Dual-Wheel Drive and Agricultural Green Development: The Co-Evolution and Impact of Digital Inclusive Finance and Green Finance

Abstract

Agricultural green development cannot be achieved without effective financial support. Based on panel data from 30 provinces in China from 2014 to 2023, this paper uses a coupling coordination model to measure and analyse the degree of coordination between digital inclusive finance and green finance; this further constructs a comprehensive evaluation system for agricultural green development, and on this basis uses a fixed-effect model and a threshold regression model to systematically examine the impact of the coordination between the two on agricultural green development. The findings are as follows: first, the coordination between digital inclusive finance and green finance shows an upward trend over time, shifting spatially from a high trend in the east to a low trend in the west to regional convergence; second, the coordination between the two has a substantial and favourable impact on green agricultural development, which is a conclusion that persists after robustness tests; third, the effect is heterogeneous, with more pronounced promotion effects in non-grain-producing regions, regions with high agricultural technology levels, and low levels of financial exclusion; fourth, the effect exhibits nonlinear characteristics, with coordination and agricultural industrial agglomeration each forming a single-threshold effect. This research lays down a foundational framework for financial coordination in supporting agricultural green development. It suggests promoting a dual-wheel coordination mechanism to effectively empower agricultural green development by strengthening technological empowerment, regional linkage, and designing differentiated policies.

1. Introduction

Against the background of the high consistency of the global sustainable development agenda and the “dual carbon” strategy, agricultural green development (AGD) has become a core proposition for restructuring agricultural production systems and optimising resource and environmental allocation. As a key hub for guiding resource flows and shaping behavioural patterns, the financial system’s functional boundaries are continuously expanding with the combined evolution of digitalisation and greening. The goal of digital inclusive finance (DIF) is to connect rural financial services to users in the last mile and allow the “long tail” of farmers who have been blocked out by traditional finance to return to the financial system, effectively alleviating financing constraints for farmers and promoting inclusive growth in the agricultural economy. Green finance (GF) emphasises the allocation of green resources, utilising financial means and instruments to encourage a greater flow of financial resources into environmentally friendly and resource-saving industries, thereby guiding the coordinated development of financial activities and environmental protection. However, when promoting green development, single-dimensional financial instruments often find it difficult to simultaneously take into account both inclusiveness and environmental performance. Therefore, the coordinated development of DIF and GF is particularly critical. The organic integration of the two enables a breakthrough in the constraints of the traditional financial framework, while also attaining the dual objectives of inclusiveness and sustainability.

Current academic research on the impact of DIF and GF on AGD focuses on the independent effects of these two. DIF can increase smallholder farmers’ savings and input utilisation, promote output and food security improvements, and build financial capacity and risk buffers for adopting green and climate-smart agriculture (CSA) [1,2,3]. In terms of operational mechanisms, big data, blockchain, and other technologies have enabled DIF to significantly enhance the coverage scope and security level of rural financial services. On one hand, it helps lenders establish more reliable farmer credit records, which reduces information gaps and moral hazard issues [4]; on the other hand, it further facilitates the AGD through raising the calibre of agricultural technological innovation and mitigating financial constraints [4,5,6]. In addition, some researchers have discovered that the effect of DIF on AGD possesses spatial spillover properties, meaning its effects can be transmitted across administrative boundaries to neighbouring regions. However, this spillover effect is not a one-way street, and it can manifest as either a positive promoting effect or an adverse inhibitory effect [7,8].

In the overall framework of promoting AGD, DIF and GF play complementary roles. The former mainly improves farmers’ financing capabilities and risk management at the micro level, while the latter promotes low-carbon and AGD at the macro level. The development of green financial instruments and mechanisms expands more funding channels for AGD, provides key support for green industries, especially low-carbon agriculture [9,10], and can also guide farmers in adopting low-carbon and sustainable production methods [11]. On this basis, GF has become an important force in promoting AGD by optimising resource allocation in rural areas, facilitating industrial structure adjustments, and promoting the transformation of economic development methods [12]. However, its role is not balanced in space. Studies have found that the promotional effect of GF is constrained by institutional and regional factors [13,14]. Meanwhile, the role of GF in boosting agricultural productivity has also attracted extensive attention. It facilitates the enhancement of green total factor productivity by guiding investment in green R&D and advancing technological advancement [15].

On this basis, the coordinated development of DIF and GF has gradually become a focus of academic attention. Collaboration generally refers to the mutual cooperation between two or more entities to achieve common goals, integrate resources, and achieve mutual benefit. Its quantitative analysis often relies on coupling coordination models [16,17,18,19]. In recent times, scholars have proposed the notion of “green digital finance”, stressing that, as a combination of digital finance and GF, it is capable of utilising innovative technologies to realise data analysis, refine investment decisions, and create employment opportunities, thereby reconciling low-carbon transformation with sustainable economic development [20,21]. Some scholars have further constructed an analytical framework connecting digital finance, GF, and social finance, pointing out that efficient GF needs to be achieved through digital financial means, thereby further promoting sustainable development [22]. Related results also show that compared with a single financial model, “green digital finance” is more conducive to achieving environmentally friendly goals [23,24]. Specifically in the field of green agricultural development, on one side, scholars have proposed the concept of “green inclusive finance” by comparing the differences and similarities between GF and inclusive finance, believing that it can provide financial support for organic and green agriculture with Chinese characteristics [25]. On the other side, from the viewpoint of financial service efficiency, studies have identified that the coordinated advancement of inclusive finance and GF not only enhances the availability and coverage of financial services in “agriculture, rural areas and farmers” regions, but also provides crucial momentum for the green transformation of agriculture [26].

In summary, existing research has laid a relatively solid theoretical foundation for this paper. However, there is still room for further research: first, regarding the impact of the coordination between DIF and GF on AGD, most relevant research is scattered, and the overall analytical framework needs to be further improved; second, although prior research has put forward the concept of “green inclusive finance”, the effects of its impacts still need to be more systematically sorted out and explained. Based on this, the marginal contribution of this paper is mainly reflected in three aspects: first, from a synergistic perspective, it investigates the influence of the coordination between DIF and GF on AGD, enriching theoretical understanding of their linkage to AGD; second, building upon the overall analysis, it further combines factors such as major grain-producing areas, agricultural science and technology levels, and the degree of financial exclusion to reveal the heterogeneous characteristics of the synergistic effects of the two; third, it introduces a nonlinear analytical framework to explore the threshold effect of the coordination between DIF and GF and agricultural industry agglomeration on AGD, providing empirical support for improving the path of the synergistic promotion of the AGD.

2. Theoretical Analysis and Research Hypotheses

2.1. The Direct Impact of the Coordination Between DIF and GF on AGD

The core goal of AGD is to achieve unity between “ecological value” and “economic value”, emphasising environmental friendliness, resource conservation, and production efficiency [27,28]. The synergistic development of DIF and GF provides complementary financial support for the realisation of this goal. From the perspective of environmental friendliness, DIF relies on tools such as online credit, digital payment and digital currency, and has the characteristics of information transparency and process traceability [29]. It can implement precise constraints on agricultural loans with high pollution and high carbon emissions, thereby forcing agricultural production methods to transform to low-carbon and clean ones, and improving the green utilisation efficiency of farmland [30]. At the same time, GF forms “hard constraints” and “positive incentives” on agricultural development through institutional tools including green credit and bonds. Thereby, it secures stable funding for the agricultural sector’s high-quality transition, while also directing investment toward eco-friendly initiatives via rigorous environmental standards to ensure ecological benefits [31]. From the perspective of resource conservation, DIF, with its low-cost and convenient financial services, lowers the threshold for farmers to access green technologies and energy-saving equipment, promotes resource utilisation efficiency, and enhances the growth of agricultural total factor productivity [32]. GF, through green credit incentives and capital allocation optimisation, guides agricultural production and management entities to focus on conservation and recycling, promotes the formation of low-consumption and sustainable production and lifestyles, and underpins the agricultural green transition with an enduring institutional framework [33]. From the perspective of production efficiency, DIF breaks through the geographical and information barriers of traditional finance, provides farmers and agricultural management entities with more accessible financing channels, helps accelerate rural capital formation and accumulation, drives rural consumption and productive investment, and further promotes agricultural investment and output efficiency [34]. At the same time, by means of capital input and industrial restructuring, GF serves to underpin the dissemination of green technologies while simultaneously driving agricultural technological innovation, thereby securing higher economic returns via enhanced production efficiency [35]. Therefore, DIF and GF complement each other in terms of function. The former improves the micro-feasibility of green development with its inclusiveness and flexibility, while the latter strengthens the macro-constraints of green development with its institutional and guiding nature. The synergistic effect of these two approaches can ensure both agricultural production efficiency and ecological benefits, thereby promoting the development of agriculture towards a low-carbon, green, and sustainable direction. Based on this, Hypothesis 1 is proposed.

H1: 

DIF and GF work together to promote green agricultural development directly.

2.2. The Nonlinear Effect of DIF and GF Coordination on AGD

The synergistic development of DIF and GF is a phenomenon in which financial resources are concentrated in the field of “green inclusive”. This focus not only promotes the rapid growth of green industries, but also accelerates the environmental modernization of traditional industries. Previous studies have shown that DIF has a threshold effect on AGD [36,37]. Similarly, studies have found that GF also showed a threshold effect in the mitigation of agricultural non-point source pollution and the promotion of sustainable agricultural practices [38]. In light of this, the difference in the coordination between the development of DIF and GF may lead to different effects in promoting AGD. By applying the degree of coordination of DIF and GF in the role of the threshold variable, this research examines whether its effect on AGD is nonlinear.

Agricultural industry agglomeration can produce economies of scale, which in turn affects the efficiency of AGD. DIF is capable of surmounting the constraints imposed by time and space, and by means of digital technology, drive the flow of capital, labour, and agricultural technology to agriculturally developed areas. The existence of the agglomeration effect and spatial interaction effect prompts the agglomerated production factors to exert diffusion effects further and drive agricultural development in surrounding areas [39]. GF can attract funds and technology to the agricultural industry, creating a spillover effect that reduces information asymmetry between various agricultural entities, lowers transaction costs in the grain industry, and promotes the development of upstream and downstream industries, thereby forming an agricultural industry agglomeration [40]. The relationship between DIF-GF coordination and AGD is likely subject to agricultural industry agglomeration, with the potential to demonstrate a nonlinear threshold pattern. Given this context, this paper incorporates the agricultural industry agglomeration factor into the research framework of the relationship between DIF, GF coordination, and AGD. It discusses whether the relationship between the two exhibits nonlinear characteristics as the intensity of agricultural industry agglomeration changes. Based on the previous analysis, hypotheses 2 and 3 are proposed.

H2: 

The coordination between DIF and GF has a threshold effect on AGD.

H3: 

The agricultural industry’s agglomeration has a threshold effect on the synergistic impact of DIF and GF on AGD.

On the basis of theoretical analysis, this paper further develops the following theoretical framework to explore the mechanism by which the coordination between DIF and GF influences AGD, as presented in Figure 1.

3. Evaluation of the Coupling Degree of DIF and GF

3.1. Construction of an Indicator System

Drawing on the connotations and characteristics of the coordination between DIF and GF, and with reference to the existing relevant literature [41,42,43,44,45,46], an indicator system was established, as shown in

Table 1. Digital inclusive finance and the green finance indicator system.

Dimensions	Specific Indicators	Measurement Description
DIF	Breadth of coverage	Digital inclusive finance coverage breadth index
	Depth of use	Digital inclusive finance depth index
	Degree of digitisation	Digital inclusive finance digitalization degree index
GF	Green credit	Total credit for environmental protection projects/total credit for the province
	Green investment	Investment in environmental pollution control/GDP
	Green insurance	Environmental pollution liability insurance income/total premium income
	Green bonds	Green bond issuance/total bond issuance
	Green support	Financial environmental protection expenditure/general fiscal budget expenditure
	Green funds	Total market value of green funds/total market value of all funds
	Green rights	Carbon trading, energy use rights trading, and emission rights trading/Total equity market transactions

. The DIF subsystem is measured using the DIF Development Index, published by the Peking University Digital Finance Research Centre, comprising three dimensions: breadth of coverage, depth of use, and degree of digitisation. The GF subsystem is constructed from seven dimensions: green credit, green investment, green insurance, green bonds, green support, green funds, and green rights, with a comprehensive evaluation conducted via the entropy method. An overview of the specific indicator system is provided in Table 1.

Based on the relevant literature [47,48,49], the entropy method is divided into four steps:

The first step is to construct an evaluation matrix, as defined by Equation (1).

$${\mathrm{Z}}_{\mathrm{M}\times \mathrm{N}}{=[\mathrm{Z}}_{\mathrm{ij}}{]}_{\mathrm{M}\times \mathrm{N}}$$

(1)

In the second step, the evaluation matrix is standardised using the maximum–minimum method, as defined by Equation (2).

$${[{\mathrm{Z}}_{ij}]}_{m\times n}={\displaystyle \frac{{\mathrm{X}}_{ij}-\underset{1\le i\le m,1\le j\le n}{{\mathrm{X}}_{\min }}}{\underset{1\le i\le m,1\le j\le n}{{\mathrm{X}}_{\max }}-\underset{1\le i\le m,1\le j\le n}{{\mathrm{X}}_{\min }}}}$$

(2)

where $i$ = 1, 2, …, m, $j$ = 1, 2, …, n, $m\times n$ is the standardised matrix.

The third step is to calculate indicator weights via the information entropy method, as defined by Equations (3)–(5).

$${\mathrm{w}}_{\mathrm{j}}={\displaystyle \frac{{1+\mathrm{E}}_{\mathrm{j}}}{\mathrm{n}-{\displaystyle \sum _{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{E}}_{\mathrm{j}}}}}$$

(3)

$${\mathrm{E}}_{\mathrm{j}}=-{\displaystyle \sum _{\mathrm{i}=1}^{\mathrm{m}}{\mathrm{P}}_{\mathrm{ij}}{\mathrm{lnP}}_{\mathrm{ij}}}$$

(4)

$${\mathrm{P}}_{\mathrm{ij}}={\displaystyle \frac{{1+\mathrm{x}}_{\mathrm{ij}}}{{\displaystyle \sum _{\mathrm{i}=1}^{\mathrm{m}}{(1+\mathrm{x}}_{\mathrm{ij}})}}}$$

(5)

Finally, the relevant indicators are aggregated into a composite score, as defined by Equation (6).

$$\mathrm{C}={\displaystyle \sum _{\mathrm{i}}{\mathrm{w}}_{\mathrm{j}}{\mathrm{x}}_{\mathrm{ij}}}$$

(6)

3.2. Coupling Coordination Model Setting

In light of the study conducted by Wang et al. [50], the coupling degree formula between the two subsystems is

$$\mathrm{C}=\sqrt{{\displaystyle \frac{{\mathrm{U}}_1{\mathrm{U}}_2}{{\left(\frac{{\mathrm{U}}_1{+\mathrm{U}}_2}{2}\right)}^2}}}={\displaystyle \frac{2\sqrt{{\mathrm{U}}_1{\mathrm{U}}_2}}{{\mathrm{U}}_1{+\mathrm{U}}_2}}$$

(7)

Among them, U₁ and U₂ are the comprehensive evaluation indices of the DIF and GF systems, as shown in Section 2.1. C represents the coupling degree of the two systems, with a value range of [0, 1]. To quantitatively characterise the interaction between DIF and GF, the coupling coordination degree approach is adopted, as detailed below:

$$\mathrm{T}=\mathrm{a}\times {\mathrm{U}}_1+\mathrm{b}\times {\mathrm{U}}_2$$

(8)

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

(9)

In Formula (8), a represents the contribution rate of the DIF subsystem to coordination, and b represents the contribution rate of the GF subsystem to coordination. Considering that the contribution rates of the two systems are equal, a = b = 0.5; COU represents the coupling coordination degree of DIF and GF, and values range from 0 to 1. The higher the number, the higher the match between the two subsystems. Specific classification rules can be seen in

Table 2. Classification criteria for coupling coordination degree.

Intervals	Disharmony				Coordination
Intervals	[0, 0.1)	[0.1, 0.3)	[0.3, 0.4)	[0.4, 0.5)	[0.5, 0.6)	[0.6, 0.7)	[0.7, 0.9)	[0.9, 1)
Coupling Coordination	Severe	Moderate	Mild	Near	Primary	Intermediate	Good	High-Quality
Classification	Antagonistic Zone			Running-in Zone		Coordinated Zone

.

3.3. Analysis of Measurement Results

3.3.1. Analysis of Time-Series Characteristics

Taking the national average as the reference, the degree of coordination between China’s DIF and GF has maintained a steady upward trend from 2014 to 2023, rising from 0.4751 in 2014 to 0.6167 in 2023, with a cumulative growth of 29.8%. This reflects a continuous enhancement in the degree of coordination between the two. The vertical evolution process comprises three primary phases. First, the cultivation phase (2014–2016), during which the coordination degree climbed from 0.4751 to 0.5123. In this phase, the policy framework was initially built, and the interactive relationship between DIF and GF was still in the stage of cultivation and exploration. Second, the improvement phase (2017–2019), where the coordination degree rose from 0.5385 to 0.5697. The launch of the GF reform and innovation pilot zones, coupled with a marked expansion in DIF coverage, jointly drove the in-depth integration of the two. Third, the stable phase (2020–2023). Despite a slight drop in 2022 compared to 2021, the overall primary coordination level remained stable. It achieved an intermediate coordination level in 2023, which suggests that the two-way promotion mechanism has become relatively mature and the system coupling has stabilised. The specific values are shown in

Table 3. The degree of coordination between DIF and GF in 30 provinces and cities from 2014 to 2023.

	2014	2015	2016	2017	2018	2019	2020	2021	2022	2023	Mean
Beijing	0.5473	0.5683	0.5813	0.6037	0.6262	0.6454	0.6550	0.6673	0.6612	0.6857	0.6241
Tianjin	0.4666	0.4881	0.4997	0.5242	0.5334	0.5504	0.5624	0.5706	0.5665	0.6023	0.5364
Hebei	0.4940	0.5306	0.5437	0.5737	0.5909	0.6054	0.6198	0.6317	0.6257	0.6611	0.5877
Shanxi	0.4447	0.4783	0.4852	0.5009	0.5199	0.5349	0.5571	0.5628	0.5599	0.5755	0.5219
InnerMongolia	0.3561	0.3870	0.4005	0.4125	0.4271	0.4273	0.4352	0.4444	0.4398	0.4490	0.4179
Liaoning	0.5199	0.5460	0.5532	0.5779	0.5920	0.6091	0.6231	0.6313	0.6272	0.6592	0.5939
Jilin	0.4479	0.4729	0.4806	0.5063	0.5213	0.5300	0.5398	0.5496	0.5447	0.5840	0.5177
Heilongjiang	0.5040	0.5367	0.5467	0.5686	0.5875	0.6025	0.6099	0.6173	0.6136	0.6480	0.5835
Shanghai	0.5927	0.6193	0.6263	0.6601	0.6835	0.7014	0.7128	0.7251	0.7189	0.7584	0.6798
Jiangsu	0.5247	0.5528	0.5616	0.5880	0.6127	0.6281	0.6432	0.6557	0.6494	0.6692	0.6085
Zhejiang	0.5383	0.5640	0.5696	0.6005	0.6249	0.6362	0.6493	0.6650	0.6571	0.6843	0.6189
Anhui	0.4555	0.4787	0.4916	0.5198	0.5341	0.5500	0.5604	0.5710	0.5657	0.5917	0.5319
Fujian	0.5199	0.5573	0.5602	0.5910	0.6090	0.6310	0.6348	0.6527	0.6437	0.6743	0.6074
Jiangxi	0.4546	0.4733	0.4861	0.5091	0.5268	0.5465	0.5598	0.5701	0.5649	0.5906	0.5282
Shandong	0.5112	0.5390	0.5496	0.5784	0.5947	0.6129	0.6216	0.6368	0.6292	0.6755	0.5949
Henan	0.4454	0.4731	0.4855	0.5125	0.5263	0.5436	0.5540	0.5655	0.5597	0.5973	0.5263
Hubei	0.5169	0.5435	0.5551	0.5822	0.6030	0.6245	0.6303	0.6414	0.6358	0.6586	0.5991
Hunan	0.4990	0.5295	0.5414	0.5714	0.5885	0.6046	0.6168	0.6316	0.6242	0.6629	0.5870
Guangdong	0.5301	0.5571	0.5651	0.5953	0.6144	0.6318	0.6467	0.6584	0.6525	0.6850	0.6136
Guangxi	0.4992	0.5292	0.5445	0.5728	0.5914	0.6004	0.6156	0.6240	0.6198	0.6603	0.5857
Hainan	0.5052	0.5500	0.5454	0.5742	0.6084	0.6095	0.6263	0.6406	0.6334	0.6444	0.5937
Chongqing	0.5140	0.5427	0.5556	0.5785	0.5971	0.6124	0.6261	0.6406	0.6333	0.6549	0.5955
Sichuan	0.4509	0.4792	0.4874	0.5171	0.5300	0.5392	0.5564	0.5589	0.5576	0.5920	0.5269
Guizhou	0.4890	0.5245	0.5349	0.5681	0.5897	0.5972	0.6067	0.6222	0.6144	0.6414	0.5788
Yunnan	0.3597	0.3741	0.4016	0.4049	0.4293	0.4343	0.4349	0.4543	0.4445	0.4665	0.4204
Shaanxi	0.5104	0.5370	0.5499	0.5744	0.5955	0.6130	0.6233	0.6326	0.6279	0.6589	0.5923
Gansu	0.4939	0.5278	0.5313	0.5610	0.5793	0.5914	0.6038	0.6153	0.6095	0.6451	0.5759
Qinghai	0.3456	0.3767	0.3730	0.4132	0.4125	0.4189	0.4169	0.4459	0.4312	0.4561	0.4090
Ningxia	0.3566	0.3844	0.3798	0.4042	0.4104	0.4331	0.4187	0.4428	0.4306	0.4788	0.4139
Xinjiang	0.3605	0.3811	0.3822	0.4102	0.4192	0.4258	0.4434	0.4411	0.4422	0.4901	0.4196
Mean	0.4751	0.5034	0.5123	0.5385	0.5560	0.5697	0.5801	0.5922	0.5861	0.6167	0.5530

.

3.3.2. Spatial Distribution Characteristics

To further examine the spatial characteristics of the coordinated development of DIF and GF, we mapped the spatial distribution of these two sectors for 2014, 2017, and 2023, as shown in Figure 2. Regional distribution revealed that, in 2014, the coordinated development of DIF and GF across China was on the verge of disharmony, presenting a clear “eastern high, western low” pattern. This pattern is primarily due to regional differences in economic development, infrastructure development, and policy support. Eastern coastal regions such as Beijing, Zhejiang, Guangdong, and Shanghai boast developed economies, well-developed digital infrastructure, and a significant clustering effect of financial institutions. Furthermore, local governments have provided strong policy support for GF, fostering the deep integration of digital technologies and GF. In contrast, western provinces such as Qinghai, Ningxia, Yunnan, and Xinjiang have lagged in economic development, with low digital infrastructure coverage, a weak green industry foundation, and insufficient financial resources, resulting in a relatively low level of coordination. In 2017, the national average entered the initial coordination stage of the running-in period, with high-value areas expanding to central China, and central and western provinces such as Chongqing, Guizhou, and Heilongjiang experiencing significant improvements. This change is driven by the advancement of regional coordinated development strategies and the rapid penetration of DIF. For example, the “Broadband China” strategy and the widespread adoption of mobile internet have significantly improved digital access in central and western regions. Simultaneously, inclusive finance pilot policies and GF incentive mechanisms are gradually shifting towards central and western regions, enhancing the accessibility and green orientation of financial services. By 2023, China enters the intermediate coordination stage overall, with the eastern coastal regions maintaining their lead and several provinces in central and western China, such as Chongqing, Shaanxi, and Guizhou, entering this stage. While still at relatively low levels, Qinghai, Ningxia, Xinjiang, and Yunnan have significantly narrowed their gap with the national average, indicating that regional coordination between the two is shifting from highly uneven to gradually converging. On the one hand, nationwide digital infrastructure coverage has narrowed the digital divide between regions. On the other hand, the “dual carbon” goals and inclusive finance reform pilot zone policies have synergized, enhancing the degree of coordination of lagging regions through mechanisms such as cross-regional ecological compensation and green fintech pilots. In summary, the degree of coordination between DIF and GF has continued to strengthen across Chinese provinces in recent years, but significant regional disparities in development remain.

4. Model Construction and Variable Selection

4.1. Model Setting

First, according to the research idea, the coordination between DIF and GF is utilised as the primary explanatory variable, and the various factors that can affect the AGD are used as control variables to set the multivariate linear regression model, as shown in Formula (10).

$${\mathrm{AGD}}_{\mathrm{it}}{=\mathsf{\alpha }}_0{+\mathsf{\alpha }}_1{\mathrm{COU}}_{\mathrm{it}}+\mathsf{\Sigma }{\mathsf{\alpha }}_{\mathrm{k}}{\mathrm{control}}_{\mathrm{it}}{+\mathrm{uit}+\mathrm{vit}+\mathsf{\epsilon }}_{\mathrm{it}}$$

(10)

wherein AGD_it represents the level of agriculture green development; COU_it is coordination, as proxied by the DIF-GF coupling coordination degree; α₀ is the intercept term; α₁ represents the regression coefficient of the explanatory variable; controlit is the control variable; α_k is the regression coefficient of each control variable; u_i represents province fixed effect; v_t represents year fixed effect; and $\mathsf{\epsilon }$_it is the error term.

Second, a threshold effect model is constructed to explore the nonlinear spillover effect of the synergistic development of DIF and GF on AGD, as shown in Formula (11).

$${\mathrm{AGD}}_{\mathrm{it}}{=\mathrm{f}}_0{+\mathrm{f}}_1{\mathrm{COU}}_{\mathrm{it}}\mathrm{I}({\mathrm{th}}_{\mathrm{it}}\le {\mathsf{\theta })+\mathrm{f}}_2{\mathrm{Cou}}_{\mathrm{it}}\mathrm{I}({\mathrm{th}}_{\mathrm{it}}>\mathsf{\theta })+\mathsf{\Sigma }{\mathrm{f}}_{\mathrm{k}}{\mathrm{control}}_{\mathrm{it}}{+\mathrm{uit}+\mathsf{\epsilon }}_{\mathrm{it}}$$

(11)

where θ denotes the threshold value to be estimated, and I(·) is the indicator function. The definitions of all other symbols hold the same meanings as in Equation (10).

4.2. Variable Selection and Data Description

4.2.1. Explained Variable

The explained variable in this paper is AGD. Integrating the existing research, this paper uses three dimensions: eco-friendliness, resource conservation, and production efficiency to measure agricultural green development [51,52,53,54,55]. Specific indicators are shown in

Table 4. Indicator system for measuring AGD.

Dimension	Specific Indicators	Measurement Description	Unit	Property
Eco-Friendly	Fertiliser input intensity	Agricultural fertiliser application/crop planting area	kg/hm2	Negative
	Film input intensity	Plant film application/crop planting area	kg/hm2	Negative
	Pesticide carbon input intensity	Pesticide application/crop planting area	kg/hm2	Negative
	Diesel input intensity	Diesel use/crop planting area	kg/hm2	Negative
Development Conditions	Land productivity	Gross agricultural output/crop planting area	%	Positive
	Urbanisation rate	Urban population/total population	%	Positive
	Machinery-use intensity	Total agricultural machinery power/crop planting area	kW/hm2	Positive
	Electricity-use intensity	Rural electricity consumption/gross agricultural output	kW·h/100 million yuan	Negative
Production Efficiency	Rural resident income	Rural residents’ disposable income	Ten thousand yuan	Positive
	Crop yield per unit area	Grain production/grain planting area	kg/hm2	Positive
	Agricultural industry Structure adjustment	Gross agricultural output value/gross agricultural, forestry, animal husbandry, and fishery output value	%	Positive
	Labour productivity	Gross agricultural output value/number of employees in the primary industry	100 million yuan/10,000 people	Positive

. Given that agricultural green development is a composite indicator, the entropy method is utilised to integrate it into a single indicator. The specific steps are shown in Formulas (1)–(6) in Section 2.1.

4.2.2. Core Explanatory Variables

The degree of DIF and GF coordination is quantified by the aforementioned coupling coordination model and is hereafter abbreviated as COU.

4.2.3. Threshold Variables

The degree of coordination (COU) refers to the degree of coupling coordination between DIF and GF, as calculated in the previous article.

Regarding the measurement of agricultural industry agglomeration, this study refers to the method proposed by Han et al. [56]. Specifically, the degree of agricultural industry agglomeration is calculated as the quotient of two ratios: one is the ratio of a province’s total output value of agriculture, forestry, animal husbandry, and fishery to the national total output value of this sector, and the other is the ratio of the province’s gross output value to the national gross output value. The specific formula is shown in Formula (12).

$${\mathrm{LQ}}_{\mathrm{it}}={\displaystyle \frac{{\mathrm{Q}}_{\mathrm{it}}/{\displaystyle \sum _{\mathrm{i}=1}^{30}{\mathrm{Q}}_{\mathrm{it}}}}{{\mathrm{G}}_{\mathrm{it}}/{\displaystyle \sum _{\mathrm{i}=1}^{30}{\mathrm{G}}_{\mathrm{it}}}}}$$

(12)

Among them, LQ_it represents the location entropy of province I; Q_it represents the total output value of agriculture, forestry, animal husbandry, and fishery in province i; ${\displaystyle \sum _{\mathrm{i}=1}^{30}{\mathrm{G}}_{\mathrm{it}}}$ represents the total output value of agriculture, forestry, animal husbandry, and fishery in the country; G_it represents the regional GDP of province I; and ${\displaystyle \sum _{\mathrm{i}=1}^{30}{\mathrm{G}}_{\mathrm{it}}}$ represents the national GDP.

4.2.4. Control Variables

Considering other factors influencing the AGD, five control variables were introduced based on the existing literature: grain planting structure, disaster severity, infrastructure development, level of openness, and economic development level [57,58,59]. Grain planting structure (Gr) is measured by the ratio of grain planting area to crop planting area; disaster severity (Dam) is measured by the ratio of crop disaster area to total crop planting area; infrastructure construction (Trans) is measured by the natural logarithm of freight volume; openness (Openness) is measured by the ratio of the value of import and export of goods to regional GDP; and regional economic development level (Lgdp) is measured by the regional gross domestic product and processed in natural logarithm. The variables used in this paper are shown in

Table 5. Variable description.

Variable Type	Variable Name	Variable Symbol	Measurement Description
Explained variable	Green agricultural development	AGD	Level of green agricultural development
Core explanatory variable	Coordination	COU	Measured the degree of coordination between DIF and GF
Control variables	Grain planting structure	Gr	Grain sown area/crop sown area
Control variables	Disaster severity	Dam	Crop disaster-affected area/total crop sown area
	Infrastructure development	Trans	Logarithmic measurement of freight volume
	Degree of openness	Openness	Value of goods imported and exported/regional GDP
	Regional economic development	Lgdp	Logarithmic measurement of regional GDP

.

4.3. Data Sources and Descriptive Statistics

The original data for this paper are sourced from the National Bureau of Statistics, the China Statistical Yearbook, the China Rural Statistical Yearbook, the China Environmental Statistical Yearbook, the China Insurance Yearbook, and provincial statistical yearbooks covering 30 provinces (excluding Hong Kong, Macao, Taiwan, and Tibet). Linear interpolation was employed for the imputation of missing data. Additionally, to prevent potential multicollinearity from exerting adverse effects on the model results, VIF test was performed to evaluate the collinearity among individual variables. The VIF values for all variables remained below 10, indicating that the model design is scientifically sound and practical, and confirming that there is no significant collinearity between the variables. Descriptive statistics and VIF values for each variable are shown in

Table 6. Descriptive statistics and VIF values for each variable.

Variable	Obs	Mean	Std.Dev.	Min	Max	VIF
AGD	300	0.3877	0.1169	0.1577	0.7143	-
COU	300	0.5547	0.0838	0.3456	0.7584	1.22
Igg	300	1.2449	0.6952	0.0454	3.5435	1.09
Gr	300	0.6482	0.1558	0.3547	1.1429	1.15
Dam	300	0.1140	0.1019	0.0014	0.6182	1.07
Trans	300	11.6955	0.8365	9.5804	12.9815	1.07
Openness	300	0.2524	0.2433	0.0076	1.1338	1.37
Lgdp	300	10.0460	0.8593	7.7421	11.8180	1.36

.

5. Empirical Analysis

5.1. Baseline Regression Analysis

Before building a model using panel data, the Hausman test is used to determine whether a fixed-effect model or a random-effect model should be used. The Hausman test statistic is 17.34, and the p-value is 0.000, so the fixed-effect model should be used. To systematically analyse the specific impact of the coordination between DIF and GF on AGD, this research incorporated pertinent data into an empirical framework, bolstered by additional control variables for empirical validation. To enhance the precision and credibility of the estimation outcomes, a trio of models was sequentially employed: initially, a baseline model sans any controls was used, followed by a model incorporating the controls, and it culminated in a model that accounted for both province-level and annual fixed effects. Columns (1) and (3) present the estimates without controls, while column (2) showcases the estimates with the controls in place and regional variations considered. Column (4) reveals the regression results with the controls factored in, accounting for both regional and temporal influences. The findings in

Table 7. Baseline regression.

	(1)	(2)	(3)	(4)
Variables	AGD	AGD	AGD	AGD
COU	1.229 *** (0.0624)	1.181 *** (0.0630)	0.645 *** (0.194)	0.633 *** (0.200)
Control variables	uncontrolled	control	uncontrolled	control
Individual/Time fixed	control	control	control	control
Constant	−0.294 *** (0.0346)	−0.213 * (0.114)	0.00587 (0.0930)	0.0293 (0.149)
N	300	300	300	300
R2	0.834	0.843	0.895	0.897

Note: * denotes p < 0.1, *** denotes p < 0.01.

underscore the model’s robust fit, both with and without the controls, affirming the soundness and scientific validity of the chosen key explanatory and control variables. As revealed by the regression analysis results, the coordination between DIF and GF produces a significant effect of enhancing AGD, and this effect is statistically significant at the 1% level, which in turn verifies Hypothesis 1.

5.2. Robustness Test

5.2.1. Adding Control Variables

A further robustness check incorporates human capital and agricultural fiscal support as additional controls, with variables intrinsically linked to DIF, GF, and AGD. The results presented in Column (1) of

Table 8. Robustness test results.

Variables	(1) Adding Control Variables	(2) Adjusting the Sample Period	(3) Excluding Municipalities
COU	0.587 *** (0.186)	0.930 *** (0.301)	0.662 *** (0.217)
Control variables	control	control	control
Individual/Time fixed	control	control	control
Constant	0.0782 (0.123)	−0.207 (0.161)	−0.0930 (0.119)
N	300	180	260
R2	0.898	0.775	0.914

Note: *** denotes p < 0.01.

reveal a high degree of consistency with the baseline regression. The positive and significant coefficients for COU confirm the robustness of the core findings.

5.2.2. Adjusting the Sample Period

Taking into account the systemic impact of public health emergencies on the macroeconomic environment, the efficiency of financial resource allocation, and agricultural production and management models since 2020, and to avoid interference from exogenous shocks on the causal relationship between core variables, this paper adjusts the sample period to 2014–2019 to isolate the core relationship. Regression results presented in Column (2) of Table 8 indicate a statistically significant coefficient (1% level) for the DIF-GF coordination, thereby validating the robustness of the prior conclusions.

5.2.3. Excluding Municipalities

Compared to the rest of the provinces, the municipalities have significant advantages with respect to policy and economy. A subsequent regression analysis excluded the four provincial-level municipalities (Beijing, Shanghai, Tianjin, and Chongqing), with results shown in Column (3) of Table 8. Regression analyses show that after removing samples of the four municipalities, the coordination between DIF and GF still exerts a significant positive impact on AGD, which validates the robustness of the study’s conclusions.

5.3. Endogenous Treatment

5.3.1. Endogeneity Treatment

As noted earlier, the baseline regression included year fixed effects and province fixed effects—this specification can alleviate endogeneity caused by omitted variables to some extent. However, a potential reverse causality issue persists: agricultural green development (AGD) may promote the coordinated development of DIF and GF, creating endogeneity between the two variables. Regarding coordinated development of DIF and GF, it demonstrates stability: the lagged-period coordination of DIF and GF shows a strong correlation with the current-period coordination, yet lagged-period coordination has no direct link to current AGD. For this reason, the lagged-period coordination of DIF and GF was chosen as an instrumental variable, and the IV-2SLS model was employed to test its impact on AGD. The estimation results from the two-stage least squares (2SLS) regression are reported in Columns (4) and (5) of Table 8. Both stages show coefficients significant at the 1% level. Furthermore, the Kleibergen–Paap rk LM statistic yields a p-value of 0.000, supporting the identification of the instrumental variable. The Cragg–Donald Wald F statistic of 147.199 surpasses the 10% critical value of 16.38, ruling out concerns regarding weak instruments. Therefore, the instrumental variable chosen is appropriate, and after addressing endogeneity, the core findings of this study continue to hold.

5.3.2. Dynamic Panel GMM Estimation Method

Given that AGD may exhibit a continuous nature, i.e., serial correlation, the dynamic panel GMM estimation method is further adopted to test the robustness of the preceding conclusions, with the results presented in Column (3) of

Table 9. Endogenous treatment results.

Variables	2SLS		GMM
	(1) First Stage	(2) Second Stage	AGD
COU	-	0.958 *** (3.51)	0.110 *** (0.035)
L.COU	0.630 *** (12.13)	-	-
L.AGD	-	-	0.960 *** (0.032)
Control variables	control	control	control
Individual/Time fixed	control	control	control
Constant	0.273 *** (7.24)	0.093 (0.47)	−0.011 (0.022)
Kleibergen–Paap rk LM	-	106.495 ***	-
Cragg–Donald Wald F	-	147.199	-
AR(1)	-	-	0.001
AR(2)	-	-	0.779
Hansen	-	-	0.105
N	270	270	270
R2	0.394	0.975	-

Note: *** denotes p < 0.01.

. The AR(2) test does not reject the null hypothesis (with a p-value > 0.1), which implies that there is no second-order serial correlation in the model. At the same time, the p-value of the Hansen test is greater than 0.1, signifying that the instrumental variables are generally effective. The regression results indicate that the coefficient of the synergistic development of DIF and GF is still significantly positive at the 1% significance level. This shows that even when the dynamic continuity of AGD is taken into consideration, the coordination between DIF and GF still has a notable promoting influence on AGD, thus validating the robustness of the conclusions here.

5.4. Heterogeneity Analysis

5.4.1. Based on the Perspective of Major Grain-Producing Areas

Due to disparities in crop varieties, economic development tiers, and social benefits between major grain-producing regions and non-major grain-producing regions, the impact of DIF-GF coordination on agricultural green development (AGD) varies across these two types of areas. Drawing on prior research, this study first examines the heterogeneity between major grain-producing areas and non-major grain-producing areas. The sample is categorised into the two groups in accordance with the Notice of the Ministry of Finance on Issuing the Opinions on Reforming and Improving Several Policy Measures for Comprehensive Agricultural Development, and the impact of DIF-GF coordination on AGD is further analysed. As shown in Columns (1) and (2) of

Table 10. Heterogeneity test based on grain production area, agricultural technology level, and degree of financial exclusion.

	Major Grain-Producing Areas		Agricultural Technology Level		Degree of Financial Exclusion
	(1) Yes	(2) No	(3) High	(4) Low	(3) Strong	(4) Weak
Variables	Agd	Agd	Agd	Agd	Agd	Agd
Cou	0.401 (0.273)	0.801 *** (0.260)	1.189 *** (0.288)	0.561 ** (0.222)	0.522 * (0.292)	0.959 *** (0.262)
Control variables	control	control	control	control	control	control
Individual/Time fixed	control	control	control	control	control	control
Constant	0.140 (0.122)	−0.0338 (0.151)	−0.261 (0.178)	−0.0198 (0.233)	0.242 (0.207)	−0.249 (0.174)
N	130	170	150	150	150	150
R2	0.963	0.869	0.919	0.905	0.898	0.908

Note: * denotes p < 0.1, ** denotes p < 0.05, *** denotes p < 0.01.

, the role of coordination in driving AGD is not statistically significant in major grain-producing regions; in contrast, this role reaches statistical significance at the 1% level in non-major grain-producing regions.

One possible cause is that major grain-producing areas have long undertaken the duty of guaranteeing national food security, and the policy system has prioritised “stabilising output and ensuring supply” as its core goal, with green development goals being in a suboptimal position. In terms of fiscal subsidies and project approvals, the government tends to favour grain-production-increasing technologies, with insufficient policy support for green agricultural projects. Furthermore, the small scale of grain production results in high service costs for GF, making it difficult for the decentralised advantages of DIF to compensate for the lack of policy incentives. The policy environment in non-grain-producing areas prioritises a synergistic balance between economic efficiency and ecological protection. Under the rural revitalisation strategy, such areas are often included in demonstration projects for “ecologically livable” and “industrially prosperous” development. Combining policy incentives for green development indicators with the technological empowerment of DIF is capable of efficiently decreasing transaction outlays for GF and promoting the visibility of synergistic effects.

5.4.2. Based on the Perspective of Agricultural Science and Technology Level

The level of agricultural science and technology may influence the role of financial instruments in AGD. Accordingly, this research employs the number of authorised agricultural patents as an indicator representing the level of agricultural science and technology, and partitions the whole sample into regions with high and low agricultural science and technology levels based on the average value for regression analysis. Findings are presented in Columns (3) and (4) of Table 10: in regions with a high level of agricultural science and technology, the role of DIF-GF coordination in advancing AGD reaches statistical significance at the 1% level; in regions with a low level of agricultural science and technology, by contrast, this role is statistically significant at the 5% level. Regions with advanced agricultural science and technology feature more complete information infrastructure, technical service systems, and talent pools, which facilitate the accurate matching of financial resources with green technologies, lower the likelihood of information asymmetry and capital misallocation, and further foster a virtuous cycle of “technology–finance–green development”. In contrast, regions with underdeveloped agricultural science and technology are restricted by insufficient technology supply and weak innovation transformation capabilities, leading to limited transformation efficiency of financial resources and a relatively diminished promotional effect of DIF-GF coordination on AGD.

5.4.3. From the Perspective of Financial Exclusion Level

Financial resources are unevenly distributed among provinces in China, and the degree of financial exclusion varies among provinces. Therefore, this paper adopts the ratio of financial institution loan balances to regional GDP as a metric to assess financial exclusion. The total sample is divided into two sub-samples with weak financial exclusion and strong financial exclusion based on the average, and a regression is performed. The results are shown in Columns (5) and (6) of Table 10. In regions with strong financial exclusion, the coordination between DIF and GF exerts a significant influence on AGD at the 10% significance level, while in regions with weak financial exclusion, the coordination exerts a significant influence on AGD at the 1% level. In regions with weak financial exclusion, the density of financial institutions is high and the credit market is sound. DIF can rely on its broad reach to lower the threshold for green financial services. In regions with strong exclusion, due to the lack of financial infrastructure and severe credit rationing, it is difficult to overcome physical barriers with the technological advantages of DIF. Green financial products are also too complex to reach demand entities due to insufficient service coverage, resulting in difficulties in transforming the coordination mechanism into actual green output.

6. Further Analysis

To further investigate the nonlinear impact of coordination between DIF and GF on AGD, before performing the threshold regression analysis, Model (11) was employed to examine the threshold effects of the coordination between DIF and GF and agricultural industrial agglomeration. To identify the number and specific values of threshold variables, the bootstrap method was adopted, which included 300 repeated sampling processes and 300 grid point searches. Results presented in

Table 11. Threshold effect test results.

Variables	Number of Thresholds	F Value	p-Value	Threshold Value	10%	5%	1%
Cou	Single-Threshold	30.3	0.0733 *	0.4997	34.3054	39.1868	45.3406
	Double-Threshold	18.54	0.2867	0.6261	25.0506	34.5154	60.2770
Igg	Single-Threshold	42.74	0.0000 ***	2.5312	24.7999	28.4726	31.3758
	Double-Threshold	16.46	0.1200	0.2416	17.4989	43.2048	86.9969

Note: * denotes p < 0.1, *** denotes p < 0.01.

reveal that the single-threshold effect of coordination was validated at the 1% significance level. In contrast, the double-threshold effect lacked statistical significance—confirming that the variable has a single-threshold effect with a threshold value of 0.4997. In addition, the single-threshold effect of agricultural industrial agglomeration passed the test at the 1% significance level. In contrast, its double-threshold effect did not reach statistical significance, indicating that this variable also has a single-threshold effect, with the threshold value being 2.5312.

Further plotting of the 95% confidence interval LR plots for the corresponding threshold values of coordination and agricultural industrial agglomeration (as shown in Figure 3 and Figure 4) shows that it is consistent with the threshold values obtained in Table 11. Therefore, the threshold value results are valid and reliable, which implies that the synergistic effect of DIF and GF on AGD shows nonlinear attributes depending on the degree of coordination and variations in agricultural industry agglomeration.

After verifying that a single-threshold effect exists, we further estimated the panel threshold effect.

Table 12. Threshold regression results.

	(1)COU	(2)Igg
Variables	Agd	Agd
Threshold Value	0.4997	2.5312
Ι(thit ≤ θ)	1.378 ***	1.173 ***
Ι(thit > θ)	1.299 ***	1.284 ***
Control variables	control	control
Individual/Time fixed	control	control
N	300	300
R2	0.8617	0.8516

Note: *** denotes p < 0.01.

shows that the coordination between DIF and GF significantly promotes the green development of agriculture, but this effect exhibits some variation. Specifically, when the coordination level is below the threshold, each unit increase in coordination leads to a 1.378-unit increase in AGD level. Beyond the threshold, the promotion effect weakens slightly, with each unit increase in coordination leading to a significant 1.299-unit increase in AGD level. This may be due to the threshold value of 0.4997 being in the “near-disharmony stage,” indicating that the two types of financial coordination are on the verge of disharmony, characterised by low interactivity and inefficient resource allocation. Coordination improvements tend to produce a strong “from-nothing-to-something” effect. However, when the coupling coordination degree transitions from the near-disharmony stage to coordination, coordination enters the “from-existence-to-excellence” stage. The initial strong effect of gap-filling weakens, requiring more systematic integrated innovation to maintain its support.

The threshold regression results of agricultural industrial agglomeration are shown in Column (2) of Table 12. When the agricultural industrial agglomeration level exceeds the threshold, the effect of coordination between DIF and GF on AGD increases from 1.173 units to 1.284 units, exhibiting a nonlinear characteristic of marginal increase. The reason may be the synergistic effect brought about by industrial agglomeration and the improvement in resource optimisation allocation capabilities. On the one hand, agglomeration leads to the concentration of agricultural production entities in geographical space, reducing the service costs of DIF and enabling green financial products to reach agricultural green projects more accurately, thereby reducing capital mismatch. On the other hand, the scale effect generated by agglomeration accelerates the diffusion and sharing of green technologies, and digital financial instruments can effectively track the efficiency of green inputs, making the incentive mechanism of GF easier to implement.

7. Conclusions and Recommendations

7.1. Conclusions

This paper uses panel data from 30 provinces in China (excluding Hong Kong, Macao, Taiwan, and Tibet) from 2014 to 2023 to evaluate the coordination between DIF and GF and the level of AGD, and further explores the effects of the coordination between the two in promoting AGD. The main findings are presented as follows:

(1) From the viewpoint of time-series features, the coupling coordination status between DIF and GF has undergone an evolution process, transitioning from a state on the brink of disharmony to primary coordination, and then to intermediate coordination. Meanwhile, the coordination mechanism between the two has been continuously maturing. In terms of spatial distribution, the coupling coordination between DIF and GF has generally exhibited a spatial pattern shift from “high in the east and low in the west” to a state of “convergence”. However, there notable regional disparities in the degree of coordination still exist.

(2) The coordination between DIF and GF exerts a notably positive influence on AGD. Moreover, this conclusion remains valid even after robustness tests and endogeneity analysis have been carried out.

(3) The impact of the coordination between DIF and GF on AGD differs significantly across various regions and under diverse conditions. Specifically, the effect is not significant in major grain-producing regions, yet it is significant in non-major grain-producing regions. Additionally, the promotional effect is more potent in regions that boast high levels of agricultural technology and low levels of financial exclusion.

(4) In the realm of agricultural green development, the coordination between DIF and GF displays remarkable nonlinear characteristics. Both the coordination itself and the agglomeration of the agricultural industry have a single-threshold effect.

7.2. Recommendations

Based on the above conclusions, the following suggestions are put forward:

(1) Coordinate regional layout and promote balanced development. Given the temporal and spatial characteristics of coordination between DIF and GF, it is necessary to implement regional policies and strategies that foster regional linkages. The eastern region will continue to leverage its technological and financial advantages to explore innovative GF and digitally inclusive integration models, forming replicable experiences. The central, western, and northeastern regions need to accelerate the construction of financial infrastructure and digital platforms, promote the adoption of financial technology, enhance the accessibility of GF, narrow regional gaps, and foster overall convergence and development across the country.

(2) Deepen financial coordination and strengthen policy support. The above research shows that the coordination between DIF and GF exerts a significant positive influence on AGD. Therefore, the interaction mechanism between the two should be further strengthened. On one hand, financial institutions should be encouraged to embed digital inclusive elements in green financial products, such as using big data, blockchain, and smart contracts to reduce the financing risks of green agriculture; on the other hand, DIF should be guided to focus on green agricultural scenarios and promote the popularisation of green insurance, green credit, and carbon finance products. At the same time, the government needs to enhance the performance assessment and incentive mechanism for green development, channel more financial resources into green agriculture, and establish a synergistic policy framework for coordination and efficiency.

(3) Adhere to the differentiation orientation and strengthen classified measures. In the main grain-producing areas, green production subsidies and financial support should be increased on the premise of ensuring food security, guiding farmers to adopt green technologies and achieving a win-win situation of stable production and green development; in non-main grain-producing areas, we should rely on the combined advantages of DIF and GF to build green agricultural industrial clusters and demonstration areas. In areas with high agricultural science and technology levels, we should accelerate the connection between scientific and technological achievements and financial resources and promote the virtuous cycle of “science, finance and green development”. In areas with low science and technology levels, we should focus on improving basic agricultural research and development, as well as technology promotion capabilities. We should also rely on financial tools to accelerate the introduction and popularisation of green technologies, and gradually enhance the conversion efficiency of financial resources. In areas with high financial exclusion, we should improve rural financial infrastructure and broaden the reach of financial services; in areas with low financial exclusion, we should further facilitate the deep integration of DIF and GF, utilising sound financial markets and rich service channels to foster the innovation and widespread adoption of green financial products.

(4) Optimise the threshold mechanism and enhance the agglomeration effect. The mechanism of action of DIF and GF presents nonlinear characteristics, and a layered promotion strategy should be formulated for different levels of coordination. When coordination is at a low level and on the verge of disharmony, the focus should be on filling institutional and service gaps. This can be achieved through increased policy incentives, optimised allocation of financial resources, improved green credit and insurance product systems, and the expansion of digital inclusive financial services to achieve a strong “from scratch” effect. For regions that have crossed the threshold and entered the initial coordination stage, systemic integration and innovation should be strengthened. This can be achieved by developing agricultural industrial clusters, promoting the precise integration of financial products with green projects, and enhancing information management and technical support to leverage the synergistic effects fully. Agricultural industry agglomeration not only reduces the cost of financial services but also improves the efficiency of capital use, accelerating the diffusion and sharing of green technologies, and strengthening the enforcement of green financial incentives through economies of scale. The government should encourage interregional experience sharing and establish dynamic supervision and assessment mechanisms to secure the continuous and stable unleashing of coordination after crossing the threshold, thus promoting sustainable development of green agriculture.

8. Research Limitations and Future Research Directions

8.1. Research Limitations

Although this paper systematically examines the effect of the coordination between DIF and GF on AGD, it still has certain limitations. First, the research sample covers China’s provincial-level panel data from 2014 to 2023, with a limited time span and spatial scale, making it difficult to fully reveal dynamic characteristics over more extended periods or at the micro level. Second, this paper employs fixed-effect and threshold regression models; however, this method remains insufficient for identifying potential policy shocks, potentially limiting the causal explanatory power of the conclusions.

8.2. Future Research Directions

Future research can be extended in the following aspects: First, extending the research scale to the micro level, and using household surveys or enterprise data to reveal better micro-transmission mechanisms of the coordination between DIF and GF; second, expanding the temporal and spatial scope, combining cross-national comparisons, and exploring the differences in coordination under different institutional environments, financial structures, and agricultural development models; third, strengthening methodological development, and incorporating tools such as spatial econometric models to identify policy effects better, thereby providing more explanatory and generalizable research conclusions on how financial coordination empowers the green development of agriculture.