Green Finance Empowers Coordinated Development of Synergistic Governance of Pollution and Carbon Reduction in Agriculture

Abstract

Fostering the coupled and coordinated progress of agricultural pollution control and carbon emission reduction (AGR) is integral to agricultural modernization and high-quality agricultural development. As China’s green finance policies have been steadily refined, green finance has emerged as a pivotal driver of AGR enhancement. To examine the influence of green finance on AGR, this study analyzes their nexus over the 2008–2023 period. The results demonstrate that green finance markedly improves AGR, and this effect exhibits heterogeneity. In terms of mechanism, green finance strengthens AGR by bolstering agricultural technological innovation (INN) and raising farmers’ income (INC). Agricultural fiscal support (FIN) and the state of rural education (EDU) serve as important external conditions for green finance to effectively promote AGR. The findings further suggest that green finance is progressively evolving into a major force for advancing AGR.

1. Introduction

The construction of a beautiful countryside embodies a distinctive path of socialist rural revitalization with Chinese characteristics [1]. As a major agricultural nation, China has been steadily advancing its agricultural modernization, with total agricultural output continuing to grow. Yet, this modernization drive is also confronted with pronounced environmental challenges—large-scale farming operations have given rise to substantial carbon emissions and environmental pollution. Agricultural carbon emissions and non-point source pollution originate from shared sources and can be addressed jointly in practice. In this study, AGR is conceptualized as an integrated production process wherein a combination of coordinated practices simultaneously mitigates non-point source pollution and greenhouse gas emissions, yielding synergistic gains in both environmental and social performance. Such a transition embodies the greening of agricultural production systems and the principles of sustainable development, and it constitutes a fundamental strategy for addressing the intertwined challenges of climate change and environmental degradation. It is, therefore, essential to balance agricultural development with environmental governance. To achieve agricultural modernization with Chinese characteristics and promote high-quality agricultural development, effective coordination must be ensured among agricultural production, pollution control, and carbon reduction. Diverging from traditional finance, green finance is closely aligned with China’s current commitment to green and sustainable development and acts as a vital instrument for advancing the country’s “dual carbon” goals. Major economies, China being among them, have set dual carbon targets (emission peak and carbon neutrality) [2]. Green finance seeks to enhance capital flows among the public, private, and non-profit sectors in order to advance for sustainable development [3]. As an innovative financial instrument, green finance bridges environmental governance and financial institutions, emphasizing environmental benefits and efficient utilization of green resources [4]. Against the backdrop of the “dual carbon” targets, green finance can promote the AGR, thereby strengthening the sustainability of agricultural production. The development of green finance alleviates the financing constraints faced by enterprises, thereby facilitating the expansion of their production scale. Particularly in heavily polluting industries, it accelerates the green transition [5]. As green regulations and standards become progressively integrated into international investment frameworks, enterprises can achieve competitive advantages within these frameworks [6]. Financing restrictions imposed on traditional energy-intensive sectors encourage them to improve production methods and curtail pollution emissions. Within the agricultural sector, green industries can obtain increased funding through green finance, thereby accelerating the sector’s green transformation and upgrading and promoting the coupled advancement of AGR.

Existing studies on AGR largely center on quantitative assessment and characteristic analysis. Building upon the traditional pollution index method, Wu et al. innovatively incorporated indicators such as atmospheric deposition, manure and straw returning to fields, livestock emissions, and rural domestic discharges, thereby offering a basis for better understanding diffuse pollution in China and formulating effective macro-level policies [7]. Zhou et al. devised a new technical system grounded in previous assessments of agricultural non-point source pollution, allowing for precise estimation of total nitrogen, total phosphorus, and ammonia nitrogen entering water bodies, thus advancing understanding of agricultural non-point source pollution [8]. By incorporating high-resolution sensing technologies with soil-water assessment tools, the comprehensive model used for agricultural non-point source pollution monitoring has established a relatively complete real-time monitoring network, substantially improving data quality and model reliability [9]. Identifying and investigating the driving forces behind the synergistic governance of AGR remains a central research focus. The rapid growth of e-commerce has drawn growing attention to its impact on agricultural sustainability. By stimulating industrial restructuring and technological innovation, the integration of e-commerce with agriculture effectively alleviates agricultural pollution [10]. Corresponding e-commerce development policies also demonstrated marked inhibitory effects on agricultural pollution. For example, China’s National E-Commerce Demonstration City initiative has been found to boost regional green technology innovation and mitigate agricultural pollution compared to control groups [11]. Agricultural industry agglomeration can also influence agricultural pollution, exhibiting an inverted U-shaped relationship; over the long term, greater agglomeration of agricultural industries can advance the green development of agriculture and ameliorate agricultural pollution [12]. Employing a revised coupling coordination degree model to measure AGR in China, and examining its spatiotemporal differences and driving factors, researchers found that agricultural economic scale, cropping structure, mechanization level, rural education, and transportation infrastructure constitute the principal determinants of the synergistic effect [13].

The AGR constitutes a vital element in advancing green and high-quality development. As green finance has undergone sustained progress and refinement, it has emerged as a novel driver for enhancing the AGR. A systematic examination of the heterogeneous impacts, underlying mechanisms, and external conditions through which green finance facilitates AGR can further advance high-quality agricultural development and contribute to the long-term sustainability of the global ecological environment.

2. Theoretical Analysis and Research Hypotheses

2.1. Green Finance and AGR

Agricultural carbon emissions and non-point source pollution are derived from common sources and follow similar processes; thus, fostering coordinated governance of AGR is a vital pathway toward green agricultural development. The predominant sources of agricultural carbon emissions and non-point source pollution lie in extensive farming practices, such as excessive fertilizer and pesticide application and improper disposal of agricultural waste. Green finance exerts multifaceted influences on AGR. Within the agricultural system, its impact is chiefly channeled through credit approval policies that incentivize agricultural producers to voluntarily curtail pollution and emissions. Owing to low returns, lengthy production cycles, and elevated risks, agricultural activities generally encounter limited financing access within conventional financial frameworks. As a financial instrument, green finance can broaden the scale of sustainable development projects and improve their financial efficiency [14]. By integrating green practices into the approval process, it grants farmers and agricultural enterprises greater access to funding. To secure more financial support, farmers and relevant businesses are motivated to adopt lower-carbon and eco-friendly production methods, thereby mitigating pollution and carbon emissions at the source. The impact of green finance on AGR also extends to the upgrading of agriculture-related external industries. Green finance hastens the green transition of heavily polluting sectors. As green regulations and standards become progressively embedded in international investment regimes, green finance enables enterprises to capture competitive advantages worldwide [6]. On one hand, low-carbon and eco-friendly enterprises within agriculture-related supporting industries can access greater financial resources through green finance, allowing them to maintain environmentally responsible operations and bolster competitiveness. On the other hand, green finance policies impose financing restrictions on energy-intensive and heavily polluting industries, placing tighter constraints on these enterprises [15]. By driving the low-carbon transition and upgrading of related industries, green finance generates conducive external conditions and industrial support for AGR, substantially enhancing the level of coordinated governance. Consequently, green finance can affect both on-farm activities and allied industries, comprehensively improving AGR performance and facilitating its low-carbon transition.

H1: 

Green finance contributes significantly to the improvement of AGR performance.

2.2. The Mediating Role of INN

Green finance can strengthen the level of INN. With respect to INN, green finance can ease financing constraints for relevant innovators by embedding green practices into loan approval criteria and lowering financing thresholds, thereby channeling greater financial support into INN activities and facilitating the research and development of INN, which directly raises investment in INN [16]. Green finance policies include government endorsement of environmentally responsible practices. As green finance evolves, the array of green financial products in the market expands, regulatory frameworks become more robust, and investors increasingly gravitate toward green assets. This environment furnishes long-term financial backing for corporate green innovation, helps diversify innovation-related risks, and relative to conventional finance, effectively mitigates financing constraints while steering capital toward green initiatives. It supplies financial resources and R&D funding for green innovation, fostering higher innovation performance [17], and alleviating financing challenges posed by innovation risks. This implies that green finance can also indirectly boost INN investment by mobilizing social capita. Beyond augmenting initial R&D expenditure in agriculture, green finance can subsequently offer innovators sustained, low-cost financial support, ensuring continuous innovation efforts and supplying ample funding for stable and efficient INN outcomes. Green finance reduces the upfront costs and uncertainties associated with new technologies, accelerating the transformation and commercialization of innovations. The ongoing evolution of green financial products, enabled by digital platforms, permits online and transparent processes, thereby reducing the information asymmetry and collateral obstacles typical of traditional finance. By effectively dispersing innovation risks through financial markets, it lowers R&D uncertainty and promotes the successful implementation and application of new technologies, thus fostering a steady output of INN. Through green finance policies, governments can guide universities toward adjusting talent cultivation models, advancing industry–education integration, and producing more highly competent professionals, injecting vitality into INN and enhancing its overall capacity.

Elevated levels of INN can effectively advance the coordinated governance of AGR. The development and deployment of novel products and technologies at the production stage can achieve source-level mitigation of agricultural pollution and carbon emissions, generating synergistic benefits in pollution control and decarbonization. As new technologies mature, the subsequent treatment of agricultural pollutants becomes more efficient and convenient, reducing transportation and disposal costs while improving processing efficiency. Strengthened innovation forms the cornerstone of the green transformation and upgrading of the entire agricultural production chain, further reducing AGR at the production, processing, and sales stages and thereby advancing high-quality agricultural development. Green finance influences both the input and output sides of INN, empowering improvements in AGR through elevated INN.

H2: 

Green finance can enhance AGR by bolstering INN.

2.3. The Mediating Role of INC

INN constitutes a critical pathway through which green finance exerts its influence on AGR. Green finance can embed green production practices into credit approval criteria, enabling farmers and agricultural processing enterprises to secure more funds from the financial market via green credit and thereby alleviating the financing difficulties they typically face in the conventional financial market. For farmers who adopt green technologies, green financial institutions further offer preferential credit policies, which lower financing costs and raise INC. Products backed by green finance are also be more likely to obtain organic certifications and similar labels. Given consumers’ growing inclination towards green consumption, acquiring such certifications can strengthen the market competitiveness of agricultural products, win consumer trust and preference, and boost INC by increasing sales volumes. As green finance continues to evolve and mature, the array of available green financial products has becoming increasingly diverse. Farmers can now purchase agricultural credit products they are more familiar with in the green finance market, improving their capital utilization efficiency and yield, which can markedly raise INC.

Rising INC can further advance AGR levels. Higher INC triggers a virtuous cycle for the green development of AGR, as it equips farmers with sufficient capital to invest in green agricultural production. This is largely because the initial investment in agricultural green farming is substantial, and certain production equipment and novel technologies often entail high risks and lengthy payback periods. With increased INC, farmers are able to allocate more funds to innovation and green production, thereby mitigating agricultural carbon emissions and pollution at the source. In addition, rising INC can contribute to an improved rural environment and upgraded rural infrastructure, which provide favorable external conditions for green agricultural production and lay the foundation for the more scientific and rational treatment and reuse of agricultural waste. This, in turn, promotes the development of AGR. In summary, green finance boosts INC through direct credit support and expanded income channels, and a higher INC facilitates the adoption of new green production models, thereby driving the improvement of AGR.

H3: 

Green finance can promote AGR by increasing INC.

2.4. The Moderating Role of FIN

The level of such support reflects a region’s prioritization of local agriculture, and its intensity moderates the specific effect of green finance on AGR improvement. Government policies typically perform a leading and guiding function. A higher level of FIN signals that the local government prioritizes agricultural development and provides greater assistance and financial backing. This, in turn, offers policy guidance for the green financial market, encouraging investors to focus more on the agricultural sector and channeling additional funds into green agricultural production. Such dynamics promote the continuous development and application of green technologies in agriculture, thereby effectively reducing AGR. The government’s emphasis on agricultural production can also steer green financial institutions to further regulate their practices, granting priority to agricultural entities in credit approval. A portion of government funds is allocated as risk compensation or used to provide guarantees for farmers, which further boosts the willingness of green financial institutions to extend credit and channels more green credit resources into the agricultural sector, thereby further enhancing AGR levels. In essence, the degree of FIN reflects the priority placed on agriculture and can encourage green finance to allocate more resources to the sector, driving improvements in AGR.

H4: 

FIN plays a moderating role in promoting AGR through green finance.

2.5. The Threshold Effect of EDU

The effectiveness of green finance in curbing AGR is not uniform; it hinges on the educational attainment of EDU directly influencing farmers’ financial literacy and their openness to green production practices. In regions where EDU levels are low, farmers tend to have limited awareness of green finance policies and are less receptive to environmentally sound farming methods. Under such circumstances, their inadequate grasp of green finance hinders their ability to tap into green credit and channel funds into agricultural production, thus weakening green finance’s contribution to AGR. Moreover, when EDU is underdeveloped, the willingness of farmers and other actors along the agricultural chain to embrace green production remains subdued. Even when green financial support is available, it is often directed more toward expanding the scale of production, while the uptake of novel green technologies and products stays low, diminishing green finance’s impact on AGR mitigation.

In areas with high EDU levels, farmers generally possess stronger financial literacy and engage more proactively with green finance support. They are better equipped to leverage green finance policies and secure greater financial resources for agricultural production. Higher EDU attainment also means that relevant producers and operators are more receptive to green production concepts, actively learn and refine their farming practices, and increasingly adopt low-carbon and green production models. This, in turn, amplifies the impact of green finance on AGR. Advances in EDU expand the pool of highly qualified professionals available to the agricultural sector, supplying the human capital necessary for agricultural scientific and technological research and innovation, as well as for the development of modern agriculture, thereby contributing more effectively to the improvement of AGR. In summary, the rise in EDU levels can holistically strengthen farmers’ own capabilities and create favorable external conditions for green finance to advance AGR.

H5: 

EDU serves a critical threshold function in promoting AGR through green finance.

3. Research Design

3.1. Variable Selection

3.1.1. Independent Variable

The independent variable is “Green Finance”. According to relevant studies [18], Green Finance is constructed as a multidimensional index encompassing green credit, green investment, green insurance, green bonds, green support, green funds, and green equity. The specific indicators are presented in

Table 1. Green Finance evaluation index system.

Primary Indicator	Secondary Indicator	Formula Mode	Indicator Attribute	Iindex Weight
Green finance	Credit	The total amount of environmental protection project loans in this province/all loans in the province	+	0.130
	Investment	Investment in environmental pollution control/ GDP	+	0.140
	Insurance	Income from environmental pollution liability insurance/total insurance premium income	+	0.129
	Bonds	Total amount of green bond issuance/total amount of all bond issuance	+	0.147
	Support	Fiscal expenditure on environmental protection/ fiscal general budget expenditure	+	0.162
	Fund	Total market value of green funds/total market value of all funds	+	0.144
	Rights and interests	Carbon trading, energy usage rights trading, pollution discharge rights trading/equity market trading volume	+	0.148

. Based on this comprehensive evaluation system, the entropy weight method is applied to determine the indicator weights, and the green finance level of each province is then calculated. The resulting weights are also reported in Table 1.

3.1.2. Dependent Variable

The dependent variable in this study is AGR, which is defined by coupling three subsystems: agricultural carbon emissions, agricultural non-point source pollution, and agricultural output value. Agricultural carbon emissions are quantified via the carbon emission coefficient method, with sources covering rice cultivation, farmland tillage, irrigation, fertilizer application, pesticide use, agricultural film use, agricultural diesel combustion, and livestock and poultry farming. Agricultural non-point source pollution is assessed through the unit survey assessment method following the approach of [12], and the pollutant sources include chemical fertilizers, livestock and poultry farming, and agricultural solid waste. Detailed information is presented in

Table 2. Agricultural carbon emissions.

Primary Indicator	Secondary Indicator	Meaning of Indicators	Reference Source
Agricultural carbon emissions	Agricultural input materials investment	Fertilizer carbon emissions	Oak Ridge National Laboratory of the United States
		Thin-film carbon emissions	Oak Ridge National Laboratory of the United States
		Pesticide carbon emissions	United Nations Intergovernmental Panel on Climate Change
	Agricultural energy utilization	Diesel carbon emissions	Institute of Agricultural Resources and Environmental Sciences, Nanjing Agricultural University
	Crop production	Soil tillage carbon emissions	[19]
		Irrigation carbon emissions	College of Biology and Technology, China Agricultural University
	Rice cultivation	Area of paddy fields	2006 IPCC National Greenhouse Gas Inventory Guidelines
Livestock farming carbon emissions	Livestock breeding	Cattle, horses, donkeys/mules, camels, goats, sheep, pigs and poultry	“Guidelines for the Compilation of Provincial Greenhouse Gas Inventories (Trial Version) (2011)”

. A modified coupling coordination model is employed to compute AGR. Compared with the traditional coupling coordination degree model, the revised formulation measures the degree of divergence based on the ratio relationships among subsystems, thereby correcting the interval compression bias inherent in the conventional model and improving its discriminatory power. The original uncorrected coupling coordination model is also applied to calculate AGR, which serves as the basis for subsequent robustness checks. The specific calculation formula is given below:

$$\mathrm{C}=\sqrt{\left[1-{\displaystyle \frac{\sqrt{{\left({\mathrm{U}}_3-{\mathrm{U}}_1\right)}^2}+\sqrt{{\left({\mathrm{U}}_2-{\mathrm{U}}_1\right)}^2}+\sqrt{{\left({\mathrm{U}}_3-{\mathrm{U}}_2\right)}^2}}{3}}\right]\times \sqrt{{\displaystyle \frac{{\mathrm{U}}_1}{{\mathrm{U}}_3}}\times {\displaystyle \frac{{\mathrm{U}}_2}{{\mathrm{U}}_3}}}}$$

(1)

$$\mathrm{T}={\mathsf{\alpha }}_1{\mathrm{U}}_1+{\mathsf{\alpha }}_2{\mathrm{U}}_2+{\mathsf{\alpha }}_3{\mathrm{U}}_3$$

(2)

$$\mathrm{D}=\sqrt{\mathrm{C}\times \mathrm{T}}$$

(3)

Among them, ${\mathrm{U}}_1, {\mathrm{U}}_2, {\mathrm{U}}_3$ denote the data of the three sectors after min–max normalization. C represents the coupling degree; D represents the coupling coordination degree; T indicates the comprehensive degree of the subsystem; ${\mathsf{\alpha }}_1, {\mathsf{\alpha }}_2, {\mathsf{\alpha }}_3$ are undetermined coefficients. The model coefficients are initially undetermined and are estimated using the entropy weight method. The resulting subsystem weights are 0.335, 0.334, and 0.331, respectively.

3.1.3. Mediating Variable

This study examines the mediating roles of INN and INC in the relationship between green finance and AGR. INN is proxied by total agricultural R&D investment, while INC is gauged by per capita disposable income in rural areas.

3.1.4. Regulated Variable

In accordance with the foregoing theoretical analysis, FIN is introduced as a moderating variable, measured as the ratio of agricultural fiscal expenditure to each province’s general public budget expenditure.

3.1.5. Control Variable

In line with [20], the following control variables are incorporated. The agricultural investment scale (INV) is proxied by fixed asset investment in agriculture; the forest coverage rate (FOR) is calculated as the proportion of forest area to total land area; the mechanization level (MEC) is represented by total power of agricultural machinery; agricultural energy consumption (ENE) is expressed as the ratio of total agricultural energy use to regional GDP; the disaster rate (DAM) is measured by the ratio of affected area to disaster-stricken area; and EDU is proxied by the average years of schooling among the rural population.

3.2. Model Building

3.2.1. Baseline Regression Model

To examine the effect of green finance on AGR, the following baseline regression model is specified:

$${\mathrm{A}\mathrm{G}\mathrm{R}}_{\mathrm{i}\mathrm{t}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{G}\mathrm{F}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_2{\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_3{\mathrm{F}\mathrm{O}\mathrm{R}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_4{\mathrm{M}\mathrm{E}\mathrm{C}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_5{\mathrm{E}\mathrm{N}\mathrm{E}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_6{\mathrm{D}\mathrm{A}\mathrm{M}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_7{\mathrm{E}\mathrm{D}\mathrm{U}}_{\mathrm{i}\mathrm{t}}+{\phi }_{\mathrm{i}}+{\mathsf{\delta }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$$

(4)

Here, ${\mathrm{A}\mathrm{G}\mathrm{R}}_{\mathrm{i}\mathrm{t}}$ is the dependent variable representing the collaborative level of AGR in province i in year t; ${\mathrm{G}\mathrm{F}}_{\mathrm{i}\mathrm{t}}$ is the core explanatory variable representing the development level of green finance; INV, FOR, MEC, ENE, DAM, and EDU are control variables; ${\mathsf{\alpha }}_0$ is the constant term; ${\phi }_{\mathrm{i}}, {\mathsf{\delta }}_{\mathrm{t}}$ represent the fixed effects of the province and the year, respectively; ${\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$ is the random error term.

3.2.2. Mediating Effect Model

Having established the impact of green finance on AGR, this study further investigates the mediating mechanisms through which green finance strengthens the synergistic performance of AGR by constructing the following mediation model:

$${\mathrm{Z}}_{\mathrm{it}} = {\mathsf{\beta }}_0 + {\mathsf{\beta }}_1{\mathrm{GF}}_{\mathrm{it}} + {\mathsf{\beta }}_2{\mathrm{INV}}_{\mathrm{it}} + {\mathsf{\beta }}_3{\mathrm{FOR}}_{\mathrm{it}} + {\mathsf{\beta }}_4{\mathrm{MEC}}_{\mathrm{it}} + {\mathsf{\beta }}_5{\mathrm{ENE}}_{\mathrm{it}} + {\mathsf{\beta }}_6{\mathrm{DAM}}_{\mathrm{it}} + {\mathsf{\beta }}_7{\mathrm{EDU}}_{\mathrm{IT}} + {\phi }_{\mathrm{i}} + {\mathsf{\delta }}_{\mathrm{t}} + {\mathsf{\mu }}_{\mathrm{it}}$$

(5)

$${\mathrm{AGR}}_{\mathrm{it}}={\mathsf{\theta }}_0 + {\mathsf{\theta }}_1{\mathrm{GF}}_{\mathrm{it}} + {\mathsf{\theta }}_2{\mathrm{Z}}_{\mathrm{it}} + {\mathsf{\theta }}_3{\mathrm{INV}}_{\mathrm{it}} + {\mathsf{\theta }}_4{\mathrm{FOR}}_{\mathrm{it}} + {\mathsf{\theta }}_5{\mathrm{MEC}}_{\mathrm{it}} + {\mathsf{\theta }}_6{\mathrm{ENE}}_{\mathrm{it}} + {\mathsf{\theta }}_7{\mathrm{DAM}}_{\mathrm{it}} + {\mathsf{\theta }}_8{\mathrm{EDU}}_{\mathrm{IT}} + {\phi }_{\mathrm{i}}+{\mathsf{\delta }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{it}}$$

(6)

Here, ${\mathrm{Z}}_{\mathrm{it}}$ is the mediating variable, which is INN and INC, ${\mathrm{A}\mathrm{G}\mathrm{R}}_{\mathrm{i}\mathrm{t}}$ is the dependent variable representing the collaborative level of AGR in province i in year t; ${\mathrm{G}\mathrm{F}}_{\mathrm{i}\mathrm{t}}$ is the core explanatory variable representing the development level of green finance; INV, FOR, MEC, ENE, DAM, and EDU are control variables; ${\mathsf{\alpha }}_0$ is the constant term; ${\phi }_{\mathrm{i}}, {\mathsf{\delta }}_{\mathrm{t}}$ represent the fixed effects of the province and the year,, respectively; ${\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$ is the random error term.

3.2.3. Modeling of the Moderating Effect

Having conducted the mechanism tests, this study specifies a moderation model to further investigate how FIN, as an external condition, shapes the effectiveness of green finance in enhancing AGR:

$${\mathrm{AGR}}_{\mathrm{it}} = {\mathsf{\gamma }}_0 + {\mathsf{\gamma }}_1{\mathrm{GF}}_{\mathrm{it}} + {\mathsf{\gamma }}_2{{\mathrm{FIN}}_{\mathrm{it}} + \mathsf{\gamma }}_3{\mathrm{FIN}}_{\mathrm{it}} \times {\mathrm{GF}}_{\mathrm{it}} + {\mathsf{\gamma }}_4{\mathrm{INV}}_{\mathrm{it}} + {\mathsf{\gamma }}_5{\mathrm{FOR}}_{\mathrm{it}} + {\mathsf{\gamma }}_6{\mathrm{MEC}}_{\mathrm{it}} + {\mathsf{\gamma }}_7{\mathrm{ENE}}_{\mathrm{it}} + {\mathsf{\gamma }}_8{\mathrm{DAM}}_{\mathrm{it}} + {\mathsf{\gamma }}_9{\mathrm{EDU}}_{\mathrm{it}} + {\phi }_{\mathrm{i}} + {\mathsf{\delta }}_{\mathrm{t}} + {\mathsf{\mu }}_{\mathrm{it}}$$

(7)

Here, ${\mathrm{FIN}}_{\mathrm{it}}$ represents the situation of FIN, ${\mathrm{A}\mathrm{G}\mathrm{R}}_{\mathrm{i}\mathrm{t}}$ is the dependent variable representing the collaborative level of AGR in province i in year t; ${\mathrm{G}\mathrm{F}}_{\mathrm{i}\mathrm{t}}$ is the core explanatory variable representing the development level of green finance; $\mathrm{I}\mathrm{N}\mathrm{V}, \mathrm{F}\mathrm{O}\mathrm{R}, \mathrm{M}\mathrm{E}\mathrm{C}, \mathrm{E}\mathrm{N}\mathrm{E}, \mathrm{D}\mathrm{A}\mathrm{M}, \mathrm{E}\mathrm{D}\mathrm{U} \mathrm{a}\mathrm{r}\mathrm{e} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l} \mathrm{v}\mathrm{a}\mathrm{r}\mathrm{i}\mathrm{a}\mathrm{b}\mathrm{l}\mathrm{e}\mathrm{s}; {\mathsf{\alpha }}_0$ is the constant term; ${\phi }_{\mathrm{i}}$, ${\mathsf{\delta }}_{\mathrm{t}}$ represent the fixed effects of the province and the year, respectively; ${\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$ is the random error term.

3.2.4. Panel Threshold Model

In order to further examine the non-linear effect of green finance on reducing AGR, drawing on the panel threshold model proposed by [21], considering the possibility of multiple threshold values, the following model is constructed:

$${\mathrm{AGR} }_{\mathrm{it}} = {\mathsf{\rho }}_0 + {\mathsf{\rho }}_1{\mathrm{GF}}_{\mathrm{it}}\times \mathrm{I}\left( \mathrm{EDU} \le {\mathsf{\omega }}_1\right) + {\mathsf{\rho }}_2{\mathrm{GF}}_{\mathrm{it}} \times \mathrm{I} \left({\mathsf{\omega }}_1 < \mathrm{EDU} \le {\mathsf{\omega }}_2\right) + {\mathsf{\rho }}_3{\mathrm{GF}}_{\mathrm{it}} \times \mathrm{I}\left(\mathrm{EDU} > {\mathsf{\omega }}_3\right) + {\mathsf{\rho }}_4{\mathrm{INV}}_{\mathrm{it}} + {\mathsf{\rho }}_5{\mathrm{FOR}}_{\mathrm{it}} + {\mathsf{\rho }}_6{\mathrm{MEC}}_{\mathrm{it}} + {\mathsf{\rho }}_7{\mathrm{ENE}}_{\mathrm{it}} + {\mathsf{\rho }}_8{\mathrm{DAM}}_{\mathrm{it}} + {\mathsf{\rho }}_9{\mathrm{EDU}}_{\mathrm{it}} + {\phi }_{\mathrm{i}} + {\mathsf{\delta }}_{\mathrm{t}} + {\mathsf{\mu }}_{\mathrm{it}}$$

(8)

Here, the EDU is the threshold variable, $\mathsf{\omega }$ is the corresponding threshold value, ${\mathrm{A}\mathrm{G}\mathrm{R}}_{\mathrm{i}\mathrm{t}}$ is the dependent variable representing the collaborative level of AGR in province i in year t; ${\mathrm{G}\mathrm{F}}_{\mathrm{i}\mathrm{t}}$ is the core explanatory variable representing the development level of green finance; $\mathrm{I}\mathrm{N}\mathrm{V}, \mathrm{F}\mathrm{O}\mathrm{R}, \mathrm{M}\mathrm{E}\mathrm{C}, \mathrm{E}\mathrm{N}\mathrm{E}, \mathrm{D}\mathrm{A}\mathrm{M}, \mathrm{E}\mathrm{D}\mathrm{U} \mathrm{a}\mathrm{r}\mathrm{e} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l} \mathrm{v}\mathrm{a}\mathrm{r}\mathrm{i}\mathrm{a}\mathrm{b}\mathrm{l}\mathrm{e}\mathrm{s}; {\mathsf{\alpha }}_0$ is the constant term; ${\phi }_{\mathrm{i}}, {\mathsf{\delta }}_{\mathrm{t}}$ represent the fixed effects of the province and the year, respectively; ${\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$ is the random error term.

3.3. Data Sources

Data on green finance, AGR, and the control variables were all collected from the China Statistical Yearbook, China Rural Statistical Yearbook, China Environmental Statistical Yearbook, China Energy Statistical Yearbook, China Industrial Statistical Yearbook, China Insurance Yearbook, various provincial statistical yearbooks, the EPS database, and the Wind database, among other sources. To preserve data completeness, missing values were imputed through interpolation methods.

4. Empirical Result Analysis

4.1. Baseline Regression Results

This study investigates the linear relationship between green finance and AGR by estimating Equation (4) with a two-way fixed effects model; the estimation results are reported in

Table 3. Baseline regression estimation results.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	AGR	AGR	AGR	AGR	AGR	AGR	AGR
GF	0.875 ***	0.898 ***	0.811 ***	0.723 ***	0.730 ***	0.719 ***	0.636 ***
	(0.046)	(0.045)	(0.058)	(0.057)	(0.058)	(0.061)	(0.071)
INV		−0.021 ***	−0.025 ***	−0.029 ***	−0.029 ***	−0.029 ***	−0.033 ***
		(0.006)	(0.006)	(0.006)	(0.006)	(0.006)	(0.006)
FOR			0.004 **	0.004 ***	0.004 ***	0.004 ***	0.004 **
			(0.002)	(0.002)	(0.002)	(0.002)	(0.002)
MEC				0.075 ***	0.075 ***	0.076 ***	0.076 ***
				(0.011)	(0.011)	(0.012)	(0.011)
ENE					0.001	0.001	0.001
					(0.002)	(0.002)	(0.002)
DAM						−0.009	−0.009
						(0.014)	(0.014)
EDU							0.023 **
							(0.011)
Constant	0.248 ***	0.350 ***	0.270 ***	−0.274 ***	−0.279 ***	−0.278 ***	−0.399 ***
	(0.015)	(0.033)	(0.047)	(0.094)	(0.095)	(0.095)	(0.109)
Province FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	480	480	480	480	480	480	480
${\mathrm{R}}^2$	0.452	0.466	0.473	0.519	0.519	0.520	0.525

The values in parentheses represent robust standard errors; ** and *** indicate significance at the 5% and 1% levels, respectively.

. Column (1) displays the baseline specification without control variables, where the coefficient of green finance is 0.875 and significant at the 1% level, providing suggestive evidence of a positive association. As shown in columns (2) through (7), control variables that may influence AGR are sequentially introduced. The coefficient on green finance remains positive and statistically significant at the 1% level across all specifications, accompanied by a gradual improvement in goodness-of-fit. This pattern reveals a stable positive correlation between green finance and AGR under different model settings. Overall, the empirical findings are in line with the hypothesis that green finance tends to improve AGR, though establishing causal relationships calls for further identification strategies.

Regarding the control variables, the coefficient on INV is negative and statistically significant, suggesting that larger INV may be associated with a disorderly expansion of production, accompanied by higher pollution and carbon emissions from agricultural activities. The coefficient on FOR is positive and statistically significant, implying that greater FOR enhances the sink capacity for agricultural carbon emissions, which is conducive to AGR. The coefficient for the MEC is also positive and statistically significant, suggesting that a greater MEC facilitates more precise application of inputs such as fertilizers and pesticides, thereby curbing the chemical pollution arising from their excessive use. The coefficients on ENE and DAM are statistically insignificant, suggesting that, under current conditions, these factors exert negligible influence on AGR. The coefficient on EDU level is positive and significant, indicating that higher EDU strengthens farmers’ receptiveness to green production concepts, raises their acceptance of new technologies and products, and encourages the voluntary adoption of low-carbon farming practices, thus promoting AGR.

4.2. Robustness Test

To assess the robustness of the baseline regression results, this study employs several strategies: quantile regression, excluding municipalities directly under the central government, replacing the dependent variable, and introducing additional policy control variables.

To further assess the robustness of the baseline association between green finance and AGR, a panel quantile regression approach is utilized, with estimations performed at multiple quantiles of the AGR distribution. The results are reported in

Table 4. Results of quantile regression.

	(1)	(2)	(3)	(4)	(5)
	0.1	0.25	0.5	0.75	0.9
GF	0.103	0.083	0.117 **	0.228 ***	0.213 ***
	(0.082)	(0.066)	(0.056)	(0.043)	(0.061)
control	Yes	Yes	Yes	Yes	Yes
Constant	−0.405 ***	−0.227 **	−0.152 *	−0.221 ***	−0.150
	(0.134)	(0.109)	(0.092)	(0.071)	(0.101)
Province FE	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes
N	480	480	480	480	480
${\mathrm{R}}^2$	0.3545	0.3872	0.3751	0.3957	0.4021

The values in parentheses represent robust standard errors; *, ** and *** indicate significance at the 10%, 5% and 1% levels, respectively.

. Columns (1) through (5) present findings for the 10th, 25th, 50th, 75th, and 90th quantiles, respectively. The estimates show that the coefficient on green finance turns positive and statistically significant at the 50th quantile and above, revealing a pronounced positive relationship at these higher quantiles. Moreover, a meaningful effect is detected only when AGR attains relatively elevated levels (the 50th quantile and beyond). One plausible interpretation is that at higher degrees of AGR synergy, the integration between agriculture and related industries becomes deeper, which enables green finance to exert a more substantial influence on AGR and reinforces its favorable contribution.

The elevated level of economic development in the four centrally administered municipalities may influence the accuracy of estimation. Therefore, observations from Beijing, Tianjin, Shanghai, and Chongqing are excluded, and the model is re-estimated. The corresponding regression results are displayed in column (1) of

Table 5. Results of robustness test.

	(1)	(2)	(3)
	AGR	AGR	AGR
GF	0.721 ***	0.451 ***	0.449 ***
	(0.071)	(0.082)	(0.077)
control	Yes	Yes	Yes
low-carbon city			0.017 **
			(0.008)
Key Control Zone of the Atmosphere			0.005 ***
			(0.001)
constant	−0.376 ***	1.305 ***	−0.278 **
	(0.111)	(0.126)	(0.115)
Province FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
N	416	480	480
${\mathrm{R}}^2$	0.633	0.169	0.565

The values in parentheses represent robust standard errors; ** and *** indicate significance at the 5% and 1% levels, respectively.

. The coefficient on green finance remains positive and statistically significant at the 1% level, indicating that the beneficial relationship between green finance and AGR is not driven by these municipalities and continues to hold in the rest of the sample.

The second column of Table 5 reports the estimates from an alternative measurement of the dependent variable. In the baseline analysis, the AGR synergy level is calculated using the modified coupling coordination model. For this robustness exercise, the uncorrected coupling coordination model is adopted to recalculate AGR, and the equation is re-estimated accordingly. The resulting coefficient on green finance is 0.451, significant at the 1% level, which aligns with the baseline findings. This confirms that the positive association between green finance and AGR remains intact when the dependent variable is reconstructed via the conventional coupling coordination approach.

As an additional robustness exercise, this study incorporates two policy variables—the low-carbon city policy and the key air pollution control zone policy—into the regression. The corresponding estimates are reported in column (3) of Table 5. After controlling for these concurrent policy interventions, the coefficient on green finance remains positive and statistically significant at the 1% level, indicating that the beneficial relationship between green finance and AGR is not confounded by these policies. Taken together, the full set of robustness checks demonstrates that the green finance coefficient is consistently positive and statistically significant, further corroborating Hypothesis 1 (H1).

4.3. Endogeneity Test

To address potential endogeneity concerns, this study instruments green finance with the average green finance level of all other provinces (excluding the focal province) and re-estimates the model using the two-stage least squares (2SLS) estimator. The choice of this instrument is motivated by the substantial “benchmark competition” and interregional imitation effects observed in environmental regulation and green finance policymaking. Green financial innovation and the scale of green credit in a given province are horizontally shaped by the policy orientations and market practices of other provinces. Consequently, the average green finance level of other provinces is strongly correlated with green finance development within the focal province. Regarding exogeneity, this instrument lacks a direct channel to affect AGR in the focal province. AGR outcomes are inherently localized, depending heavily on region-specific factors such as fertilizer application intensity, livestock farming systems, land use patterns, and agricultural technological progress. Compared with industrial pollution, the spatial spillover effects of AGR are relatively limited. Moreover, the design and implementation of green finance policies in other provinces cannot directly influence agricultural production practices or environmental outcomes within a given province. To address potential endogeneity, green finance is instrumented by its mean value across all other provinces, excluding the focal province.

Table 6. Results of endogeneity test.

	(1)	(2)	(3)
	2SLS First	2SLS Second	GMM
	GF	AGR	AGR
L.GF	0.032 ***
	(0.003)
L.FSR			0.934 ***
			(0.057)
GF		0.227 **	0.035 *
		(0.106)	(0.021)
Control	Yes	Yes	Yes
constant	−0.238 ***	−0.078	−0.111
	(0.067)	(0.082)	(0.109)
Province FE	Yes	Yes	Yes
Year FE	Yes	Yes	Yes
LM		91.734
		(0.000)
F		111.517 (16.38)
AR (1)			0.000
AR (2)			0.472
N	480	480	450
${\mathrm{R}}^2$	0.489	0.524

The values in parentheses represent robust standard errors; *, ** and *** indicate significance at the 10%, 5% and 1% levels, respectively.

presents the estimation results. Columns (1) and (2) report the first-stage and second-stage estimates from the two-stage least squares (2SLS) procedure, respectively. The first-stage result shows that the instrument enters with a positive and statistically significant coefficient at the 1% level, confirming its relevance. In the second stage, the coefficient on green finance remains positive and significant at the 1% level, aligning with the baseline findings and indicating a robust positive association. The underidentification test (Kleibergen–Paap rk LM statistic) yields 91.734 (p = 0.000), and the weak identification test (Cragg–Donald Wald F statistic) is 111.517, well above the Stock–Yogo critical value of 16.38, thereby rejecting the null of weak instruments and supporting instrument validity. To further examine endogeneity, the model is also estimated via the system generalized method of moments (GMM); the results are shown in column (3). The GMM estimate of the green finance coefficient remains positive and statistically significant. The Arellano–Bond test yields a p-value below 0.1 for AR(1) and above 0.1 for AR(2), indicating no second-order serial correlation and an appropriate specification. Overall, both the 2SLS and GMM estimates consistently reveal a significant positive relationship between green finance and AGR.

4.4. Heterogeneity Analysis

The influence of green finance on AGR may differ by geographic location and level of agricultural development. From a geographical perspective, regions across China differ markedly in economic development and infrastructure. The eastern region enjoys a higher level of economic development, more rapid growth of green finance, and more comprehensive supporting policies, with green concepts more deeply embedded and industrial transformation and upgrading further advanced. Under these conditions, green finance can play a more substantial role in advancing AGR. In contrast, regions with relatively lower levels of economic development have seen green finance develop later, and its link with AGR tends to be weaker than in the east. The level of agricultural development reflects a region’s agricultural progress and the extent of industrial integration. Classifying regions by agricultural development level permits a further investigation of green finance’s relationship with AGR under varying conditions. Accordingly, this study conducts heterogeneity analysis along two dimensions: geographic location and agricultural development level.

We first examine regional heterogeneity. The sample is divided into four regions—eastern, central, western, and northeastern—and the corresponding estimates are reported in Table 6. The coefficient of green finance is found to be positive and statistically significant across all regions. This finding implies that the beneficial relationship between green finance and AGR is not region-specific; rather, green financial instruments may serve as a broad-based mechanism that facilitates AGR in heterogeneous contexts. In terms of magnitude, the estimated coefficient in the eastern region is substantially larger than those in other regions, implying that the association is more pronounced in the east. A plausible explanation is that green finance in the eastern region emerged earlier, offers a more diverse set of financial instruments, and is supported by more comprehensive policies, including financial regulation. These conditions enable green finance to better leverage its advantages and are associated with a stronger link to AGR.

Heterogeneity also arises from the level of agricultural development. Using the provincial average agricultural output value as the classification criterion, this study divides the full sample into two groups—high and low agricultural development levels—and re-estimates the model for each group separately. The results are reported in columns (5) and (6) of

Table 7. Estimation results of regional heterogeneity.

	(1)	(2)	(3)	(4)	(5)	(6)
	East	Middle	West	Northeast	High	Low
GF	1.121 ***	0.114 ***	0.171	0.510 ***	0.153 ***	0.134
	(0.098)	(0.032)	(0.105)	(0.060)	(0.034)	(0.160)
Control	Yes	Yes	Yes	Yes	Yes	Yes
Constant	1.145 ***	−0.428 ***	−1.577 ***	−0.128	−0.076	−0.638 **
	(0.218)	(0.063)	(0.234)	(0.195)	(0.051)	(0.317)
Province FE	Yes	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
N	160	96	176	48	171	309
${\mathrm{R}}^2$	0.692	0.751	0.836	0.896	0.414	0.749

The values in parentheses represent robust standard errors; ** and *** indicate significance at the 5% and 1% levels, respectively.

. In regions with a higher level of agricultural development, the coefficient on green finance is positive and statistically significant at the 1% level, whereas in regions with a lower level of agricultural development, the coefficient does not reach statistical significance. A plausible explanation is that higher agricultural development is typically accompanied by greater mechanization, wider adoption of new products and technologies, and deeper integration between agriculture and related industries. Under these conditions, relevant actors are more receptive to green, low-carbon production and environmental protection concepts and are better positioned to utilize green financial instruments, which contributes to a stronger positive association between green finance and AGR.

4.5. Mediation Effect Test

Given that the baseline coefficient of green finance on AGR is significantly positive, thus lending preliminary support to H1, this study further introduces INN and INC as mediating variables. A sequential regression procedure is adopted to examine the indirect pathways through which green finance enhances the synergistic governance of AGR.

Table 8. Results of mediation effect test.

	(1)	(2)	(3)	(4)	(5)
	AGR	INN	AGR	INC	AGR
GF	0.636 ***	4.128 ***	0.085 *	1.063 ***	0.087 *
	(0.071)	(0.405)	(0.051)	(0.191)	(0.047)
INN			0.013 **
			(0.005)
INC					0.048 ***
					(0.011)
Control	Yes	Yes	Yes	Yes	Yes
Constant	−0.399 ***	3.900 ***	−0.157 **	6.493 ***	−0.418 ***
	(0.109)	(0.668)	(0.079)	(0.316)	(0.103)
Sobel			0.018 ***		0.005 ***
Province FE	Yes	Yes	Yes	Yes	Yes
Year FE	Yes	Yes	Yes	Yes	Yes
N	480	480	480	480	480
${\mathrm{R}}^2$	0.525	0.644	0.534	0.472	0.547

The values in parentheses represent robust standard errors; *, ** and *** indicate significance at the 10%, 5% and 1% levels, respectively.

presents the results. Columns (2) and (3) report the estimates for the mediation channel via INN. The finding in column (2) indicates that green finance exhibits a significantly positive relationship with INN at the 1% level, suggesting that green finance development substantially fosters INN. Column (3) reveals that both green finance and INN are positively and significantly associated with AGR, confirming that green finance contributes to AGR partly by promoting INN. Hence, H2 is supported. Columns (4) and (5) display the mediating role of INC. As shown in column (4), the coefficient of green finance is positive and significant at the 1% level, implying that green finance can markedly increase INC. In column (5), both green finance and INC enter with significantly positive coefficients for AGR, validating the mediating function of INC in the green finance–AGR nexus. Therefore, H3 is supported. Moreover, In

Table 9. Bootstrap test.

	Effect	Coefficient		95%
INN	Indirect	0.004 **	p = 0.036	[0.000, 0.007]
	direct	0.013 **	p = 0.014	[0.003, 0.023]
INC	Indirect	0.005 **	p = 0.030	[0.000, 0.009]
	direct	0.048 ***	p = 0.000	[0.027, 0.069]

** and *** indicate significance at the 5% and 1% levels, respectively.

Sobel and bootstrap tests confirm that the indirect effects through INN and INC are statistically significant.

4.6. Test of Moderation Effect

To explore whether FIN moderates the relationship between green finance and AGR, an interaction term between FIN and green finance is incorporated into the model. The estimation results are presented in

Table 10. Results of moderation effect test.

	(1)	(5)
	AGR	AGR
GF	0.636 ***	0.431 ***
	(0.071)	(0.078)
FIN		0.062 ***
		(0.008)
FIN $\times$ GF		0.082 ***
		(0.031)
Control	Yes	Yes
Constant	−0.399 ***	−0.889 ***
	(0.109)	(0.125)
Province FE	Yes	Yes
Year FE	Yes	Yes
N	480	480
${\mathrm{R}}^2$	0.525	0.577

The values in parentheses represent robust standard errors; *** indicates significance at the 1% levels.

. With the interaction term included, both the interaction term and the green finance variable display positive and statistically significant coefficients. This finding indicates that FIN acts as a moderating factor: as FIN expands, the positive association between green finance and AGR intensifies. To characterize this moderating effect more precisely, a simple slope analysis is conducted to evaluate the marginal effect of green finance on AGR at different levels of FIN (

Table 11. Marginal effects of green finance on AGR at different levels of FIN.

Level of FIN	Value	Marginal Effect (b)	SE	t	p	95%Cl
Low (−1 SD)	9.010	0.028	0.044	0.06	0.948	[−0.084, 0.089]
Mean	10.423	0.144	0.039	3.65	0.000	[0.066, 0.221]
High (+1 SD)	11.836	0.285	0.066	4.32	0.000	[0.155, 0.415]

). The results demonstrate that both the magnitude and statistical significance of this marginal effect are conditional on the level of FIN. When FIN is low (M − 1 SD = 9.010), the marginal effect is a negligible 0.028 and statistically nonsignificant (p = 0.948), suggesting that green finance does not yet contribute meaningfully to AGR without adequate fiscal backing. At the mean level of FIN (10.423), the marginal effect increases to 0.144 and becomes significant at the 1% level (p < 0.001), indicating that green finance begins to yield discernible AGR improvements. At a high level of FIN (M + 1 SD = 11.836), the marginal effect rises further to 0.285 (p < 0.001), reflecting the strongest AGR enhancement. Overall, the marginal effects display a monotonic upward trend across FIN levels, shifting from nonsignificant to significant. This pattern aligns with the significantly positive interaction term estimated in the baseline model, offering strong evidence that FIN positively moderates the effect of green finance on AGR. Importantly, this moderating role becomes effective and progressively strengthens only after FIN surpasses a certain threshold. Thus, H4 is supported.

4.7. Threshold Effect Test

EDU significantly shapes farmers’ voluntary adoption of green production practices. To further investigate how the relationship between green finance and AGR varies with educational conditions, this study treats the EDU level as a threshold variable.

Table 12. Results of threshold effect test.

Variable of Threshold	Model	F	p	BS	1%	5%	10%
EDU	Single	41.56	0.003	300	24.058	29.546	34.488

presents the threshold effect test results, and

Table 13. Estimation results of threshold values.

Variable of Threshold	Model	The Threshold Estimate	95%
EDU	Single	6.7173	[6.6779, 6.7360]

reports the estimated threshold value. As indicated in Table 12, the single-threshold test produces a p-value of 0.003, thereby verifying its statistical significance. The model thus exhibits a single threshold effect and is suitable for threshold analysis. Table 12 also presents the estimated threshold value and its 95% confidence interval derived from the likelihood ratio (LR) test. Table 13 gives the single-threshold estimate of 6.7173. These results indicate that the EDU level significantly passes the single-threshold test. Accordingly, the single-threshold model is employed in the subsequent empirical analysis.

Having identified the threshold, a piecewise regression is conducted with AGR as the dependent variable and EDU as the threshold variable. The estimation results are presented in

Table 14. Threshold regression results.

	(1)
	AGR
EDU ≤ 6.7173	0.411 **
	(0.164)
EDU > 6.7173	0.531 ***
	(0.140)
control	Yes
Constant	−0.401 *
	(0.207)
Province FE	Yes
Year FE	Yes
N	480
${\mathrm{R}}^2$	0.564

The values in parentheses represent robust standard errors; *, ** and *** indicate significance at the 10%, 5% and 1% levels, respectively.

. In the regime where the EDU level lies below the estimated threshold of 6.7173, green finance exhibits a positive coefficient of 0.411. A possible explanation is that lower EDU attainment is associated with limited financial literacy and weaker receptiveness to new technologies and ideas, making it difficult for farmers to effectively leverage green finance instruments and policies to secure funding. Once EDU surpasses the threshold, farmers tend to possess stronger literacy, allowing them to actively acquire green finance knowledge, utilize their financial understanding to obtain more credit, and more readily embrace green production concepts. Under these conditions, green finance exhibits a stronger positive association with AGR. Consequently, a higher level of EDU enhances the AGR-related benefits associated with green finance.

5. Discussion

The essential role of green finance in advancing green development has been amply recognized. Green finance can facilitate the sustainability transition of agriculture [22], exert a favorable influence on the green growth trajectory of emerging economies [23], and contribute to the improvement of national ESG outcomes [24]. Prior investigations addressing the effects of green finance on agricultural carbon emissions and agricultural non-point source pollution have predominantly examined these two issues in isolation. A large body of work has scrutinized the green finance–agricultural carbon emission nexus, with findings indicating that green finance can meaningfully restrain agricultural carbon emissions. The principal channels identified involve green technology innovation [25], input composition optimization [26], and industrial structure upgrading [27]. The potency of green finance in mitigating agricultural non-point source pollution has similarly been corroborated. The existing literature broadly indicates that, through pathways such as environmental regulatory instruments, land transfer [28], the expansion of large-scale farming, and the deployment of green technologies [29], green finance can markedly reduce agricultural non-point source pollution. These insights furnish a solid theoretical and empirical underpinning for the present investigation. The central contribution of this paper rests on integrating green finance, agricultural carbon emissions, and agricultural non-point source pollution into a coherent analytical framework. Because agricultural carbon emissions and non-point source pollution share common roots and can be addressed through analogous governance strategies, this study regards them as an interrelated whole and adopts a holistic lens to examine agricultural pollution abatement and carbon mitigation. The concrete contributions are as follows:

We make an empirical contribution. On the empirical front, drawing on Chinese provincial panel data covering the period 2008–2023, this study offers systematic evidence that substantiates the promoting effect of green finance on AGR through theoretical reasoning and empirical testing. It further delineates region-specific heterogeneous effects, furnishing practical references for the deployment of green finance and the tackling of agricultural green development challenges.

We provide a mechanism of action. This incorporates INN and INC as mediating variables and clarifies the transmission channels through which green finance exerts its influence on AGR, supported by both theoretical logic and empirical verification.

We also make a conditional-effect contribution. In terms of conditional effects, a further notable contribution lies in the systematic investigation of the external conditions that moderate the green finance–AGR relationship. Specifically, the impact of green finance on AGR is shown to be contingent on FIN, regional factors, agricultural development levels, and EDU.

6. Conclusions

Utilizing a balanced panel dataset of 30 provincial-level administrative regions in China from 2008 to 2023, this study empirically investigates the effect of green finance on AGR. The principal findings are summarized as follows. First, green finance is significantly and positively associated with AGR. Second, this positive association is more pronounced in the eastern region and in regions with higher agricultural development levels. Third, INN and INC serve as mediating channels between green finance and AGR. Fourth, FIN plays a significant moderating role in the green finance–AGR nexus. Fifth, the influence of green finance on AGR exhibits a single-threshold effect with respect to the level of EDU.

To accelerate the improvement of China’s green financial system and facilitate the coordinated development of AGR, the following policy recommendations are proposed. The government should continue to strengthen the green financial market infrastructure and formulate unified national green standards. Specifically, it is necessary to establish uniform definitions for various green financial products and gradually align the domestic criteria for standard-setting and project recognition with international benchmarks. Furthermore, the development of green bonds, green funds, green equities, and green insurance can be advanced through preferential policies and market-based incentives, such as tax reductions and risk compensation mechanisms. In practice, dynamic adjustments should be made according to the specific conditions of each province. Given its higher level of economic development, earlier initiation of green finance, and more complete market system, the eastern region should focus on steering green finance toward a more sophisticated and high-end trajectory, while actively encouraging financial institutions to engage in international green finance cooperation. The northeastern region, endowed with abundant agricultural resources and extensive arable land, should encourage financial institutions to refine green financial products and steer capital flows toward the agricultural sector through tax incentives and other policy instruments; the central and western regions, where agricultural production methods remain comparatively extensive and agro-ecological environments are relatively fragile, should prioritize environmental protection in agricultural development to prevent ecosystem degradation. The government should boost investment in agricultural R&D, establish dedicated research funds, and invigorate INN. The commercialization and application of INN achievements should be accelerated. Tax relief policies should be offered to entities that develop and utilize INN outcomes, thereby lowering the costs of new technology R&D and extension. Green financial institutions should continuously innovate and launch a wider array of green financial products tailored to agricultural production, thus enabling farmers to constantly diversify their income channels through green finance and achieve income growth. More fiscal resources should be channeled into green agricultural production to advance pollution abatement and carbon mitigation. The government should further strengthen its support for agriculture by continuously refining fiscal assistance mechanisms and allocating a larger share of public expenditure to agricultural production. It is also essential to improve the efficiency with which fiscal funds are utilized, ensuring that these resources yield tangible benefits in agricultural production. In particular, a greater proportion of fiscal funds should be directed to farmers who adopt green production technologies, and dedicated subsidies should be granted to those employing clean production methods, thereby encouraging wider adoption of green, low-carbon, and environmentally sustainable farming practices. Additionally, targeted training programs, such as specialized lectures on green finance, should be organized to educate farmers on how to access funds through green financial instruments and to clarify the relevant application procedures. Such measures can enhance farmers’ financial literacy and enable green finance to operate more effectively.

A key limitation of this study is that data constraints restrict the sample period to 2023, making it difficult to capture the recent evolution of green finance and AGR over the past three years. Furthermore, this study relies on provincial panel data, and the state of AGR at more micro levels—such as the county level—has not been examined in detail. The selection of instrumental variables also poses inherent challenges, which prevents a full resolution of endogeneity concerns. Future research will update the dataset promptly as official statistics become available and extend the analysis to finer spatial scales. Moreover, more suitable instrumental variables will be sought to enable a more rigorous investigation of the green finance–AGR nexus.