In view of a high-quality life, improving farmers’ enthusiasm for farming is also one of the urgent problems to be solved at present. Finally, the regression coefficient of the indicator of rural electricity consumption is 0.126, which is significant at the 1% significance level. In comparison with all other variables, this variable has the smallest influence. For every one percentage point increase in rural electricity consumption, China’s agricultural economy will also increase by 0.126%. Therefore, everything should be done to maximize the development of China’s agricultural economy and improve the current rural electricity consumption efficiency. The use of electrical energy in agricultural production still has an objective role in promoting the development of China’s agricultural economy.

5.3.6. Direct, Indirect, and Total Effects

In the estimation of spatial econometric models, the total effect can be decomposed into direct and indirect effects [38]. Specifically, the agricultural economic development of a region is affected by regional control variables and by adjacent regional control variables. The direct effect refers to the changes in agricultural economic development caused by the relevant influencing factors in the region and then through spatial correlation. In turn, the development of the agricultural economy in the region changes. The indirect effect refers to the changes in the development of the agricultural economy caused by the related factors in the adjacent regions and the changes in the development of the agricultural economy in the region through the spatial indirect effect. The results are more credible because the SDM incorporates the spatial lags of all control variables into the model for estimation. The regression results of the model are decomposed into direct, indirect, and total effects, as shown in

Table 11. Effect decomposition of time fixed SDM.

Index	Direct Effect	Indirect Effect	Total Effect
LnInv	0.286 ***	0.185 *	0.471 ***
	(0.000)	(0.070)	(0.000)
LnPop	0.570 ***	−0.223 *	0.347 **
	(0.000)	(0.089)	(0.021)
lnEle	0.095 **	0.102	0.197 **
	(0.010)	(0.158)	(0.031)
lnPow	0.175 ***	0.097 *	0.271 ***
	(0.000)	(0.098)	(0.000)
lnInc	−0.550 ***	−0.747 **	−1.297 ***
	(0.000)	(0.012)	(0.000)

Note: *, **, and *** indicate that they passed the significance test at the 10%, 5%, and 1% levels, respectively.

.

Total Investment in Fixed Assets. For China’s agricultural economy, the direct effect is positive and highly significant, indicating that the increase in total investment in fixed assets leads to the growth of the agricultural economy in adjacent regions, which, in turn, leads to the growth of the agricultural economy in the region and is highly significant. The indirect effect, which is significant at the 10% significance level, indicates that the increase in the total investment in fixed assets in adjacent regions will also improve the agricultural economy of the region and ultimately make the total effect positive. From this point of view, the improvement of agricultural infrastructure in rural areas is significant for the development of the agricultural economy. It will not only promote the development of the local agricultural economy but also play a leading role in the agricultural economy of adjacent areas.

Number of Rural Population. Its direct effect is positive and highly significant. Therefore, when the local rural population increases, the agricultural economy of the surrounding areas will also grow, and, in turn, the agricultural economy of the local area will also be promoted. The indirect effect is negative and significant, indicating that the increase in the rural population in the surrounding areas will also reduce the development of the agricultural economy in the region. Therefore, the introduction of relevant policies to retain the local labor force and attract the urban population will promote the development of the agricultural economy to a greater extent. The final total effect is still positive; thus, the influence of the direct effects seems greater than that of the indirect effects.

Rural Electricity Consumption. The direct effect is significantly positive, and the indirect effect does not pass the significance test, indicating that the direct effect is mainly at work. The increase in rural electricity efficiency will lead to the improvement of the agricultural economy in adjacent areas, which, in turn, will promote the agricultural economy of the region. The final total effect is positive. Therefore, promoting the improvement of the current agricultural technology and improving the efficiency of rural electricity consumption will have a better effect on promoting the local agricultural economy.

Total Power of Agricultural Machinery. The direct effect is significantly positive, indicating that the increase in the total power of agricultural machinery in this region will promote the development of the agricultural economy in adjacent regions. The indirect effect is also highly significant and positive, indicating that the increase in the total power of agricultural machinery in adjacent areas will promote the development of the agricultural economy in this area. The final total effect is positive and highly significant.

Rural per Capita Disposable Income. Under the direct effect, the regression coefficient is −0.550, the effect is negative, and the p-value is 0.000, which is significant at the 1% significance level. The difference from other variables is that the increase in rural per capita disposable income in this region will not promote the agricultural economic development of the surrounding areas. On the contrary, it will have a restraining effect on the agricultural economic development of the surrounding areas. The reason may be that the growth of the rural disposable income in this area comes from the local farmer-friendly policy. A good policy will further attract the surrounding areas, and farmers will come to the local area, which leads to a shortage of productive labor in the surrounding areas. The indirect effect is also negative, and the regression coefficient is −0.747, which is significant at the 5% significance level. Thus, the increase in rural per capita disposable income in adjacent areas will also reduce the local area and agricultural economy, ultimately making the total effect negative.

6. Discussion

On the issue of how to coordinate the development of China’s agricultural economy and the environment, most scholars have used the entropy weight TOPSIS model to calculate the comprehensive score between regions, and further combined the coupling coordination degree model to evaluate the coupled coordinated development of each region. The relevant suggestions are often more biased toward the macro. On the basis of the aforementioned non-parametric methods, this study comprehensively uses parametric methods to analyze the spatial correlation effects between regions from the perspective of spatial effects. Finally, the SDM is selected to analyze the influencing factors of China’s agricultural economic development from the perspective of production, technology, and society. The results show that investment in fixed assets, total power of agricultural machinery, rural electricity consumption, rural population, and rural per capita disposable income all have an important influence on China’s agricultural economy. Through the decomposition of the spatial effect, the impact path of each factor on China’s agricultural economy can be analyzed in detail. Therefore, macro-perspective and micro-suggestions can be proposed for the conditions of different regions. The coupled and coordinated development of China’s agricultural economy and environment should also be promoted. However, there still exist areas that need to be further deepened in this research. First, certain subjective factors exist when looking for relevant indicators of China’s agricultural economy. Therefore, in future research, further field study will be conducted to collect more microscopic data, from which more reasonable conclusions can be drawn. Second, due to the limitations of many factors, such as manpower, time, and statistical data, the panel data used in this study are only in the period between 2011 and 2020. Years can be stretched to forecast developments over a short period of time.

7. Conclusions and Suggestions

On the basis of the panel data of 31 provinces in China from 2011 to 2020, this paper selects the impact indicators of China’s agricultural environment from the pressure, state, and response layers through a PSR model and builds an index system, namely, agro-environmental composite score. Moreover, the overall environmental changes in the eastern region are not large, and the overall level is relatively backward compared with the central and western regions. The agricultural environment in the central region is also mixed, and the environmental conditions in each region are good from 2011 to 2020. In the western region, the north and the south also showed great differences. The environmental conditions in Tibet, Gansu, Qinghai, Ningxia, and Xinjiang were all at a relatively low level, and no environmental condition was evident in the past 10 years. By contrast, the levels of the southwestern regions, such as Sichuan and Guizhou, were moderate and made great progress in the past 10 years [39].

Furthermore, the per capita output value of agriculture, forestry, animal husbandry, and fishery are used to represent the development of China’s agricultural economy, combined with China’s agricultural environment score from 2011 to 2020, to establish a coupling coordination degree model and calculate the coupling coordination degree of agricultural economy and environment in each province in the past 10 years. The results show that compared with 2020, the coupling coordination status in 2011 showed great progress in general, and the coupling coordination level in all regions basically improved greatly. In particular, the central and eastern regions made great progress. Some regions have reached the level of high-quality coordination on the level of coupling coordination. The regions with a high level of coupling coordination are more concentrated in the central and southern regions, and the performance is relatively dense. The western region generally shows extreme incoordination between the agricultural economy and the environment. The eastern region still shows a certain degree of coupling coordination, but still needs further improvement compared with the central region.

To promote the coordinated development of China’s agricultural economy and the environment, an index system of the influencing factors of the agricultural economy is established. After confirming the existence of spatial effects, the SDM under the time-fixed effect is established through verification, and the spatial effects are considered. The fitting results show that the infrastructure construction of agricultural production will play a significant role in promoting the growth of the agricultural economy. The role of production equipment in improving agricultural production efficiency is still extremely large and has a significant effect on the improvement of productivity. The rural population has the greatest effect on the positive improvement of the agricultural economy. The rural per capita disposable income has a negative effect on the agricultural economy, and improving farmers’ enthusiasm for farming is also one of the urgent problems to be solved. Therefore, to promote the development of China’s agricultural economy, the current rural electricity efficiency should be improved, and electricity energy should be used in agricultural production as much as possible. The role of promoting the development of China’s agricultural economy is still objective.

Finally, the spatial effect is decomposed into direct, indirect, and total effects, and the improvement of agricultural infrastructure in rural areas is significant for the development of the agricultural economy. The economy will also play a leading role. The direct effect of the rural population is greater than the indirect effect, and the final total effect is positive; population growth has a considerably significant role in promoting local agricultural economic growth. It promotes the improvement of current agricultural technology and improves rural areas. Electricity efficiency also has a certain effect on the local agricultural economy. The upgrading of equipment in the region will attract farmers in the surrounding areas to introduce advanced equipment, which will directly improve agricultural productivity. The growth of local rural per capita disposable income has not contributed to the growth of China’s agricultural economy effect.

Combined with the current imbalance in most areas where the degree of coupling is high but the degree of coordination is not high, reducing the cutting of trees in forestry areas, maintaining good forest coverage, and using physical methods as much as possible are economically recommended to eliminate agricultural pests, reduce the use of pesticides, further increase the efficiency in the farming process, improve the agricultural output value, and build an ecological and efficient agricultural farming system. A farming waste treatment department should also be established to treat and recycle the waste in the sowing process in an ecological manner. Moreover, SO₂ and harmful gas emissions should be reduced, and the environmental construction of agriculture should be strengthened.

For agricultural economic development, the government should increase capital investment, increase investment in fixed production facilities, and further improve and upgrade local agricultural production systems; at the same time, promote agricultural exchanges in surrounding areas, promote cross-regional exchanges and cooperation, and introduce advanced technologies from various regions. High-tech talents and excellent management modes, according to the actual situation in various places, improve relevant policies and increase farmers’ income. In addition, they provide a development path for China’s agricultural economy and environment and promote the construction of regional ecological civilization and rural revitalization.