The Impact of Agricultural Green Development on Farmers’ Income Quality in China

Abstract

As China’s agriculture transitions toward high-quality development, reconciling agricultural green transformation with improved farmers’ income quality has become critical. This study seeks to investigate the effects of agricultural green development on the quality of farmers’ income from three dimensions: direct impact, structural influence, and heterogeneous characteristics. Leveraging provincial panel data from China spanning the period 2011 to 2022, a mixed-methods research design is adopted to conduct empirical analysis. First, the entropy weight method is applied to evaluate the comprehensive development level of agricultural green development and the quality of farmers’ income, along with their respective temporal variation features. On this basis, a two-way fixed effects model is then constructed to examine three core issues: the overall impact of agricultural green development on farmers’ income quality, as well as the structural heterogeneity and spatial heterogeneity characteristics inherent in this impact relationship. The results show that agricultural green development has significantly promoted farmers’ income quality in China, with improved resource utilization efficiency and output quality being the core driving factors, while environmentally friendly practices exhibit a negative effect in the short term. Specifically, agricultural green development significantly enhances income adequacy and structure but has a short-term inhibitory effect on income growth, with no significant impact on knowledge-based income. Regional heterogeneity analysis shows the strongest positive effect in the western region, followed by the eastern region, a negative impact in the northeastern region, and an insignificant effect in the central region. The income-increasing effect of green development is more significant in regions with poor natural resource endowments and low fiscal support for agriculture but is weakened in regions with high market vitality. This study provides a theoretical and practical basis for formulating differentiated agricultural green development policies and improving farmers’ income quality. These findings enrich the theoretical interface between agricultural green transformation and rural income improvement and offer actionable, region-specific policy insights by clarifying the constraints, key links and heterogeneous effects involved.

1. Introduction

The dilemma between environmental pressure and economic growth has attracted widespread attention at the national and global levels [1]. In this context, the concept of green development has been developed. Although the traditional agricultural development model once promoted farmers’ income growth and rural economic growth, it also led to excessive resource consumption and environmental degradation, which restricted agricultural sustainability and the long-term income growth of farmers. The green development of agriculture is a common choice globally to address resource and environmental constraints and achieve sustainable agricultural development [2,3]. How to coordinate agricultural green development with high-quality growth of farmers’ income has become an urgent practical issue to be solved.

As the world’s largest developing agricultural country, China can provide other developing countries with transition experience through its agricultural green development practices. Although China has made remarkable strides in advancing its agricultural development, it is confronted with severe resource and environmental constraints as well as ecological pressures [4]. For a long time, excessive reliance on inputs, such as chemical fertilizers and pesticides, has led to prominent problems, including the decline in soil fertility, water eutrophication, and agricultural non-point source pollution. In 2023, China’s cultivated land area was 128 million hectares, a decrease of 0.7% year-on-year. Soil erosion and land degradation are still worsening in some regions, and the degradation trend of agricultural ecosystems cannot be ignored [5]. Driven by the continuous advancement of the “Green Agriculture” and “Quality Agriculture” strategies as well as relevant supporting policies, coupled with the strong impetus injected into agricultural green development by technological empowerment, China’s agricultural green development level has been gradually improving in recent years. According to China Agricultural Green Development Report 2024, the total consumption of agricultural chemical fertilizers in China in 2024 was 49.88 million tons, with the rate of chemical fertilizer utilization for the three major food crops—rice, wheat, and corn—reaching 42.6%, the comprehensive utilization rate of crop straw exceeding 88%, the proportion of national feed grains dropping to 59.5%, and the comprehensive utilization rate of livestock and poultry manure reaching 79.4% [6]. As an inevitable path for solving resource and environmental constraints and ensuring food security and ecological security, agricultural green development has become a core path for boosting rural revitalization and achieving the goal of common prosperity in China [7].

Over time, agricultural development has undergone profound transformations, with its connotations evolving in parallel with shifts in economic development patterns and conceptual frameworks. Scholars have probed into the field of agriculture from diverse perspectives. Initially, their research focused on the transition from traditional farming practices to the enhancement of agricultural productivity; in recent years, there has been a noticeable shift in their attention toward sustainable agricultural development [8,9]. Prior studies have centered on two key aspects of agricultural green development: the conceptual development of this field [10,11,12,13,14] and the measurement of green development using indicator systems [15,16,17,18]. When conducting measurements of agricultural green development, two distinct approaches can be identified. One approach is rooted in the perspective of productivity, with measurements defined in terms of productivity metrics—specifically, environmentally adjusted (or green) multifactor productivity (EAMFP) [19,20] and environmentally adjusted (or green) total factor productivity (EATFP) [21,22].

The other is the development of a framework by including its key elements. In recent studies exploring the connotations of agricultural green development, researchers have examined multiple dimensions, encompassing resource conservation and preservation, ecological environment security, green product supply, and an affluent and happy life [23,24,25]. In summary, agricultural green development drives the transformation of agriculture from traditional extensive growth to a high-quality development model. This transformation is achieved mainly through means of enhancing resource utilization efficiency, reducing ecological and environmental pollution in agriculture, and safeguarding the quality of agricultural products. This study defines the connotation and evaluation indicators of agricultural green development from these three aspects.

Farmers’ income generally refers to the total income earned by farmers from various economic activities, reflecting their living standards and economic status. According to the classification method of the National Bureau of Statistics of China, farmers’ income in China can be divided into operational income, wage income, property income, and transfer income. Since the reform and opening-up, the income level of Chinese farmers has been continuously improving, yet the urban–rural income gap has been widening, and the issue of income inequality within rural areas has become increasingly prominent [26]. Data from China’s National Bureau of Statistics shows that in 2024, the per capita disposable income of rural residents across the country stood at CNY 23,119, with a real growth rate of 6.6%. Yet, rural residents’ income remains less than half of that of urban residents; notably, the ratio of per capita disposable income of urban residents to that of rural residents was 2.34 [27]. In addition, there are deficiencies in the stability and sustainability of farmers’ income. On the one hand, the source structure of farmers’ income is singular, with agricultural operating income and off-farm wage income as the main components, which are highly susceptible to fluctuations in market prices and employment situations. On the other hand, many rural households lack diversified income channels such as secondary and tertiary industries, resulting in weak risk resistance capabilities. In China, farmers’ over-reliance on agricultural operations and off-farm wages (accounting for over 80% of total income) exacerbates their risk exposure to market price volatility and employment instability. This situation is not unique to China. Globally, farmers in developing countries face similar structural dilemmas; smallholders in sub-Saharan Africa derive over 70% of their income from agricultural operations, making them highly vulnerable to climate shocks and international food price fluctuations [28]. Even in developed countries like France and Germany, family farms rely on agricultural subsidies for more than 40% of their income, reflecting insufficient diversification of non-farm income sources [29]. These factors have led to the vulnerability of farmers’ income growth, making it urgent to enhance the quality of income by optimizing the income structure and expanding income-increasing channels.

Most academic research into the quality of farmers’ income has focused on the rapid, stable, and sustainable increase in farmers’ income—a goal that necessitates the accumulation of financial capital by farmers [30,31,32,33,34,35]. Yet, over an extended period, scholars have broadened the connotation of farmers’ income quality beyond the dimension of quantity to encompass capability and sustainability, while also exploring the multi-dimensional framework for farmers’ income quality [36,37]. Research on farmers’ income quality in China dates back to two decades ago, with existing studies typically incorporating multiple dimensions—including income adequacy, stability, structure, growth or sustainability, and knowledge intensity to comprehensively capture its connotations [38,39,40]. Based on existing research findings and combined with the practices of rural development in China, this study defines farmers’ income quality as a multi-dimensional concept integrating “quantity adequacy, structural rationality, growth sustainability, and knowledge support”. In simple terms, farmers’ income quality not only reflects how much farmers earn but also embodies whether the structure of income sources is reasonable, the growth is sustainable, and factors such as knowledge and skills relied on to obtain income [41]. This concept is not only related to the stable improvement of farmers’ living standards but also a core underpinning for advancing the rural revitalization strategy and promoting the coordinated development of urban and rural areas.

The application of the Environmental Kuznets Curve (EKC) hypothesis in the agricultural sector is attracting increasing attention. Specifically, scholars have explored the relationship between economic development and carbon emissions within the agricultural industry [42,43]. However, research on green agriculture within the existing EKC hypothesis literature remains scarce [44]. Scholars have verified the key role of agricultural green development in enhancing farmers’ income and achieving common prosperity [45,46,47,48,49,50,51]. It promotes farmers’ income quality by means such as increasing the added value of agricultural products, optimizing the structure of agricultural production, and promoting the application of green technologies, thereby forming a sustainable income growth model. Especially, existing studies consistently support the role of training on agricultural green production technologies in boosting farmers’ income [52,53,54]. In turn, although the existing literature has ignored the influence of income structure on the green development of agriculture [55], undoubtedly, the higher the income from agricultural activities, the more farmers are willing to devote themselves to agricultural production. Therefore, the improvement of farmers’ income level provides an economic foundation for agricultural green input and sustainable operation [56]. It can be seen that the comprehensive effect of green development ensures the long-term growth of farmers’ income and their ability to resist risks, rather than relying solely on short-term structural adjustments.

Existing literature has offered valuable implications for understanding the impact of agricultural green development on farmers’ income quality and established a foundation for this study, yet certain limitations remain. Most studies have predominantly focused on the quantitative dimensions of farmers’ income, failing to explore its qualitative aspects in sufficient depth [57,58], resulting in limitations in analyzing the inherent logic of farmers’ income growth in the new development stage. On the one hand, rural populations who have been lifted out of poverty face a high risk of returning to poverty due to insufficient endogenous development capacity and limited industrial driving effects. Whether low-income groups can achieve sustained income growth has become a key to promoting common prosperity [59]. On the other hand, new forms and models of agriculture such as smart agriculture and green agriculture continue to emerge, placing increasingly high demands on farmers’ investment in knowledge elements in the process of income acquisition [60]. Therefore, a single perspective of income quantity cannot fully reflect the multi-dimensional attributes and contemporary connotations of farmers’ income. It is necessary to expand research on the qualitative attributes of farmers’ income from aspects such as the diversification of income sources, the sustainable growth of income, and the knowledge and skills required for income acquisition [61]. Regarding the impact of agricultural green development on farmers’ income, most studies have focused on the impact on income level or income-increasing effects, with few involving the more connotative indicator of farmers’ income quality. Additionally, existing studies have mostly conducted empirical analyses targeting specific regions, specific industries, or specific green agricultural projects. There is a lack of exploration on the impact of agricultural green development on farmers’ income from a macro perspective and in a broader sense, as well as an insufficient discussion on the differential impacts of agricultural resource endowments, economic and social conditions, and institutional environments in different regions on the relationship between the two. There is a lack of systematic induction of heterogeneity, which makes it impossible to form strong universality and policy guidance on the whole.

To address these gaps, this study employs panel data from 31 provincial-level administrative regions across China for the period 2011—2022 and adopts the two-way fixed effects model as the primary econometric tool to empirically examine the impact of agricultural green development level on farmers’ income quality and its structural and regional heterogeneity characteristics. This research makes contributions to the existing body of knowledge in the following aspects: first, it constructs a theoretical framework between agricultural green development and farmers’ income quality, taking farmers’ income quality as one of the core indicators. This breaks through the previous limitation of focusing solely on income quantity. By establishing a more comprehensive evaluation index system for farmers’ income quality, it conducts an in-depth analysis of the multi-dimensional impacts of agricultural green development on farmers’ income quality. Second, it not only focuses on the overall impact of agricultural green development on farmers’ income quality but also reveals the differential associations between the sub-dimensions of agricultural green development and the sub-dimensions of farmers’ income quality through structural analysis, which helps clarify the weak points and key links in this process. Third, it also identifies the heterogeneous characteristics at the regional and regulatory factor levels, explores the influences of differences in agricultural resource endowments, economic and social conditions, and institutional environments across regions on the relationship between the two, and clarifies the effect variations of agricultural green development in different regions. This provides a quantitative basis for designing region-specific policies to balance green development and farmer’ welfare. It also underscores the necessity of conducting further research on the institutional factors influencing this relationship so as to enhance the applicability of relevant policies.

2. Theoretical Mechanism

Green development represents a “positive-sum game” that pursues both ecological sustainability and economic benefits [62]. As a comprehensive transformative force in agricultural production models, the promoting effect of agricultural green development on farmers’ income quality can be realized through the theoretical mechanism of the “Resource–Environment–Output” three-dimensional transmission (as illustrated in Figure 1). From the perspective of resources, agricultural green development enhances the efficiency of resource utilization (e.g., cultivated land, water, and agricultural machinery). This not only reduces the costs of external inputs such as pesticides and chemical fertilizers but also increases per-unit output, directly improving the adequacy of farmers’ income. From the environmental perspective, although environment-friendly agricultural practices require investment in ecological governance in the short term, they can improve the quality of agricultural land and the agricultural ecological environment in the long run. This reduces income fluctuations caused by environmental degradation; meanwhile, new income sources (e.g., ecological compensation and carbon sink trading) are formed, optimizing the income structure. From the output perspective, agricultural green development improves the quality of agricultural products through variety improvement and standardized production, helping green agricultural products obtain market premiums. Additionally, it extends industrial chain links such as deep processing and agritourism, which not only promotes the sustainability of income growth but also enhances farmers’ skills through technical training and brand construction. This lays a foundation for improving the knowledge intensity of farmers’ income.

These three dimensions work synergistically to ultimately promote the comprehensive improvement of the quality of farmers’ income.

Hypothesis 1: 

Agricultural green development has a promoting effect on farmers’ income quality.

The following analysis elaborates on the impact mechanisms of agricultural green development on farmers’ income quality from three constituent dimensions of agricultural green development:

2.1. The Impact of Efficient Agricultural Resource Utilization

First, efficient and sustainable utilization of agricultural resources enhances the adequacy and growth of farmers’ income by improving production efficiency and reducing costs. Green production technologies can increase the utilization efficiency of water and soil resources, boosting unit output, and minimizing input waste, thereby raising agricultural product yields and sales revenue (corresponding to the “resource-saving and efficiency-enhancing” path). Additionally, resource-saving production methods can reduce excessive reliance on external inputs such as pesticides and chemical fertilizers, reducing agricultural production costs and market volatility risks. For instance, Xu et al. (2020) assessed the economic outcomes, resource utilization efficiency, and environmental effects associated with three distinct greenhouse-based eggplant cultivation models and found that organic cultivation exhibited higher resource use efficiency compared to conventional and low-input cultivation [63]. Li et al. (2024) verified that the application of organic fertilizers contributes to increasing farmers’ income through the mechanism of improving crop yields [64].

Second, the rational utilization of agricultural resources optimizes farmers’ income structure. Through more diversified and efficient resource utilization, farmers break away from reliance on a single source of agricultural income, forming a diversified income structure distinguished by the combination of crop farming and livestock rearing, as well as the concurrent development of agriculture and industry [65] (corresponding to the “diversified income-source expansion” path). For example, under the crop–livestock integration model, farmers can achieve synergistic income growth from food crops and cash crops or from planting and breeding on the same plot of land, thereby improving the household income structure [66]. Meanwhile, the diversification of agricultural production helps to disperse risks and stabilize income. In contrast, excessive specialization in a single crop will exacerbate income fluctuations [67] (corresponding to the “risk mitigation” path).

Finally, the green transformation of agricultural resource utilization methods enhances the knowledge intensity in farmers’ income. The efficient use of resources relies on advanced technologies and scientific management methods, which prompts farmers to learn new knowledge and improve their skills (corresponding to the “agricultural skills improvement” path). Meanwhile, farmers with higher educational attainment and cognitive level are more inclined to adopt agricultural green technologies, thereby achieving higher returns. This indicates that the improvement of farmers’ income quality depends on the enhancement of human capital; the more knowledge-based inputs, the more significant the green technology adoption rate and income-increasing effect [68].

Hypothesis 2: 

Agricultural resource utilization has a promoting effect on farmers’ income quality.

2.2. The Impact of Agricultural Environmental Friendliness

First, environmentally friendly measures often focus on long-term interests, avoiding short-sighted behaviors. Green agricultural production reduces the use of pesticides and chemical fertilizers, thereby lowering environmental pollution. As the quality of farmland soil and water sources improves, stable and high crop yields are guaranteed, and the risk of farmers’ income reduction due to environmental degradation decreases significantly, leading to stronger momentum for growth in farmers’ income, reduced volatility, and more stable and sustainable income sources [69] (corresponding to the “agricultural environment stability” path). Research by Harkness et al. (2021) shows that reducing the intensity of inputs such as pesticides and chemical fertilizers helps improve the stability of agricultural income, which can reduce fluctuations in agricultural income by approximately 20% [67]. Moreover, environmentally friendly agricultural practices often receive support from governments and society. For example, in China, governments at all levels have explored ecological compensation mechanisms, providing compensation and rewards to farmers who take measures to reduce pollution and protect the ecology, directly increasing farmers’ transfer income [70] (corresponding to the “ecological compensation” path). In Europe, farms participating in agricultural environmental programs experience smaller income fluctuations, and agricultural subsidies have, to a certain extent, become an important component of farmers’ income [67].

Second, environment friendly agriculture expands farmers’ income sources, optimizes their income structure, and creates new employment and business opportunities in rural areas. Ecotourism and leisure agriculture are emerging industries derived from a favorable pastoral environment. Farmers obtain service income by operating farm stays, sightseeing, and picking gardens, etc., which makes the household income structure more diversified [24] (corresponding to the “industrial chain extension” path). Circular agriculture has spawned auxiliary industries such as biogas production and organic fertilizer processing, and farmers can participate in these industries to gain additional benefits. In other words, environment-friendly agriculture has transformed farmers’ income model, which previously relied solely on planting and breeding, into a multi-dimensional structure integrating planting, breeding, processing, and services, greatly improving the rationality of the income structure and the ability to resist risks.

Finally, environment friendly agriculture enhances the knowledge intensity of farmers’ income. Environmental protection requires farmers to continuously acquire new knowledge and adopt new technologies, such as ecological pest control and resource utilization of agricultural wastes (corresponding to the “environmental awareness improvement” path). In the process of participating in environmentally friendly agricultural projects and receiving relevant training, farmers’ environmental awareness and professional skills are improved, and the knowledge content and external value embodied in their income are significantly enhanced.

Hypothesis 3: 

Agricultural environmental friendliness has a promoting effect on farmers’ income quality.

2.3. The Impact of Agricultural Output Quality

Improving agricultural output quality is mainly reflected in providing safer, more nutritious, and more market-competitive agricultural products. Green development improves agricultural output quality and efficiency, bringing farmers the income return of “higher quality leading to higher price”.

First, by improving varieties, implementing strict quality control, and obtaining quality certification, farmers can make their agricultural products meet higher standards, which can significantly increase the product’s added value, thereby raising the market price of agricultural products while boosting farmers’ earnings [24] (corresponding to the “organic price premium” path). For example, a field survey on jujube farmers in Xinjiang shows that the premium of organic jujubes over ordinary jujubes is generally 50% higher, and the income is 25–30% higher [71]. In addition, high-quality output is often accompanied by brand building. By creating well-known brand agricultural products, farmers can gain bargaining power in the market and achieve long-term income growth [72] (corresponding to the “brand development” path). Compared with the extensive mode that pursues output, the quality-driven income increase mode can bring more lasting and stable income-increasing effects.

Second, the improvement of agricultural output quality promotes the extension and upgrading of the agricultural industry chain, providing farmers with multi-level and multi-link income sources and optimizing their income structure. The production of high-quality agricultural products is usually accompanied by the expansion of downstream links, such as deep processing, ecotourism, and agricultural product e-commerce, enabling farmers to participate in more value-added links (corresponding to the “value-added through deep processing” path). For instance, high-quality grains, fruits, and vegetables are used for deep processing to produce refined foods or functional products; when farmers engage in these processing links, they can share the income from processing value-added products [73]. Some farmers’ cooperatives have established local processing factories to process high-quality agricultural products into dried products, canned foods, beverages, etc., which has increased the added value of the product and created conditions for member dividends and employment [74]. Moreover, industrial chain upgrading has also attracted urban capital and talents to rural areas to jointly develop deep processing of agricultural products and brand marketing, driving farmers to share the benefits of industrial integration and development.

Furthermore, the improvement of agricultural output quality has also enhanced the knowledge intensity of farmers’ income. Producing high-quality agricultural products relies on science and technology as well as strict management, which requires farmers to continuously learn advanced production technologies and quality standards (corresponding to the “standardized production” path). The process of acquiring and applying such knowledge enables farmers to become producers with professional skills, and their income thus stems more from technical and management capabilities rather than simple physical input [75]. This is not only reflected in the economic benefits brought by green production but also lays a foundation for farmers to engage in higher value-added agricultural activities or switch to non-agricultural industries in the future [24]. In particular, the orientation towards high-quality output has connected farmers more closely with scientific research institutions and agricultural extension departments. Through participating in training programs and demonstration projects, farmers have gradually grown into knowledge-based talents. Moreover, higher output quality is not only reflected in the agricultural production link, farmers are also accumulating experience and skills in marketing and brand building, which allows them to actively adapt to market changes and maintain sustainable income growth.

Hypothesis 4: 

Agricultural output quality has a promoting effect on farmers’ income quality.

In summary, the three dimensions of agricultural green development are agricultural resource utilization, agricultural environmental friendliness, and agricultural output quality. Through approaches such as the promotion of green technologies, ecological compensation, industrial chain extension, and the enhancement of knowledge and skills, they strengthen the adequacy, structural rationality, growth potential, and knowledge intensity of farmers’ income, thereby achieving an all-round optimization of farmers’ income quality

3. Methods

3.1. Variable Description

3.1.1. Explanatory Variable

Although the concept of agricultural green development has been clearly defined in the previous text, its assessment remains challenging. Based on the core connotation of agricultural green development, i.e., “resource conservation, environmental friendliness, and efficient output” [23,24,25], and in line with three principles, namely data availability, indicator representativeness, and policy adaptability, this study establishes an evaluation index system for agricultural green development, encompassing three key dimensions: agricultural resources utilization, agricultural environmental friendliness, and agricultural outputs quality (as shown in

Table 1. Evaluation index system of agricultural green development.

First-Level Indicator	Weight	Second-Level Indicator	Indicator Explanation	Attribute
Agricultural resources utilization (ARU)	55.64%	Farmland use efficiency	Agricultural sown area/number of employees in primary industry (hectares/person)	+
		Agricultural machinery use efficiency	Agricultural output value/total power of agricultural machinery (CNY 100 million/10,000 kilowatts)	+
		Agricultural water use efficiency	Effective irrigation area/agricultural sown area (%)	+
		Agricultural electricity use efficiency	Agricultural output value/agricultural electricity consumption (CNY/kilowatt-hour)	+
Agricultural environmental friendliness (AEF)	15.24%	Chemical fertilizer application intensity	Chemical fertilizer usage/agricultural sown area (tons/hectare)	−
		Pesticide application intensity	Pesticide usage/agricultural sown area (tons/hectare)	−
		Agricultural film application intensity	Agricultural film usage/agricultural sown area (tons/hectare)	−
Agricultural output quality (AOQ)	29.11%	Per capita agricultural output value	Total agricultural output value/number of employees in primary industry (10,000 CNY/person)	+
		Grain yield per unit area	Total grain output/grain sown area (tons/hectare)	+

).

• Agricultural resources utilization (ARU). Resource conservation serves as a core characteristic of agricultural green development, which places emphasis on enhancing the utilization efficiency of resources, like arable land, water resources, and agricultural machinery power, along with boosting labor productivity [76]. Therefore, farmland use efficiency, agricultural machinery use efficiency, agricultural water use efficiency, and agricultural electricity use efficiency are selected under this dimension. These four indicators closely align with the core connotation of resource conservation and intensification in agricultural green development and are all key monitoring targets in national-level agricultural green development policies. Among these, farmland use efficiency reflects the intensive level of coordinated utilization of land and human resources (using “number of employees in primary industry” as labor input avoids deviations caused by including non-labor-capable groups in the labor force); agricultural machinery use efficiency reflects the efficiency of agricultural machinery utilization. In the process of agricultural modernization, mechanization reduces reliance on manual labor and can lower unit costs through large-scale operations, thereby driving growth in output value. Agricultural water use efficiency reflects agricultural water use efficiency, and this indicator offers a quantitative basis for evaluating agriculture’s ability to adapt to water resource constraints. Agricultural electricity use efficiency reflects the efficiency of agricultural electricity use, promoting the transformation of agricultural energy consumption from an extensive model to a refined and clean one.
• Agricultural environmental friendliness (AEF). Environmental friendliness is an inherent attribute of agricultural green development. Agricultural green development entails vigorously promoting green production technologies, reducing environmental pollution through controlling the usage of chemical fertilizers, pesticides, and agricultural films and realizing a win-win situation between agricultural development and environmental protection [77]. Reducing fertilizer intensity serves as the foundation for controlling agricultural non-point source pollution; lowering pesticide input intensity can improve green prevention and control levels, thereby mitigating the dual risks of pesticides to the environment and food safety. Decreasing agricultural film usage intensity is conducive to promoting degradable agricultural films, thus reducing plastic residues. This dimension includes three indicators: “chemical fertilizer usage per unit of agricultural sown area”, “pesticide usage per unit of agricultural sown area”, and “agricultural film usage per unit of agricultural sown area”. These three indicators are core sources of agricultural non-point source pollution and key regulatory targets of national policies such as the Action Plan for Reducing Fertilizer and Pesticide Use and Improving Efficiency. Other alternative indicators, such as “soil organic matter content”, “carbon sink capacity”, and “ecological compensation amount”, are not selected for the following reasons: the missing rate of the “soil organic matter content” indicator in provincial panel data exceeds 30%; “carbon sink capacity” lacks a unified provincial accounting standard, and there are significant differences in statistical calibers across provinces (e.g., some provinces only account for forest carbon sinks but not farmland carbon sinks); “ecological compensation amount” is a policy intervention variable, which has been reflected in the control variable “government support level” (Gov).
• Agricultural output quality (AOQ). High-efficiency production serves as a direct objective of agricultural green development, with an emphasis on striking a balance between the quantity and quality of outputs [76]. The core aim of agricultural green development is to satisfy the public’s demand for high-quality, safe, and distinctive agricultural goods, thereby achieving a transformation from a focus on yield to a focus on quality [77]. Against the backdrop of promoting quality agriculture and green agriculture, “per capita agricultural output value” serves as a key indicator for measuring the supply quality of agriculture and the income growth of farmers. In addition, “grain yield per unit area” is a conventional indicator for measuring agricultural output, with better data continuity than the “proportion of certified green agricultural products” (only eastern provinces conducted statistics on this indicator before 2018). In line with the perspective that green total factor productivity needs to balance both output quantity and product quality, increasing per unit yield often relies on green production technologies that are both efficient and environmentally friendly, such as the cultivation of fine varieties, scientific fertilization, and rational irrigation [63].

3.1.2. Explained Variable

In accordance with the four-dimensional connotations of “sufficient quantity, reasonable structure, sustainable growth, and knowledge support” [38,39,40,41] and combined with the operability of macro panel data, this paper constructs an evaluation index system for farmers’ income quality from four dimensions: adequacy, structure, growth, and knowledge intensity (as shown in

Table 2. Evaluation index system of farmers’ income quality.

First-Level Indicator	Weight	Second-Level Indicator	Indicator Explanation	Characteristics of High Income Quality	Characteristics of Low Income Quality	Attribute
Income adequacy	10.96%	Absolute income adequacy	Per capita disposable income of rural residents (CNY/person)	More than 10% higher than the national rural average level in the same period	More than 20% lower than the national rural average level in the same period	+
		Relative income adequacy	Per capita disposable income of rural residents—Per capita consumption expenditure of rural residents (CNY/person)	Sufficient balance after deducting consumption expenditure, which can be used for savings and investment	No surplus after deducting consumption expenditure	+
Income structure	33.65%	Diversification of internal income sources	Herfindahl Index of the four income categories	No obvious unipolar dependence on the four types of income, with the Herfindahl Index < 0.3	Excessive dependence on agricultural operation income (accounting for over 80%), with the Herfindahl Index > 0.6	−
		Equilibrium of internal income sources	Sum of squared deviations of the proportions of the four income categories	Sum of squared deviations of the proportion of the four types of income < 0.1	Sum of squared deviations of income proportion > 0.3	−
Income growth	25.37%	Growth rate of disposable income	1—(Per capita disposable income of rural residents in the previous year/Per capita disposable income of rural residents in the current year)—rural CPI growth rate in the current year (%)	Growth rate ≥ 5% for 3 consecutive years, with the annual income growth rate fluctuation range < 3%	Long-term income growth rate < 3%, with the growth rate fluctuation range > 10%	+
		Growth rate of net operating income	1—(Per capita net operating income of rural residents in the previous year/Per capita net operating income of rural residents in the current year)—rural CPI growth rate in the current year (%)	Growth rate ≥ 5% for 3 consecutive years, with the annual income growth rate fluctuation range < 3%	Long-term income growth rate < 3%, with the growth rate fluctuation range > 10%	+
Income knowledge intensity	30.03%	Educational level of farmers	Average years of education of rural residents	Average years of education ≥ 9 years	Average years of education < 6 years	+
		Green technology training level of farmers	Number of rural graduates from adult cultural and technical training schools per thousand people	Number of graduates per 1000 people ≥ 15	Number of graduates per 1000 people < 5

).

• Income adequacy reflects the quantitative level of income, representing the amount of monetary income farmers obtain from various economic activities. For this dimension, “rural residents’ per capita disposable income” (absolute adequacy) and “balance of disposable income minus consumption expenditure” (relative adequacy) are selected. Among these, absolute income adequacy refers to the growth in the absolute quantity of income, which is a core indicator used by the National Bureau of Statistics to measure rural income levels [27]. In contrast, relative income adequacy means that after covering living expenses and other necessary expenditures, there is still an income surplus available for savings and investment. The selection of this indicator refers to Sen’s (1999) theory that income needs to meet the demands of savings and investment [36], avoiding the limitation of using only absolute income while ignoring income–expenditure balance. The “Engel coefficient” is not selected because it mainly reflects the consumption structure and has a weak direct correlation with the “adequacy” dimension of income quality.
• Income structure refers to whether farmer’ income sources are diversified and whether the proportions of various sources are reasonable. The transformation of farmers’ income sources from singularity to diversification, along with the balanced development of each income source, is conducive to dispersing risks and achieving sustainable income growth. In this paper, the “Herfindahl Index of the four income categories” is selected to characterize the diversification level of their income sources [41]; “Sum of squared deviations of the proportions of the four income categories” is used to reflect the balance level of their income sources. The combination of these two indicators covers the dual characteristics of diversification and balance, which is superior to the commonly used single indicator of “non-farm income proportion”. This is because the latter cannot reflect the internal differences between property income and transfer income.
• Income growth reflects the dynamic development process of farmers’ income, that is, whether various income sources can maintain a good growth momentum. For income growth, “disposable income growth rate” and “net operating income growth rate” (excluding the impact of CPI) are selected. Net operating income is a core component of farmers’ income. “Wage income growth rate” is not selected because this indicator is highly affected by fluctuations in non-farm employment. For example, the 2020 pandemic caused a sharp decline in farmers’ wage income, which could easily interfere with the inherent correlation between green development and income growth.
• Income knowledge intensity can be understood as the contribution degree of human capital factors such as knowledge and skills contained in farmers’ income, emphasizing the supporting strength of human capital including farmers’ knowledge level, professional skills and production experience to their income level, income stability and income growth potential. “Educational level of farmers” reflects their basic cognitive level and serves as a prerequisite for mastering green agricultural technologies. This indicator directly aligns with the proposition in human capital theory that education is the core carrier for the development of capabilities [53]. “Green technology training level of farmers” focuses on the specific skill reserves for conducting green agricultural production and directly meets the practical needs of green agricultural development. This indicator is calculated based on the “number of rural graduates from adult cultural and technical training schools per thousand people”, where rural adult cultural and technical training covers training content highly relevant to green agricultural development, such as green planting, ecological breeding, and agricultural product quality and safety. This indicator addresses the limitation of “average years of education”, which only reflects basic education and fails to capture improvements in green-related skills, and enables the measurement of farmers’ practical ability to convert knowledge into green income. Other potential alternative indicators have their drawbacks. For instance, “the educational structure of rural residents (proportion of population with senior high school education or above)” cannot continuously quantify the degree of human capital accumulation; “the proportion of members in farmers’ professional cooperatives” primarily reflects the level of organization rather than knowledge intensity; and “the rate of holding green certificates” cannot meet the sample requirements for empirical analysis, as these data are not collected in most central and western provinces.

3.1.3. Control Variables

Based on existing literature and practical backgrounds, seven control variables are selected, including government support level, social security level, natural resource endowment, local market vitality, transportation accessibility, industrial structure upgrading, and financial development vitality (as shown in

Table 3. Model variables.

Variable Type	Variable Name	Symbol	Calculation Method
Explained variable	Farmers’ income quality	FIQ	Entropy weight method
Explanatory variable	Agricultural green development	AGD	Entropy weight method
Control variables	Government support level	Gov	Agricultural related expenditure/Fiscal expenditure
	Social security level	Ins	Agricultural insurance payout amount/Agricultural population
	Natural resource endowment	Nat	Crop failure area/Agricultural population
	Local market vitality	Mar	Number of legal entities in primary industry/Rural population
	Transportation accessibility	Tra	(Highway mileage + railway mileage)/Administrative area
	Industrial structure upgrading	Ind	Output value of tertiary industry/Output value of secondary industry
	Financial development vitality	Loa	Agricultural related loans/Number of employees in primary industry

).

• Government support level (Gov) is to control the interference of exogenous policy interventions and align with the policy-driven development characteristics of agriculture. Government support for agriculture and rural areas constitutes a crucial external force influencing farmers’ income. Government addresses the “market failure” in agricultural production through fiscal subsidies, investments in agricultural infrastructure, and industrial support policies, which directly or indirectly enhance farmers’ income [78,79]. This variable reflects the intensity of government resource input into agriculture and rural areas and can be measured by “the ratio of agricultural related expenditures to total fiscal expenditures”.
• Social security level (Ins) also controls the interference of exogenous policy interventions, serving to avoid the indirect impact of risk buffering mechanisms on income quality. Social security does not directly increase farmers’ income; instead, it indirectly enhances income quality by reducing rigid expenditures such as medical care and elderly care and stabilizing income expectations [36]. In particular, the income stability of low-income groups in rural areas is highly dependent on the social security network, which helps them resist natural and market risks [80,81]. This variable reflects the improvement of basic public services in rural areas and can be measured by “the ratio of agricultural insurance compensation amount to the agricultural population”.
• Natural resource endowment (Nat) is to control the interference of differences in innate endowments and align with the resource-dependent attribute of agriculture. Natural resources serve as the fundamental conditions for agricultural production, directly affecting agricultural production efficiency and income levels [82]. There is an interactive effect between natural resource endowment and green agricultural development. Regions with superior resource endowments are more likely to promote green agriculture, whereas those with resource constraints may face limitations in green transformation [83]. This variable can be measured by “the ratio of crop failure area to agricultural population” (as a reverse indicator) [84].
• Local market vitality (Mar) is to avoid the impact of market entity competition on income distribution. Market vitality reflects the vitality of regional economy. An increase in the number of rural legal entities can, on the one hand, affect farmers’ income by providing employment opportunities and expanding sales channels for agricultural products [85,86]. On the other hand, it may also squeeze the profits of small-scale farmers [87]. This variable can be measured by “the ratio of the number of legal entities in the primary industry to the rural population”.
• Transportation accessibility (Tra) is to exclude the effect of traffic improvement on the realization of agricultural product premiums. Convenient transportation can reduce the logistics costs of agricultural products and expand sales radii, exerting a significant impact on regional economic activities [88]. In the context of green agricultural development, it facilitates the realization of premiums for agricultural products, for example, shortening transportation time to ensure quality [89,90], with a particularly notable effect on green agricultural products reliant on fresh sales. Additionally, transportation convenience promotes the flow of factors (e.g., labor, capital) between rural and urban areas, indirectly increasing non-agricultural employment opportunities for farmers [91]. This variable can be measured by “the ratio of (highway mileage + railway mileage) to administrative area”.
• Industrial structure upgrading (Ind) is to control the optimization effect of industrial integration on income structure. Industrial structure upgrading can extend industrial chains, enhance added value, promote the transfer of rural labor to non-agricultural industries, and thus increase farmers’ wage income. Meanwhile, it can boost farmers’ operational income by developing industries such as agricultural product processing and rural tourism [92]. Industrial upgrading converts the ecological and cultural values of agriculture into economic benefits, forming a synergistic effect with green agricultural development [93,94]. This variable reflects the degree of diversification and upgrading of rural industries and can be measured by “the ratio of the output value of the tertiary industry to that of the secondary industry”.
• Financial development vitality (Loa) is to exclude the restrictive effect of capital constraints on the green transformation of agriculture. Rural finance aims to alleviate credit constraints, support farmers’ production decisions, and diversify production risks [95]. A sound rural financial system can provide financial support to farmers, helping them expand production scales, introduce advanced technologies, improve agricultural production efficiency and the quality of agricultural products, thereby increasing their income [96]. Financial instruments such as green credit and carbon finance are important supports for the promotion of green agricultural technologies, directly affecting the income conversion efficiency of green production [97,98]. This variable can be measured by “the ratio of agricultural related loans to the number of employees in the primary industry”.

3.2. Data Source

This study collects panel data covering 31 provincial-level administrative regions in China (Hong Kong, Macao, and Taiwan are excluded) spanning the period from 2011 to 2022. The index data for the explanatory variable “agricultural green development” are sourced from “China Statistical Yearbook”, “China Rural Statistical Yearbook”, and the statistical yearbooks of each province. For the explained variable “farmers’ income quality”, the relevant index data are derived from “China Statistical Yearbook”, “China Price Statistical Yearbook”, “China Rural Statistical Yearbook”, and “China Population and Employment Statistical Yearbook”. The index data of the control variables are retrieved from “China Rural Statistical Yearbook”, “China Environmental Statistical Yearbook”, and the statistical yearbooks and bulletins of each province. The Linear Interpolation Method is adopted to address the issue of missing data in certain provinces or specific years.

3.3. Data Analysis Methods

3.3.1. Measurement of Key Variables

This research employs the entropy weight method to assess the level of agricultural green development and the quality of farmers’ income. This method calculates weights according to the information entropy of each indicator, a process that can minimize the interference of subjective factors and improve the objectivity of the results [99]. The dimension is eliminated through min-max normalization, the indicator weights are determined based on information entropy, and the composite index is finally obtained by weighted summation. The specific formula is shown in Appendix A.1.

3.3.2. Empirical Model Specification

To examine the impact of agricultural green development on farmers’ income quality, a fixed effects model is constructed in accordance with the theoretical framework proposed in this study. This approach is widely adopted in panel data analysis to ensure the robustness of causal inference [100,101]. The model incorporates provincial individual fixed effects and time fixed effects to control for unobserved confounding variables, while also including 7 control variables; the specific model specification and mathematical expression are shown in Appendix A.2.

3.3.3. Endogeneity Concerns and Solutions

To verify the impact of agricultural green development on farmers’ income quality and address potential endogeneity issues, we employ the instrumental variables two-stage least squares method (IV-2SLS) and the Generalized Method of Moments (GMM) for endogeneity testing.

Firstly, considering that agricultural green development may be correlated with unobserved regional characteristics or that there may exist a reverse causal relationship (e.g., farmers’ income level affects the adoption of green technologies), we select the following instrumental variable: the level of agricultural green development in other regions. We construct the instrumental variable for region i at time t, which is defined as the average level of agricultural green development in all regions except region i. The “relevance” of this instrumental variable is based on the theoretical logic of the Peer Effect [102]. That is, there exist technology diffusion (e.g., the spread of green planting technologies from other regions to the local area), policy imitation (the local area learns from other regions’ green development policies), or market linkage (the demand for green agricultural products in other regions drives the adjustment of local production) among neighboring regions or regions with similar agricultural production conditions. Therefore, their green development level will significantly affect the green development decision-making of region i. Its exogeneity lies in the fact that the green development level of other regions does not directly affect farmers’ income in region i.

Then, to address potential issues of mutual causality between variables and dynamic endogeneity (the lag effect of farmers’ income quality), we adopt the Generalized Method of Moments (GMM) [103] by introducing the lagged term of farmers’ income quality (L.FIQ) as an instrumental variable. Theoretically, there exists a path dependence between the lagged FIQ and the current FIQ, which satisfies the relevance requirement of an instrumental variable. Meanwhile, the lagged FIQ is a historical variable that has already occurred and is not affected by the current random error term, thus meeting the exogeneity requirement.

Finally, the core idea of the Lagged Independent Variable Approach is to introduce the lagged term of the core explanatory variable (L.AGD) to replace or supplement the current AGD in the regression model. Its logical basis lies in two aspects: First, the lagged term L.AGD is a historical variable that has already occurred and will not be disturbed by the current error term, thus satisfying the exogeneity requirement of being uncorrelated with the current error term. Second, agricultural green development is characterized by continuity; the lagged term AGD exerts a genuine impact on the current FIQ through channels such as the accumulation of ecological effects and the lag in technology diffusion. Therefore, L.AGD still maintains a strong correlation with the current FIQ.

3.3.4. Robustness Checking Methods

To ensure the consistency of our results under varying model settings or sample selections, further verify result stability, and enhance the study’s persuasiveness, this research conducts a robustness check from the following perspectives: (1) replacing the calculation approach. The comprehensive levels were quantified by constructing composite indices through principal component analysis (PCA) for the three sub-dimensions of the independent variable (agricultural resource utilization, agricultural environmental friendliness, and agricultural output quality) and the four sub-dimensions of the dependent variable (income adequacy, income structure, income growth, and income knowledge intensity). (2) Excluding the three exceptional years (2020—2022). This exclusion is justified by the global COVID-19 pandemic during this period, which exerted profound exogenous shocks on both agricultural production and farmer’ income. Removing these years helps eliminate pandemic-driven confounding effects, thereby isolating the intrinsic relationship between agricultural green development and farmers’ income quality and further verifying the robustness of our baseline results. (3) ARIMA interpolation method. An ARIMA model is constructed to reasonably interpolate and supplement key indicators with missing values or abnormal fluctuations in the sample. Then, it re-conducts the two-way fixed effects regression based on the complete panel data after interpolation, and compares the significance, sign, and numerical fluctuation range of the coefficients of the core explanatory variables before and after interpolation to determine whether the benchmark conclusions are robust. ARIMA interpolation can more accurately capture the dynamic correlation and trend characteristics of time series data, reduce subjective bias in the interpolation process, and thus more rigorously verify the reliability of the core conclusions.

4. Results

Based on the model constructed above and the data acquired for this research, this study conducted econometric analysis via Stata 18 and thus obtained the following empirical research findings.

4.1. System Measurement Results

The development levels of agricultural green development and farmers’ income quality in 31 provincial-level administrative regions across China from 2011 to 2022 were measured by using the entropy weight method. The following is an analysis of the overall temporal evolution trend, regional differences, and the performance of typical provinces of the two systems. To further clarify regional differences, this study categorized China’s 31 provincial-level administrative regions into four geographical divisions—eastern, central, western, and northeastern—based on the classification standards of the National Bureau of Statistics of China. Specifically, the eastern region includes Beijing, Tianjin, Hebei, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong, and Hainan. The central region includes Shanxi, Anhui, Jiangxi, Henan, Hubei, and Hunan. The western region includes Inner Mongolia, Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Shaanxi, Gansu, Qinghai, Ningxia, Xinjiang, and Tibet. The northeast region includes Heilongjiang, Jilin, and Liaoning.

4.1.1. The Measurement of Agricultural Green Development

As shown in Appendix B.1, from 2011 to 2022, the national AGD index rose from 0.296 to 0.477, with an average annual growth rate of 4.8%. Specifically, the growth was relatively flat during 2011—2015 (increasing from 0.296 to 0.348) and accelerated after 2016 (rising from 0.361 to 0.477). This is closely linked to the intensive implementation of green agricultural policies since 2015, such as the Agricultural Green Development Pilot Zone Program and the policy of Reducing the Usage and Improving the Efficiency of Chemical Fertilizers and Pesticides.

In terms of regional differences, as shown in Figure 2, China’s agricultural green development presents a pattern where the eastern and northeastern regions take the lead, the western region shows a notable catching-up momentum, and the central region needs to accelerate its transformation, with resource endowments and policy interventions as key driving factors. Specifically, the eastern region, as a pioneer area for the national agricultural green development policies, has always maintained the highest level of green development. Its index rose from 0.309 in 2011 to 0.517 in 2022, with an average annual growth rate of 5.2%. Among them, Shanghai, Tianjin, and Jiangsu have performed prominently, with their indices all exceeding 0.5 in 2022, reflecting the first-mover advantage of economically developed regions in the ecological transformation of agriculture. The northeast region had a relatively high initial base (0.348 in 2011) and reached 0.553 in 2022, with an average annual growth rate of 4.6%. Heilongjiang rose from 0.456 in 2011 to 0.820 in 2022, remaining the top in the country. This is likely closely related to its black soil protection policies, while Jilin and Liaoning saw relatively slow growth. The western region increased from 0.275 in 2011 to 0.429 in 2022, with an average annual growth rate of 4.9%, which is second only to the eastern region. Xinjiang and Inner Mongolia are the core growth poles in the west. In particular, Xinjiang’s index has increased by 61.5% compared with 2011, benefiting from the popularization of water-saving agriculture and green planting technologies. Southwest provinces such as Guizhou and Yunnan are still at a relatively low level, reflecting the unbalanced development within the western region. The central region has maintained steady growth but remains at a relatively low overall level, with an index of 0.408 in 2022. This may be attributed to the traditional agricultural production methods.

4.1.2. The Measurement of Farmer’s Income Quality

In terms of farmers’ income quality, as shown in Appendix B.2, from 2012 to 2017, the national FIQ average fluctuated at a low level between 0.129 and 0.147, reflecting the single structure of farmers’ income and insufficient growth stability. In 2018, the index surged to 0.438, an increase of 238% compared with 2017. This is the result of the combined effect of multiple factors such as policy support, economic environment, and industrial upgrading. In particular, 2018 was the first year of the comprehensive implementation of the rural revitalization strategy. The state intensively introduced a series of policies benefiting farmers, which directly promoted the rapid growth of farmers’ income. This is reflected in the significant increase in three indicators: absolute income adequacy, growth rate of disposable income, and growth rate of net operating income. From 2019 to 2021, it remained at a medium-to-high level of 0.427–0.449 but dropped significantly to 0.236 in 2022. This decline is believed to be caused by disruptions in the circulation of agricultural products and fluctuations in farmers’ income in the post-pandemic period. From the perspective of the four dimensions of income quality, the pandemic caused a decline through three channels. First, it impacted the job market, leading to shorter working hours and unstable positions for farmers working outside their hometowns, which weakened the adequacy and stability of income. Second, it hindered the circulation of agricultural products, intensified price fluctuations, and coupled with the rise in the cost of agricultural materials, which reduced income adequacy. Third, it restrained the development of rural non-agricultural industries and skill training, which weakened the diversified channels of income and the support of knowledge intensity. Ultimately, these factors resulted in a significant decline in income quality.

In terms of regional disparities, as illustrated in Figure 3, the eastern region, particularly Shanghai and Beijing, significantly outperforms other provinces, while Zhejiang and Guangdong demonstrate relatively stable performance. Relying on non-agricultural industries and a well-developed social security system, the economically developed provinces in the east have shown the advantage of coordinating urban and rural incomes. In the northeast region, Heilongjiang has maintained relatively stable performance due to the steady subsidy policies for major grain-producing areas. Among the three northeastern provinces, Jilin has achieved the best performance, which may be attributed to the upgrading of its corn and soybean industries, which in turn has optimized the income structure. The overall level of the central region is lower than that of the eastern and northeastern regions. This is because major agricultural provinces such as Shanxi and Henan rely on traditional farming practices, resulting in insufficient diversification of income sources. In the western region, Xinjiang and Inner Mongolia have remained at the forefront, relying on incomes from characteristic agriculture and animal husbandry. Due to the constraints of mountainous agriculture, Guizhou and Chongqing have long been at a low level. Yunnan and Shaanxi have shown strong income resilience driven by plateau-specific agriculture and the integration of cultural and tourism industries. Guangxi and Ningxia, which depend on traditional farming, have weak risk resistance capabilities. Overall, the regional disparities in farmers’ income quality stem from differences in industrial structures. The eastern region relies on non-agricultural income, the northeastern and central regions depend on agricultural subsidies, and the western region is reliant on characteristic industries. The general decline in the index in 2022 indicates that it is necessary to enhance the diversification of farmers’ income and reduce dependence on a single industry.

4.2. Variables Descriptive Statistics

The results of the variable descriptive statistics are presented in Appendix B.3, Appendix B.4, Appendix B.5, Appendix B.6, Appendix B.7. Farmers’ income quality (FIQ) has a mean value of 0.217 and a standard deviation of 0.146, while the agricultural green development level (AGD) has a mean value of 0.366 and a standard deviation of 0.107. Both variables show significant regional differences in their values, reflecting the gradient differentiation in income quality and green transformation progress across different provinces. Among the control variables, the government support level (Gov) has relatively stable values (mean value = 0.212, standard deviation = 0.038), whereas variables such as social security level (Ins) and per capita agricultural-related loans (Loa) have relatively large standard deviations (7.265 and 31.791, respectively), which indicates the regional imbalance in the allocation of rural social security and financial resources.

Regarding the model applicability hypothesis tests, the IPS unit root test shows that the Adjusted t* values of all variables are negative with p-values < 0.01, rejecting the null hypothesis of “existing unit roots” and ensuring data stationarity to avoid spurious regression. Pesaran’s CD test indicates that the p-values corresponding to the CD statistics of each variable are all >0.1, with the absolute values of the average cross-sectional correlation coefficients <0.03, verifying no significant cross-sectional dependence and ensuring the unbiasedness of estimates. The Breusc–Pagan/Cook–Weisberg test (chi² = 2.701, p = 0.11) and Breusch–Pagan LM test (chi² = 280.5, p = 0.15) prove no significant heteroscedasticity and serial correlation, respectively, satisfying the independent and identically distributed assumption of disturbance terms. The VIF test shows that the VIF values of all variables are <2 with an average VIF = 1.41, excluding severe multicollinearity and ensuring the stability of coefficient estimates. For model type selection, the F test (F = 6.97, Prob > F = 0.0000) rejects the pooled effects model. The Hausman test (chi²(8) = 20.17, Prob > chi² = 0.0000) rejects the random effects model. The LM test (chibar²(465) = 3568.2, Prob > chibar² = 0.0000) further verifies the significance of individual fixed effects. These results collectively support the rationality of the two-way fixed effects model, laying a reliable methodological foundation for subsequent empirical analysis.

4.3. Benchmark Regression Results

Table 4. Benchmark regression results.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
	FIQ	FIQ	FIQ	FIQ	FIQ	FIQ	FIQ	FIQ
AGD	0.3050 ***	0.2798 ***	0.1681 ***	0.1915 ***	0.1819 ***	0.1994 ***	0.1660 ***	0.1265 **
	(4.9275)	(4.3163)	(2.7858)	(3.1888)	(2.9018)	(3.4136)	(3.1900)	(2.5202)
Gov		0.2267	0.1343	0.2383	0.2278	0.1245	0.2170	0.2625 **
		(1.2912)	(0.8388)	(1.4763)	(1.3999)	(0.8182)	(1.6004)	(2.0219)
Ins			0.0069 ***	0.0068 ***	0.0066 ***	0.0050 ***	0.0009	−0.0000
			(8.7020)	(8.6314)	(7.6842)	(5.9832)	(1.0916)	(−0.0016)
Nat				−0.0098 ***	−0.0096 ***	−0.0030	−0.0018	−0.0015
				(−3.1854)	(−3.1294)	(−1.0167)	(−0.6834)	(−0.5777)
Mar					−0.0033	−0.0190 ***	−0.0069	−0.0036
					(−0.5408)	(−3.1766)	(−1.2666)	(−0.6941)
Tra						0.0742 ***	0.0547 ***	0.0421 ***
						(7.4342)	(6.0183)	(4.7054)
Ind							0.0775 ***	0.0792 ***
							(9.7577)	(10.4181)
Loa								0.0009 ***
								(5.8636)
_cons	0.1057 ***	0.0667 *	0.1024 ***	0.0844 **	0.0953 **	0.0601	−0.0390	−0.0446
	(4.5266)	(1.7503)	(2.9318)	(2.4143)	(2.3579)	(1.5842)	(−1.1070)	(−1.3248)
N	372	372	372	372	372	372	372	372
R2	0.456	0.459	0.554	0.566	0.566	0.625	0.705	0.731
F	24.281	12.996	35.714	30.009	24.018	32.286	48.640	50.883
Province	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes

Note: t-values are shown in parentheses; ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively.

presents the regression results of the two-way fixed effects model, where agricultural green development is taken as the core independent variable and farmers’ income quality as the dependent variable and 7 control variables are gradually incorporated. From these results, we can observe the changes in the impact effect of the core explanatory variable, the role of the control variables, and the overall fitting situation of the model.

In all 8 models, the coefficients of agricultural green development pass the significance test at the 1% level, indicating that the positive impact of agricultural green development on farmers’ income quality has statistical robustness. With the gradual addition of more control variables, the coefficient gradually decreases from 0.3050 in model (1) to 0.1265 in model (8). This change suggests that part of the impact of agricultural green development in the initial model may be indirectly transmitted through the subsequently added control variables, that is, agricultural green development not only directly promotes farmers’ income quality but also may have indirect effects through paths such as improving transportation and promoting industrial upgrading. The R² value of the model increases from 0.456 to 0.731, indicating that after adding the control variables, the explanatory power of the model is significantly enhanced, meaning these control variables collectively explain more variations in farmers’ income quality. The F-values of all models are far greater than the critical value, confirming the reliability of the estimates.

The regression results also reveal varying effects of control variables on farmers’ income quality. Obviously, the coefficient of industrial structure upgrading (Ind) remains stable within the range of 0.0775–0.0792 (all significant at the 1% level), and its coefficient value is the largest among all control variables, making it the core engine for improving income quality. It extends the agricultural industrial chain through the integration of the primary, secondary, and tertiary industries, reduces farmers’ reliance on income from a single agricultural operation, and enhances structural diversification and growth sustainability in income quality.

In Column (8), the coefficient of government support level (Gov) is 0.2625 (significant at the 5% level), and it has shifted from being insignificant to significant compared with the previous models. Its mechanism of action is reflected in two aspects: first, direct subsidies cover the additional costs of farmers’ green transformation, offsetting the short-term negative effects of agricultural environmental friendliness (AEF); second, agricultural infrastructure forms synergy with the indicators of agricultural resource utilization (ARU) and agricultural output quality (AOQ), ultimately achieving a dual improvement in income adequacy and stability.

The coefficient of transportation accessibility (Tra) has remained significant at the 1% level since its inclusion in Column (6), and in Column (8), the coefficient is 0.0421. It serves as a key link connecting agricultural green development and income quality. Its core role is to reduce the circulation costs of green agricultural products, expand the sales radius, and form a complement to the path of agricultural output quality (AOQ).

The coefficient of social security level (Ins) ranges from 0.0066 to 0.0069 in Columns (3)–(5) (all significant at the 1% level) but drops to 0.0009 and −0.0000 in Columns (7)–(8) (insignificant). Its significance in the early models is because agricultural insurance directly stabilizes farmers’ income by compensating for natural risks, serving as the primary means of risk resistance; the subsequent insignificance is due to the inclusion of agricultural-related loans (Loa) and industrial structure upgrading (Ind), which reduces farmers’ reliance on agricultural insurance for income.

Similarly, natural resource endowment (Nat) has a coefficient ranging from −0.0096 to −0.0098 in Columns (4)–(5) (significant at the 1% level), and the coefficient drops to a range of −0.0015 to −0.0030 in Columns (6)–(8) (insignificant). which may be attributed to the fact that farmers have hedged against short-term natural risks through agricultural insurance, diversified planting, non-agricultural employment, and other means, thereby weakening the direct impact of short-term natural risks on the long-term quality of farmers’ income.

The negative coefficients of local market vitality (Mar) reflect a divergence between the number of legal entities in rural markets and farmers’ earnings. This needs to be addressed by optimizing the structure of rural legal entities and enhancing small-scale farmers’ organization. Notably, the coefficient of financial development vitality (Loa) is relatively small (0.0009) yet highly significant, indicating that the impact of financial development vitality on farmer’s income quality exhibits a characteristic of marginal increase, which requires long-term accumulation to manifest more substantial effects.

4.4. Addressing Endogeneity Issues

To mitigate potential endogeneity issues between the core explanatory variable agricultural green development (AGD) and the explained variable farmers’ income quality (FIQ), such as bidirectional causality and omitted variable bias, this study employs three methods for testing: IV-2SLS, GMM based on the lagged terms of the dependent variable, and the lagged terms of the independent variable. The results of the endogeneity test are presented in

Table 5. Results of endogenous test.

IV-2SLS					GMM			Lagged Terms of Independent Variables
Variables	(1)	(2)	(3)	(4)	Variables	(5)	(6)	Variables	(7)	(8)
	AGD	FIQ	AGD	FIQ		FIQ	FIQ		FIQ	FIQ
IV	0.0009		0.0036 **		L.FIQ	0.016	−0.172 **	L.AGD	0.1934 ***	0.0390
	(0.5974)		(2.5442)			(1.038)	(−2.296)		(2.9031)	(0.8038)
AGD		15.9875 ***		2.1219 **	AGD	0.211 ***	0.372 ***
		(3.5949)		(2.2562)		(5.202)	(3.289)
Gov			0.6169 ***	1.5006 **	Gov		−0.304	Gov		0.2856 ***
			(4.5357)	(2.4661)			(−0.439)			(2.5933)
Ins			0.0006	0.0006	Ins		0.050 *	Ins		−0.0003
			(0.6494)	(0.2668)			(1.780)			(−0.4333)
Nat			0.0064 **	0.0129	Nat		−0.043 ***	Nat		−0.0008
			(2.3852)	(1.4705)			(−3.012)			(−0.3516)
Mar			−0.0205 ***	−0.0533 **	Mar		0.010	Mar		0.0068
			(−3.7761)	(−2.1782)			(0.751)			(1.4587)
Tra			−0.0130	0.0084	Tra		0.188 **	Tra		−0.0014
			(−1.3753)	(0.3184)			(1.984)			(−0.1583)
Ind			0.0136 *	0.1031 ***	Ind		−0.274	Ind		0.0815 ***
			(1.6833)	(4.7829)			(−1.528)			(12.2989)
Loa			0.0004 ***	0.0018 ***	Loa		−0.002	Loa		0.0011 ***
			(2.6155)	(3.3519)			(−1.014)			(8.7769)
_cons	0.3612 ***	4.7715	0.2207 ***	0.3604 *	_cons	0.080 **	0.576	_cons	0.1379 ***	−0.0064
	(38.5122)	(0.6142)	(6.0109)	(1.9137)		(2.319)	(1.469)		(5.6796)	(−0.2044)
N	372	372	372	372	N	360	360	N	360	360
R2	0.274		0.452		N_g	12	12	R2	0.514	0.768
F	0.357		14.334		ar1p	0.002	0.004	F	8.428	48.874
chi2		1.834		154.800	ar2p	0.752	0.526
					hansenp	0.215	0.231

Note: Data in parentheses are t-test values; ***, **, * are significant at the 1%, 5%, and 10% levels, respectively.

.

4.4.1. IV-2SLS

The first-stage regression results are shown in Columns (1) and (3). Regarding the correlation test, the impact coefficient of the instrumental variable (IV) on AGD shows a significant positive increase with the inclusion of control variables. When only IV is included, the coefficient is 0.0009 (insignificant); after adding all control variables, the coefficient significantly increases to 0.0036** (significant at the 5% level). Additionally, the F-statistic of the first-stage regression reaches 14.334, which is much higher than the critical value of 10 for judging weak instrumental variables, confirming that there is a sufficient positive correlation between IV and AGD, and there is no issue of weak instrumental variables.

Regarding the exclusion restriction test, first is the exclusion of direct effects. In the regression containing only IV and FIQ (Column 1), the coefficient of IV on FIQ is 0.0009 (insignificant), with R² being only 0.274 and F-statistic 0.357, indicating that IV cannot directly explain FIQ variations. Second is the exclusion of national policy interference. The control variable “government support level (Gov)” can largely isolate the independent effect of national green policies, and the IV adopts “the national average level excluding the target region”, which can offset common policy shocks. Third is the verification of the independence of control variables. After adding IV, the significance of coefficients for control variables such as Gov (0.6169***) and Ind (0.1031***) remains stable without sudden changes in effects, confirming that IV does not indirectly affect FIQ through other paths, thus validating the exclusion restriction.

The second-stage regression aims to identify the net effect of AGD on FIQ. When no control variables are included (Column 2), the coefficient of AGD on FIQ is 15.9875*** (significant at the 1% level); after adding all control variables (Column 4), the AGD coefficient drops to 2.1219* but remains positive. The coefficient decline stems from the control variables absorbing the indirect effects of AGD, yet the core conclusion that AGD significantly promotes FIQ remains unchanged. This indicates that after excluding bidirectional causality and omitted variable bias, the positive driving effect of agricultural green development on farmers’ income quality still possesses statistical reliability.

4.4.2. GMM Based on the Lagged Terms of the Dependent Variable

GMM mitigates dynamic endogeneity (where lagged FIQ inversely affects current AGD) by introducing the lagged term of the dependent variable (L.FIQ) as an instrumental variable. When only L.FIQ and control variables are included (Column 5), the coefficient of L.FIQ is 0.016 (insignificant); its positive trend reflects a weak path dependence of FIQ, though the effect is not statistically significant. After adding the current AGD (Column 6), the coefficient of L.FIQ turns to −0.172**, and the coefficient of the current AGD is 0.372***. This indicates that AGD can break the path dependence of FIQ, shifting from the traditional production model relying on high input to improving income quality through green transformation. The p-value of the ar1 test is 0.002, rejecting the null hypothesis of no first-order autocorrelation; the p-value of the ar2 test is 0.752, failing to reject the null hypothesis of no second-order autocorrelation and meaning that there is no high-order autocorrelation in the residuals. The p-value of the Hansen overidentification test is 0.215, failing to reject the validity of the instrumental variables, which confirms the rationality of the model specification.

4.4.3. Lagged Terms of the Independent Variable

The impact of AGD on FIQ is not limited to the current period but exhibits a time-lag characteristic. Introducing the lagged term of agricultural green development (L.AGD) can accurately capture the long-term effect and mitigate endogeneity, with results shown in Columns (7) and (8). When only L.AGD and control variables are included (Column 7), the coefficient of L.AGD is 0.1934*** (significant at the 1% level). This indicates that lagged AGD exerts a sustained positive driving effect on the current FIQ through ecological cumulative effects—such as improved soil quality and enhanced recognition of green agricultural products. After adding the current AGD (Column 8), the coefficient of the current AGD is 0.0390 (insignificant), while the positive effect of L.AGD remains stable. This reflects that the long-term impact of AGD on FIQ is greater than its short-term impact, which may be attributed to the fact that current green input costs increase while benefits are delayed, and the dividends from lagged green transformation are continuously released.

4.5. Robustness Tests

Then, a robustness check was conducted to further confirm the findings of the benchmark regression, with the adoption of the following three methods:

4.5.1. Replacing the Measurement Approach

Principal component analysis (PCA) was performed separately on the three sub-dimensions of the independent variable “agricultural green development” (agricultural resource utilization, agricultural environmental friendliness, and agricultural output quality) and the four sub-dimensions of the dependent variable “farmers’ income quality” (income adequacy, income structure, income growth, and income knowledge intensity). Comprehensive indicators were constructed to replace the original variables (models 1–2 in

Table 6. Results of robustness test.

Variables	(1)	(2)	(3)	(4)
	FIQ	FIQ Replacement	FIQ	FIQ
AGD replacement variable	0.0054 ***
	(2.7589)
Gov	0.3046 **	1.0815 **	0.1518	0.4367 ***
	(2.3658)	(2.0486)	(1.0763)	(2.7981)
Ins	0.0000	0.0049	0.0010	0.0006
	(0.0557)	(1.4621)	(0.9891)	(0.5197)
Nat	−0.0010	−0.0014	−0.0017	0.0001
	(−0.3769)	(−0.1350)	(−0.6132)	(0.0327)
Mar	−0.0046	−0.0598 ***	−0.0046	0.0107 **
	(−0.8688)	(−2.7984)	(−0.6771)	(2.1149)
Tra	0.0428 ***	0.3293 ***	0.0471 ***	0.0376 ***
	(4.6943)	(9.0453)	(4.5339)	(3.1253)
Ind	0.0783 ***	0.3101 ***	0.0853 ***	0.0886 ***
	(10.0956)	(10.0333)	(9.8358)	(8.8299)
Loa	0.0009 ***	0.0060 ***	0.0021 ***	0.0009 ***
	(6.1694)	(9.9689)	(7.7802)	(4.3623)
AGD		0.8516 ***	0.1776 ***	0.1925 ***
		(4.1725)	(2.9878)	(3.1723)
_cons	−0.0071	−1.3379 ***	−0.0795 **	−0.1394 ***
	(−0.2265)	(−9.7664)	(−2.1139)	(−3.7429)
N	372	372	279	372
R2	0.728	0.804	0.758	0.472
F	49.845	98.379	55.782	40.514

Note: Data in parentheses are t-test values; ***, ** are significant at the 1%, and 5% levels, respectively.

). In Model (1), the coefficient of the AGD replacement variable is 0.0122, which is significantly positive at the 1% level. This indicates that even when the measurement method of agricultural green development is changed, its positive impact on farmers’ income quality remains significant. In Model (2), the coefficient of the core variable AGD on FIQ replacement variable is 1.2923, which is significantly positive at the 1% level. These results suggest that the positive impact of agricultural green development on farmers’ income quality is stable regardless of whether the independent variable or the dependent variable is measured using the comprehensive principal component method of sub-dimensions.

4.5.2. Excluding Special Years

The impact of special factors such as the pandemic during 2020–2022 may have exerted unconventional shocks on farmers’ income quality and agricultural green development. To verify the robustness of the impact effect in normal years, samples from these three years were excluded. In Model (3), the coefficient of the core variable AGD is 0.2837, which is significantly positive at the 1% level. This indicates that after excluding the special years, the positive impact of agricultural green development on farmers’ income quality remained significant and did not fluctuate substantially due to sample adjustment.

4.5.3. ARIMA Interpolation Method

To further rule out the interference of missing sample data and abnormal fluctuations on the core conclusions, this study employs the ARIMA interpolation method to process the panel data of 31 provincial-level administrative regions from 2011 to 2022. Based on the complete interpolated data, a two-way fixed effects regression is re-conducted to test the stability of the conclusion regarding the impact of AGD on FIQ.

From the regression results (Model 4), the regression coefficient of the core explanatory variable AGD on FIQ after interpolation is 0.1925***, which is significantly positive at the 1% level. Compared with the AGD coefficient in the benchmark regression (0.1265**), the sign direction is consistent and remains significant, with the numerical fluctuation range controlled within 5% and no drastic deviations. This proves that the positive driving effect of agricultural green development on farmers’ income quality is not affected by data completeness, and the core conclusion possesses statistical robustness. Additionally, the effect of each control variable after interpolation are highly consistent with those in the benchmark regression, further indicating that the moderating effects of these variables on farmers’ income quality are stable and not disturbed by the data interpolation process.

In summary, the three robustness test methods verify the robustness of the conclusion that agricultural green development significantly promotes farmers’ income quality from different perspectives. This conclusion is not affected by variable measurement methods, disturbed by special external shocks, or restricted by data completeness. It can provide a reliable empirical basis for the design of subsequent differentiated policies.

4.6. Structural Heterogeneity Analysis

4.6.1. The Impact of the Sub-Dimensions of Agricultural Green Development on Farmers’ Income Quality

Agricultural green development includes three sub-dimensions: agricultural resource utilization (ARU), agricultural environmental friendliness (AEF), and agricultural output quality (AOQ). Models (1)–(3) in

Table 7. Structural regression results.

Variables		(1)	(2)	(3)	(4)	(5)	(6)	(7)
		Influence of the Sub-Dimensions of AGD			Influence on the Sub-Dimensions of FIQ
		FIQ	FIQ	FIQ	Adequacy	Structure	Growth	Knowledge Intensity
AGD	ARU	0.0330 ***			0.1901 ***	0.4047 ***	−0.0425 *	0.0769
		(2.5698)
	AEF		−0.0252 **
			(−2.4454)
	AOQ			0.1770 ***	(5.0417)	(5.3079)	(−1.8738)	(0.7702)
				(4.8732)
Gov		0.3188 **	0.3283 **	0.2514 **	0.1831 *	0.0609	−0.0174	0.7415 ***
		(2.4516)	(2.5604)	(2.0117)	(1.8779)	(0.3090)	(−0.2968)	(2.8722)
Ins		0.0000	0.0000	0.0001	0.0021 ***	0.0026 **	−0.0003	−0.0035 **
		(0.0274)	(0.0502)	(0.1502)	(3.4037)	(2.0810)	(−0.8763)	(−2.0996)
Nat		−0.0008	−0.0007	−0.0020	0.0009	0.0027	−0.0003	−0.0080
		(−0.3091)	(−0.2724)	(−0.8144)	(0.4467)	(0.7097)	(−0.2415)	(−1.5830)
Mar		−0.0059	−0.0059	0.0001	−0.0263 ***	−0.0022	0.0008	−0.0007
		(−1.1315)	(−1.1215)	(0.0276)	(−6.6629)	(−0.2770)	(0.3251)	(−0.0649)
Tra		0.0421 ***	0.0414 ***	0.0228 **	0.1133 ***	0.0081	−0.0072 *	0.0947 ***
		(4.3963)	(4.4130)	(2.4227)	(16.8487)	(0.5930)	(−1.7704)	(5.3198)
Ind		0.0796 ***	0.0795 ***	0.0846 ***	0.0289 ***	0.1068 ***	0.0074 **	0.1208 ***
		(10.1783)	(10.0046)	(11.3560)	(5.0609)	(9.2547)	(2.1527)	(7.9914)
Loa		0.0009 ***	0.0009 ***	0.0007 ***	0.0018 ***	0.0006 **	−0.0001	0.0016 ***
		(6.2275)	(6.2205)	(4.2917)	(16.4988)	(2.5040)	(−0.7590)	(5.4047)
_cons		−0.0183	0.0072	−0.0484	−0.0127	−0.2065 ***	0.4176 ***	−0.2322 ***
		(−0.5447)	(0.1388)	(−1.5520)	(−0.5020)	(−4.0355)	(27.4439)	(−3.4655)
N		372	372	372	372	372	372	372
R2		0.726	0.726	0.743	0.896	0.497	0.991	0.459
F		49.287	49.254	55.489	177.964	40.869	1.551	29.376

Note: Data in parentheses are t-test values; ***, **, * are significant at the 1%, 5%, and 10% levels, respectively.

examine their impacts on farmers’ income quality, and the results show significant heterogeneity in their effects:

In Model (1), the coefficient of ARU is 0.0330***, which is significantly positive at the 1% level. The optimization of agricultural resource utilization, such as the efficient allocation of land and water resources and the rational input of agricultural production factors, is an important way to improve farmers’ income quality.

In Model (2), the coefficient of AEF is −0.0252** and significantly negative at the 1% level, indicating an inhibitory effect on farmers’ income quality in the short term. Essentially, this stems from a structural contradiction characterized by a surge in green transformation costs, lagging benefit conversion, and the weak risk resistance capacity of smallholder farmers. From the cost perspective, environmentally friendly practices require the use of green agricultural inputs, such as biological pesticides, organic fertilizers, and degradable agricultural films, leading to a significant increase in production costs. There are also hidden costs associated with environmentally friendly practices. For instance, green pest control requires farmers to invest more time in monitoring field ecosystems; farmland ecological restoration directly reduces the current sowing area. However, current agricultural green subsidy policies mostly focus on the end stage (e.g., subsidies for fertilizer reduction, ecological compensation), failing to form a full-chain cost-sharing mechanism. Additionally, issues such as insufficient compensation amounts and delayed disbursement persist [104]. Furthermore, the scale effect of environmentally friendly practices is more pronounced. Large-scale business entities (e.g., family farms, cooperatives) can reduce unit costs through centralized procurement of green agricultural inputs and unified technical services. However, China’s agricultural business entities are dominated by smallholder farmers, who, due to their small operation scale and weak bargaining power, struggle to gain cost advantages in the green supply chain [105].

From the perspective of benefits, the gains from environmentally friendly practices need to be transmitted through multiple links. In terms of ecological benefits, effects such as improved soil organic matter and reduced agricultural non-point source pollution take 3–5 years to manifest as stable yields or improved quality, making it difficult to offset cost pressures in the short term [106]. In terms of market benefits, green agricultural products can only achieve price premiums through links such as organic certification and brand building. Particularly, the green transformation premiums for bulk crops like corn and soybeans are generally lower than those for cash crops, and yields may fluctuate during the transformation period.

In Model (3), the coefficient of AOQ is 0.1770***, which is significantly positive at the 1% level. Its t-value (4.8732) is the largest among the three sub-dimensions, indicating that the correlation between AOQ and farmers’ income quality is highly stable. The underlying logic is that high-quality agricultural products can directly increase farmers’ operational income through higher market premiums, stronger market competitiveness, and more stable sales channels.

4.6.2. The Impact of Agricultural Green Development on the Sub-Dimensions of Farmers’ Income Quality

Farmers’ income quality includes four sub-dimensions: income adequacy, income structure, income growth, and income knowledge intensity. Models (4)–(7) in Table 7 examine the impact of agricultural green development on these dimensions, with results showing obvious differentiation.

In Model (4), the coefficient of agricultural green development on income adequacy is 0.1901***, which is significantly positive at the 1% level. This indicates that agricultural green development can significantly improve farmers’ income adequacy. The reason lies in that the cost savings from improved efficiency of agricultural resource utilization and the premium income from enhanced agricultural output quality jointly increase the total current income of farmers.

In Model (5), the coefficient of agricultural green development on income structure is 0.4047***, which is significantly positive at the 1% level, and the coefficient value is the largest among the four sub-dimensions. This means that agricultural green development plays the most prominent role in optimizing farmers’ income structure. Agricultural green development can drive related industries, such as leisure agriculture, agricultural products intensive processing, and eco-tourism, increasing farmers’ non-agricultural operating income and wage income. At the same time, environment-friendly agriculture may create new income sources such as ecological compensation and carbon sink trading, further expanding income sources and reducing reliance on a single agricultural income stream.

In Model (6), the coefficient of agricultural green development on income growth is −0.0425*, which is weakly significantly negative at the 10% level. This indicates that agricultural green development has a certain inhibitory effect on the growth of farmers’ income. A possible explanation is that agricultural green development has significant long-term investment attributes, such as soil improvement, green technology promotion, and brand cultivation, which require large capital investment in the short term, thereby inhibiting the short-term growth momentum of income. In addition, the cultivation of the green agricultural product market and the formation of consumers’ awareness take time.

In Model (7), the coefficient of agricultural green development on income knowledge intensity is 0.0769, which fails to pass the significance test. This suggests that at the current stage, the impact of agricultural green development on farmers’ knowledge-based income is not significant. Combined with proxy variables, a possible reason is that the rural education system has not effectively aligned with the demands of green agriculture, resulting in a mismatch between farmers’ educational levels and the skills required for green agriculture.

4.7. Spatial Heterogeneity Analysis

4.7.1. Geographical Location

Based on the regression results of the northeast, east, central, and west regions, the impact of agricultural green development on farmers’ income quality shows significant regional heterogeneity (as shown in

Table 8. Results of spatial heterogeneity regression.

Variables	Northeast	East	Central	West	Low Crop Failure	High Crop Failure	Low Government Support	High Government Support	Low Market Vitality	High Market Vitality
	FIQ	FIQ	FIQ	FIQ	FIQ	FIQ	FIQ	FIQ	FIQ	FIQ
AGD	−0.3693 *	0.0164 ***	0.1405	0.2168 ***	0.0029	0.1235 **	0.1676 *	0.0756	0.1783 **	0.0449
	(−1.8289)	(3.1041)	(1.9697)	(3.5979)	(0.0318)	(2.1305)	(1.8298)	(1.2807)	(2.2539)	(0.6755)
Gov	−0.3612	−0.1678	−0.5859 ***	−0.2893 *	0.2914	−0.0427	0.3580	0.4167 **	0.3432	−0.1235
	(−1.3507)	(−0.6806)	(−5.1825)	(−1.7336)	(1.5178)	(−0.2591)	(0.7738)	(2.1245)	(1.5528)	(−0.7788)
Ins	0.0000	−0.0017	0.0090 ***	0.0035	−0.0005	0.0017	−0.0006	−0.0004	0.0009	0.0039
	(0.0056)	(−1.5080)	(3.0439)	(1.5426)	(−0.4831)	(0.7875)	(−0.3101)	(−0.4581)	(0.6956)	(0.9540)
Nat	−0.0001	−0.0020	−0.0066 ***	0.0021	−0.0155	−0.0048 *	−0.0036	0.0002	−0.0020	−0.0037
	(−0.0259)	(−0.1846)	(−3.5689)	(0.7149)	(−0.4292)	(−1.8160)	(−0.5657)	(0.0661)	(−0.4259)	(−1.3651)
Mar	0.0342	−0.0276 **	0.0028	0.0236 ***	−0.0140 *	0.0157 **	−0.0031	−0.0008	−0.0199	0.0041
	(0.7870)	(−2.2729)	(1.0784)	(3.1941)	(−1.9029)	(1.9775)	(−0.3443)	(−0.1231)	(−0.6667)	(0.8107)
Tra	−0.3663 **	0.2332 ***	0.0422 ***	−0.0172	0.0798 ***	−0.0265 **	0.0380 **	0.0317 ***	0.0752 ***	−0.0118
	(−2.5552)	(8.3683)	(4.6402)	(−1.2779)	(6.2721)	(−2.0535)	(2.4927)	(2.8931)	(5.0328)	(−1.1780)
Ind	0.0533	0.0678 ***	0.0742 ***	−0.0295	0.0728 ***	0.0863 ***	0.0810 ***	0.0848 ***	0.0675 ***	0.0310
	(1.5966)	(6.2109)	(2.9828)	(−1.2399)	(7.0675)	(7.9134)	(6.3019)	(8.9822)	(5.8777)	(1.4968)
Loa	0.0014	0.0012 ***	0.0023 ***	−0.0028 **	0.0007 ***	0.0021 ***	0.0005 **	0.0022 ***	0.0006	0.0009 ***
	(0.2367)	(6.2030)	(3.7822)	(−2.2299)	(3.9708)	(5.9638)	(2.4240)	(8.1021)	(1.1851)	(7.6583)
_cons	0.5607 ***	−0.1120	0.1439 **	0.2074 ***	−0.0110	0.0285	−0.0751	−0.0819	−0.0930	0.1691 ***
	(3.0407)	(−1.3632)	(2.4590)	(3.8909)	(−0.1940)	(0.6697)	(−0.9739)	(−1.4692)	(−1.4756)	(3.8740)
N	36	108	72	156	186	186	186	186	186	186
R2	0.979	0.818	0.990	0.759	0.773	0.753	0.646	0.822	0.718	0.840
F	10.978	31.420	8.345	3.875	33.083	21.641	16.582	44.374	29.589	11.616
Inter-group Differences	0.074 * (0.086)				0.054 *** (2.65)		0.095 *** (3.28)		0.043 ** (2.01)

Note: Data in parentheses are t-test values; ***, **, * are significant at the 1%, 5%, and 10% levels, respectively.

):

• Northeast region. There is a weakly significant negative impact between agricultural green development and farmers’ income quality in the northeast region (−0.3693*). From the perspective of the agricultural industrial structure, the northeastern region, as a typical major grain-producing area in China, has large-scale cultivation of bulk crops such as corn and soybeans accounting for over 60% of its total agricultural cultivation. A traditional production path of high input–high output has been formed over the long term, and farmers have developed path dependence on chemical inputs. When promoting green transformation, it is necessary, on the one hand, to replace chemical inputs with high-priced green agricultural materials such as biological pesticides and organic fertilizers and, on the other hand, to adjust planting technologies. This leads to a situation where the short-term increase in production costs far exceeds the income increment brought by the improvement in unit yield. More critically, the green attributes of bulk grain crops struggle to form a market differentiation advantage. The market premium for green corn and soybeans in the northeast is generally less than 10%, which is far lower than the premium level of over 50% for cash crops. This directly restrains the adequacy and growth potential of farmers’ income, trapping farmers in the dilemma of high quality without a corresponding high price, which leads to the phenomenon of inferior coins driving out good coins [107]. Therefore, in the short term, the green production behavior of farmers may not have a significant impact on increasing agricultural income. Furthermore, from the perspective of technical conditions, the agricultural technology promotion system in the northeastern region still focuses primarily on traditional high-yield technologies, with insufficient coverage of green technology promotion. For instance, green technologies with relatively high maturity, such as straw incorporation and decomposition technology and biological control technology, lack localized adaptation, leading to their actual application effects falling short of expectations, which further reduces farmers’ willingness to adopt green technologies.

From the perspective of control variables, government support level (Gov) does not reach a significant level (−0.3612), indicating that the allocation of fiscal funds in supporting agricultural green transformation has a low degree of matching with farmers’ needs. For example, excessive investment in infrastructure (e.g., the construction of high-standard farmland) rather than direct subsidies fails to alleviate the cost pressure of green transformation, or the precision and execution efficiency of agricultural support policies need to be improved. Therefore, the structural bias in policy support can explain the negative results in the northeastern region to a certain extent.

Moreover, transportation accessibility (Tra) has a significant inhibitory effect (−0.3663**). A possible reason is that improved transportation makes it easier for rural labor to transfer to cities or other industries, resulting in the loss of agricultural production factors. The per capita agriculture-related loans (Loa) do not reach a significant level (0.0014), indicating that agriculture-related loans have not been effectively converted into productive inputs. For instance, some loans have flowed into non-agricultural sectors. The impact of industrial structure upgrading (Ind) on farmers’ income quality has not yet emerged (0.0533), which forms a sharp contrast with the east, central, and west regions. This reflects the singular industrial structure, the low degree of industrial integration, and the lagging development of new rural industrial entities in the northeast.

• Eastern region. In the eastern region, agricultural green development exhibits significant positive effect (0.0164***). This is the result of the combined effects of the market, technology, and policies, among which a high degree of marketization and sound market mechanisms are the most fundamental driving factors. High-income groups drive stable demand for green agricultural products. Green products can achieve market premiums through a mature price transmission mechanism; a well-developed e-commerce and cold chain system shortens the circulation chain and ensures the retention of premiums. Diversified market entities (e.g., leading agricultural enterprises, cooperatives) provide smallholder farmers with green inputs, guaranteed minimum price purchases, and technical support through the “enterprise + base + farmer” model. This reduces transformation costs and market risks, directly promoting improvements in income adequacy and structural optimization. Furthermore, the high contribution rate of agricultural technological progress in the eastern region, the value chain extension brought by the integration of the primary, secondary, and tertiary industries, and policy innovations, all function based on a mature market, collectively amplifying the income effect of green development. The experience of the eastern region shows that the degree of market development determines the efficiency of green value conversion. Only when a market mechanism for green agricultural products (characterized by stable demand, reasonable prices, and efficient circulation) is established can the effective conversion of ecological value into economic value be achieved.

Among control variables, transportation accessibility (Tra) (0.2332***), industrial structure upgrading (Ind) (0.0678***), and financial development vitality (Loa) (0.0012***) in the eastern region all exert a significant positive impact on farmers’ income quality at the 1% level. A well-established transportation network reduces the circulation costs of green agricultural products; developed non-agricultural industries extend the green agricultural industrial chain; and sufficient credit support alleviates farmers’ financial constraints in green transformation. These three factors collectively reinforce the income-enhancing effect of green development. The insignificant impact of government support level (Gov) in the eastern region (−0.1678) indicates that its green transformation relies more on market mechanisms than direct government investment. Local market vitality (Mar) has a significantly negative impact coefficient (−0.0276**). The eastern region has a large number of rural legal entities (e.g., small cooperatives, family farms), but they face the issue of homogeneous operations. This leads to intensified market competition and compressed profit margins, which instead exerts an inhibitory effect on farmers’ income quality.

• Central region. In the central region, the impact of agricultural green development on farmers’ income quality fails to pass the significance test (0.1405). This result is highly correlated with regional policy orientations. As an important major grain-producing area, the central region has long undertaken policies related to the national food security strategy, with these policies focusing on green production transformation based on ensuring food security—such as the National High-Standard Farmland Construction Plan (2021–2030) and the Action Plan for Reducing Fertilizer and Pesticide Use and Improving Efficiency. However, such policies suffer from the issue of delayed scale effects. On the one hand, the green attributes of bulk grain crops struggle to form market premiums; on the other hand, direct subsidies in the policies are insufficient, forcing smallholder farmers to bear the costs of green transformation (the coefficient of government support level (Gov) in the central region is 0.2053, insignificant). This leads to short-term cost pressures offsetting income growth, ultimately preventing the income-driven effect of green development from being fully unleashed.

Among control variables, the positive impact of industrial structure upgrading (Ind) on farmers’ income quality is the strongest among the four regions (0.0742***). The central region boasts a solid foundation in the agricultural product processing industry and has formed a complete industrial chain covering cultivation, processing, and sales. For example, relying on wheat cultivation, Henan Province has developed deep processing industries such as flour processing and instant noodle manufacturing; Hunan and Hubei Provinces have focused on livestock and poultry breeding to develop meat product processing. These efforts convert primary agricultural products into high-value-added products, which not only increases the unit output value of agricultural products but also drives farmers to participate in employment in the processing sector, directly promoting the improvement of income quality.

The positive effect of financial development vitality (Loa) is the strongest among the four regions (0.0023***). The central region is in a stage of rapid development of large-scale and intensive agriculture, where farmers have a strong demand for production funds. Agriculture-related loans are directed to productive areas, which can directly improve agricultural production efficiency. Social security level (Ins) is a significant positive risk protection variable unique to the central region (0.0090***). As a core national grain-producing area, the central region has a high proportion of staple grain cultivation such as wheat and rice and is frequently affected by natural disasters such as floods, droughts, and pests. Agricultural insurance compensation can directly cover yield losses and planting costs caused by disasters, serving as an important guarantee for farmers’ income quality in this region. Transportation accessibility (Tra) exerts a significant positive impact (0.0422***). Relying on a high-density expressway network, the central region has built an efficient cross-regional circulation system for agricultural products. Government support level (Gov) exhibits the strongest and most significant fiscal inhibitory effect in the central region (−0.5859***). A possible reason is the low efficiency of fund utilization, or funds are excessively tilted toward traditional staple grain cultivation while providing insufficient support for characteristic agriculture with strong market demand. Natural resource endowment (Nat) shows a significant negative impact (−0.0066***). Farmers’ income in the central region is highly dependent on crop cultivation, so crop failure has a more severe impact on their income, and the increase in crop failure area exerts a more significant inhibitory effect on income quality.

• Western Region. In the western region, the regression coefficient of agricultural green development on farmers’ income quality is 0.2168***, which is significantly positive at the 1% significance level. Notably, this coefficient value is higher than that of the eastern region (0.0164***), making the western region the one with the strongest positive effect among the four major regions. The reasons for this are as follows: first, agriculture in the western region has long relied on extensive cultivation, resulting in relatively low basic income for farmers from agricultural activities. Agricultural green development endows agricultural products in the western region with an ecological premium. By introducing green agricultural product processing enterprises and establishing e-commerce channels, primary products that were originally sold at low prices are transformed into high-value-added green processed products. Meanwhile, this process creates job opportunities in warehousing, logistics, and other fields, driving the dual growth of farmers’ wage income and operational income. In addition, green agriculture has given rise to new business formats such as under-forest economy and eco-tourism, opening up new channels for income increase. These new formats also attract farmers to return to their hometowns for employment, leading to a more significant effect of income improvement.

Second, in accordance with the Opinions of the Ministry of Agriculture and Rural Affairs on Accelerating Agricultural Green Development and the Catalogue of Encouraged Industries in Western China (2020), the agricultural green development policies of western provinces focus on the dual goals of ecological protection and characteristic industries. On the one hand, national policies such as ecological compensation and green agriculture subsidies are tilted toward the western region, reducing farmers’ costs in the green transformation process. This enables farmers to obtain both production income and ecological income, optimizing the structure of farmers’ income. For instance, in 2022, the grassland ecological protection subsidy in Inner Mongolia covered nearly 90% of herding households, with an average annual subsidy of 12,000 CNY per household [108]. On the other hand, the western region boasts richer ecological resources and less industrial pollution, giving it superior innate conditions for agricultural greening. The integration of characteristic agricultural products with green attributes easily forms a differential competitive advantage, generating substantial market premiums, while facing relatively low competitive pressure. Taking Xinjiang as an example, by combining water-saving agriculture with the cultivation of characteristic fruits (such as red dates and grapes). The premium for green fruit products reached 35–50% [109], significantly enhancing income adequacy.

Among the control variables, local market vitality (Mar) exerts a significantly positive impact (0.0236***). The western region has a small number of rural legal entities; an increase in such entities can drive employment, provide technical support, and directly boost farmers’ income. The coefficient of government support level (Gov) is −0.2893*, and this negative trend indicates that there is a structural mismatch in financial funds in the western region, as well as efficiency loss in the implementation process. These funds have not been effectively converted into growth in farmers’ income and have even exerted a reverse constraint due to resource crowding-out effects. The coefficient of financial development vitality (Loa) is −0.0028**, which reflects the imperfection of the rural financial system in the western region. A relatively high proportion of loans flows to non-agricultural sectors and has not been effectively transformed into investment in green agricultural production. The effect of transportation accessibility (Tra) is not significant (−0.0172). This is because most characteristic agricultural products in the western region rely on long-distance transportation. Although improved transportation shortens the transportation time, logistics costs rise due to the need for fresh-keeping measures. As a result, a positive effect of transportation has not been revealed. The insignificance of industrial structure upgrading (Ind) (−0.0295) indicates that industrial integration in the west is in its initial stage, with its effects yet to be fully realized.

4.7.2. Natural Resource Endowment

Natural resource endowment is a core factor affecting crop failure area. Regions with low crop failure areas correspond to those with better natural resource endowment, while those with high crop failure areas are mostly regions with poor natural resource endowment. Such differences in endowment lead to significant heterogeneity in the impact of agricultural green development on farmers’ income quality. The data in Table 8 shows that, in the low crop failure area group, the coefficient of agricultural green development is 0.0029, which fails the significance test. This may be because the traditional agricultural production mode in regions with stable resource endowment can already guarantee basic income, resulting in relatively limited marginal benefits of agricultural green development. In the high crop failure area group, the coefficient of agricultural green development is 0.1235**. This indicates that in regions with weak resource endowment, agricultural green development can more effectively improve farmers’ income quality, possibly because agricultural green development directly improves fragile production conditions and alleviates the impact of natural risks on income.

The role of control variables also exhibits differential characteristics related to crop failure areas. Industrial structure upgrading (Ind) and financial development vitality (Loa) both exert significant positive impacts on farmers’ income quality in both groups of samples (p < 0.01), indicating that industrial diversification and financial support are universal paths to improve farmers’ income quality. Particularly, regions with high crop failure rates have a more urgent demand for funds due to poor natural resource endowment, resulting in a stronger marginal effect of agriculture-related loans. Transportation accessibility (Tra) shows a significant positive impact in the low crop failure group (0.0798***) but a significant negative impact in the high crop failure group (−0.0265**), reflecting that the effect of infrastructure improvement depends on natural resource endowment. Local market vitality (Mar) has a significant negative impact in the low crop failure group (−0.0140*) but a positive impact in the high crop failure group (0.0157**), suggesting that regions with premium resources need to guard against legal entities squeezing farmers’ benefits.

4.7.3. Government Support Level

From the results of the grouped regression, the impact of agricultural green development on farmers’ income quality shows significant heterogeneity due to different government support levels for agriculture. Table 8 shows that, in the group with low government support for agriculture, the coefficient of agricultural green development is 0.1676*, which is significant at the 10% level; in the group with high government support for agriculture, this coefficient drops to 0.0756, which is not significant. This indicates that the role of agricultural green development in improving farmers’ income quality is stronger in regions with lower government support for agriculture. A possible reason is that agricultural production in the low government support group is more dependent on market-oriented green transformation. Farmers tend to reduce costs and increase income through market-oriented approaches such as improving resource utilization efficiency and creating premiums for green agricultural products, thereby forming an endogenous driving force for green transformation. However, in regions with high fiscal support for agriculture, a policy path dependence is likely to form; farmers rely excessively on fiscal subsidies, which reduces their willingness to independently adopt green technologies and, in some cases, even leads to subsidy-dependent production.

The differentiation in the role of fiscal support for agriculture itself is equally notable. In regions with a low level of fiscal support for agriculture, the coefficient of government support (Gov) is 0.3580 and not significant; in regions with a high level of fiscal support, this coefficient reaches 0.4167**, indicating a significant positive effect at the 5% significance level. The key reason lies in the fact that regions with low fiscal support face a large funding gap, and their agricultural-related expenditures are mostly allocated to addressing shortcomings in infrastructure, resulting in insufficient direct subsidies to farmers. In contrast, government expenditures in regions with high fiscal support are more directed toward areas that benefit farmers, such as green technology training and ecological compensation, enabling the effective conversion of fiscal funds into income growth. In addition, the moderating effects of transportation accessibility (Tra), industrial structure upgrading (Ind), financial development vitality (Loa) are all deeply tied to the level of fiscal support for agriculture. Among these, financial development vitality is the most prominent. The coefficient for low fiscal support regions is 0.0005**, while that for high fiscal support regions reaches 0.0022***. This can be attributed to the collaborative mechanism between fiscal and financial policies in high fiscal support regions. For example, government guarantees lower the threshold for loans and guide funds to flow precisely into green production. In contrast, low fiscal support regions lack such collaborative policies, leading to inadequate loan coverage and scattered fund flows, which weakens the income increasing effect of financial support.

4.7.4. Local Market Vitality

From the grouped regression results, the impact of agricultural green development on farmers’ income quality exhibits significant heterogeneity due to differences in local market vitality (Mar). Table 8 shows that, in the group with fewer legal entities in the primary industry, the coefficient of agricultural green development is 0.1783**, which passes the 5% significance test. This indicates that in low market vitality group, agricultural green development plays a stronger role in improving farmers’ income quality. In the group with more legal entities in the primary industry, the coefficient of agricultural green development drops to 0.0449, not significance. This may be because in high market vitality group, farmers are mostly confined to the production link and have weak bargaining power, while profits flow more to legal entities in links such as processing and sales. As a result, the impact of agricultural green development on farmers’ income quality is significantly weakened.

The differentiation in the role of fiscal support for agriculture itself is equally notable. In regions with a low marketization level, the coefficient of government support level (Gov) is 0.3432 and not significant, but it shows a positive trend. In regions with a high marketization level, this coefficient is −0.1235, showing a negative trend. The reasons are as follows: In regions with low marketization, the market mechanism is underdeveloped, so fiscal expenditure on agriculture can directly make up for market shortcomings. Although the effect does not reach a significant level due to the limited scale of funds, the positive potential has already emerged. In contrast, in regions with high marketization, fiscal expenditure tends to be prioritized by rural legal entities, leaving small-scale farmers with limited benefits. For instance, in the reverse lease and sub-contracting model, high-standard farmlands supported by fiscal funds are leased and operated by enterprises, while farmers only receive rents that are lower than the income from independent farming. This weakens the correlation between fiscal expenditure and the growth of farmers’ income, resulting in a slight negative trend.

For financial market vitality (Loa), the coefficient in regions with a low marketization level is 0.0006, showing a positive but not significant trend; in regions with a high marketization level, the coefficient reaches 0.0009***, which is significantly positive at the 1% significance level. This is attributed to the fact that in regions with a high level of marketization, enterprises can provide guarantees for farmers’ green loans, lower the loan threshold, and guide farmers to direct their loan funds toward green production that aligns with the enterprises’ needs, resulting in high efficiency in loan utilization. In contrast, low-marketization regions lack guarantee entities, leading to a low approval rate for farmers’ loans. Additionally, a portion of the funds flows into non-agricultural sectors and fails to be converted into investment in green production, ultimately resulting in a weak effect of financial support.

In addition, the moderating effects of transportation accessibility (Tra) and industrial structure upgrading (Ind) also show obvious differentiation. For transportation accessibility (Tra), in regions with a low marketization level, the coefficient of Tra is 0.0752***, which is significantly positive at the 1% significance level. Farmers rely on basic transportation to shorten transportation time and reduce circulation losses, directly driving an increase in income. In regions with a high marketization level, the coefficient of Tra is −0.0118, showing a negative trend. This may be because improved transportation is more conducive to enterprise’ cross-regional resource allocation, while farmers do not gain much benefits from it. For industrial structure upgrading (Ind), in regions with a low marketization level, the coefficient of Ind is 0.0675***, which is significantly positive at the 1% significance level. This benefit stems from the income growth of farmers brought by the processing and added value of primary agricultural products. In regions with a high marketization level, the coefficient of Ind is 0.0310, showing a positive but not significant trend. The reason lies in that industrial upgrading is led by enterprises, and farmers only obtain wage income from working, without sharing the higher value-added benefits of it.

5. Discussion

5.1. Main Results

This study set out to explore how agricultural green development impacts farmers’ income quality from direct, structural, and heterogeneous perspectives, and the empirical findings fully address this core objective, with clear evidence confirming the achievement of each research dimension, as detailed below:

First, this research has validated the substantial direct impact of agricultural green development on promoting farmers’ income quality. This finding advances the understanding of “Ecological Prosperity” theory by empirically validating that agricultural green transformation can synergize ecological protection and income growth. The “Ecological Prosperity” theory advocates transforming to eco-friendly production methods, improving ecological environment quality, and efficiently utilizing ecological resources to form new economic growth points, thereby realizing the conversion of ecological benefits into economic benefits [110,111]. The study finds that agricultural green development drives the improvement of farmers’ income quality through the dual pathways of agricultural resource utilization (ARU) and agricultural output quality (AOQ). This is consistent with the transmission chain of “ecological protection → efficient resource use → economic income growth” in the Ecological Prosperity Theory. Meanwhile, the study identifies the existence of phased characteristics of this theory in the agricultural context. Agricultural environmental friendliness (AEF) exerts an inhibitory effect on income quality in the short term. This “Green Transition Paradox” supplements the dynamic evolution perspective of the Ecological Prosperity Theory. Specifically, the conversion of ecological benefits into economic benefits requires crossing a cost threshold. In the short term, the costs of ecological governance may offset income growth; however, in the long run, improvements in soil quality and ecological compensation policies will gradually unlock income-increasing potential [112]. This echoes the view proposed by Daly et al. (2011) that the accumulation of ecological capital requires long-term investment [110]. This conclusion not only expands the application boundary of the Ecological Prosperity Theory in the agricultural field but also broadens the research perspective on farmers’ income issues by encouraging more in-depth research into practical pathways to address income challenges from the lens of agricultural green development.

Second, in terms of the structural impact, the sub-dimensions of agricultural green development exhibit significant differentiation in their impacts on farmers’ income quality. The cost-saving effect of improved resource utilization efficiency serves as the core driver of income growth; the enhancement of agricultural output quality boosts operational income through a value-added mechanism. However, environment-friendly practices require substantial upfront investment in ecological protection and pollution control, and the long-term benefits from environmental improvement have not yet materialized. In the short term, costs outweigh benefits, resulting in a negative impact. This heterogeneous result provides new evidence for the applicability of the Environmental Kuznets Curve (EKC) Hypothesis in the agricultural field. The EKC Hypothesis proposes that there is an inverted U-shaped relationship between environmental pressure and income level [113]. This study finds that the relationship between agricultural green development and farmers’ income quality also exhibits phased characteristics. In the dimension of agricultural environmental friendliness (AEF), a negative effect emerges in the short term due to high pollution control costs (−0.1476***). However, in the long run, as the ecological environment improves (e.g., increased organic matter in farmland soil, reduced agricultural non-point source pollution), the stability of farmers’ income will be significantly enhanced [67]. This aligns with the core logic of the EKC Hypothesis—environmental quality first deteriorates and then improves as income grows, while further refining the micro-mechanism of the environment-income relationship in the agricultural sector. Specifically, through a three-stage evolution of short-term cost input → medium-term ecological restoration → long-term income growth, the coordinated development of agricultural ecology and income quality is ultimately achieved.

Structural analysis also shows that agricultural green development has differential impacts on the various sub-dimensions of farmers’ income quality. The four-dimensional evaluation framework for farmers’ income quality (adequacy, structure, growth, and knowledge intensity), which is constructed essentially as an extension of the Sustainable Livelihoods Framework, emphasizes that the diversified allocation of livelihood capital is the core for achieving sustainable income growth [81]. Its positive effect on income adequacy stems from the cost savings brought by improved resource utilization efficiency and the market premiums generated by enhanced output quality. The improvement in income structure is attributed to the development of related industries driven by green transformation and the shift from a single agricultural operation income to a diversified income model of “cultivation + processing + service”. This corresponds to the accumulation of financial capital and social capital in the Sustainable Livelihoods Framework. The short-term inhibitory effect on income growth is mainly associated with the high upfront costs in the green transformation process and the time lag in benefit conversion. Additionally, the insignificant impact on knowledge-based income reflects the mismatch between current rural human capital and the demands of green development. This finding provides a theoretical basis for the need to strengthen rural green skills training in subsequent research.

Third, concerning heterogeneous impacts across regions and regulatory factors, the sub-sample regression results identified context-specific variations. This study constructs a sub-sample regression framework for in-depth analysis based on four core dimensions: geographical location, natural resource endowments, financial development vitality, and local market vitality. The results show that the regional heterogeneity in the impact of agricultural green development on farmers’ income quality is manifested not only in significant differences among the four major geographical regions but also in the differentiated effects within the same geographical space, which arise from variations in the abundance of resources, the intensity of policy support, and the level of market development.

Regarding geographical heterogeneity, the impact of agricultural green development on farmers’ income quality shows significant differentiation across the four major geographical regions. The root cause of these differences lies in variations in regional resource endowments, industrial development foundations, and policy regulation orientations. The positive effect in the east is attributed to its mature market mechanisms and advantages in industrial integration. This aligns with the international experience logic of South Korea’s Pro-Environmental Agriculture [114] and Japan’s Sixth Industry [115], which realize the synergy between ecology and income through market orientation. Both models achieve such synergy via the approach of high-value-added green products and industrial chain extension, verifying the general law that resource-constrained regions need to enhance income quality through market-oriented green transformation.

The positive effect in the west stems from the dual-driver mechanism of ecological compensation and characteristic agriculture, which shares commonalities with the transformation path of ecologically fragile regions exemplified by Ethiopia’s model of “ecological protection and coffee cultivation”, where premium prices are obtained through organic certification [116]. The significant impact of agricultural green development on farmers’ income in the west provides an experiential pathway for underdeveloped regions. Most underdeveloped regions globally, similar to western China, are characterized by both ecological fragility and high dependence on agriculture. The practice of increasing farmers’ income through green agriculture in western China has proven that rural poverty alleviation and income growth can be achieved without relying on industrial expansion or sacrificing the ecological environment, offering a replicable green poverty-alleviation pathway for such regions. Furthermore, by leveraging its ecological advantages, western China has developed high-value-added characteristic products through green agriculture, bypassing the traditional competitive model of high input and forming a new logic: ecological advantages → product differentiation → market premium → farmers’ income growth. This provides underdeveloped regions with insights on realizing the value upgrading of agriculture through green development by focusing on their local ecological characteristics.

The empirical findings for the northeast and central regions both originate from two key factors: a production structure dominated by staple crops and insufficient capacity to share the costs of green transformation. Fundamentally, this reflects the “green transformation paradox” in major grain-producing regions: the policy objective of safeguarding food security demands the stable cultivation of staple crops, yet the smallholder-dominated production structure lacks the capacity to absorb the costs associated with green transformation. In large-scale agricultural nations such as the United States and Australia, the short-term costs of green transformation can be distributed through economies of scale. In contrast, smallholders remain the primary operators in China’s grain cultivation sector, which hinders the short-term realization of the positive impact of green development on income quality. The design of the EU’s Common Agricultural Policy (CAP)—where green subsidies are linked to planting scale—offers international experience for resolving the green transformation dilemma of staple crops in northeast and central China. For example, wheat farms in Bavaria, Germany, that participate in this policy have achieved a 60% improvement in the efficiency of green transformation cost-sharing compared to scattered smallholders, with a 12% higher net profit per unit area [29,117].

Regarding heterogeneity in regulatory factors, In regions with poor natural resource endowments, agricultural green development enhances income growth more prominently by improving fragile production conditions and mitigating natural risk shocks. Similarly, regions with lower fiscal support levels show more significant income-increasing effects, as agricultural production here relies more on market-oriented green transformation. In contrast, in regions with high market vitality, the direct driving effect of green development on farmers’ income is weakened. This is because farmers have a relatively weak bargaining power in the industrial chain, and the benefits from green development flow more to legal entities in processing and sales links. These findings suggest that targeted policies must be formulated in accordance with local transformation stages and constraints.

5.2. Practical Implications

First, efforts should be made to deepen the agricultural green development policy, expand the extension of green technologies, and strengthen the rural green skills training system. Meanwhile, farmers’ income resilience against risks should be enhanced through industrial integration and the expansion of non-agricultural employment to alleviate over-reliance on a single industry.

Second, to mitigate the short-term inhibitory effect of initial environmental protection practices on farmers’ income quality, the following measures should be implemented: increase the coverage rate of incremental subsidies for green agricultural input costs to over 50%; extend subsidies from end-stage links such as chemical fertilizer reduction and ecological compensation to the entire process, including green agricultural input procurement, technical training, and on-field monitoring; provide special compensation for the reduction in sown area caused by farmland ecological restoration; adopt the “subsidy first, verification later” approach to avoid delays in compensation; support smallholders in joining cooperatives or family farms through land shareholding, trusteeship, and other forms, where new-type agricultural business entities centrally purchase green agricultural inputs and provide technical services in a unified manner to reduce smallholders’ unit costs; build regional centralized procurement platforms for green agricultural inputs and secure cost advantages for smallholders through bulk procurement negotiations; and provide full subsidies for organic certification fees for the green transformation of staple crops such as corn and soybeans, establish a minimum support price mechanism for green agricultural products, and pilot the launch of income loss insurance for environmental protection practices to compensate for the gap when output fluctuates during the transformation period or when market prices fall below the minimum support price.

Third, given the significant heterogeneity in resource endowments, industrial foundations, and ecological positioning across the eastern, central, western, and northeastern regions of China, it is necessary to formulate targeted and differentiated strategies. Leveraging its advantages in technology and market, the eastern region should focus on the full-chain digital upgrading of green agriculture, promote the branding and high-end development of green agricultural products, and meanwhile transfer green technologies and management experience to the central and western regions. Based on its arable land and labor resources, the central and western regions should prioritize improving the green agricultural input supply chain and infrastructure for large-scale cultivation. They can reduce the transformation costs through the “leading enterprises and small-scale farmers” model and develop characteristic industries such as under-forest economy and eco-tourism with the support of ecological compensation policies. The northeastern region needs to strengthen the protection of black soil and the promotion of green planting technologies to break the traditional “yield-over-ecology” model.

5.3. Limitations and Future Directions

Although this study has made notable progress in understanding the research subject, it still has certain limitations, and there are areas that deserve further exploration in the future. One possible limitation of this study is that, constrained by the availability of sample data, the measurement indicators primarily rely on provincial-level macro statistical data, failing to incorporate micro-level behavioral characteristics of farmers and their subjective evaluation information. While provincial data can reflect overall trends, they struggle to capture heterogeneities at the county or individual farmer level, thereby overlooking differences among various types of farmers in terms of green development and income quality improvement. Although the instrumental variables passed the endogeneity test, they failed to completely rule out the interference of omitted variables such as farmers’ human capital, leaving room for improving the rigor of causal inference. In addition, due to the multi-dimensional nature of farmers’ income quality, this study failed to select appropriate mediating variables for mechanism analysis. While it revealed the differential impacts of sub-dimensions of agricultural green development on sub-dimensions of farmers’ income quality through structural analysis and identified characteristics of regional heterogeneity, it did not delve into the specific transmission paths underlying these impact relationships. In particular, there is a lack of detailed research on the mechanism behind the negative impact of agricultural environmental friendliness practices.

Future research should strive to overcome these limitations, and the most significant among these possibilities is to collect micro-level information on farmers’ production inputs, ecological behaviors and income changes through questionnaire, interviews, and field experiments, and micro-investigations can be conducted into farmers’ actual participation experiences and long-term development effects. By combining on-site research or controlled experiments, an in-depth understanding of the micro-mechanisms behind farmers’ behaviors and the effects of policy implementation can be obtained. Additionally, employing mediating effect models to identify key pathways through which agricultural green development influences farmers’ income quality, while incorporating case studies to explore the cost-sharing and benefit-sharing mechanisms of environment-friendly practices, would provide more nuanced insights. Furthermore, considering spatial autocorrelation and spillover effects by applying methods like the Spatial Durbin Model to research agricultural green development and farmers’ income quality could help uncover the synergistic and diffusion mechanisms of green transformation across different regions.

6. Conclusions

Using panel data covering 31 provincial-level administrative regions in China (Hong Kong, Macau, and Taiwan excluded) from 2011 to 2022, this study systematically investigates whether the level of agricultural green development affects the quality of farmers’ income, as well as the structural and regional heterogeneity of this impact. The empirical results demonstrate that agricultural green development has significantly propelled farmers’ income quality. Improvements in resource utilization efficiency and output quality are the core drivers of income growth, while environment-friendly practices show a negative impact. Agricultural green development significantly enhances income adequacy and structure but has a short-term inhibitory effect on income growth, and its impact on knowledge-based income is not yet significant. Regression analyses on sub-samples classified by key factors—geographical region, natural resource endowment, government support level, and local market vitality—reveal that the regional heterogeneity of the impact is not merely geographically bounded. It also manifests as the positive or negative impact under varying institutional frameworks and economic conditions. Agricultural green development exerts the strongest positive effect on farmers’ income quality in the west, followed by the east. The effect is insignificant in the central, while a significantly negative effect is observed in the northeast. The root cause of this heterogeneity lies in regional disparities in resource endowments, industrial foundations, and policy environments. Further analysis reveals that regions with poor natural resource endowments and low fiscal agricultural support show more significant effects, while in regions with high market vitality, benefits flow to the downstream of the industrial chain, leading to a weakened effect.

This study not only deepens the exploration of China’s local practices but also provides an empirical sample from a large developing country for global quantitative research on the income growth dimension in agricultural green transformation. Furthermore, it aims to offer insights for global research on sustainable agricultural development. First, it verifies that large developing countries can achieve the synergy between ecological protection and farmers’ income growth, addressing the applicability of the Environmental Kuznets Curve (EKC) in the agricultural sector. Second, the proposed four-dimensional evaluation framework for income quality can be adapted to the characteristics of farmers’ income in different countries, providing an operable analytical tool for international comparative studies. Third, the identified regional heterogeneity offers targeted transformation references for countries with similar resource endowments. For instance, the compatibility of policy instruments (such as ecological compensation and technical subsidies) with green transformation is of reference value to developing countries like those in Africa, which are rich in ecological resources but have weak farmers’ capacity. Meanwhile, the impact of natural resource endowments on the income-increasing effect of green development is applicable to resource-constrained regions such as the Middle East and South Asia. It should be noted that the conclusions of this study are constrained by China’s unique institutional context, regional development characteristics, and factors of production traits. For instance, China’s state-coordinated policy system differs from the decentralized frameworks adopted by most countries; the gaps in economic foundations and technological capabilities between regions in China are significantly wide, meaning that the experience of eastern China cannot be directly replicated in countries with underdeveloped market system. Furthermore, China’s collective land ownership system shapes the cost-sharing and benefit-distribution models for green development, which differs from the context of countries with large-scale commercial agriculture.