Financing constraints refer to excessively high external investment costs resulting from underdeveloped financial markets, directly impacting enterprises' ability to secure sufficient capital to meet internal financing needs. This prevents enterprises from achieving optimal financing levels (Farre-Mensa and Ljungqvist, 2016). According to the pecking order theory, enterprises should prioritize internal financing (Myers and Majluf, 1984). However, most enterprises cannot generate sufficiently high profits to meet internal financing demands, necessitating external financing (Denis and Sibilkov, 2010). For China, most companies remain under significant financing constraints (Zhuo et al., 2020). As a prominent example, agribusinesses face unstable internal cash flow due to high volatility in agricultural product prices and low profit margins (averaging 4–6 percentage points lower than manufacturing), making it difficult to fund technological upgrades or production expansion through retained earnings (Zhang and Jin, 2025). Whether financing constraints arise depends on the borrower's perspective—analyzing whether firms lack financing channels or methods—and the lender's perspective—assessing whether firms are worthy investments, including investment risks, profits, and firm value. Thus, resolving corporate financing constraints can be achieved through expanding market financing (Cornaggia et al., 2015), utilizing internet finance (Hau et al., 2024), reducing information asymmetry between parties, and enhancing corporate transparency (Chen et al., 2011; Menkhoff et al., 2012). However, there are many problems in external financing for agriculture-related enterprises.

Duoguang Bei define digital inclusive finance as a model capable of achieving broad financial coverage, reducing service costs, and maintaining long-term sustainability. They emphasize that by reducing barriers to financial service access, stimulating price discovery mechanisms, and enhancing information flow, digital inclusive finance effectively addresses the “last mile” challenge in financial services. Furthermore, tailored to China's national context, the “Peking University Inclusive Finance Index” has been developed. Aligned with the characteristics of finance in the new era, this digital inclusive finance index better captures trends in China's digital inclusive finance landscape (Guo et al., 2020). Using the cash-cash flow sensitivity model (Fazzari et al., 1988), research indicates that inclusive financial innovations can effectively alleviate financing constraints for SMEs (Demirgüç-Kunt et al., 2017). These innovations impact SME financing constraints across multiple dimensions of digital inclusive finance.

Agribusinesses, constrained by factors like land and products, struggle to form group-based risk mitigation strategies. Their low profit margins and slow return on investment further deter investors. The inherent nature of agriculture makes agribusinesses more susceptible to obstacles in risk management and product sales. Even leading enterprises in the agricultural sector may encounter financing difficulties during critical phases of technological upgrades in cultivation and breeding, or new product development (Deng et al., 2025). And because of the constraints of various objective factors, the difficulty of agricultural risk prediction increases, and the information asymmetry is aggravated (Alli, 2025). Therefore, analyzing the financing constraints of agribusinesses requires not only a focus on digital inclusive finance but also guidance and support from public policy.

Policy-based agricultural insurance serves not only as a vital tool for safeguarding corporate interests but also as a significant national support policy. As a crucial means of agricultural risk management, it acts as an important lever for advancing agricultural development (Tang and Zhang, 2020). By transferring risks, it enhances total factor productivity and boosts agricultural economic growth (Karlan et al., 2014). At the same time, agricultural insurance faces significant information asymmetry issues such as adverse selection and moral hazard, which may distort the behavior of market participants in the agricultural insurance sector (Liao et al., 2023). This information asymmetry, as the root cause of adverse selection and moral hazard, increases the likelihood of farmers engaging in passive loss prevention, lax damage mitigation, or insurance fraud (Clarke, 2016). So, when the research focus shifts to agribusinesses, what impact does digital inclusive finance have on them? What mechanisms underlie this impact? Are there regional or equity-based heterogeneities? These are the key issues we should focus on.

Through reviewing existing literature, this study identifies gaps in research on financing constraints for agribusinesses. Most studies lack targeted analysis of agribusinesses—such as their low marketization and high risk levels—and rarely examine the combined effects with policy-based agricultural insurance.

Therefore, this paper utilizes provincial-level digital inclusive finance data from 2011 to 2022 and data on agriculture-related enterprises listed on the NEEQ. Insurance data is calculated based on the China Insurance Yearbook and regional statistical yearbooks. A two-way fixed effects model is constructed to examine the relationship between digital inclusive finance and financing constraints of agriculture-related enterprises. A multi-period DID model is further employed to measure the impact at the inclusive finance policy level. While using Tobin's Q as an mediating variable, the study introduces policy-based agricultural insurance as a moderator variable and conducts heterogeneity analysis across dimensions such as ownership structure, enterprise location, and investment efficiency.

Existing literature indicates that most Chinese enterprises face financing constraints, which can be analyzed from both borrowers' and lenders' perspectives. From the enterprise's standpoint, factors such as equity structure, scale, and ability to obtain investor information influence financing channels (Allen et al., 2005). Limited financing opportunities and high financing costs increase the difficulty of external financing. Combined with low profits and insufficient internal financing among agribusinesses, this creates financing constraints (Berger and Udell, 1998). From the perspective of lenders, imperfect market mechanisms create significant information asymmetry and moral hazard issues (Giné and Karlan, 2014), Under incomplete information, lenders may restrict the number of participants in high-credit-rated markets and require full collateral to curb the pretense of low-risk borrowers. This market mechanism's imperfection makes credit rationing a primary tool for adverse selection management, yet it may also displace high-quality borrowers (Cable and Turner, 2021). Simultaneously, the value and growth potential of investee companies are key considerations for investors. The combined impact of these factors within traditional financial markets—which are often inefficient and lack liquidity—results in low financing efficiency, corporate funding difficulties, and widespread misallocation of resources under conventional financial models.

The continuous advancement of innovative technologies like big data and the internet has propelled inclusive finance to a new level, breaking through numerous constraints of traditional finance and creating a financing platform better suited to serve the “long tail” population (Suri, 2025). This has brought marginal borrowing groups excluded by traditional financial models closer while expanding diversified investment options for household investors (Berg et al., 2020) and enhancing capital circulation efficiency. The emergence of digital inclusive finance alleviates financing constraints by enhancing funding efficiency, reducing financing costs, and broadening financing channels. Furthermore, it significantly eases corporate financing constraints by reducing information asymmetry between transaction parties (Berg et al., 2020) and improving corporate transparency (Tian et al., 2022). The operational characteristics of agribusinesses make it difficult for them to establish credit. Coupled with the high risks inherent in agricultural production, even enterprises that pass credit institution reviews often face substantial interest burdens. These factors collectively exacerbate financing constraints for agribusinesses. Digital inclusive finance holds promise in alleviating these financing constraints to some extent, leading to the first hypothesis of this paper.

Digital inclusive finance not only enhances information transparency and reduces information asymmetry, but also helps enterprises more accurately grasp market operations and policy changes (Obschonka and Audretsch, 2020), influence corporate value, and improve credit efficiency (Huang et al., 2018). As shown in Figure 1, on one hand, innovations in digital inclusive finance have revolutionized traditional financial sectors. They not only expand the coverage of financial services and create broader platforms but also enhance corporate financing channels while effectively reducing information asymmetry between lenders and borrowers. On the other hand, they effectively attract numerous small-scale investors in financial markets, helping to expand the investment demand of the “long-tail” group in stock and bond markets. Simultaneously, by reducing information asymmetry between enterprises and investors, as well as among enterprises themselves, search costs are correspondingly lowered. Enterprises' operational and profit conditions become more directly exposed to investors. From this perspective, transparency in high-value enterprises would attract more investment interest from lenders.

As shown in Figure 2, at this level, increased transparency in less market-oriented enterprises reduces enterprise value (Ma and Meng, 2015), the increase of transparency of enterprises with low market value will make investors lose interest in investment, thereby diminishing investors‘ willingness to invest. Most rural areas are in the low marketization stage, and the marketization degree of agriculture-related enterprises is relatively low (Ding and Fan, 2024). For instance, in the “Western Dairy Companies 'Biological Asset Revaluation Case,” a dairy farming listed company disclosed the valuation model for transitioning from the cost method to fair value measurement for biological assets (cows) as required by accounting standards. The audit report revealed that disease risks caused annual fluctuations in cow fair value reaching 35% (compared to just 5% under the cost method). Banks deemed the valuation too uncertain, reducing the collateral ratio from 60% to 40% and cutting the credit line by 120 million yuan (based on Accounting Standard for Business Enterprises No.5—Biological Assets and the bank's internal risk control documents). Meanwhile, this paper categorizes enterprises based on marketization data into three groups: agriculture-related enterprises, comparison enterprises, and total enterprises.

The annual marketization index is averaged as shown in Figure 3. It is evident that agricultural enterprises (agribusinesses) exhibit inherently low marketization levels, whether compared to other comparison enterprises or the overall enterprise average. Enterprises with higher growth and investment value attract more investor capital. Tobin's Q serves as a crucial metric for assessing corporate value and growth potential (Tobin, 1969). It also measures investment decision-making tendencies, making it a key indicator investors consider when making investment decisions, leading to the second hypothesis.

Nowadays, the climate change is strong, which has a great degree of damage to the stability of agricultural production (Musa, 2025). Agricultural insurance can guide farmers in adjusting production structures, helping enhance agribusinesses' risk resilience and improve agricultural output quality. It serves not only as a vital tool for agricultural risk management but also as a key national support policy for the agricultural sector, enhancing enterprise development stability and increasing risk-bearing capacity (Walters and Preston, 2017). However, agricultural insurance also faces significant issues of information asymmetry, including adverse selection and moral hazard (Le et al., 2020). While these problems are not unique to agricultural insurance, the challenge lies in the higher costs associated with obtaining agricultural production information and conducting oversight, making it difficult for insurers to reduce the probability of adverse selection and moral hazard behaviors. It is highly probable that the existence of agricultural insurance further increases the probability of moral hazard, which was originally reduced by digital inclusive finance. Studies have found that insured farmers tend to reduce field management inputs (e.g., a 12% decrease in fertilization frequency and a 25% increase in delayed pest and disease control rates), and these behaviors significantly increase loan default risks (Rosalia, 2023). After lenders detected such behaviors through satellite remote sensing technology, the agricultural loan disbursements in high-risk areas were reduced by 30%. Moreover, due to its inherent characteristics, agriculture typically faces systemic risks. Once disasters occur, such as typhoons or droughts, they often result in excessive scale and severity, rendering insurance ineffective (Miranda and Glauber, 1997). Under these circumstances, investors may reassess their investment decisions in agribusinesses located in regions with robust agricultural insurance coverage, reducing investments to avoid associated risks. Thus, the implementation of agricultural insurance may partially undermine the role of digital inclusive finance in alleviating corporate financing constraints (as illustrated in Figure 1, leading to the third hypothesis.

From the perspective of ownership structure, China's banking system is predominantly composed of state-owned banks, leading to a greater propensity for banks to favor state-owned enterprises in financing. Consequently, banks are more inclined to provide credit support to state-owned enterprises, while private enterprises often face “financial discrimination” (Allen et al., 2005). Regarding financing through alternative channels, SOEs also frequently receive “preferential treatment” from securities regulators (Luo et al., 2025). Consequently, SOEs face fewer financing constraints than private enterprises. Thus, SOEs are less responsive to digital inclusive finance, making private enterprises—which face greater financing difficulties—the primary beneficiaries of its impact.

Geographically, while some scholars argue that inclusive finance exists to bridge gaps in areas underserved by traditional finance, the Matthew Effect suggests that digital inclusive finance yields more pronounced results in regions where traditional digital finance is already well-developed. Remote areas, hindered by immature infrastructure and technology, now face a wider gap with advanced regions despite greater digital maturity. This exacerbates the financing difficulties of the “long-tail” population to some extent.

Regarding corporate investment efficiency, inefficient corporate investments can be broadly categorized into overinvestment and underinvestment. Depending on cash flow utilization, overinvesting firms often exhibit an “empire-building” mentality, favoring aggressive investment strategies. Underinvesting firms, typically driven by prudence, may forgo investment opportunities due to reluctance to bear excessive risk (Balakrishnan et al., 2016). As vulnerable enterprises, agribusinesses struggle to bear the costs of overly aggressive strategies. With fixed technologies and limited transformation capabilities, they require relatively prudent investment approaches. Investors, predominantly risk-averse, may be more inclined to finance underinvesting enterprises under digital inclusive finance conditions, leading to the fourth hypothesis.

This study references the China Securities Regulatory Commission's Guidelines for Industry Classification of Listed Companies and the Specialized Statistical System for Agriculture-Related Loans. The scope of agriculture-related industries includes agriculture, forestry, animal husbandry, and fisheries, as well as enterprises in the manufacturing sector such as agricultural and sideline food processing, food manufacturing, alcoholic beverages, beverages, and refined tea manufacturing, and wood processing. The study focuses on agriculture-related enterprises from 2011 to 2022, with data sourced from the Guotai An Database (CSMAR), Peking University Digital Inclusive Finance Index, China Insurance Yearbook, and regional statistical yearbooks. Sample selection criteria are as follows: exclusion of enterprises with incomplete data; exclusion of enterprises under ST or PT status during the sample period, as well as those newly listed or listed during the year. Ultimately, 1,585 agribusinesses were selected for observation, with data processing conducted using STATA 17.0.

Concurrently, this study employs a multi-period DID analysis of digital inclusive finance policies implemented in Lankao County, Henan Province (2016), Ningbo City, Zhejiang Province, and Ningde City and Longyan City, Fujian Province (both 2019). The policy selection is appropriate: first, the policies were implemented in staggered phases (2016 vs. 2019), meeting the requirements for a multi-period DID design. Second, the policies cover representative regions at different development levels—from a national pilot zone (Lankao) to eastern coastal areas (Ningbo) and revolutionary base areas (Ningde, Longyan)—combining regional diversity with relevance to agriculture-related enterprises. Finally, while all policies centered on digital inclusive finance, their specific focuses—such as Lankao's “county-level poverty alleviation,” Ningbo's “micro and small enterprises,” and Fujian's “poverty eradication”—closely aligned with the research theme. This alignment enables precise identification of their heterogeneous impacts on financing constraints for agribusinesses, providing a reliable exogenous policy shock for causal inference. Municipal-level digital inclusive finance policies (Lankao County, Henan in 2016; Ningbo City, Zhejiang and Ningde City/Longyan City, Fujian in 2019) represent provincial-level impacts. This is primarily because these pilot zones required provincial government approval and support, with their outcomes directly serving provincial inclusive finance development goals.

Where subscripts i and t denote individual and year, respectively, and d represents different agricultural insurance indicators; FC_{i, t} represents the financing constraint for firm i in year t; DID_{i, t} is the product of the treatment group dummy variable (treatment group = 1, control group = 0) and the policy timing dummy variable (post-policy = 1, pre-policy = 0);α₀, b₀, c₀, β₀, δ₀, and γ₁are the intercept terms; Index_{i, t} is the inclusive finance index for firm i in year t; Tobin′s Qi,t denotes the Tobin's Q ratio for firm i in year t; Controls_{i, t} represents the above series of control variables; Insurance_{i, t, d} denotes a series of agricultural insurance indicators; β₁ Tj_{i, t, d} represents the product of different Insurance indicators and Index; f_i, y_t respectively represent individual fixed effects and time fixed effects; ε₁, ε₂, ε₃, ε₄, ε₅, ε₆ are random error terms.

The benchmark model employed in this study is a two-way fixed effects model, simultaneously accounting for time and firm-level fixed effects. The results are presented in Table 1. Column (1) reveals that digital inclusive finance significantly alleviates financing constraints for enterprises. The coefficient for the overall index (Index) is −0.002 and is statistically significant at the 1% level, indicating that agribusinesses in cities with higher levels of digital inclusive finance development face relatively fewer financing constraints. In other words, digital inclusive finance helps alleviate corporate financing constraints. Further analysis of the digital inclusive finance index in columns (2), (3), and (4) reveals that, compared to the depth and breadth of digital inclusive finance, the level of convenience provided to customers is the indicator most likely to influence corporate financing constraints. Its coefficient is −0.001, significant at the 1% level. This may be because, for agribusinesses—the “long-tail” segment of the investment market—the depth and breadth of digital adoption are less effective than more convenient usage methods in reducing investors' costs, boosting their interest, and thereby easing financing constraints for these enterprises. Column 5 shows that after controlling for individual and time fixed effects, the average effect coefficient of policy implementation (DID) is −0.041, significantly negative. This indicates that, on average, the outcome variable Y for individuals exposed to the policy treatment decreased by 0.041 units compared to unexposed individuals, thus confirming Hypothesis 1.

As shown in Table 5's empirical results, digital inclusive finance demonstrates more pronounced effects for enterprises facing underinvestment. Typically, agribusinesses carry higher inherent risks, leading investors to favor underfunded enterprises that tend to adopt relatively conservative operational approaches. A key objective of digital inclusive finance is to bridge resource gaps for the long-tail population, thereby increasing investment opportunities for small household investors traditionally excluded from formal financial systems. This demographic prioritizes stable and sustainable investment products, meaning overly aggressive enterprises may not be significantly impacted by digital inclusive finance.

Secondly, state-owned enterprises (SOEs), benefiting from ample capital chains and facing relatively low financing pressures, may experience less pronounced impacts from financial model reforms. While private enterprises prioritize profit maximization, SOEs shoulder social responsibilities such as maintaining employment stability, promoting public welfare, and ensuring food security. Consequently, SOEs often receive greater resources and preferential treatment from central banks and financial institutions. Private enterprises, particularly those in agriculture, inherently possess weaker corporate characteristics. Lacking financial market favor while needing external financing, these entities are precisely the focus of digital inclusive finance policies.

Finally, as shown in Table 5, digital inclusive finance proves more effective in regions with stronger traditional economic development. Agriculture-related enterprises, which already enjoy easier access to financing, may secure even greater capital through digitalization. We categorize enterprises by region based on GDP, such as the eastern region encompassing cities like Liaoning Province and Beijing, which collectively account for ~60% of China's GDP. Heterogeneity analysis by region indicates that digital inclusive finance has the most pronounced effect on enterprises in the more developed eastern region, with a coefficient of −0.002, significant at the 1% level. The development of digital inclusive finance exhibits a “Matthew effect.” agribusinesses in less economically developed regions may struggle to compete for financing with other industries due to sector-specific characteristics, even when digital inclusive finance is implemented. Thus, Hypothesis 4 is confirmed.

The robustness analysis in this paper is broadly categorized into three approaches: (1) altering the dependent variable; (2) modifying the fixed-effects model by upgrading from individual-level to industry-level fixed effects; (3) excluding enterprises with relatively limited agricultural involvement, such as those in food processing, catering, textiles, and apparel. As shown in Table 6, Column (1) demonstrates that after replacing the dependent variable FC with another financing constraint indicator SA, the mitigating effect of digital inclusive finance persists, with a coefficient of −0.000 that is significant at the 5% level. Column (2) indicates that to avoid industry bias inherent in individual-level fixed effects, industry-level fixed effects were employed, revealing that the results remain robust and significant at the 1% level. Column (3), after refining the sample, yields a coefficient of −0.003 that is significant at the 1% level. This not only demonstrates the robustness of the findings but also underscores the critical relevance of this study's focus on agricultural analysis.