Front. Sustain. Food Syst. Frontiers in Sustainable Food Systems Front. Sustain. Food Syst. 2571-581X Frontiers Media S.A. 10.3389/fsufs.2025.1611018 Sustainable Food Systems Original Research Research on the spatial spillover effect of increasing farmers’ income on environmental efficiency: evidence from China Chen Yahui Deng Xuelin Sun Bingteng * School of Humanities and Social Science, University of Electronic Science and Technology of China, Zhongshan Institute, Zhongshan, China

Edited by: Amar Razzaq, Huanggang Normal University, China

Reviewed by: Muhammad Zubair Chishti, Zhengzhou University, China

Chun Hu, Wuhan Polytechnic University, China

 *Correspondence: Bingteng Sun, bingteng888@163.com 10 09 2025 2025 9 1611018 15 04 2025 22 08 2025 Copyright © 2025 Chen, Deng and Sun. 2025 Chen, Deng and Sun

This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

Given the growing scarcity of global resources and the continuous deterioration of the ecological environment, protecting ecosystems while increasing farmers’ income has become a critical challenge. Based on panel data from 30 Chinese provinces between 2010 and 2019, this study examines the impact of increasing farmers’ income on provincial environmental efficiency by exploring its underlying mechanism and employing a spatial Durbin model. The results reveal significant spatial agglomeration effects in provincial environmental efficiency, primarily manifesting as high-high clusters and low-low clusters. Increasing farmers’ income not only constrains environmental efficiency improvement within the local province but also generates negative spatial spillovers on the environmental efficiency of neighboring provinces. This finding fills a gap in previous research that overlooked the impact of spatial interactions. Consequently, have put forward policy recommendations to promote tech sharing, pollution co-control and differentiated governance. Our research provides novel insights and a decision-making basis for achieving coordinated progress between agricultural economic growth and ecological protection.

 increasing farmers’ income environmental efficiency spatial agglomeration spatial spillover effect Chinese province section-at-acceptance Agricultural and Food Economics Introduction

In recent years, the Chinese economy has transitioned from high-speed growth to high-quality development, accompanied by a continuous increase in rural residents’ income (Luo et al., 2020). According to survey data from the National Bureau of Statistics of China, the per capita disposable income of rural residents reached ¥17,131.5 in 2020, representing an increase of ¥1,110.8 compared to 2019. However, enterprises characterized by high energy consumption, high pollution, and high emissions (“three-high” enterprises) persist in China. Concurrently, environmental pollution and excessive resource consumption remain severe issues. These resultant problems seriously hinder the green development of the socio-economy, exacerbating the increasingly prominent tension between ecological civilization construction and economic growth (Zhang et al., 2022). Against this backdrop, investigating the impact of increasing farmers’ income on environmental efficiency, particularly the spatial agglomeration effects of provincial environmental efficiency in China, holds significant importance. Such research can provide an empirical basis for proactively improving institutional policies and enhancing the green transformation of the socio-economy.

Farmers’ income remains an active area of academic and policy research (e.g., Wang et al., 2022). Geographically, studies primarily operate at either national or regional scales. National-scale analyses typically examine long-term temporal trends in farmers’ income (e.g., He et al., 2021; Lu and Li, 2021; Liu et al., 2020; Chen et al., 2024), while regional studies focus on localized income enhancement strategies (e.g., Ceballos et al., 2020; Irvan and Yuliarmi, 2019; Bhuiyan et al., 2022). Thematically, research predominantly identifies determinants of farmers’ income. Education level is widely recognized as a critical factor (Lockheed, 1987; Panda, 2015). Other significant drivers include digital economy development (Li and Jiang, 2023), accelerated urbanization (Huang et al., 2021), agricultural industrial restructuring (Wang et al., 2024), and human capital investment (Djomo and Sikod, 2012).

Global warming represents a long-standing and urgent challenge for policymakers worldwide. This phenomenon is accelerating alongside global economic expansion (Zhu et al., 2021), creating mounting environmental pressures through persistent growth trajectories (Chishti and Sinha, 2022). Consequently, maintaining high environmental quality is imperative for achieving sustainable development at both regional and global scales (Chishti et al., 2021). Furthermore, trade dynamics and financial development have been identified as critical determinants of environmental outcomes (Salam et al., 2025). Environmental efficiency quantifies a society’s capacity to utilize ecological resources effectively. Current research employs diverse perspectives, predominantly utilizing Data Envelopment Analysis (DEA) under the assumption of constant returns to scale (e.g., Farrell, 1957; Charnes et al., 1984; Song et al., 2012; Chen and Jia, 2017). However, the accuracy of calculated environmental efficiency values can be compromised by stochastic factors, including environmental fluctuations and policy shifts. Common environmental pressures stem from energy consumption and excessive emissions of pollutants like SO₂ and CO₂ (Shi et al., 2023). Furthermore, incorporating such environmental variables as undesirable outputs within efficiency models typically results in lower regional average environmental efficiency scores, reflecting the true cost of pollution (Zheng et al., 2021). In addition, empirical studies examining industrial production and environmental efficiency across sectors reveal a distinct spatial pattern in China: higher environmental efficiency levels in the eastern regions and lower levels in the west (Li et al., 2021). The development of high-tech industries also demonstrates a substantial positive effect on enhancing regional environmental efficiency (Peng et al., 2022a).

While the existing research provides a valuable theoretical foundation and empirical insights into farmers’ income and environmental efficiency, two significant gaps remain. First, studies specifically examining the impact of increasing farmers’ income on environmental efficiency are notably scarce. Second, research addressing the spatial dimensions of provincial environmental efficiency, particularly its spillover effects, is underdeveloped. Consequently, this study investigates the spatial spillover effects of regional increasing farmers’ income on provincial environmental efficiency. This focus is critical within China’s dual objectives of enhancing rural livelihoods and advancing green socio-economic development. To achieve this, we employ a spatial Durbin model and provincial panel data from China to analyze the spatial linkages between increasing farmers’ income and environmental efficiency. The marginal contributions of this study are twofold. First, it advances the theoretical framework for socio-economic green development by synthesizing three foundational theories: the Environmental Kuznets Curve, externality theory, and path dependence. This integrated lens provides a novel mechanism to analyze the complex interplay between increasing farmers’ income and provincial environmental efficiency. Second, it generates actionable insights for revitalizing rural economies, offering evidence-based pathways to simultaneously boost farmers’ incomes, enhance rural vitality, and promote environmentally sustainable development practices.

 Theoretical analysis and research hypotheses

As a vital component of the national economy, the rural economy has undergone significant advancement, driven by improvements in financial infrastructure, service systems, and the diversification of financial products. Rural financial development facilitates the crucial transformation of rural savings into productive investments, thereby stimulating farmers’ income growth (Kambali and Panakaje, 2022). Consequently, fostering the high-quality development of the rural economy and augmenting farmers’ incomes constitutes a core imperative for China’s green ecological transition and sustainable economic development (Chi et al., 2021).

The development of the rural economy has increased farmers’ income. This, in turn, has spurred consistent annual growth in investment in agricultural infrastructure, improving production conditions and boosting funding for agricultural technology research and promotion. Governments provide direct financial subsidies or interest subsidies to farmers cultivating high-quality seeds, purchasing agricultural machinery, and constructing high-standard farmland (Góral et al., 2015). Collectively, these measures have promoted agricultural technological progress, enhanced mechanization levels, and accelerated the sector’s green transformation (Peng et al., 2022b). However, alongside these income gains, rural industrial development characterized by high inputs, high energy consumption, and high pollution has inflicted severe damage on the local ecological environment. This has resulted in a rural economic development pattern marked by high economic growth but low ecological quality. The Environmental Kuznets Curve (EKC) posits that during the initial stages of economic development, agricultural intensification increases resource consumption and pollution emissions (Grossman and Krueger, 1993). When farmers’ incomes are low, agricultural production primarily expands in a survival-driven manner, leading to deteriorating environmental efficiency as income rises. Path dependence theory further suggests that following income growth, farmers are more likely to invest in conventional intensive technologies—familiar and offering higher short-term returns—rather than adopting eco-friendly alternatives requiring greater knowledge and skills (Song et al., 2022). This path dependence can lock technology into pollution-intensive pathways (Cecere et al., 2014), thereby reducing environmental efficiency. Additionally, the externalization of environmental costs means farmers do not bear the full cost of the pollution or resource depletion they cause. Consequently, when incomes rise, the private benefits of expanded production for farmers exceed the associated social costs, promoting excessive utilization of environmental resources (Muñoz, 2020). This dynamic inhibits improvements in local environmental efficiency. Furthermore, rural enterprises often operate with outdated equipment, simplistic production methods, and limited technological investment. These factors constrain progress in ecological environment development. Amid increasingly fierce market competition, local governments frequently prioritize economic growth—overemphasizing its role in driving farmers’ income—at the expense of environmental protection (Duanmu et al., 2018). This focus results in inadequate environmental infrastructure, thereby inhibiting improvements in local environmental efficiency.

According to the first law of geography, economic phenomena and factor flows exhibit strong spatial interdependence, with neighboring regions experiencing significant spillover effects (Tobler, 2004). When planning rural economic development, local governments often fail to account for the region’s actual conditions. This unscientific planning drives unsustainable exploitation of surrounding resources, resulting in extensive land idling and waste (Liotta et al., 2020). Furthermore, insufficient environmental awareness during development—coupled with an excessive focus on economic scale and profits—has degraded pre-existing ecosystems in adjacent areas. Such damage is not only temporally irreversible but also undermines regional environmental carrying capacity. Concurrently, inadequate planning foresight has led to disorganized residential layouts and reckless mineral extraction in development zones. These practices accelerate rural environmental deterioration and create irreconcilable tensions between ecological preservation and economic growth. Based on this analysis, we propose Hypotheses 1 and 2:

Hypotheses 1: The increasing farmers’ income is not conducive to promoting the development of the ecological environment, that is, the increasing farmers’ income can significantly inhibit the improvement of environmental efficiency in the local area.

Hypotheses 2: The spatial spillover effect of increasing farmers’ income has a significant negative impact on the environmental efficiency of neighbouring areas.

 Model construction Spatial autocorrelation test

The commonly used methods for testing spatial correlation in economic activities are Global Moran’s I and Local Moran’s I. Global Moran’s I examines the overall spatial differences and correlations across regions, indicating the degree of spatial agglomeration. Local Moran’s I, in contrast, examines the differences and spatial correlations between each region and its neighboring areas, revealing local spatial heterogeneity. The specific expression for Global Moran’s I is as follows:

Moran′sI=n∑i=1n∑j=1nwij(xi−x¯)(xj−x¯)s2∑i=1n∑j=1n(xi−xj)2

In Equation 1, s2=1n∑i=1n(xj−x¯)2,x¯=1n∑i=1nxi, i,j = 1,2,3…n. Xi represents the observed value for region i, and n denotes the total number of regions. The spatial weight matrix is denoted by Wij. When Moran’s I is greater than 0, it indicates that the measured variable has a spatial positive correlation. The correlation increases with the increase of Moran’s I. When Moran’s I is less than 0, indicating that the measured variable has a spatial negative correlation. If Moran’s I is equal to 0, then there is no spatial autocorrelation in the measured variable.

The Moran scatter plot is widely employed to illustrate local Moran’s I values. In this plot, the first quadrant corresponds to high-high (HH) clustering areas, indicating that both the local area and its surrounding regions exhibit high environmental efficiency. The second quadrant represents low-high (LH) clustering areas, where the local environmental efficiency is low, while that of the surrounding areas is relatively high. The third quadrant denotes low-low (LL) clustering areas, characterized by low environmental efficiency in both the local area and its surroundings. The fourth quadrant corresponds to high-low (HL) clustering areas, signifying high local environmental efficiency amidst surrounding regions with low environmental efficiency.

 Spatial weight matrix

When calculating Moran’s I, a spatial weight matrix must be defined. Following the methodology of Janatabadi and Ermagun (2024), this study employs an adjacency-based spatial weight matrix. The matrix, denoted as W, is constructed such that:

Wij={1,There is an adjacent boundary between regioniand regionj;0,There isnoadjacent boundary between regioniand regionj.

In Equation 2, W represents the proximity relationship among regions. When W = 1, it indicates that the region has adjacent boundaries. If there are no adjacent boundaries, W = 0. all elements on the main diagonal are 0.

 Spatial panel measurement model

Spatial econometric models introduce spatial factors into traditional regression models, mainly including spatial error models, spatial lag models, and spatial Durbin models. The general form of a spatial panel model is as follows:

LnGeeit=b+β1Lnincomeit+β2∑j=1nXitj+θ1∑j=1nWijLnGeeit+θ2∑j=1nWijLnincomeit+θ3∑j=1nWijXitj+μi+vi+εi;

In Equation 3, Geeit represents the environmental efficiency of province i in year t. incomeit refers to the income increase of farmers in province i in year t. X represents a series of control variables. W represents the spatial weight matrix. Β, θ is the parameter to be estimated. μ and v represent spatial and temporal effects, respectively. b is a constant term, and ε is a random interference term.

The spatial Durbin model is one of the important models in spatial econometrics, which extends the spatial lag model (SLM) and spatial error model (SEM). Spatial lag model only considers the spatial lag term of the dependent variable, while spatial error model considers the spatial correlation of the error term. The spatial Durbin model not only includes the spatial lag term of the dependent variable but also includes the spatial lag term of the explanatory variable. This means that the spatial Durbin model can not only analyze the impact of farmers’ income increase in a region on local environmental efficiency (i.e., direct effect) but also explore how farmers’ income increase in that region affects the environmental efficiency of adjacent regions through spatial transmission (i.e., indirect effect or spatial spillover effect).

 Spatial effect analysis model

According to the theory of spatial econometric models, the dependent variable of a region is not only influenced by the independent variable of that region (i.e., direct effect), but also by the independent and dependent variables of other regions (i.e., indirect effect). This article refers to the partial differential method in existing literature (Lesage and Pace, 2008), and writes the spatial econometric model in vector form:

Yt=(I−δW)−1(βXt+WXtγ)+(I−δW)−1)ε∗

In Equation 4, Yt represents the provincial environmental efficiency. ε∗ is the error term. Deriving Equation 4 with the k-th explanatory variable as the independent variable, then obtain the partial differential matrix:

[∂Y∂X1k.∂Y∂XNk]=[∂Y1∂X1k⋯∂Y1∂XNk.⋯.∂YN∂X1k⋯∂YN∂XNk]=(I−δW)−1[βkw12λk⋯w1Nλkw21λkβk⋯w2Nλk..⋯.wN1λkwN2λk⋯βk]

In the matrix on the right, the direct effect represents the mean of all elements on the main diagonal, while the indirect effect represents the mean of all elements on the non main diagonal.

 Variable selection and data sources Explained variables

Environmental efficiency is denoted as Gee. This article uses the DEA method to calculate environmental efficiency considering energy input and environmental pollution output. Tone (2002) improved the information of effective decision-making units based on traditional measurement methods and proposed a super-efficient SBM model. The formula is as follows:

minρ=1m∑i=1mx¯xik1q1+q2(∑i=1q1ya¯yaIk+∑i=1q2yb¯yrkb)s.t.{x¯≥xk;x¯≥∑j=1,j≠knxijμj;∑j=1,j≠knyrjaμj≤ya¯≤∑j=1,j≠knyIjaμj;ya¯≤yak;yb¯≥ybk;μj≥0;i=1,2,⋯,m;j=1,2,⋯,n;I=1,2,⋯,q1;r=1,2,⋯,q2;θ

In Equation 6, all n evaluation units contain input quantity m, expected output quantity q1, and unexpected output quantity q2. Correspondingly, the elements of the input matrix, expected output matrix, and unexpected output matrix are represented by x, ya, and yb, respectively. This model considers both the reduction of input and the increase of output, avoiding the choice of measurement perspective. However, the environmental efficiency measured by the SBM model is a static analysis that can only reflect the relative relationship between various production units and production boundaries. Therefore, based on the method proposed by Chung et al. (1997), we introduces the concept of intertemporal dynamics and draws on the idea of geometric mean to construct ML indices for adjacent references t and t + 1 for two consecutive years as follows:

MLtt+1={[1+D→t(xt+1,yt+1,dt+1,gt+1)][1+D→t(xt,yt,dt,gt)]×[1+D→t+1(xt+1,yt+1,dt+1,gt+1)][1+D→t+1(xt,yt,dt,gt)]}12

In Equation 7, D represents the production unit DMU, and x, y, d, and g represent production input, expected output, unexpected output, and direction vector, respectively. The indicator processing is as follows:

(1) Input indicators

About the capital input, we used the perpetual inventory method to estimate capital stock. The specific calculation formula is Kit=Kit−1(1−δit)+Iit/Pit. Among them, Kit and Kit−1 represent the capital stock of i province in years t and t−1. δit,Iit,Pit represent the capital depreciation rate, investment price index, and total investment amount for year t in province i. Labor input is expressed as the sum of the year-end employment figures of urban units, private enterprises, and individuals in each province. Energy input is measured by selecting the total energy consumption.

(2) Output indicators

Expected output is represented by regional Gross Domestic Product (GDP). Undesirable outputs (unexpected outputs) are measured using the “industrial three wastes” framework, encompassing industrial wastewater discharge, industrial waste gas emissions, and industrial solid waste generation (Li et al., 2020). Among these, the predominant pollutants are sulphur dioxide (SO₂) in waste gas and chemical oxygen demand (COD) in wastewater (Wang et al., 2010). Therefore, SO₂ and COD were selected as proxy indicators for industrial waste gas emissions and industrial wastewater emissions, respectively. The complete set of input and output variables used in this study is detailed in Table 1.

Evaluation index system of environmental efficiency in Chinese provinces.

Indicator Specific indicator Indicator description
Input indicator Capital investment Physical capital stock
Human input Labor force
Energy input Energy consumption
Expected output Economic output GDP
Undesirable output Wastewater discharge COD discharge in industrial wastewater
Exhaust emissions SO2 emissions in industrial waste gas
Waste discharge Industrial solid waste generation

It is worth noting that the ML productivity index represents the rate of change in green total factor productivity, rather than green total factor productivity. Therefore, we used the month on month method to process the ML index in order to obtain the environmental efficiency (Qiu et al., 2008).

 Explanatory variables

The core explanatory variable in this study is increasing farmers’ income. Disposable income per capita of rural residents serves as a reliable indicator of changes in their income level, living standards, and consumption capacity (Xu and Xue, 2022). Therefore, we utilize rural residents’ disposable income per capita as the measure of income growth. This variable is denoted as income.

 Control variables

To isolate the impact of increased farmers’ income on environmental efficiency, we introduce control variables capturing six key dimensions: (1) Economic Development Level (pgdp): Economic development provides the foundational resources for environmental protection investments, pollution governance, and green technology adoption (Zhang et al., 2019). (2) Openness to Foreign Investment (open): Foreign direct investment may facilitate technology transfer to less developed regions, potentially improving ecological management practices (Cole et al., 2017). (3) Energy Consumption Structure (ec): The dominance of coal in the energy mix significantly affects emission intensity, particularly under energy conservation and emission reduction constraints (Muhammad et al., 2022). (4) Government Environmental Engagement (gov): Public expenditure on environmental governance exerts substantial influence on regional pollution control and ecological conservation (Gonçalves et al., 2021). (5) Industrial Structure (ind): Higher proportions of secondary industry correlate negatively with environmental efficiency due to increased resource intensity and pollution (Zhou et al., 2019). (6) Population Pressure (ps): Large population scales intensify regional environmental burdens through elevated energy demand and waste generation (Song et al., 2013). Specific indicator descriptions are shown in Table 2.

List of variables.

Variables Variable name Symbol Variable definition
Explained variable Environmental efficiency Gee Estimation of Unexpected Output Super SBM Model
Explanatory variable Increasing farmers’ income income Disposable income of rural residents
Control variable Level of economic development pgdp Per Capita GDP
Degree of openness to the outside world open Total amount of foreign investment/GDP
Energy consumption structure ec Total coal consumption/Total energy consumption
Degree of government intervention gov Fiscal expenditure/Real GDP
Industrial structure ind Value added of the secondary industry/GDP
Population size ps Total population at the end of the year
Sample selection and data sources

China’s vast territory encompasses significant disparities in resource endowments, economic development levels, and agricultural practices across its provinces. The 30 provinces studied encompass economically developed coastal areas in the east, major agricultural production zones in the central region, and ecologically fragile areas in the west. This sampling strategy comprehensively captures the relationship between farmer income growth and environmental efficiency across diverse development gradients, enhancing the generalizability of the research findings.

To ensure data comparability and support comprehensive analysis of core mechanisms, we utilized panel data from 30 Chinese provinces spanning 2010 to 2019. This dataset provides robust empirical foundations, demonstrating scientific validity, generalizability, and practical relevance. Data were predominantly sourced from official Chinese statistical publications, including the China Statistical Yearbook, China Environmental Statistical Yearbook, China Energy Statistical Yearbook, China Population and Employment Statistical Yearbook, National Bureau of Statistics and provincial statistical yearbooks and statistical bulletins. Minor missing values were addressed through interpolation.

 Empirical result analysis

This article is based on the regional division of China’s traditional economic belt, dividing China into eastern, central, and western regions. The eastern region includes provinces such as Beijing, Tianjin, Hebei, Liaoning, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong, and Hainan. The central region includes provinces such as Shanxi, Heilongjiang, Jilin, Anhui, Jiangxi, Henan, Hubei, and Hunan. The western region includes provinces such as Inner Mongolia, Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Shaanxi, Gansu, Qinghai, Ningxia, and Xinjiang.

Spatial distribution characteristics of environmental efficiency in China

Utilizing ArcGIS 10.7, we visualized provincial environmental efficiency trends (Figure 1). Overall, most provinces exhibited an upward trajectory during the study period. Fujian Province demonstrated significant improvement between 2010 and 2015. This progress stemmed from policy-driven reductions in resource exploitation and minimization of surface ecological damage, substantially enhancing its ecological environment. Economically advantaged regions like Beijing and Jiangsu actively developed eco-environmental protection technologies. Leveraging their financial and human capital, they alleviated regional ecological issues through improved resource utilization and reduced pollutant emissions. From 2015 to 2019, environmental efficiency improved markedly in most provinces. However, significant disparities persisted, with central and western provinces like Xinjiang, Guangxi, Hunan, and Heilongjiang exhibiting notably lower efficiency levels in 2019. The main reason is that, on the one hand, central and western provinces possess less robust economic development capacity compared to eastern coastal cities, limiting investment in and access to advanced eco-environmental technologies, funding, and talent. On the other hand, the diffusion and integration of environmental policies across regions remain inadequate. Current policy implementation often remains fragmented, focusing predominantly on localized ecological issues with insufficient cross-regional coordination.

Spatial distribution pattern of environmental efficiency in various provinces in China.

Three maps of China show the spatial distribution of some data from 2010 to 2019. Each map uses shades of purple to represent levels from low to high, with a legend indicating "low level" to "high level." The maps display changes over time: 2010-2011, 2014-2015, and 2018-2019. Major regions are labeled, and an inset shows the South China Sea Islands. Spatial correlation test Global spatial correlation test

This study employs global Moran’s I to analyze the spatial autocorrelation of environmental efficiency across Chinese provinces (Table 3). Results indicate positive Moran’s I values (greater than 0) for inter-provincial environmental efficiency from 2011 to 2019, with statistical significance achieved in most years. This demonstrates significant positive spatial autocorrelation, implying that environmental efficiency improvements in neighboring provinces positively influence local provincial efficiency. These findings underscore the importance of incorporating spatial spillover effects into further analysis.

Global autocorrelation test of the mean environmental efficiency index of each province.

Year I E(I) SD(I) Z p value
2010–2011 −0.031 −0.034 0.081 0.039 0.484
2011–2012 0.045 −0.034 0.084 0.944 0.172
2012–2013 0.080 −0.034 0.103 1.106 0.134
2013–2014 0.213 −0.034 0.112 2.213 0.013
2014–2015 0.118 −0.034 0.110 1.381 0.084
2015–2016 0.240 −0.034 0.122 2.246 0.012
2016–2017 0.137 −0.034 0.123 1.391 0.082
2017–2018 0.133 −0.034 0.123 1.367 0.086
2018–2019 0.153 −0.034 0.123 1.519 0.064
Local spatial correlation test

While global Moran’s I captures average spatial autocorrelation across the study area, it fails to reveal heterogeneity among regional units or dynamic evolutionary patterns in environmental efficiency. To address these limitations, we construct Moran scatter plots for key intervals (2010–2011, 2014–2015, and 2018–2019), enabling localized spatial analysis and temporal evolution tracking (Figure 2; Table 4).

Moran scatter plots of environmental efficiency in 30 provinces of China from 2010 to 2011, 2014 to 2015, and 2018 to 2019.

Three Moran scatterplots display spatial autocorrelation for different time periods. Top left: 2010-2011 with Moran's I of -0.031. Top right: 2014-2015 with Moran's I of 0.118. Bottom: 2018-2019 with Moran's I of 0.153. Points represent locations identified by initials and are distributed along Z-axes with trend lines.

Moran scatter plot distribution.

Year 2010–2011 2014–2015 2018–2019
Quadrant I YN, QH, GZ, SN, GS, HE, JX, HA SH, IM, JX, HB, GZ, HE, HN SH, JX, AH, GD, YN, IM, GZ
Quadrant II SH, JS, ZJ, CQ, JL, HL, AH, SD, FJ JL, ZJ, BJ, TJ, JS, SD, FJ, AH, HA JL, TJ, SD, HE, CQ, BJ, LN, HB, ZJ, FJ, HL
Quadrant III HI, HN, HB, BJ, NX HI, YN, SX, GD, SC, SN, NX, QH, GS, XJ SX, SN, HI, QH, NX, GS, XJ
Quadrant IV LN, GX, CQ, HL LN, GX, CQ, HL JX, HA, GX, SC, HN

Specifically, from 2010 to 2011, provincial environmental efficiency exhibited significant positive spatial autocorrelation. Provinces in Quadrant I (high-high cluster) included economically advanced coastal regions (e.g., Zhejiang, Jiangsu, Shanghai, Fujian), indicating localized hotspots of high environmental efficiency. Conversely, provinces like Hainan and Ningxia occupied Quadrant III (low-low cluster), reflecting regions with persistently low efficiency surrounded by similar neighbors. From 2014 to 2015, with the improvement of economic development level, communication among provinces became increasingly frequent, making production factors more mobile among provinces. The environmental efficiency scores of most provinces in China are in Quadrant I and Quadrant III, showing a clear spatial positive correlation. From 2018 to 2019, the spatial agglomeration trend of green energy efficiency in various provinces in China has become increasingly significant, with most provinces and cities still in Quadrant I and Quadrant III. The Moran scatter plots reveal substantial quadrant transitions among provinces from 2010 to 2019. Notably, the dominance of eastern coastal provinces in Quadrant I and western provinces in Quadrant III persisted throughout the period. This persistent divergence is primarily attributable to the eastern coastal region’s inherent advantages: strategic geographical positioning (e.g., maritime access), robust economic foundations, advanced infrastructure, and mature technological and educational ecosystems. Consequently, while inland provinces maintained resource-intensive (high-input, high-consumption) production models, eastern regions leveraged national policy frameworks to pioneer industrial restructuring and economic transformation toward technology-intensive and sustainable paradigms. Eastern regions operate near the production possibility frontier in environmental efficiency, maintaining national leadership. Conversely, western provinces centered on Xinjiang, Qinghai, Ningxia, and Gansu persistently occupy Quadrant III (low-low clusters). This is mainly because western regions have relatively backward economic development, and their production methods are more prone to environmental pollution and energy waste. In the long run, they lack technological innovation capabilities and do not pay enough attention to environmental and energy issues, resulting in low environmental efficiency.

 Model recognition and fitting results

The previous Moran’s I test results indicate that there is a spatial correlation in environmental efficiency among different provinces. Therefore, if LM and Robust LM tests indicate the need for spatial econometric analysis, further Wald test, LR test, and Huasman test can be used to determine the specific spatial econometric model and its specific form. Table 5 shows the inspection results.

Model recognition test results.

Test Statistic p-value
Moran’s I (error) 9.975 0.000
Lagrange multiplier (error) 91.856 0.000
Robust Lagrange multiplier (error) 6.325 0.012
Lagrange multiplier (lag) 89.495 0.000
Robust Lagrange multiplier (lag) 3.963 0.047
LR_spatial (lag) 51.48 0.000
LR_spatial (error) 76.68 0.000
Wald_spatial lag 53.70 0.000
Wald_spatial error 76.51 0.000
Hausman 17.08 0.017

In Table 5, the test results of Robust LM error and Robust LM lag indicate that spatial econometric models should be used for analysis. The results of LR test and Wald test both indicate the use of spatial Durbin model. In addition, the Hausman test results suggest that a fixed effects spatial Durbin model should be chosen, and the regression results are shown in Table 6.

Results of different estimation methods for spatial Durbin model.

Variables Time fixed effects Individual fixed effects Two-way fixed effects WX Time fixed effects Individual fixed effects Two-way fixed effects
Lnincome −0.061(0.373) −0.337(0.205) −0.728***(0.001) Lnincome 0.201(0.168) 0.473*(0.082) −0.907**(0.011)
pgdp −0.012***(0.000) −0.017**(0.047) −0.017**(0.014) pgdp −0.015**(0.014) −0.030*(0.079) −0.034**(0.023)
open −0.172***(0.000) −0.104(0.152) −0.113*(0.063) open 0.147*(0.095) 0.648***(0.000) 0.452***(0.000)
ec −0.175***(0.000) −0.188**(0.034) −0.266***(0.000) ec 0.113*(0.097) −0.402**(0.046) −0.606***(0.001)
Lngov −0.310***(0.000) 0.257**(0.039) 0.360***(0.002) Lngov 0.184(0.130) 0.195(0.305) 0.598**(0.034)
Lnind −0.018(0.793) −0.504***(0.000) −0.074(0.576) Lnind 0.370***(0.006) −0.151(0.558) 0.906***(0.003)
Lnps −0.070***(0.004) −0.332(0.435) −0.980***(0.007) Lnps −0.159***(0.002) −2.257***(0.003) −4.354***(0.000)
rho 0.264***(0.001) 0.411***(0.000) −0.309***(0.001) sigma2_e 0.021***(0.000) 0.013***(0.000) 0.009***(0.000)
R2 0.079 0.001 0.064 N 300 300 300
AIC −269.4465 −395.3341 −516.5632 Log_L 150.7232 213.6670 274.2816

The p-value is in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1. Same below.

Table 6 shows that the spatial autoregressive coefficients of the spatial Durbin model under individual, temporal, and two-way fixed effects all passed the 1% significance level. Among them, in the statistical values of the AIC criterion, the absolute value of the bidirectional fixed effects spatial Durbin model is the largest, and its Log_L value is also the largest. Based on the size of the dispersion (sigma_2), it can be seen that compared to individual and time-fixed effects, the spatial Durbin model under bidirectional fixed effects has a significantly better fitting degree. Therefore, this article ultimately chooses the spatial Durbin model with bidirectional fixed effects for regression analysis.

 Decomposition of spatial effects

According to Lesage and Pace’s (2008) study, regression coefficients cannot be directly used to measure the influence of explanatory variables and dependent variables. Therefore, it is necessary to use the partial differential method of regression models to decompose the effect of explanatory variables on dependent variables. Table 7 presents the decomposition results of spatial effects.

Spatial effect decomposition of the spatial Durbin model.

Variables Direct effect Indirect effect Total effect
Lnincome −0.673***(0.007) −0.583*(0.079) −1.255***(0.000)
pgdp −0.016**(0.030) −0.022*(0.082) −0.038***(0.001)
open −0.141**(0.024) 0.406***(0.000) 0.265***(0.001)
ec −0.229***(0.004) −0.444***(0.005) −0.674***(0.000)
Lngov 0.326***(0.004) 0.404*(0.073) 0.730***(0.002)
Lnind −0.132(0.287) 0.776***(0.001) 0.644**(0.021)
Lnps −0.707*(0.061) −3.380***(0.000) −4.087***(0.000)

In Table 7, the decomposition results indicate that there are differences in the spatial impact and mechanism of farmers’ income increase on provincial environmental efficiency. Under the direct effect, the regression coefficient of farmers’ income increase on environmental efficiency is −0.673, passing the significance level of 1%. For every 1% increase in farmers’ income, the environmental efficiency of the province decreases by 0.673%. The reason is that while farmers pursue economic benefits, they lack corresponding awareness of ecological environment protection. In their daily lives, household garbage is indiscriminately piled up and sewage is discharged, seriously polluting the ecological environment and not conducive to the improvement of local environmental efficiency. Hypothesis 1 has been validated. Under the indirect effect, the regression coefficient of farmers’ income increase on environmental efficiency is −0.583, passing the significance level of 10%. It indicates that for every 1% increase in farmers’ income, it will lead to a 0.583% reduction in environmental efficiency in the surrounding provinces, indicating that the spatial spillover effect caused by the increase in farmers’ income has a significant hindering effect on the improvement of environmental efficiency in neighboring areas. The reason is that there may be a siphon effect during the development of agricultural economy. Affected by the local agricultural economic environment, the income level of farmers will develop first in areas with high factor returns, promoting the inflow of talent, capital and other factors. This polarization phenomenon may be detrimental to the development of the surrounding ecological environment, which verified the Hypothesis 2.

Among the controlled variables, the direct and indirect effects of economic development level (pgdp) on environmental efficiency are significantly adverse. In the early stages of economic development, regional governments typically prioritize pursuing high economic growth rates and invest a substantial amount of resources into industrialization and urbanization processes. This extensive development model is often accompanied by excessive energy consumption and massive emissions of pollutants, directly leading to a decline in environmental efficiency in the local area. In addition, the flow of factors among regions will encourage surrounding areas to relax environmental regulatory standards in order to attract resources, thereby exacerbating environmental pollution in neighboring areas and leading to a suppressive effect of economic development on the environmental efficiency of neighboring areas.

The direct effect of openness (open) on environmental efficiency is significantly adverse, while the indirect effect is significantly positive. Under the logic of GDP competition among local governments, some regions actively reduce the intensity of environmental regulations to attract foreign investment, resulting in the influx of high-pollution-intensive industries and a decline in local environmental efficiency. However, foreign-invested enterprises with advanced clean production technologies generate technology spillover effects through industrial supply chains. This technology diffusion transcends administrative boundaries, enabling neighboring provinces and industry enterprises to rapidly adopt and imitate, ultimately leading to an increase in environmental efficiency in the surrounding areas.

The direct and spillover effects of changes of energy consumption structure (ec) on environmental efficiency are both significantly adverse. On the one hand, coal-rich areas have formed a “resource curse” due to the lock-in of resource endowments, making it challenging to achieve industrial structural transformation in the short term. On the other hand, the ambiguity of responsibilities and interests in environmental governance among regions makes it difficult to effectively coordinate pollution control measures, leading to the continuous erosion of environmental efficiency in adjacent areas.

The improvement of government policies (gov) has a significant inhibitory effect on the improvement of environmental efficiency in both the local and neighboring areas. A possible reason is that, in order to attract investment and promote employment, local governments often adopt supportive measures, such as a policy tilt towards high-polluting enterprises, which directly exacerbates local environmental pollution and reduces environmental efficiency. In addition, local governments often only consider the economic interests of the local area when formulating industrial policies, neglecting the coordination of environmental governance with surrounding areas. This results in inconsistent standards for air pollution control and ultimately leads to a dilemma in improving overall environmental efficiency.

The direct effect of industrial structure (ind) on environmental efficiency is not significant, but the indirect effect is significant and positive. Industrial restructuring, led by the secondary industry, makes it challenging to improve local environmental efficiency in the short term significantly. However, with the rise of the environmental service industry, the level of regional environmental governance technology has been improved. This cross-regional industrial collaboration and technology spillover have optimized the production mode and resource allocation in surrounding areas, effectively improving environmental efficiency.

The direct and indirect effects of population size (ps) on environmental efficiency are significantly adverse. The reason is that population expansion directly exacerbates resource consumption and pollutant emissions, which in turn further increases the environmental burden of the region and reduces environmental efficiency. In addition, the spread of lifestyle changes caused by population mobility among regions can lead to a synchronous spillover of environmental pressure, resulting in significant adverse effects of population growth on environmental efficiency in surrounding areas.

 Robustness test

To ensure the reliability of the research conclusions in this article, the following robustness tests were conducted (Table 8).

Robustness test.

Variables Replace control variables Exclude municipalities directly under the central government Replace the spatial weight matrix
Direct effects Indirect effect Total effects Direct effects Indirect effect Total effects Direct effects Indirect effect Total effects
Lnincome −0.645***(0.008) −0.943***(0.003) −1.588***(0.000) −0.908***(0.000) −1.114***(0.002) −2.023***(0.000) −0.5118**(0.036) −0.8726***(0.005) −1.3844***(0.000)
pgdp −0.013*(0.088) −0.041***(0.002) −0.053***(0.000) −0.000(0.493) 0.000(0.525) 0.000(0.774) −0.0253***(0.000) −0.0090(0.516) −0.0343**(0.031)
open −0.021*(0.070) 0.055***(0.001) 0.034**(0.010) −0.025(0.733) 0.322***(0.004) 0.297***(0.002) −0.1301**(0.030) 0.2810***(0.001) 0.1509*(0.077)
ec −0.234***(0.003) −0.277*(0.074) −0.511***(0.000) −0.293***(0.000) −0.637***(0.000) −0.930***(0.000) −0.2988***(0.000) −0.2212(0.123) −0.5200***(0.001)
Lngov 0.331***(0.001) 0.549***(0.005) 0.879***(0.000) 0.553***(0.000) 0.666***(0.010) 1.219***(0.000) 0.4211***(0.000) 0.3280*(0.088) 0.7492***(0.001)
Lnind −0.240*(0.055) 0.450*(0.066) 0.210(0.450) −0.014(0.911) 1.242***(0.000) 1.228***(0.000) −0.1823(0.171) 0.0170(0.943) −0.1652(0.587)
Lnps −0.761**(0.040) −3.280***(0.000) −4.041***(0.000) −1.542***(0.000) −3.845***(0.000) −5.388***(0.000) −0.7013*(0.073) −1.8310***(0.003) −2.5324***(0.000)
Replacing control variables

In order to avoid differences in conclusions due to the selection of control variables, this article refers to existing research (Deng and Sun, 2022; Zeng et al., 2022). The degree of openness (open) uses the total import and export volume to replace the ratio of foreign direct investment to GDP, and the degree of government intervention (gov) uses the difference between fiscal expenditure and science and technology education to replace the proportion of fiscal expenditure to actual GDP.

 Exclude municipalities directly under the central government

Due to differences in the development of farmers’ income growth among different provinces in China. Therefore, to verify the universality of the research conclusions in this article, referring to the study by Zeng et al. (2022), panel data from Beijing, Tianjin, Shanghai, and Chongqing were excluded.

 Replacing the spatial weight matrix

Considering that the binary adjacency spatial weight matrix may overly simplify the interaction of space, we further utilize a nested spatial weight matrix that combines economic distance and geographical distance, whose elements are the product of each element in the two spatial weight matrices.

According to Table 8, after replacing controlling for the variables, excluding municipalities directly under the central government, and replacing the spatial weight matrix, the coefficients of the direct, indirect, and total effects of increasing farmers’ income on environmental efficiency remain significantly negative, consistent with the previous conclusion. From this, it can be seen that the econometric results of the previous model have good robustness.

 Endogenous treatment

Considering the possible causal relationship between farmers’ income increase and environmental efficiency, which will lead to endogeneity issues. Based on this, we use the first-order lagged term of farmers’ income increase as the instrumental variable, employs the two stage least square (2SLS) method to correct the estimation error caused by endogenous factors, and conducts correlation tests to provide explanations. The regression results are shown in Table 9.

Endogenous test results and treatment.

Variables First stage Second stage
lnincome LnGee
L.lnincome 0.811***(0.000)
incomehat −1.3802*** (0.000)
Control variables Yes Yes
Time fixed effect Yes Yes
Individual fixed effect Yes Yes
Constant 2.086***(0.003) 23.2234*** (0.000)
F 428.835
Hausman test 35.58
Prob>chi2 0.000

According to Table 9, the p-value of Hausman’s test is less than 0.05, indicating that the model has endogeneity issues. According to the results of the first stage regression, the regression coefficient of the first-order lagged term of the instrumental variable for increasing farmers’ income (L.lnincome) is significantly positive. The F-value is 428.835, which exceeds 10. It has passed the weak instrumental variable identification test and verified the rationality of the instrumental variable selection. Furthermore, in the second stage, the regression coefficient of the predicted value of farmers’ income increase (incomehat) on environmental efficiency is significantly negative. This is consistent with the estimated values in the previous section, thus proving the reliability of the conclusions obtained based on the spatial Durbin model with bidirectional fixed effects.

 Heterogeneity analysis

There are certain differences in the income levels of farmers among regions, and the impact of increasing farmers’ income on environmental efficiency may have gradient differences. Therefore, we examine the effect of increasing farmers’ income on environmental efficiency within the eastern, central, and western regions, based on China’s traditional economic belt division. The results are presented in Table 10.

Heterogeneity analysis results.

Variables Eastern region Central region Western region
Direct effect Indirect effect Total effect Direct effect Indirect effect Total effect Direct effect Indirect effect Total effect
Lnincome 0.6013*(0.050) 0.3931(0.216) 0.9943**(0.013) −1.9066**(0.017) 0.3169(0.708) −1.5897***(0.000) −0.3480(0.513) 0.2972(0.569) −0.0508(0.722)
pgdp −0.0805**(0.017) 0.0607(0.417) −0.0198(0.821) 1.0361***(0.000) −0.0245(0.918) 1.0117***(0.000) −0.0167**(0.026) 0.0052(0.786) −0.0116(0.593)
open 0.2459***(0.001) 0.5474***(0.000) 0.7933***(0.000) 0.5567*(0.080) −0.9094*(0.097) −0.3527(0.588) 0.1975(0.272) 0.3469(0.295) 0.5444(0.124)
ec −0.8290***(0.008) −1.2778**(0.034) −2.1068***(0.004) −0.1324(0.471) 0.5381(0.205) 0.4057(0.393) 0.2218**(0.036) −0.3399(0.125) −0.1180(0.619)
Lngov −0.5231***(0.005) 0.4561(0.225) −0.0670(0.894) 0.6259(0.110) −0.0768(0.883) 0.5491(0.334) −0.1055(0.549) −0.9237***(0.004) −1.0292***(0.001)
Lnind 1.1972***(0.000) 0.2613(0.690) 1.4585*(0.086) −0.8140***(0.002) −0.3760(0.320) −1.1900**(0.030) −0.4939**(0.011) 1.0054***(0.006) 0.5115(0.282)
Lnps 0.0262(0.973) −3.5595***(0.007) −3.5333**(0.024) −4.3334**(0.032) 6.9332***(0.004) 2.5998*(0.086) 2.4080***(0.000) 8.8594***(0.000) 11.2674***(0.000)

Table 10 reveals that in the eastern region, the direct effect of increasing farmers’ income on environmental efficiency is 0.6013, and the total effect reaches 0.9943, both demonstrating a statistically significant positive impact. This indicates that in this economically developed region, rising farm incomes positively drive local environmental efficiency. Conversely, in the central region, the direct effect is −1.9066 and the total effect is −1.5897, showing a significant negative impact. This suggests that income growth in the central region exerts an inhibitory effect on environmental efficiency. In the western region, the direct effect (−0.3480) and total effect (−0.0508) are both negative, but significantly weaker than those observed in the central region. This pattern implies that the relationship between farmers’ income and environmental efficiency in the west may be in a transitional phase.

This “positive in the east and negative in the west” dichotomy intuitively reflects the structural transformation of farmers’ income impacts along the regional development gradient. Furthermore, the existence of these gradient differences highlights that the relationship between increasing farmers’ income and environmental efficiency is not simply linear. Instead, it is moderated by multiple factors such as regional development stage, industrial structure, and infrastructure.

 Research conclusion and countermeasures suggestions Research conclusion

This article takes panel data from 30 provinces in China from 2010 to 2019 as samples. Based on spatial autocorrelation tests of provincial environmental efficiency levels, a spatial Durbin model is used to empirically analyze the spatial effects of increasing farmers’ income in various regions on provincial environmental efficiency. The empirical results indicate that, firstly, the environmental efficiency of each province shows a certain degree of agglomeration in space, but the agglomeration patterns vary. Among them, the high-high agglomeration areas of provincial environmental efficiency account for a large proportion, with a wide distribution in the eastern coastal areas. In contrast, the low-low agglomeration areas are mainly concentrated in the western regions, presenting an unreasonable polarization distribution.

This study utilizes panel data from 30 Chinese provinces spanning the period 2010–2019. Based on spatial autocorrelation tests conducted on provincial environmental efficiency levels, a spatial Durbin model was employed to empirically analyze the spatial effects of increasing farmers’ income on provincial environmental efficiency across different regions. The empirical results reveal, firstly, that provincial environmental efficiency exhibits significant spatial clustering, although the specific patterns vary. Specifically, provinces characterized by high-high clusters of environmental efficiency constitute a substantial proportion and are widely distributed across the eastern coastal areas. In sharp contrast, low-low clusters are predominantly concentrated in the western regions, indicating a distinct regional disparity in the spatial distribution. Secondly, increasing farmers’ income not only inhibits environmental efficiency improvement within the province itself but also exerts a significant negative spatial spillover effect on the environmental efficiency of neighboring provinces. Thirdly, the mechanisms and magnitudes through which factors—such as economic development level, degree of openness, coal consumption share, government policies, industrial structure, and population size—affect provincial environmental efficiency exhibit significant heterogeneity.

 Policy suggestions

Based on the above conclusions, it is recommended to take the following measures at the policy level.

Firstly, eastern coastal provinces should jointly create an environmental technology sharing platform. This platform should regularly share advanced pollution treatment methods and energy-saving equipment R&D results. Western provinces experiencing environmental efficiency low-low clusters should assign dedicated personnel to access this platform, acquire technologies tailored to local needs, and implement targeted incentives to foster growth of indigenous environmental protection industries.

Secondly, adjacent provinces should establish coordinated agricultural pollution prevention mechanisms, implementing regular cross-border environmental monitoring to address transboundary contamination from unsustainable farming practices. Concurrently, all participating provinces should clearly demarcate responsibilities and enhance governance coordination through interprovincial linkages to mitigate negative environmental impacts stemming from increasing farmers’ income.

Thirdly, provinces should develop tailored governance strategies accounting for differential impacts of local factors on environmental efficiency. Economically advanced provinces must prioritize transforming industries toward green high-end production. Regions with high foreign exposure should enhance environmental audit standards for foreign investment, while coal-dependent provinces need to accelerate clean energy transitions. All provinces should optimize environmental regulations based on industrial structure and population characteristics, implementing precision governance to improve local environmental efficiency.

 Ideas for future research

This paper offers a new idea for the study of the increasing increasing farmers’ income and environmental efficiency, as well as references for nations around the globe to use in developing disposable income of farmers and ecological economy.

However, several limitations warrant further investigation. Firstly, we construct a spatial weight matrix based solely on geographical proximity, overlooking economic and technological distance factors. Future studies should incorporate trade volumes and R&D investment gaps to better characterize interregional environmental efficiency linkages. Secondly, the provincial-level analysis of 30 Chinese administrative regions cannot capture county-level heterogeneity. Subsequent research should employ county-level panel data to examine how resource endowments mediate the farmer income-environmental efficiency relationship. Thirdly, while the 2010–2019 study period ensures data comparability and analytical coherence, it excludes the post-2020 period. It is difficult to reflect the dynamic changes of relevant variables after the COVID-19 epidemic. Subsequent research can extend the time series beyond 2020 to validate the robustness of the conclusions drawn in this study and delve deeper into the long-term effects brought about by the pandemic.

 Data availability statement

The datasets presented in this article are not readily available because the datasets used and analyzed during the current study available from the corresponding author on reasonable request. Requests to access the datasets should be directed to bingteng888@163.com.

 Author contributions

YC: Funding acquisition, Writing – review & editing, Conceptualization, Data curation, Formal analysis. XD: Conceptualization, Formal analysis, Funding acquisition, Writing – review & editing. BS: Conceptualization, Data curation, Methodology, Software, Writing – original draft.

 Funding

The author(s) declare that financial support was received for the research and/or publication of this article. The research was supported by the 2024 Guangdong Province Key Research Base for Humanities and Social Sciences in Ordinary Universities “Guangdong Rural Revitalization High Quality Development Grassroots Governance Innovation and Local Legislation Research Center” (2024WZJD018).

 Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

 Generative AI statement

The authors declare that no Gen AI was used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

 Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

 References BhuiyanM. A. DavitM. ZengX. ZhangZ. (2022). The impact of agricultural insurance on farmers’ income: Guangdong Province (China) as an example. PLoS One 17:e0274047. doi: 10.1371/journal.pone.0274047 CeballosF. KannanS. KramerB. (2020). Impacts of a national lockdown on smallholder farmers’ income and food security: empirical evidence from two states in India. World Dev. 136:105069. doi: 10.1016/j.worlddev.2020.105069 CecereG. CorrocherN. GossartC. OzmanM. (2014). Lock-in and path dependence: an evolutionary approach to eco-innovations. J. Evol. Econ. 24, 1037–1065. doi: 10.1007/s00191-014-0381-5 CharnesA. CooperW. W. LewinA. Y. MoreyR. C. RousseauJ. (1984). Sensitivity and stability analysis in DEA. Ann. Oper. Res. 2, 139–156. doi: 10.1007/BF01874736 ChenH. HoH. W. JiC. ZhengH. ZhangS. (2024). Spatiotemporal evolution and driving factors of agricultural land transfer in China. PLoS One 19:e0310532. doi: 10.1371/journal.pone.0310532, PMID: 39292710 ChenL. JiaG. (2017). Environmental efficiency analysis of China’s regional industry: a data envelopment analysis (DEA) based approach. J. Clean. Prod. 142, 846–853. doi: 10.1016/j.jclepro.2016.01.045 ChiY. XuY. WangX. JinF. LiJ. (2021). A win–win scenario for agricultural green development and farmers’ agricultural income: an empirical analysis based on the EKC hypothesis. Sustainability 13:8278. doi: 10.3390/su13158278 ChishtiM. Z. AhmadM. RehmanA. KhanM. K. (2021). Mitigations pathways towards sustainable development: assessing the influence of fiscal and monetary policies on carbon emissions in BRICS economies. J. Clean. Prod. 292:126035. doi: 10.1016/j.jclepro.2021.126035 ChishtiM. Z. SinhaA. (2022). Do the shocks in technological and financial innovation influence the environmental quality? Evidence from BRICS economies. Technol. Soc. 68:101828. doi: 10.1016/j.techsoc.2021.101828 ChungY. H. FäreR. GrosskopfS. (1997). Productivity and undesirable outputs: a directional distance function approach. J. Environ. Manag. 51, 229–240. doi: 10.1006/jema.1997.0146 ColeM. A. ElliottR. J. ZhangL. (2017). Foreign direct investment and the environment. Annu. Rev. Environ. Resour. 42, 465–487. doi: 10.1146/annurev-environ-102016-060916 DengY. SunH. (2022). The impact and mechanism of industrial agglomeration on economic resilience. Soft Sci. 36, 48–54. doi: 10.13956/j.ss.1001-8409.2022.03.07 DjomoJ. M. N. SikodF. (2012). The effects of human capital on agricultural productivity and farmer’s income in Cameroon. Int. Bus. Res. 5:134. doi: 10.5539/ibr.v5n4p149 DuanmuJ. L. BuM. PittmanR. (2018). Does market competition dampen environmental performance? Evidence from China. Strateg. Manage. J. 39, 3006–3030. doi: 10.1002/smj.2948 FarrellM. J. (1957). The measurement of productive efficiency. J. R. Stat. Soc. Series A (Gen.) 120, 253–290. doi: 10.2307/2343100 GonçalvesF. PernaR. LopesE. MacielR. TovarL. LopesM. (2021). Strategies to improve the environmental efficiency and the profitability of sugarcane mills. Biomass Bioenergy 148:106052. doi: 10.1016/j.biombioe.2021.106052 GóralJ. KamboK. KulawikJ. OsuchD. PłonkaR. Poczta-WajdaA., … & WąsA. (2015). Subsidies versus economics, finances and income of farms (1). Monographs of multi-annual Programme. GrossmanG. M. KruegerA. B. (1993). Environmental impacts of a north American free trade agreement. The Mexico-US Free Trade Agreement 11:13. HeY. WangJ. GaoX. WangY. ChoiB. R. (2021). Rural Tourism: Does It Matter for Sustainable Farmers’ Income? Sustainability, 13:10440. doi: 10.3390/su131810440 HuangX. WangZ. RenZ. (2021). Research on the influence of urbanization structure on farmers’ income increase—based on the perspective of income structure. Open J. Soc. Sci. 9, 540–552. doi: 10.4236/jss.2021.95029 IrvanI. P. YuliarmiN. N. (2019). Analysis of impact factors on farmers income. Int. Res. J. Manag. IT Soc. Sci. 6, 218–225. doi: 10.21744/irjmis.v6n5.731 JanatabadiF. ErmagunA. (2024). Access weight matrix: a place and mobility infused spatial weight matrix. Geogr. Anal. 56, 746–767. doi: 10.1111/gean.12395 KambaliU. PanakajeN. (2022). A review on access to agriculture finance by farmers and its impact on their income. Available at SSRN 4104741. LesageJ. P. PaceR. K. (2008). Spatial econometric modeling of origin-destination flows. J. Reg. Sci. 48, 941–967. doi: 10.1111/j.1467-9787.2008.00573.x LiX. N. FengY. WuP. Y. ChiuY. H. (2021). An analysis of environmental efficiency and environmental pollution treatment efficiency in China’s industrial sector. Sustainability 13:2579. doi: 10.3390/su13052579 LiH. JiangH. (2023). Effect of the digital economy on farmers’ household income: county-level panel data for Jilin Province, China. Sustainability 15:4450. doi: 10.3390/su15054450 LiD. WangM. Q. LeeC. (2020). The waste treatment and recycling efficiency of industrial waste processing based on two-stage data envelopment analysis with undesirable inputs. J. Clean. Prod. 242:118279. doi: 10.1016/j.jclepro.2019.118279 LiottaC. KervinioY. LevrelH. TardieuL. (2020). Planning for environmental justice-reducing well-being inequalities through urban greening. Environ. Sci. Pol. 112, 47–60. doi: 10.1016/j.envsci.2020.03.017 LiuY. ZouL. WangY. (2020). Spatial-temporal characteristics and influencing factors of agricultural eco-efficiency in China in recent 40 years. Land Use Policy 97:104794. doi: 10.1016/j.landusepol.2020.104794 LockheedM. E. (1987). Farmers’ education and economic performance-science direct. Econ. Educ. 12, 110–116. LuY. LiB. (2021). Research on regional differences of the leisure agriculture’s impact on farmers’ income—an empirical analysis based on nonlinear threshold regression. Sustainability 13:8416. doi: 10.3390/su13158416 LuoC. LiS. SicularT. (2020). The long-term evolution of national income inequality and rural poverty in China. China Econ. Rev. 62:101465. doi: 10.1016/j.chieco.2020.101465 MuhammadS. PanY. AghaM. H. UmarM. ChenS. (2022). Industrial structure, energy intensity and environmental efficiency across developed and developing economies: the intermediary role of primary, secondary and tertiary industry. Energy 247:123576. doi: 10.1016/j.energy.2022.123576 MuñozL. (2020). The road towards sustainability markets: linking cost externalization to market structure and price structure using qualitative comparative means. Int. J. Latest Res. Humanit. Soc. Sci. 3, 20–32. PandaS. (2015). Farmer education and household agricultural income in rural India. Int. J. Soc. Econ. 42, 514–529. doi: 10.1108/IJSE-12-2013-0278 PengF. ZhangX. ZhouS. (2022a). The role of foreign technology transfer in improving environmental efficiency: empirical evidence from China’s high-tech industry. Front. Environ. Sci. 10:855427. doi: 10.3389/fenvs.2022.855427 PengJ. ZhaoZ. LiuD. (2022b). Impact of agricultural mechanization on agricultural production, income, and mechanism: evidence from Hubei province, China. Front. Environ. Sci. 10:838686. doi: 10.3389/fenvs.2022.838686 QiuB. YangS. XinP. (2008). Research on FDI technology spillover channels and productivity growth in China's manufacturing industry: analysis based on panel data. World Econ. 8, 20–31. SalamM. YingzhiX. ChishtiM. Z. KhanM. K. (2025). Trade, financial development and the environment: analysis of BRI countries having direct connectivity with China. Financ. Innov. 11:104. doi: 10.1186/s40854-025-00775-y ShiQ. LiangQ. HuoT. YouK. CaiW. (2023). Evaluation of CO2 and SO2 synergistic emission reduction: the case of China. J. Clean. Prod. 433:139784. doi: 10.1016/j.jclepro.2023.139784 SongM. AnQ. ZhangW. WangZ. WuJ. (2012). Environmental efficiency evaluation based on data envelopment analysis: a review. Renew. Sust. Energ. Rev. 16, 4465–4469. doi: 10.1016/j.rser.2012.04.052 SongB. RobinsonG. M. BardsleyD. K. (2022). Multifunctionality and path dependence: farmer decision-making in the peri-urban fringe. J. Rural. Stud. 96, 64–77. doi: 10.1016/j.jrurstud.2022.10.012 SongM. SongY. AnQ. YuH. (2013). Review of environmental efficiency and its influencing factors in China: 1998–2009. Renew. Sust. Energ. Rev. 20, 8–14. doi: 10.1016/j.rser.2012.11.075 ToblerW. (2004). On the first law of geography: a reply. Ann. Assoc. Am. Geogr. 94, 304–310. doi: 10.1111/j.1467-8306.2004.09402009.x ToneK. (2002). A slacks-based measure of super-efficiency in data envelopment analysis. Eur. J. Oper. Res. 143, 32–41. doi: 10.1016/S0377-2217(01)00324-1 WangJ. PengL. ChenJ. DengX. (2024). Impact of rural industrial integration on farmers’ income: evidence from agricultural counties in China. J. Asian Econ. 93:101761. doi: 10.1016/j.asieco.2024.101761 WangB. WuY. YanP. (2010). Regional environmental efficiency and total factor productivity growth in China. Econ. Res. 45, 95–109. WangL. ZhangF. WangZ. TanQ. (2022). The impact of rural infrastructural investment on farmers’ income growth in China. China Agric. Econ. Rev. 14, 202–219. doi: 10.1108/CAER-09-2020-0211 XuZ. XueM. (2022). Research on prediction of per capita disposable income of rural residents in Shandong Province based on GM (1, 1) model. In 2022 3rd international conference on computer science and management technology (ICCSMT) (48–51). IEEE. ZengY. JiangC. CuiZ. (2022). The impact of digital finance on high quality economic development: a study based on spatial Durbin model. Tech. Econ. 41, 94–106. ZhangY. ShenL. ShuaiC. BianJ. ZhuM. TanY. . (2019). How is the environmental efficiency in the process of dramatic economic development in the Chinese cities? Ecol. Indic. 98, 349–362. doi: 10.1016/j.ecolind.2018.11.006 ZhangL. WangH. ZhangW. WangC. BaoM. LiangT. . (2022). Study on the development patterns of ecological civilization construction in China: an empirical analysis of 324 prefectural cities. J. Clean. Prod. 367:132975. doi: 10.1016/j.jclepro.2022.132975 ZhengW. L. WangJ. W. JiangA. D. Rehman KhanS. A. YangX. Q. ZhangX. . (2021). Study on environmental performance evaluation of different linkage development types of the logistics and manufacturing industries considering the unexpected output. J. Air Waste Manage. Assoc. 71, 1025–1038. doi: 10.1080/10962247.2021.1920516, PMID: 33900887 ZhouY. KongY. ShaJ. WangH. (2019). The role of industrial structure upgrades in eco-efficiency evolution: spatial correlation and spillover effects. Sci. Total Environ. 687, 1327–1336. doi: 10.1016/j.scitotenv.2019.06.182, PMID: 31412466 ZhuW. ChishtiM. Z. RehmanA. AhmadM. (2021). A pathway toward future sustainability: Assessing the influence of innovation shocks on CO2 emissions in developing economies. Environ. Dev. Sustain., 1–24. doi: 10.1007/s10668-021-01634-3