Research on the Mechanisms and Effects of Digital Finance on High-Quality Development of Foreign Trade

Abstract

The deep integration of digital finance and foreign trade can drive the steady and rapid development of a high-quality economy. This article conducts an empirical analysis of the mechanisms and impact effects of digital finance on high-quality development of foreign trade (HDFT) using panel data from 30 provinces in China (excluding Tibet, Hong Kong, Macao, and Taiwan) from 2011 to 2021. The research demonstrates that digital finance significantly promotes HDFT and facilitates the achievement of sustainable development goals while enhancing financial development levels and infrastructure to advance the quality of foreign trade further. Heterogeneity test results reveal that the effectiveness of digital finance varies across regions with different financial development levels. The study also identifies threshold effects in the relationship between digital finance and HDFT, with investment environments and internet penetration rates dynamically reinforcing the sustainability of digital finance in driving HDFT.

1. Introduction

Since the beginning of the reform and opening-up, China has leveraged its demographic dividend and resource endowment advantages to deeply integrate into the global value chain, achieving rapid expansion of its foreign trade scale. However, in recent years, the global economy has experienced a slowdown, and trade protectionism has been on the rise, posing severe challenges to the traditional growth model of foreign trade. According to data from the World Trade Organization (WTO), global merchandise trade growth significantly slowed from 2020 to 2023, with developed economies promoting “nearshoring” and “friend-shoring,” leading to a restructuring of global supply chains. Against this backdrop, China’s foreign trade development model urgently needs to shift from scale expansion to quality improvement to adapt to the new landscape of international competition. The introduction of the new development philosophy has provided strategic guidance for China’s high-quality development of foreign trade (HDFT), emphasizing innovation-driven growth, green and low-carbon development, and digital transformation to enhance international competitiveness. The 2024 Central Economic Work Conference further emphasized the need to expand high-level opening-up, foster new business formats such as green trade and digital trade, and inject new momentum into the sustainable development of foreign trade.

Meanwhile, the rise of digital finance offers new opportunities for global economic governance and trade model innovation. As a new business model that deeply integrates digital technology with finance, digital finance is becoming a vital engine for promoting high-quality economic development. The 2023 Central Financial Work Conference, for the first time, listed digital finance as one of the five core areas for building a strong financial sector, highlighting its synergy with green finance and technology finance and providing new impetus for the low-carbon transition of the real economy. By lowering transaction costs, optimizing resource allocation, and improving service efficiency, digital finance not only compensates for the shortcomings of traditional finance but also provides critical support for the digital transformation and intelligent upgrading of foreign trade enterprises.

However, although the role of digital finance in promoting industrial upgrading and rural revitalization has been widely discussed [1,2,3,4,5,6], its impact mechanism on HDFT still lacks systematic research. The existing literature has accumulated considerable findings on the role of the digital economy [7,8,9,10] and other financial factors [11] in trade development. With the rapid development of digital finance, its role in promoting trade has become increasingly prominent. Relevant studies show that digital finance, through channels such as supply chain finance platforms, digital payments, and intelligent services, significantly improves the financing accessibility and risk resistance capacity of foreign trade enterprises [12]. Moreover, leveraging digitalization and artificial intelligence technologies effectively reduces trade costs and enhances the quality and competitiveness of export products [13,14,15]. In addition, digital finance also plays a positive role in driving green transformation and sustainable trade [16] and further promotes export growth by alleviating financing constraints and stimulating enterprise innovation [17,18]. However, these studies primarily focus on the enterprise or platform level, lacking systematic analysis of the overall impact mechanisms and heterogeneity at regional, industry, or national levels.

Based on existing research, scholars have adopted diverse methodologies to measure the level of HDFT. The first approach expands the dimension of trade sustainability by incorporating indicators such as trade scale, trade performance, and competitiveness [19], or by adding metrics like trade development foundations, trade openness, and trade innovation capabilities [20]. The second approach constructs a comprehensive evaluation system based on the new development philosophy, encompassing five dimensions: innovation, coordination, greenness, openness, and sharing [21]. Regarding the influencing factors of HDFT, studies indicate that key determinants include industrial structure [22], financial development [23], and regional trade agreements [24]. However, the role of digital finance has not yet received sufficient attention. Furthermore, existing studies are mostly confined to linear analyses, neglecting the heterogeneous effects and nonlinear constraints that may arise from factors such as regional financial development levels and internet penetration rates, which can potentially generate threshold effects.

Based on this, this paper innovatively approaches HDFT from the perspective of digital finance, utilizing Chinese provincial panel data from 2011 to 2021 to systematically examine the direct effects, heterogeneous impacts, and mechanisms of digital finance on HDFT. The potential marginal contributions are as follows: (1) constructing a theoretical framework linking digital finance and HDFT to reveal their intrinsic mechanisms; (2) introducing financial development and infrastructure as mediating variables to analyze the transmission path through which digital finance empowers improvements in trade quality; (3) employing a threshold model to explore the nonlinear moderating effects of investment environment and internet penetration rate, thereby providing differentiated policy implications. This study not only broadens the theoretical horizon at the intersection of finance and trade but also offers practical guidance for China’s transition from a “major trading nation” to a “strong trading nation.”

Against the backdrop of global value chain restructuring, this research carries significant international implications: for developing countries, the innovative application of digital finance offers breakthrough solutions to longstanding structural challenges such as weak financial infrastructure and underdeveloped trade financing channels; for developed countries, the findings of this study provide valuable insights into optimizing digital trade governance systems. Moreover, this study echoes the United Nations 2030 Sustainable Development Goals (SDGs) and contributes Chinese perspectives to trade policy innovation under the global climate governance framework.

2. Theoretical Analysis and Research Hypotheses

2.1. The Linear Impact of Digital Finance on HDFT

Digital finance promotes HDFT by alleviating financial exclusion. Traditional trade models rely on conventional financial institutions, which struggle to meet the needs of China’s vast number of “small, scattered, and weak” financial demand entities. With limited financing channels for enterprises, digital finance addresses financing constraints through innovative financial models. Leveraging big data and artificial intelligence, these models establish efficient credit risk assessment systems, conduct comprehensive analyses of corporate credit risks, and provide tailored financial services to more foreign trade enterprises. This lowers financing barriers and diversifies funding avenues.

Digital finance boosts HDFT by optimizing resource allocation. Its robust data processing capabilities integrate domestic and international trade information, advancing the informatization of foreign trade processes. This ensures resources flow precisely to areas of greatest need and productivity, reducing mismatches caused by information asymmetry. As a result, foreign trade transactions become smoother and more efficient, with significantly reduced information costs. Additionally, digital finance leverages technological innovations like internet and mobile payments to transcend temporal and geographical limitations, streamlining payment procedures, shortening transaction times, and substantially lowering transaction fees.

By utilizing big data analytics to assess consumer behavior and purchasing power, digital finance tailors financial services to client needs, enabling precise alignment between products and market demand, thereby boosting trade efficiency. Furthermore, the growth of digital finance drives the integration of traditional industries with digital technologies, fostering new export categories and enriching trade markets. Through digital platforms such as the internet and mobile data, digital finance breaks spatial constraints on China’s foreign trade, spurring novel trade activities. It provides robust support for emerging trade formats like cross-border e-commerce and market procurement trade. By establishing online international trade collaboration platforms, digital finance expands the scale of foreign trade, facilitates innovation and upgrading of trade models, optimizes trade structures, and ultimately catalyzes transformative shifts in trade paradigms.

Based on this analysis, we propose:

Hypothesis H1.

Digital finance positively promotes HDFT.

2.2. Mechanisms of Digital Finance for HDFT

2.2.1. Level of Financial Development

Digital finance has spurred the emergence of numerous new financial institutions, driving revolutionary transformations in the business scope, product offerings, and service models of traditional finance. This evolution has partially mitigated financial exclusion, enhanced the efficiency of financial resource allocation, and improved the diversity, inclusiveness, and accessibility of financial products and services. By addressing structural deficiencies and supply shortages in traditional finance, digital finance compensates for its limitations [25]. More advanced organizational processes in financial institutions and a more robust financial system have accelerated the application of finance in foreign trade, invigorated financial markets, optimized financial structures, and further elevated overall financial development. As financial development progresses, it effectively alleviates long-standing issues such as information asymmetry, homogeneous products, and cumbersome credit approval procedures in financial markets. This lowers the barriers to financing constraints, expands credit scales, accelerates capital circulation, and facilitates the rational allocation of financial resources. Collectively, these improvements enhance trade efficiency and promote HDFT.

Based on this analysis, we propose:

Hypothesis H2.

Digital finance enhances the level of HDFT by improving financial development.

2.2.2. Level of Infrastructure

Digital finance directly supports regional infrastructure development through its capital while also enhancing government fiscal revenue, thereby increasing investments in local transportation, education, technology, and other infrastructure projects. This fosters comprehensive infrastructure advancement. Furthermore, digital finance effectively integrates the characteristics of information infrastructure and traditional finance. On one hand, infrastructure provides a data network-sharing platform for digital finance, strengthening data security and reliability while enhancing mutual trust between trading parties—a core foundation for trade-driven growth. This helps create the necessary facilitation for HDFT. On the other hand, improved infrastructure enhances the accessibility and controllability of financial services, reduces trade costs, and amplifies the inclusive effects of digital finance. By offering more comprehensive and equitable financial services for trade activities, digital finance facilitates the growth of emerging trade formats like e-commerce, ultimately elevating the level of HDFT.

Based on this analysis, we propose:

Hypothesis H3.

Digital finance enhances the level of HDFT by improving infrastructure.

2.3. The Non-Linear Impact of Digital Finance on HDFT and Its Constraint Mechanisms

2.3.1. Non-Linear Impact

Digital finance inherently follows a nonlinear developmental trajectory, exhibiting distinct characteristics at different stages, which may consequently exert nonlinear impacts on HDFT. In the initial phase of digital finance development, substantial financial, temporal, and human resources are required to establish network platforms and infrastructure. During this stage, foreign trade entities face high costs in accessing information resources and financial services, while digital finance generates limited economic benefits, thereby constraining its capacity to drive high-quality foreign trade growth. As digital finance matures, the costs of information acquisition and financial product development decline, leading to rising marginal benefits. This progress creates a demonstration effect, attracting broader participation from financial institutions. Consequently, information asymmetry diminishes, transparency improves, and the costs of developing personalized financial services and products decrease, while their speed and quality of innovation increase. The breadth and depth of financial services expand, facilitating the upgrading of trade structures and more effectively advancing HDFT.

Based on this analysis, we propose:

Hypothesis H4.

Digital finance exerts a nonlinear impact on HDFT.

2.3.2. Non-Linear Constraint Mechanisms

Both investment environments and internet penetration rates impose nonlinear constraints on the impact of digital finance on HDFT. Regarding the investment environment dimension, a favorable investment environment can stimulate market vitality, enhance public and private sector investment enthusiasm, facilitate resource mobility and allocation efficiency, and attract greater capital inflows into infrastructure development—particularly digital infrastructure. Increased investment scale and amplified investment effects enable digital finance to more effectively elevate the level of HDFT. Conversely, low internet penetration rates hinder information exchange between foreign trade enterprises and overseas markets, restricting the development and service capacity of digital finance. In regions with limited internet access, reliance on intermediaries prolongs trade cycles, reduces profit margins, and introduces trade risks, thereby constraining HDFT. With the advancement of new infrastructure, improved internet penetration facilitates timely and efficient information exchange, unleashes digital dividends, and fosters innovative trade models such as e-commerce. This optimizes trade structures, enhances the competitiveness of foreign trade enterprises, and amplifies the efficacy of digital finance in driving HDFT.

Based on this analysis, we propose:

Hypothesis H5.

The impact of digital finance on HDFT is nonlinearly constrained by investment environments and internet penetration rates.

3. Research Design

3.1. Variable Selection

3.1.1. Explained Variable

The explained variable is the level of high-quality foreign trade development (HDFT). Aligning with the characteristics of current foreign trade development and the key objectives outlined for the High-quality Development of Foreign Trade during the 14th Five-Year Plan period, this study integrates the new development philosophy of “innovation, coordination, greenness, openness, and sharing” into the evaluation criteria. Considering data availability, we construct a comprehensive indicator system for HDFT by referencing the indicator selection framework of Cao et al. [26,27,28,29]. Five primary dimensions are selected: openness, coordination, sustainability, innovation, and international competitiveness. Corresponding secondary indicators include the foreign trade dependency ratio, trade balance index, value-added contribution rate of trade, R&D investment intensity, and trade competitiveness (TC) index. Following the approach of Wei et al. [30,31], the entropy-weighted TOPSIS model is employed to calculate indicator weights (see

Table 1. Indicators for high-quality foreign trade development.

Primary Indicator	Secondary Indicator	Measurement	Weight	Direction
Openness	Foreign Trade Dependency Ratio	Total Import & Export Volume/GDP	36.71%	Positive
Coordination	Trade Balance Index	Import Volume/Export Volume	34.56%	Positive
Sustainability	Value-Added Contribution Rate of Trade	Export Value-Added/GDP	4.91%	Positive
Innovation	R&D Investment Intensity	R&D Expenditure/Tertiary Industry GDP	15.13%	Positive
International Competitiveness	Trade Competitiveness (TC) Index	(Exports − Imports)/(Exports + Imports)	8.70%	Positive

) and derive a composite score for HDFT level across provinces.

3.1.2. Core Explanatory Variable

The core explanatory variable is the digital finance development level (lnDif), measured using the Digital Inclusive Finance Index compiled by Peking University’s Digital Finance Research Center [32]. This index covers 31 provincial-level regions in China and includes three sub-dimensions: coverage breadth (lnCov), usage depth (lnUse), and digitalization level (lnDig). Adopting this index ensures methodological rigor and stability while enabling analysis of heterogeneous effects across sub-dimensions. All indices are logarithmically transformed to eliminate scale effects.

3.1.3. Control Variables

To mitigate potential endogeneity issues caused by omitted variables and accurately reflect other factors influencing HDFT, this study draws on existing research to select human capital level (Hr), foreign direct investment (Fdi), technological innovation (lnTec), and industrial structure (Str) as control variables. Human capital is measured by the number of college graduates divided by the total population. Foreign direct investment is quantified using the ratio of actual foreign direct investment inflows to GDP. Technological innovation is proxied by the number of invention patent grants per 10,000 population, with logarithmic transformation applied to mitigate heteroscedasticity effects. Industrial structure is calculated as the proportion of tertiary industry value-added in GDP.

3.1.4. Mediating Variables

The financial development level (Fin) is measured by the ratio of the year-end balance of financial institution loans to GDP. The infrastructure level (Inf) is characterized by road area per capita and subjected to logarithmic transformation. Generally, regions with greater road area per capita exhibit higher infrastructure development levels.

3.1.5. Threshold Variables

Threshold variables include digital finance (lnDif), consistent with the core explanatory variable; investment environment (Inv), quantified by the proportion of fixed asset investment as the ratio of fixed asset investment to regional GDP; and internet penetration rate (Int), measured by the ratio of internet broadband subscribers to the registered population in each province.

The variable definition table is shown in

Table 2. Variable definitions.

Variable Type	Variable	Description	Symbol
Explained Variable	High-Quality Foreign Trade	Measured via the entropy-weighted method	HDFT
Explanatory Variables	Digital Finance	Logarithm of Digital Inclusive Finance Index	lnDif
	Coverage Breadth of Digital Finance	Logarithm of Coverage Breadth Index	lnCov
	Usage Depth of Digital Finance	Logarithm of Usage Depth Index	lnUse
	Digitalization Level of Digital Finance	Logarithm of Digitalization Level Index	lnDig
Mediating Variables	Financial Development Level	Ratio of financial institutions’ loan balance to GDP	Fin
	Infrastructure Level	Logarithm of road area per capita	Inf
Control Variables	Human Capital	Number of undergraduate/college graduates per 100 people	Hr
	Foreign Direct Investment	Ratio of FDI to GDP (scaled by 10)	Fdi
	Technological Innovation	Logarithm of invention patents granted per 10,000 people	lnTec
	Industrial Structure	Share of tertiary industry value-added in GDP	Str
Threshold Variables	Digital Finance	Logarithm of Digital Inclusive Finance Index	lnDif
	Investment Environment	Ratio of fixed asset investment to GDP	Inv
	Internet Penetration Rate	Ratio of broadband users to registered population	Int

below.

3.2. Model Construction

To explore the linear impacts, mediating mechanisms, and potential nonlinear impacts of digital finance on HDFT, the following econometric models are constructed.

3.2.1. Baseline Regression Model

To test whether digital finance positively promotes HDFT, the baseline model is specified as

$${HDFT}_{it}={\alpha }_0+{\alpha }_1{lnDif}_{it}+{\alpha }_2{X}_{it}+{\mu }_i+{\rho }_t+{\epsilon }_{it}$$

(1)

where ${HDFT}_{it}$: explained variable, representing the high-quality foreign trade development level of region i in year t; ${lnDif}_{it}$: core explanatory variable, the logarithm of the digital finance index. $X_{it}$: control variables, including human capital, foreign direct investment, technological innovation, and industrial structure. ${\mu }_i$ and ${\rho }_t$: region and time fixed effects. ${\epsilon }_{it}$: random error term.

3.2.2. Mediation Effect Model

To further explore the mechanism through which digital finance promotes HDFT, this paper builds on the theoretical analysis and considers that digital finance may indirectly influence HDFT through mediating variables such as the level of financial development and infrastructure development. This mechanism aligns with the multi-layered pathways of digital finance in facilitating information flows, reducing transaction costs, and diversifying financial products, and it is also consistent with the theoretical hypotheses proposed earlier. Based on this, the paper adopts the three-step mediation analysis method proposed by Wen Zhonglin et al. [33], incorporating financial development level and infrastructure level into the empirical model. Taking the financial development level as an example, the mechanism model is constructed as follows:

$${Fin}_{it}={\beta }_0+{\beta }_1{lnDif}_{it}+{\beta }_2{X}_{it}+{\mu }_i+{\rho }_t+{\epsilon }_{it}$$

(2)

$${HDFT}_{it}={\gamma }_0+{\gamma }_1{lnDif}_{it}+{\gamma }_2{Fin}_{it}+{\gamma }_3{X}_{it}+{\mu }_i+{\rho }_t+{\epsilon }_{it}$$

(3)

Here, ${Fin}_{it}$ is a mediator variable that indicates the level of financial development in region i in period t. ${\beta }_1$ captures the effect of digital finance on financial development, and ${\gamma }_2$ reflects the mediating effect of financial development on HDFT.

3.2.3. Threshold Effect Model

Furthermore, the impact of digital finance on HDFT is not necessarily linear, which is closely related to the stage-specific characteristics and regional disparities inherent in the development of digital finance itself. Simultaneously, factors such as the investment environment and internet penetration rate may also exert a threshold effect on the influence of digital finance at different stages of development. To systematically identify this potential non-linear relationship, this study extends the panel threshold model proposed by Hansen (1999) [34], incorporating restrictions on the two threshold variables of investment environment and internet penetration rate. The specific model is set as follows:

$$\begin{gathered} {HDFT}_{it}={\kappa }_0+{\kappa }_1{lnDif}_{it}\times {\rm I}\left({Adj}_{it}\le \theta \right)+{\kappa }_2{lnDif}_{it}\times {\rm I} \\ \left({Adj}_{it}>\theta \right)+{\alpha }_j{X}_{it}+{\mu }_i+{\rho }_t+{\epsilon }_{it} \end{gathered}$$

(4)

where ${\rm I}\left(·\right)$: indicator function. ${Adj}_{it}$: threshold variables (digital finance, investment environment, internet penetration). $\theta$: estimated threshold value. The value is 1 when the condition in parentheses is satisfied, and 0 otherwise; Equation (4) considers only a single threshold effect and can be extended as needed in practical applications.

In addition, the specific threshold values and their significance were determined through repeated sampling using the Bootstrap method to ensure the robustness and scientific validity of the model. A detailed explanation of the threshold tests and the selection of thresholds is provided in the empirical results section, thus ensuring the rigor and completeness of the model.

3.3. Data Description

The study uses panel data from 30 Chinese provinces (excluding Tibet, Hong Kong, Macau, and Taiwan) from 2011 to 2021 (330 observations). The digital finance data are sourced from the Peking University Digital Inclusive Finance Index (https://idf.pku.edu.cn/yjcg/zsbg/513800.htm, accessed on 20 September 2024), while other variables are primarily derived from the China Statistical Yearbook (https://www.stats.gov.cn/sj/ndsj/, accessed on 20 September 2024), the official website of the Development Research Center of the State Council (https://www.drcnet.com.cn/, accessed on 20 September 2024), and the National Bureau of Statistics (https://www.stats.gov.cn/sj/, accessed on 20 September 2024). Given the small proportion of missing data, this study uses linear interpolation to input the missing values in order to minimize its impact on the empirical analysis. Empirical analysis was conducted using Stata 18.0, with descriptive statistics for key variables summarized in

Table 3. Descriptive statistics.

Variable Type	Variable	N	Mean	p50	SD	Min	Max
Explained Variable	HDFT	330.000	0.249	0.203	0.110	0.106	0.745
Explanatory variables	lnDif	330.000	5.283	5.474	0.669	2.909	6.129
	lnCov	330.000	5.149	5.374	0.817	0.673	6.072
	lnUsa	330.000	5.266	5.433	0.652	1.911	6.236
	lnDig	330.000	5.556	5.809	0.681	2.026	6.136
Mediating Variables	Fin	330.000	1.435	1.361	0.454	0.650	2.759
	Inf	330.000	1.657	1.700	0.415	0.095	2.573
Control Variables	Hr	330.000	0.518	0.506	0.145	0.205	1.045
	Fdi	330.000	0.187	0.163	0.179	0.001	1.210
	lnTec	330.000	−0.951	−0.907	1.435	−4.962	2.331
	Str	330.000	0.495	0.487	0.089	0.327	0.837
Threshold Variables	Inv	330.000	0.832	0.855	0.287	0.201	1.597
	Int	330.000	0.225	0.215	0.103	0.052	0.479

.

4. Empirical Results Analysis

4.1. Baseline Regression Results

To ensure model accuracy and stability, we conducted regressions by sequentially adding control variables. Results are presented in

Table 4. Benchmark regression results of digital finance on HDFT.

	(1)	(2)	(3)	(4)	(5)
	HDFT	HDFT	HDFT	HDFT	HDFT
lnDif	0.114 ***	0.0938 ***	0.0780 ***	0.0755 ***	0.0766 ***
	(0.0172)	(0.0178)	(0.0176)	(0.0176)	(0.0173)
Hr		0.183 ***	0.224 ***	0.231 ***	0.238 ***
		(0.0526)	(0.0517)	(0.0516)	(0.0508)
Fdi			0.0735 ***	0.0645 ***	0.0701 ***
			(0.0164)	(0.0169)	(0.0167)
lnTec				0.0216 **	0.0221 **
				(0.0105)	(0.0103)
Str					0.289 ***
					(0.0915)
_cons	−0.133 **	−0.145 **	−0.124 **	−0.0769	−0.209 ***
	(0.0626)	(0.0615)	(0.0597)	(0.0637)	(0.0753)
N	330	330	330	330	330
R-sq	0.347	0.373	0.414	0.423	0.442

Notes: Results displayed in the table were obtained using STATA 17. ** and ***, respectively, represent p < 0.05, and p < 0.01; standard errors are in parentheses.

. Column (1) shows that without controls, the coefficient of digital finance (lnDif) on HDFT is 0.114, significant at the 1% level, indicating that digital finance development facilitates HDFT. Columns (2)–(5) progressively incorporate control variables—human capital (Hr), foreign direct investment (Fdi), technological innovation (lnTec), and industrial structure (Str). While the inclusion of controls dilutes the explanatory power of the effect (reducing the coefficient), the significance of the core explanatory variable (lnDif) remains stable, consistently passing the 1% significance test. As shown in Column (5), holding other variables constant, the coefficient of digital finance is 0.0766, implying that a 1% increase in digital finance corresponds to a 0.0766% rise in HDFT. These findings confirm that digital finance actively promotes HDFT, suggesting policymakers should further explore and harness its potential to advance trade transformation. Hypothesis H1 is thus validated.

4.2. Heterogeneity Analysis

4.2.1. Structural Heterogeneity Analysis

This study incorporates three sub-dimensions of digital finance—coverage breadth, usage depth, and digitalization level of inclusive finance—into the research framework to conduct a deeper analysis and investigate the specific impacts of digital finance on HDFT across these dimensions. The regression results, presented in

Table 5. Analysis of structural heterogeneity.

	(1)	(2)	(3)	(4)
	HDFT	HDFT	H DFT	HDFT
lnDif	0.0766 ***
	(0.0173)
lnCov		0.0258 ***
		(0.00693)
lnUsa			0.0249 **
			(0.0126)
lnDig				−0.0256 **
				(0.0112)
Hr	0.238 ***	0.263 ***	0.299 ***	0.284 ***
	(0.0508)	(0.0502)	(0.0497)	(0.0508)
Fdi	0.0701 ***	0.0734 ***	0.0799 ***	0.0802 ***
	(0.0167)	(0.0168)	(0.0170)	(0.0169)
lnTec	0.0221 **	0.0243 **	0.0260 **	0.0279 ***
	(0.0103)	(0.0104)	(0.0106)	(0.0106)
Str	0.289 ***	0.306 ***	0.298 ***	0.282 ***
	(0.0915)	(0.0926)	(0.0943)	(0.0937)
_cons	−0.209 ***	−0.0302	−0.0515	0.154 **
	(0.0753)	(0.0542)	(0.0691)	(0.0721)
N	330	330	330	330
R-sq	0.442	0.431	0.412	0.414

Notes: Results displayed in the table were obtained using STATA 17. ** and ***, respectively, represent p < 0.05, and p < 0.01; standard errors are in parentheses.

, show that the regression coefficients for coverage breadth and usage depth are 0.0258 and 0.0249, respectively, both statistically significant at the 1% and 5% levels. This indicates that both coverage breadth and usage depth of digital finance positively contribute to HDFT. Specifically, the expansion of digital financial coverage enhances regional financial accessibility, thereby fostering a more supportive financial environment for HDFT. Concurrently, the deepening of usage breadth enriches diversified financial services—such as payments, credit, and investments—leading to significant improvements in total usage, activity levels, and per capita transaction volumes, all of which further advance HDFT. Notably, coverage breadth exhibits the most pronounced facilitating effect among the three sub-dimensions. Moreover, from the regression coefficients, the level of digitalization is significantly negative at the 5% level, indicating that, to some extent, digitalization has a restraining effect on the HDFT. This may be due to the fact that the overall level of digitalization is still relatively low, and the convenience and efficiency of digital finance have not yet been fully realized, resulting in limited resource allocation efficiency, thereby imposing certain constraints on HDFT. This also suggests that the construction of a supportive digital environment has not yet effectively facilitated HDFT. In particular, in some relatively underdeveloped regions, institutional frameworks and policy support remain insufficient, making it difficult for enterprises to promptly access digital financial services. Consequently, the efficiency of financial resource allocation and utilization is low, leading to some degree of resource idleness and waste. Furthermore, digital finance inherently carries financial attributes, and the associated risks have certain contagion and spillover effects. In some regions, risk prevention and regulatory capacities are still inadequate, which can lead to the spread of financial risks, increasing operational uncertainty for enterprises and thus negatively affecting their production, exports, and HDFT.

4.2.2. Regional Heterogeneity Analysis

Significant economic development disparities and uneven digital finance advancement across different regions in China collectively result in varying impacts of digital finance on HDFT. To further investigate the effects of digital finance on HDFT, this study divides the sample into eastern, central, and western regions based on regional development differences and conducts grouped regressions to examine regional heterogeneity. As shown in Columns (1) to (3) of

Table 6. Regional heterogeneity analysis.

	(1)	(2)	(3)	(4)	(5)
	Eastern Region	Central Region	Western Region	Northern Region	Southern Region
lnDif	0.122 **	−0.0517	−0.0480 **	0.1000 ***	0.0748 ***
	(0.0516)	(0.0574)	(0.0219)	(0.0308)	(0.0199)
Hr	0.144	−0.119	0.0332	0.239 **	0.213 ***
	(0.126)	(0.0802)	(0.0556)	(0.0928)	(0.0541)
Fdi	0.0554 **	−0.0218	−0.00382	0.0738 ***	0.0641
	(0.0259)	(0.0496)	(0.0454)	(0.0224)	(0.0407)
lnTec	0.0566 **	−0.00893	0.0151	0.00565	0.0317 ***
	(0.0235)	(0.0146)	(0.0107)	(0.0230)	(0.00998)
Str	0.836 ***	0.123	0.0411	0.361 **	0.139
	(0.279)	(0.116)	(0.102)	(0.150)	(0.135)
_cons	−0.519 **	0.398 **	0.349 ***	−0.349 ***	−0.120
	(0.205)	(0.199)	(0.0921)	(0.129)	(0.100)
N	121	99	110	165	165
R-sq	0.704	0.217	0.274	0.405	0.578

Notes: Results displayed in the table were obtained using STATA 17. ** and ***, respectively, represent p < 0.05, and p < 0.01; standard errors are in parentheses.

, which present regression results for the eastern, central, and western regions, respectively, it is evident that the impact of digital finance on HDFT exhibits relatively pronounced differences across these regions. For the eastern region, the coefficient is 0.122, significantly positive at the 5% level, indicating that digital finance enhances HDFT in this region. A 1% increase in digital finance corresponds to a 0.122% improvement in HDFT. In the central region, the coefficient is −0.0517 and statistically insignificant, suggesting a mild but nonsignificant inhibitory effect of digital finance on HDFT. For the western region, the coefficient is −0.0480, significantly negative at the 5% level, demonstrating that digital finance suppresses HDFT in this area. Specifically, a 1% increase in digital finance leads to a 0.048% decline in HDFT. This regional disparity may stem from multifaceted differences between eastern and central/western China. The eastern region, with its abundance of coastal cities, well-developed transportation networks, and extensive port infrastructure, provides optimal conditions for international trade, which predominantly relies on maritime shipping. Additionally, the eastern region boasts advanced digital infrastructure, robust economic foundations, and leading digital infrastructure construction, application scenarios, and digital literacy. These advantages enable Eastern foreign trade firms to deeply understand and leverage digital technologies and financial expertise, facilitating access to advanced digital financial services. Such capabilities optimize capital flows, mitigate financing risks, enhance the value-added of export products, improve trade quality, and ultimately drive HDFT. In contrast, the central region lacks geographical advantages for attracting foreign investment, faces higher transportation costs in international trade, and thus exhibits limited trade engagement. The western region, constrained by harsh natural environments, technological backwardness, mono-industrial economic structures, and low trade competitiveness, experiences counterproductive effects: the development of the digital economy inadvertently hinders HDFT.

The division between southern and northern China holds critical significance in reflecting regional disparities and economic development levels across the country. In this study, we adopt the classification criterion proposed by Ouyang et al. [35], which defines the boundary at 35° N latitude. Based on long-standing disparities in resource endowments between the two regions, the 30 provinces are classified into southern and northern groups for subgroup empirical regression analyses. Columns (4) to (5) of Table 6 present regression results for the northern and southern regions, respectively, revealing notable differences between the two areas. The coefficients for both regions are significantly positive at the 1% level, indicating that digital finance promotes HDFT in both regions. However, this effect is more pronounced in the northern region. This divergence may arise because, despite the northern region’s lower informatization level compared to the south, it benefits from central macroeconomic policies and abundant local resources, granting it late-mover advantages. In recent years, accelerated digital economy construction in the north has spurred rapid growth in its financial and digital industries, significantly enhancing the coverage breadth and usage depth of digital finance. This progress provides a robust digital foundation for HDFT, actively driving regional trade transformation. In contrast, the southern region, with its stronger economic and industrial resource endowments, leverages inherent advantages and mature financial infrastructure to empower digital finance as a catalyst for HDFT. However, as digital finance permeates various industries and sectors, its marginal promotional effect on HDFT is due to saturation.

4.2.3. Heterogeneity Analysis of Financial Development Levels

Based on the ranking of financial development levels across 30 provinces, the sample is divided into high- and low-financial-development groups using the median as the threshold. Subsequent subgroup regression analyses are conducted, with results summarized in

Table 7. Heterogeneity analysis of financial development levels.

	(1)	(2)	(3)	(4)
	High-Financial-Development Regions	High-Financial-Development Regions	Low-Financial-Development Regions	Low-Financial-Development Regions
lnDif	0.118 ***	0.0712 ***	0.0452 *	0.0262
	(0.0242)	(0.0244)	(0.0246)	(0.0245)
Hr		0.419 ***		0.0829
		(0.0856)		(0.0505)
Fdi		0.0976 ***		0.0767 ***
		(0.0272)		(0.0230)
lnTec		0.0344 *		0.00140
		(0.0192)		(0.00915)
Str		0.379 **		0.141
		(0.178)		(0.105)
_cons	−0.114	−0.268 **	0.0759	0.0368
	(0.0902)	(0.132)	(0.0867)	(0.0888)
N	165	165	165	165
R-sq	0.482	0.589	0.212	0.305

Notes: Results displayed in the table were obtained using STATA 17. *, ** and ***, respectively, represent p < 0.1, p < 0.05, and p < 0.01; standard errors are in parentheses.

. The findings reveal that both high- and low-financial-development regions positively contribute to HDFT. However, the coefficient for the high-financial-development group is significantly positive at the 1% level, passing the statistical significance test, indicating a more pronounced enhancement effect on HDFT in these regions. Specifically, a 1% increase in the digital finance index raises HDFT by 0.0712% in high-financial-development areas, compared to 0.0262% in low-financial-development areas. This disparity may stem from the fact that regions with advanced financial development benefit from superior market environments and robust financial infrastructures. Coupled with digital technologies, these advantages accelerate the growth of digital finance, which effectively compensates for deficiencies in traditional financial services. The synergistic interplay between digital and traditional finance fosters a more conducive financial ecosystem, thereby amplifying the positive impact on HDFT.

4.3. Robustness Tests

4.3.1. Variable Substitution

Following Wu et al. [36], we replaced the core explanatory variable by normalizing the digital finance index. As shown in Column (1) of

Table 8. Robustness tests.

	(1)	(2)	(3)	(4)	(5)
	Variable Substitution	Lagging One Period	Lagging Two Periods	Adding Control Variables	Removing Municipalities
Dif	0.247 ***
	(0.0557)
L.lnDif		0.0879 ***
		(0.0166)
L2.lnDif			0.0972 ***
			(0.0157)
lnDif				0.0722 ***	0.0382 **
				(0.0173)	(0.0156)
Gov				0.0611 **
				(0.0293)
Hr	0.238 ***	0.144 ***	0.0558	0.231 ***	0.0929 **
	(0.0508)	(0.0530)	(0.0550)	(0.0506)	(0.0470)
Fdi	0.0701 ***	0.0751 ***	0.0705 ***	0.0916 ***	0.0522 ***
	(0.0167)	(0.0170)	(0.0174)	(0.0196)	(0.0174)
lnTec	0.0221 **	0.00908	0.000949	0.0240 **	0.0115
	(0.0103)	(0.0109)	(0.0118)	(0.0103)	(0.00858)
Str	0.289 ***	0.382 ***	0.406 ***	0.308 ***	0.123
	(0.0915)	(0.0984)	(0.104)	(0.0914)	(0.0773)
_cons	0.0140	0.285 ***	0.308 ***	0.284	0.0308
	(0.0507)	(0.0738)	(0.0743)	(0.248)	(0.0660)
N	330	300	270	330	286
R-sq	0.442	0.439	0.441	0.450	0.318

Notes: Results displayed in the table were obtained using STATA 17. ** and ***, respectively, represent p < 0.05, and p < 0.01; standard errors are in parentheses.

, the coefficient remains positive and significant at the 1% level, confirming that digital finance significantly promotes HDFT. This validates the robustness of our estimates.

4.3.2. Lagged Explanatory Variables

To account for potential time lag effects in digital finance development, we lagged the digital finance index by one and two periods in the baseline model. Columns (2) and (3) of Table 8 show that the coefficients for both lagged terms remain positive and significant at the 1% level, further demonstrating the robust positive impact of digital finance on HDFT.

4.3.3. Inclusion of Additional Control Variables

If unobserved variables are omitted, the estimated regression coefficients may become biased, thereby failing to accurately capture the causal relationship between digital finance and HDFT. To address this concern, we introduce the degree of government intervention (Gov) as a control variable in the regression model, measured by the logarithm of local government general budget expenditures. Government support enables firms to secure sufficient funding and resources for innovation, enhancing their core competitiveness and facilitating the transformation of foreign trade structures toward optimization. This shift promotes the transition from traditional trade models to modernized, digitized, and intelligent paradigms. As shown in Column (4) of Table 8, the coefficient for Gov is 0.0722, significantly positive at the 1% level, consistent with baseline regression results. This alignment confirms the robustness of our findings.

4.3.4. Exclusion of Regional Samples

The four direct-controlled municipalities in China—Beijing, Tianjin, Shanghai, and Chongqing—exhibit distinct economic characteristics. Their digital finance development levels, HDFT trajectories, and innovation capacities often diverge significantly from other provinces, reflecting unique advantages in international trade openness and policy frameworks. To mitigate potential biases caused by these inherent disparities and unobserved policy-specific factors, we exclude data from these municipalities and re-estimate the regression using the remaining provincial samples. Results in Column (5) of Table 8 demonstrate that the regression coefficients remain statistically significant at the 5% level, further validating the robustness of our conclusions.

4.4. Endogeneity Tests

To address the endogeneity problem caused by potential reverse causality between the core explanatory variable and the dependent variable, we follow the Bartik research approach and use the lagged one-period digital finance variable and its first-order time difference as instrumental variables (IV) for the core explanatory variable. We then conduct a 2SLS test [37]. According to the data in column (1) of

Table 9. Endogeneity tests.

Variable	First Stage	Second Stage
	(1)	(2)
IV	0.364 ***
	(13.60)
lnDif		0.306 ***
		(4.64)
Control Variables	Yes	Yes
Time Fixed Effects	Yes	Yes
Region Fixed Effects	Yes	Yes
_cons	5.3933 ***
	(80.29)
Kleibergen-Paap rk LM	125.81 ***
	[0.00]
Kleibergen-Paap rk Wald F	184.89
	[16.38]
N	300	300
R2	0.994	0.121

Notes: Results displayed in the table were obtained using STATA 17. ***, represent p < 0.01; standard errors are in parentheses (), p-values are in brackets [].

, the coefficient of the instrumental variable is significantly positive at the 1% level, indicating that the instrumental variable satisfies the relevance condition. Furthermore, both the LM and Wald F tests confirm the appropriateness of the instrumental variable selection. The data in column (2) show that the empirical results are largely consistent with those of the benchmark regression, further supporting the robustness and reliability of the benchmark regression results.

4.5. Mechanism Analysis

After establishing the direct effects, this study further employs a mediation model to dissect the mechanisms through which digital finance promotes HDFT, specifically testing whether the financial development level (Fin) and infrastructure level (Inf) serve as mediators. Regression results are presented in

Table 10. Results of the intermediary mechanism test.

Variables	Financial Development Level			Infrastructure Level
	(1)	(2)	(3)	(4)	(5)	(6)
	HDFT	F	HDFT	HDFT	Inf	HDFT
lnDif	0.0766 ***	0.410 ***	0.0665 ***	0.0766 ***	0.263 ***	0.0604 ***
	(0.0173)	(0.0835)	(0.0179)	(0.0173)	(0.0513)	(0.0178)
F			0.0246 **
			(0.0122)
Inf						0.0618 ***
						(0.0197)
Hr	0.238 ***	0.539 **	0.225 ***	0.238 ***	1.051 ***	0.173 ***
	(0.0508)	(0.245)	(0.0510)	(0.0508)	(0.151)	(0.0542)
Fdi	0.0701 ***	−0.316 ***	0.0779 ***	0.0701 ***	0.0572	0.0666 ***
	(0.0167)	(0.0807)	(0.0171)	(0.0167)	(0.0495)	(0.0165)
lnTec	0.0221 **	−0.152 ***	0.0259 **	0.0221 **	0.0532 *	0.0188 *
	(0.0103)	(0.0499)	(0.0104)	(0.0103)	(0.0306)	(0.0102)
Str	0.289 ***	1.570 ***	0.250 ***	0.289 ***	−0.259	0.305 ***
	(0.0915)	(0.442)	(0.0930)	(0.0915)	(0.271)	(0.0903)
_cons	−0.209 ***	−1.487 ***	−0.172 **	−0.209 ***	0.104	−0.215 ***
	(0.0753)	(0.364)	(0.0771)	(0.0753)	(0.223)	(0.0742)
N	330	330	330	330	330	330
R-sq	0.442	0.729	0.450	0.442	0.865	0.461

Notes: Results displayed in the table were obtained using STATA 17. *, **, and ***, respectively, represent p < 0.1, p < 0.05, and p < 0.01; standard errors are in parentheses.

.

Columns (1) and (4) of Table 10 present the baseline regression models without mediating variables, where the explained variable is high-quality foreign trade development (HDFT), reflecting the total effect of digital finance on HDFT. Columns (2) and (3) examine the financial development level: Column (2) uses financial development level (Fin) as the explained variable to explore the correlation between digital finance and financial development, while Column (3) incorporates both the core explanatory variable (digital finance, lnDif) and the mediating variable (Fin) into the regression model for explanation. Columns (5) and (6) focus on infrastructure level: Column (5) employs infrastructure level (Inf) as the dependent variable to investigate the relationship between digital finance and infrastructure, and Column (6) includes both digital finance (lnDif) and the mediating variable (Inf) in the regression for explanation. From Columns (2) and (5), it is evident that digital finance positively impacts both financial development and infrastructure levels. The coefficients for the explanatory variable (lnDif) are 0.410 and 0.263, respectively, both statistically significant at the 1% level. This confirms that digital finance promotes financial development and infrastructure improvement. Columns (3) and (6) present regression results after introducing two mediating variables—financial development level and infrastructure level—into the baseline model. Even after sequentially incorporating these mediators, digital finance retains a statistically significant positive effect on HDFT. Both mediating variables pass significance tests, indicating that advancing financial development and infrastructure construction actively contribute to HDFT. Furthermore, compared to the baseline models in Columns (1) and (4), the total effect of digital finance on HDFT is 0.0766. After separately adding the mediating variables (financial development level and infrastructure level), the direct effects of digital finance decrease to 0.0665 and 0.0604, respectively. The reduction in coefficients suggests that the total effect outperforms the isolated inclusion of either mediator, highlighting financial development and infrastructure as two critical mechanisms through which digital finance influences HDFT. Additionally, the mediating effects are quantified as follows: Financial development level: 0.0101 (0.410 × 0.0246), Infrastructure level: 0.0163 (0.263 × 0.0618). These results confirm the presence of mediating effects. Specifically, a 1% increase in digital finance elevates HDFT by 0.0101% and 0.0163% through financial development and infrastructure channels, respectively. This underscores the role of these mediators in amplifying digital finance’s impact on trade quality. Hypotheses H2 and H3 are thus empirically validated.

4.6. Further Analysis

To gain deeper insights into the nonlinear effects of digital finance on HDFT and the constraining roles of investment environments and internet penetration rates, this study employs a panel threshold model, with digital finance, investment environment, and internet penetration rate serving as threshold variables.

First, to identify the number of thresholds and specify the model structure, we conduct 300 bootstrap replications for each threshold variable. As shown in

Table 11. Threshold effect test results.

Threshold Variable	Model	RSS	MSE	F-Value	p-Value	Critical Value
						10%	5%	1%
lnDif	Single Threshold	0.2989	0.0009	25.25	0.0233	18.1030	22.5214	26.6235
	Double Threshold	0.2839	0.0009	16.77	0.0400	13.4363	16.4633	25.3639
	Triple Threshold	0.2773	0.0009	7.59	0.3833	18.7017	24.6730	48.7264
Inv	Single Threshold	0.2815	0.0009	46.44	0.0133	29.3224	33.3204	49.4997
	Double Threshold	0.2837	0.0009	−2.44	1.0000	26.1535	31.1238	53.4706
Int	Single Threshold	0.2814	0.0009	46.63	0.0067	26.4271	32.3584	40.1150
	Double Threshold	0.2710	0.0008	12.26	0.4067	53.3063	73.3142	92.1807

, when digital finance is the threshold variable, both single- and double-threshold effects are statistically significant at the 5% level, passing the double-threshold test. However, the triple-threshold test yields a p-value of 0.3833, failing significance. Thus, a double-threshold model is adopted for digital finance. For the investment environment, the single- and double-threshold tests produce p-values of 0.0133 and 1.0000, respectively, indicating only a single-threshold effect. Similarly, the internet penetration rate exhibits single- and double-threshold p-values of 0.0067 and 0.4067, also supporting a single-threshold model. Following the threshold effect validation, we estimate the threshold values and their 95% confidence intervals for each variable.

Table 12. Threshold estimation results.

Threshold Variable	Threshold	Threshold Value	95% Confidence Interval
lnDif	Single Threshold	5.8049	(5.7903, 5.8124)
	Double Threshold	6.0169	(6.0081, 6.0352)
Inv	Single Threshold	0.2938	(0.2794, 0.3213)
Int	Single Threshold	0.4414	(0.4374, 0.4432)

presents these estimates, which quantify the nonlinear relationships between the threshold variables and HDFT.

Corresponding to Table 12, according to the principle of the threshold model, the threshold estimate corresponds to the value where the likelihood ratio statistic (LR) approaches zero. Figure 1 presents the likelihood ratio function graph with digital finance as the threshold variable, showing two threshold estimates (5.8049 and 6.0169) and their 95% confidence intervals. Figure 2 and Figure 3, respectively, illustrate the likelihood ratio function graphs with investment environment and internet penetration rate as the threshold variables. In the figures, the lowest point of the likelihood ratio statistic represents the corresponding threshold estimate, while the dashed line indicates the critical value of 7.35. From the figures, it can be observed that the critical value of 7.35 is significantly higher than the threshold estimates for digital finance, investment environment, and internet penetration rate, indicating that these threshold estimates are statistically significant and robust.

When digital finance serves as the threshold variable, a double-threshold effect is observed, with threshold values of 5.8049 and 6.0169, falling within confidence intervals of (5.7903, 5.8124) and (6.0081, 6.0352), respectively. The regression coefficients for digital finance remain consistently significant and positive across all stages, reflecting its sustained positive influence on HDFT, though the magnitude of impact varies across stages. We categorize the development level of digital finance into three intervals based on the threshold regression results in Column (1) of

Table 13. Panel threshold model parameter estimation results.

Variable	(1)	(2)	(3)
	Digital Finance	Investment Environments	Internet Penetration Rates
Threshold type	double threshold	single threshold	single threshold
Threshold value (q1 vs. q2)	5.8049	0.2938	0.4414
	6.0169
lnDif·I (HDFT ≤ q1)	0.00970 *	−0.00484	0.0107 **
	(0.00488)	(0.0113)	(0.00517)
lnDif·I (q1 < HDFT < q2)	0.0135 **	0.0106 **	0.0219 ***
	(0.00551)	(0.00497)	(0.00786)
lnDif·I (HDFT ≥ q2)	0.0222 ***
	(0.00780)
Constant term	0.120	0.219 **	0.153
	(0.128)	(0.0834)	(0.111)
Control variable	Yes	Yes	Yes
N	330	330	330
R-squared	0.366	0.362	0.355

Notes: Results displayed in the table were obtained using STATA 17. *, **, and ***, respectively, represent p < 0.1, p < 0.05, and p < 0.01; standard errors are in parentheses.

. When the digital finance index is below the first threshold value of 5.8049, it exhibits a positive effect on HDFT, with an impact coefficient of 0.00970, passing the 10% significance level test. Specifically, a 1% increase in the digital finance index promotes a 0.00970% growth in HDFT. When the digital finance index lies between the two thresholds (5.8049 to 6.0169), the impact coefficient rises to 0.0135, validated at the 5% significance level. This indicates that the driving effect of digital finance on HDFT strengthens as digital finance advances. Upon exceeding the second threshold value of 6.0169, the impact coefficient further increases to 0.0222. The regression coefficient of the core explanatory variable on the dependent variable grows substantially, more than doubling the effect observed in the first stage, reaching an optimal impact on HDFT. These results demonstrate that the effect of digital finance on HDFT is nonlinear and exhibits a characteristic of “increasing marginal effects.”

Additionally, Table 13 demonstrates the nonlinear impacts of digital finance on HDFT under the constraints of the investment environment and internet penetration rate. When the investment environment serves as the threshold variable, a single-threshold effect is observed, with a threshold value of 0.2938 and a confidence interval of (0.2794, 0.3213). As shown in Column (2) of Table 13, the estimated coefficient after surpassing the threshold is 0.0106, which passes the 5% significance level test, significantly larger than the pre-threshold coefficient of 0.00484 and more statistically significant. This indicates that as the investment environment continuously improves, the positive impact of digital finance on HDFT exhibits an intensifying trend.

When the internet penetration rate is used as the threshold variable, a significant single-threshold effect is identified, with a threshold value of 0.4414. Column (3) of Table 13 reveals that when the internet penetration rate does not exceed the threshold of 0.4414, a 1% increase in digital finance promotes a 0.0107% growth in HDFT. After surpassing the threshold, the estimated coefficient rises from 0.0107 to 0.0219, further enhancing HDFT. This suggests that the nonlinear impact of digital finance on HDFT is constrained by the internet penetration rate and exhibits a characteristic of “increasing marginal effects.”

These findings demonstrate that the influence of digital finance on HDFT is not only shaped by its own development level but also dynamically constrained by the developmental stages of investment environment and internet penetration rate. Both factors can progressively amplify the positive effects of digital finance on HDFT. Consequently, Hypotheses H4 and H5 proposed in this study are empirically validated.

5. Conclusions and Suggestions

5.1. Main Conclusions

This study constructs a comprehensive evaluation system to measure HDFT across 30 Chinese provinces using panel data from 2011 to 2021. It empirically examines the linear impact, mechanisms, heterogeneity, and nonlinear characteristics of digital finance on HDFT through a two-way fixed effects model and a mediation effect model. The findings are as follows.

First, digital finance significantly promotes HDFT, primarily achieved through financial development and infrastructure improvement. This finding aligns at the micro level with the research results of Jin and Zhang [38], who found that digital finance significantly promotes enterprise export behavior by alleviating financing constraints, providing micro-enterprise-level support for the macro-level conclusions of this paper. Furthermore, Li and Wang [39] studied the enhancement effect of digital finance on the quality of agricultural product exports, further supplementing the positive role of digital finance in specific segments of foreign trade. This suggests that digital finance can not only promote HDFT overall but can also assist in improving product quality within specific industries.

Second, the positive impact of digital finance on HDFT is more pronounced in eastern and northern regions compared to central, western, and southern regions, and its facilitating effect is stronger in areas with higher financial development levels. This is consistent with the research of Xiao and Pan [40], who pointed out that the enhancement effect of digital finance on international trade competitiveness is particularly prominent in the Eastern region. Similarly, studies by Wang [41] and Zhang Lin et al. [42] also revealed the positive driving role of the digital economy in the Eastern region. Meanwhile, Nie [43] found that the promoting effect of digital finance on technological innovation in SMEs is more significant in high-tech industries and state-owned enterprises. This suggests that differences in enterprise characteristics and industries may amplify the positive impact of digital finance, indirectly corroborating the conclusions of this paper.

Third, threshold effects exist in digital finance’s impact on HDFT, where the marginal effects increase with digital finance development levels, constrained by investment environments and internet penetration rates. This conclusion aligns with the findings of Xiao and Pan [40] regarding threshold effects in international trade competitiveness, further emphasizing the dynamic nature of the role of digital finance in promoting HDFT.

5.2. Research Limitations and Future Research Directions

Although this study constructs an indicator system and uses the fixed effects model and mediation effects model to empirically analyze the impact of digital finance on HDFT, it should be noted that HDFT is a complex, systematic process influenced by multiple factors. This study primarily focuses on the role of digital finance and does not cover all potential influencing factors, which may affect the comprehensiveness and depth of the conclusions.

Additionally, this study uses provincial panel data for empirical analysis, which is common in research on digital finance and high-quality development. However, the accuracy and comparability of provincial data across years may have certain limitations, potentially affecting the precision of the results. Due to the focus on the provincial level, this study is limited to available provincial data and cannot incorporate more granular city-level or micro-level analyses.

Future research could introduce more economic and social variables, build a more systematic, multi-dimensional analytical framework, and explore the interactions between various factors and their overall impact on HDFT. Additionally, future studies could incorporate city-level or even micro-level data to improve the robustness and explanatory power of empirical analyses, providing more precise policy recommendations for HDFT.

5.3. Policy Suggestions

Based on the findings of this study, the following recommendations are proposed to further enhance HDFT.

5.3.1. Encourage Deep Integration Between Digital Finance and Foreign Trade Enterprises

Promote the precise alignment of digital financial products and services with the needs of foreign trade enterprises, focusing on the digital upgrading of services such as cross-border trade financing, export credit insurance, and supply chain finance. Expand the depth and breadth of digital financial services in areas such as international settlement, credit evaluation, and logistics informatization. Building digitalized industrial chains and cross-border financial ecosystems enhances the competitiveness and risk resilience of foreign trade enterprises.

5.3.2. Strengthen Digital Finance Governance and Regulatory Capacity

Establish and improve the legal and regulatory framework for digital finance and enhance the standardized management of emerging businesses such as cross-border payments, digital currencies, and smart contracts. Promote the development of a dynamic, tiered, and sector-specific regulatory framework that both supports innovation and effectively mitigates financial risks. Strengthen supporting systems such as data security protection and anti-money laundering to lay a solid foundation for the healthy and sustainable development of digital finance and ensure the trust and participation of foreign trade enterprises in the digital finance environment.

5.3.3. Optimize the Construction of Digital Finance Infrastructure

Accelerate the development of cross-border financial networks, improve digital payment and settlement systems, and promote the establishment of national-level public service platforms for digital finance that integrate core functions such as cross-border trade settlement, credit evaluation, and supply chain finance. Emphasize improvements in investment environments and internet penetration and enhance capabilities in network bandwidth, computing power, and information security to ensure efficient operation of digital financial infrastructure in supporting cross-border trade. This will fully unleash the potential of digital finance in driving HDFT.

5.3.4. Promote Regionally Differentiated Digital Finance Development

Implement regionally differentiated digital finance development strategies based on the economic foundation and level of digital finance development in each region. In eastern and economically developed regions, focus on supporting the innovation of high-end digital financial services such as digital asset trading and supply chain finance and promote deep integration between digital finance and foreign trade. In central, western, and less-developed regions, increase the supply of inclusive digital financial services to support the digital transformation of small and medium-sized enterprises, gradually narrow regional gaps in digital finance development, and enhance the overall level of HDFT.

5.3.5. Emphasize Policy Dynamism and Flexibility

Fully consider the phased characteristics of digital finance development and implement supportive policies and institutional arrangements in a stepwise manner. Focus on optimizing complementary factors such as the investment environment, internet penetration, and industrial digitalization. By improving the policy environment and strengthening information infrastructure, gradually eliminate bottlenecks and constraints to the development of digital finance, and ensure that digital finance can better support HDFT.