Impact of digital trade on the technological complexity of China's exports

Yuan Huang; Chenggang Wang; Yuan Huang; Chenggang Wang

doi:10.48130/stati-0026-0004

Against the backdrop of the rapid development of globalization and digitalization, digital trade, as a new form of trade, is profoundly changing the pattern and operation mode of international trade. As an important participant in world trade, studying the impact of China's digital trade on the complexity of its export technology has significant practical significance. Based on the panel data of various provinces in China from 2011 to 2023 and using the fixed effects model, this paper explores the impact of digital trade on the complexity of China's export technology and the mediating role of industrial structure in this context. The research results show that: (1) Digital trade has significantly increased the technological complexity of China's exports. This means that as digital trade continues to develop, the technological content and added value of China's export products are constantly increasing. (2) The industrial structure plays a significant positive mediating role in the impact of digital trade on the complexity of export technologies. (3) Digital trade has a significant positive impact on the technical complexity of exports in different regions. This article can provide important references for relevant governments and export enterprises to formulate management policies.

HTML

Introduction

In the era of accelerating globalization and digital integration, digital trade has emerged as a prominent area in the world economy. Simultaneously, with the rapid advancement of information technology, digital trade is profoundly reshaping the landscape and patterns of international trade. The State Council of China issued the '14^th Five-Year Plan for Digital Economic Development', emphasizing the need to enhance the application of digital technology in the field of trade. It indicates the Chinese government's awareness of the importance of digital technology in driving trade development and improving trade efficiency. It would help enhance the technological content and value-added of export products, thereby increasing export technological complexity. Furthermore, the General Office of the State Council of China issued the 'Implementation Opinions on Promoting Innovative Development of Foreign Trade'. This policy encourages enterprises to utilize digital means to expand markets, optimize supply chains, and enhance trade efficiency and competitiveness. It would contribute to raising China's export technological complexity and achieving higher-quality exports. Therefore, advancing China's digital trade requires accelerating the digital transformation of traditional industries to expand channels for digital trade. Moreover, it also necessitates improving the quality of China's export structure. Research in relevant fields should focus on enhancing the quality of export products in the current information technology and economic environment to provide strategic guidance for economic development.

Reviewing existing literature on digital trade, early studies primarily focused on the foundational aspects of the concept and scope of digital trade. However, as research progressed and economic and technological environments evolved, attention shifted towards the impact of digital trade on economic growth, trade structure, and other aspects. Furthermore, digital trade exerts significant influence on both domestic and international economies. In terms of the domestic economy, digital trade profoundly affects a country's industrial structure, technological innovation, labor force, and more. Additionally, digital trade strengthens global economic connections, facilitating the formation of a more refined and specialized division of labor, thereby enhancing the overall trade environment and efficiency. Factors such as trade barriers, trade liberalization, institutional quality, and environmental regulations, among others, all have varying impacts on export technological complexity^[1−5]. For instance, Zhang et al. conducted empirical tests on 16 manufacturing industries across 43 countries from 2000 to 2014^[6]. Zhang et al. suggested that population aging significantly influences export technological complexity, while artificial intelligence can mitigate the adverse effects of aging. Furthermore, existing scholars often consider factors such as technological innovation, human capital, etc., as mediating effects in the relationship between digital trade and export technological complexity^[7,8]. However, literature that utilizes industrial structure as a mediating variable is relatively scarce.

To sum up, existing research has mainly focused on studying the impact of digital trade on the domestic macroeconomy, with less consideration given to its effects on exports. Furthermore, studies on export technological complexity are relatively limited, with few exploring the relationship between digital trade and export technological complexity. Therefore, this paper employs empirical analysis and fixed-effects models to delve into the influence of digital trade on export technological complexity, elucidating the mediating role of industrial structure. Additionally, this study conducts research on regional heterogeneity to examine how digital trade affects export technological complexity across different regions.

The marginal contributions of this paper are as follows: (1) It has enriched the evaluation index system of digital trade. This paper discards the limitations of previous literature that mostly started from 2013 and only contained a few indicators, and constructs an evaluation system consisting of three levels and 15 indicators. Moreover, the samples cover the period from 2011 to 2023, effectively expanding the depth and breadth of theoretical research on the complexity of digital trade and export technology. (2) Application of innovative empirical research methods. This paper comprehensively employs the entropy method, fixed effects model, etc., to analyze the relationship between the two, filling the gap that previous literature did not comprehensively use these methods, enhancing the scientificity and accuracy of the research, and providing new research ideas for related fields. (3) Optimize the design of regional heterogeneity research, no longer following the previous fixed division method of four major regions. Instead, based on the differences in digital trade development, China is divided into inland and coastal regions for analysis, which is more in line with the current development situation. The conclusions drawn can provide highly valuable references for the government to formulate relevant policies.

However, this paper also has some research deficiencies: (1) Only a single model is used to test the mediating effect of the industrial structure. This method can test the existence and positive significance of the mediating effect of industrial structure, but the quantitative analysis of the mediating effect is insufficient. (2) The research subjects are limited to China only, making it difficult to comprehensively understand the universal laws and differences on a global scale.

Variables and methods

Measurement of variables

Explanatory variables

The explanatory variable is digital trade. The digital trade development levels of 30 provinces in China from 2011 to 2023 are measured to maximize the display of horizontal differences in each province's digital trade development while ensuring dynamic comparability. A system of digital trade measurement indicators was established based on the framework proposed by Li^[35], taking into account the availability of regional data. Therefore, a system of digital trade measurement indicators is constructed, consisting of four primary indicators, eight secondary indicators, and 15 tertiary indicators, as shown in Table 1.

Table 1. Digital trade measurement indicators.

Primary indicator	Secondary indicator	Tertiary indicator	Code
Digital trade infrastructure	Information network infrastructure	Length of long-haul fiber-optic cables (10,000 km)	X1
		Number of internet broadband access ports (10,000)	X2
		Population penetration rate of mobile phones (%)	X3
	Logistics transport infrastructure	Number of operational freight trucks on highways (10,000)	X4
		Number of civil transport ships/units	X5
		Express delivery volume/million items	X6
Digital technological innovation	Technological innovation input	Number of patents granted (1,000)	X7
	Technological innovation input	R & D expenditure/10 billion RMB	X8
	Digital innovation output	Technology market transaction volume (billion RMB)	X9
Digital trade market	Digital supply	Total telecommunications services volume (100 billion RMB)	X10
	Digital supply	Total postal services volume/billion RMB	X11
	Digital demand	Software services revenue (billion RMB)	X12
	Digital demand	E-commerce sales volume/billion RMB	X13
Trade potential	Economic strength	GDP (100 million RMB)	X14
Trade potential	Trade scale	Total import and export trade volume (${\text{\$}} $10,000 USD)	X15
Source: compiled by the author.

Explained variables

The dependent variable in this study is the export technological complexity. Following the approach of Hausmann et al.^[12], the export technological complexity is calculated for each province and city. This method involves two steps. Firstly, the technological complexity of each export product category (PRODY_j) is computed using the following formula:

$ PRODY_\mathrm{j}=\sum\limits_{\text{i}}^{ }\dfrac{x_{\text{ij}}/X_{\text{i}}}{\sum\limits_{\text{i}}^{ }(x_{\text{ij}}/X_{\text{i}})}Y_\mathrm{i} $

(1)

where, x_ij represents the trade value of province i's exports in category j, X_i represents the total export value of province i, and Y_i represents the per capita GDP of each province, adjusted using the consumer price index and standardized to the real value based on the year 2000. Among them, i represents the 30 provincial research samples of this paper. While j represents the 22 types of goods classified by HS codes.

Next, the export technological complexity of each province and city is calculated. Building on the results from the first step, the export technological complexity of each province is computed by weighted calculation. The formula for this calculation is as follows:

$ {Expy}_\mathrm{i}=\sum\limits_{\mathrm{j}}^{ }\dfrac{x_{\mathrm{ij}}}{X_{\mathrm{ij}}}PRODY_\mathrm{j} $

(2)

To mitigate the impact of heteroscedasticity in the data, a logarithmic transformation is applied to the final export technological complexity values.

Intermediary variables

In this study, the mediating variable is industrial structure. There are various indicators in existing literature for measuring industrial structure, such as the proportion among the three industries, Theil index, and other metrics. On one hand, considering the current output status of the three industries in China, the output of the primary industry is relatively small compared to the secondary and tertiary industries. On the other hand, the secondary industry mainly represents industrial production, reflecting the capacity for material production and processing. The tertiary industry, or the service industry, signifies the level of social services and consumption^[36]. The ratio between these two sectors can intuitively reflect a region's material production and service conditions, providing essential insights for analyzing the optimization direction of industrial structure and development patterns. Therefore, drawing on the work of Gan et al.^[37], and based on Clark's theorem, this study employs the ratio of the secondary industry to the tertiary industry to measure the industrial structure of each province and city.

Control variables

Building on prior research^[38,39], this study incorporates five control variables to comprehensively assess the relevant environment. As shown in Table 2, the specific variables selected are as follows. The level of economic development and social consumption partly reflects China's economic progress. Tax burden level signifies the rationality of social economic resource allocation and national economic output. Foreign direct investment showcases the extent of external capital utilization in regions. Degree of openness serves as a crucial indicator of a country or region's economic integration with the global economy. Human capital level, industrialization level, and urbanization level illuminate societal human resources, industrial development status, and social structure.

Table 2. Control variables.

Index	Variable name (Abbreviation)	Calculation formula
1	Economic development level (LNPgdp)	LNPgdp Logarithm of GDP per capita
2	Human capital level (HC)	Number of students in higher education/Total population
3	Industrialization level (IL)	Industrial value added/Gross regional product
4	Openness degree (OPEN)	Total import and export trade volume/GDP
5	Social consumption level (SCL)	Total retail sales of consumer goods/Gross regional product
Source: compiled by the author.

Model construction

Referencing the research of Zhang^[40], this study establishes a fixed effects model. To investigate the relationship between digital trade and export technological complexity, the model (3) is formulated. Taking into account the influence of various economic environments on the model, this study incorporates control variables on top of model (3), as demonstrated in model (4).

$ Expy_\mathrm{{it}}=\alpha_0+\alpha_1Dtra_\mathrm{{it}}+v_\mathrm{i}+\varepsilon_\mathrm{{it}} $

(3)

$ Expy_\mathrm{{it}}=\beta_0+\beta_1Dtra_\mathrm{{it}}+\beta_2Control_{\text{it}}+v_{\mathrm{i}}+\varepsilon_{\mathrm{i\mathrm{t}}} $

(4)

In the model, i represents the province, t represents the year, Expy represents export technological complexity, Dtra represents digital trade, Control represents the control variables, v_i characterizes the province-specific fixed effects that do not change over time, and ε_it stands for the random disturbance term.

To investigate whether industrial structure mediates the relationship between digital trade and export technological complexity, the approach of Fan et al.^[41] was followed, and models (5) and (6) specified.

$ Ind_\mathrm{{it}}=\gamma_0+\gamma_1Dtra_\mathrm{{it}}+\gamma_2Control_\mathrm{{it}}+v_\mathrm{{it}}+\varepsilon_\mathrm{{it}} $

(5)

$ Expy_{\mathrm{it}}=\lambda_0+\lambda_1Dtra_{\mathrm{it}}+\lambda_2Ind_\mathrm{{it}}+\lambda_3Control_\mathrm{{it}}+v_\mathrm{i}+\varepsilon_\mathrm{{it}} $

(6)

where, Ind represents industrial structure.

Variable testing and model selection

The study employs a broad set of indicators, some of which may overlap in their measurement. To ensure the reliability of the selected variables, the accuracy of parameter estimates, and the validity of the conclusions, a multicollinearity test is conducted before performing baseline regressions. As shown in Table 3, all VIF values are below five, both for individual variables and overall. This indicates that severe multicollinearity is not present among the variables. Additionally, an F-test is performed on the model. The result is statistically significant at the 1% level, supporting the existence of a linear relationship in the regression equation and confirming that the model as a whole is meaningful. Finally, a Hausman test is conducted. The results show that the p-value is less than 0.05, indicating the rejection of the null hypothesis, providing strong evidence in favor of using a fixed-effects model rather than a random-effects model.

Table 3. Multicollinearity test.

	VIF	1/VIF
Dtra	3.131	0.319
HC	1.535	0.651
IND	4.305	0.232
IL	2.562	0.390
LNPgdp	4.482	0.223
OPEN	1.893	0.528
SCL	1.478	0.676
Mean VIF	2.770
Source: compiled by the author.

Data

Based on the research hypotheses outlined above, panel data were analyzed from 30 provinces in China from 2011 to 2023. It ensures the scientific rigor, availability, and comprehensiveness of the data for constructing the digital trade index system, export technological complexity, industrial structure, and the control variables. The data for the digital trade indicators, industrial structure, and the indicators in the control variables are mainly sourced from the official department. They are the China Statistical Yearbook, provincial statistical yearbooks, and the Ministry of Commerce. Data on export technological complexity and trade-related information are obtained from the General Administration of Customs, China Statistical Yearbook, and National Research Network.

In this study, Xizang, where data is severely missing, was not included in the research sample. For the missing data of some provinces in a certain year, this paper adopts the interpolation method to complete the data. For sample data with relatively stable data changes, the missing values are mainly calculated through the linear relationship between two adjacent valid data points. However, if the data fluctuates greatly, Lagrange interpolation is adopted, and a polynomial equation is constructed using multiple known data points for estimation to enhance the accuracy of filling.

Results

Basic regression analysis

First, the impact of digital trade on the export technological complexity of China is examined. Table 4 presents the baseline regression results. Column (1) reports the estimated effect of digital trade alone. The coefficient is 2.207, which is positive and statistically significant at the 1% level. This suggests that, without controlling for other variables, a 1% increase in digital trade is associated with a 2.207% rise in export technological complexity. Control variables that may affect regional export technological complexity are then incorporated. The results are shown in Column (2) of Table 4. As additional controls are included, the coefficient on digital trade declines consistently. After controlling for human capital in Model (2), the coefficient drops to 1.471. When the economic development level is added in Model (3), it further decreases to 0.702. Subsequent models introduce social consumption level, industrialization level, and openness to trade. The coefficient stabilizes between 0.704 and 0.767. This pattern implies that some control variables exhibit complementary or substitution effects with digital trade. The initial estimate may capture indirect influences from other economic factors. Even after including all controls, the coefficient on digital trade remains positive and significant at the 1% level. This indicates a robust positive effect of digital trade on export technological complexity.Therefore, Hypothesis H1 is supported.

Table 4. Basic regression results.

	(1) Expy	(2) Expy	(3) Expy	(4) Expy	(5) Expy	(6) Expy
Dtra	2.209***	1.471***	0.702***	0.735***	0.767***	0.704***
	(39.796)	(21.551)	(7.745)	(8.119)	(8.505)	(7.806)
HC		0.663***	0.387***	0.381***	0.346***	0.405***
		(14.235)	(8.210)	(8.145)	(7.271)	(8.188)
LNPgdp			0.790***	0.777***	0.810***	0.768***
			(11.171)	(11.067)	(11.519)	(10.974)
SCL				−0.079***	−0.058**	−0.045
				(−2.799)	(−2.013)	(−1.600)
IL					0.113***	0.131***
					(2.978)	(3.475)
OPEN						−0.229***
						(−3.673)
Observations	390	390	390	390	390	390
R²	0.815	0.882	0.913	0.914	0.917	0.920
F	1583.7	1337.9	1242.0	951.3	779.6	674.8
, , ** indicates significant at the significance level of 10%, 5%, and 1% respectively, the content inside () represents the t-statistic.

Columns (2) to (6) in Table 4 progressively introduce all five control variables. The regression results indicate the following:

(1) The coefficients of per capita GDP, human capital, and industrialization level are all positive and statistically significant. First, after including human capital in Column (2), the coefficient on digital trade decreases considerably. This suggests a strong complementary relationship between human capital and digital trade. The technology diffusion effect of digital trade relies on a certain level of human capital. Only when workers possess relevant knowledge and skills can they effectively absorb and apply digital technologies. These skills help transform technological advantages into higher export technological complexity. Second, when per capita GDP is added in Column (3), its coefficient is 0.790 and statistically significant. Meanwhile, the coefficient on digital trade declines further. This implies that the economic development level serves as an important foundation for digital trade to exert its influence. Higher per capita GDP reflects better infrastructure and a more favorable innovation environment. They provide sufficient funding and resources for technological upgrading in exports. Thus, economic development strengthens the positive effect of digital trade on export technological complexity. Finally, Column (5) introduces the industrialization level. Its coefficient is 0.113 and significant. In this specification, the coefficient on digital trade increases slightly to 0.767. This result supports the presence of synergy between industrialization and digital trade. Industrialization provides broad application scenarios for digital technologies. For example, smart manufacturing and industrial internet integration combine digital tools with traditional production. Such integration enhances the technological content and value-added of products. It further amplifies the role of digital trade in driving technological upgrading.

(2) The estimated coefficients for both social consumption level and the degree of openness are statistically significant and negative. When social consumption level is introduced in Column (4), its coefficient is –0.079 and significant. Concurrently, the coefficient on digital trade experiences a slight increase. This result indicates that there is a potential mismatch between the domestic consumption structure and the technological upgrade driven by digital trade. Specifically, if social consumption remains concentrated in mid- and low-tier products, firms may prioritize production that caters to this lower-tech domestic demand. This reallocation of resources could crowd out investment in R & D and production of high-technology export products, thereby exerting a modest dampening effect on export technological complexity in the short term. Regarding openness to trade, its introduction in Column (6) yields a coefficient of –0.229, which is significantly negative. The coefficient on digital trade remains stable at 0.704. This finding is consistent with the baseline regression results and indicates potential risks within the current pattern of openness. If trade openness is predominantly focused on sectors with lower technological content, such as processing trade, firms may become overly reliant on external technology inputs while underinvesting in indigenous innovation capabilities. This reliance could anchor China's export industries in lower segments of the global value chain, consequently inhibiting the enhancement of export technological complexity.

Robustness test

Endogeneity test

Potential endogeneity in the explanatory variables may bias the regression results. To address this concern and verify the accuracy of the empirical findings, a 2SLS approach was employed to mitigate endogeneity issues in the model. Following the method of Wang & Zheng^[42], the first lag of the digital trade development level is used as an instrument variable. The results of the re-estimated model are presented in Column (1) of Table 5.

Table 5. Results of robustness tests.

	(1) Expy	(2) Expy	(3) Expy	(4) Expy
Dtra	0.294***	1.074***	0.649***
	(5.011)	(10.012)	(6.576)
DT				0.651***
				(6.803)
HC	0.241***	0.371***	0.352***	0.471***
	(5.623)	(6.357)	(7.095)	(8.312)
LNPgdp	0.660***	0.646***	0.704***	0.928***
	(10.607)	(7.994)	(10.789)	(11.928)
SCL	−0.077**	−0.019	−0.050*	−0.060*
	(−2.022)	(−0.635)	(−1.707)	(−1.825)
IL	−0.098***	0.121***	0.094**	0.204***
	(−2.727)	(2.728)	(2.485)	(4.518)
OPEN	−0.682***	−0.283***	−0.203***	−0.227***
	(−14.743)	(−3.959)	(−3.131)	(−2.904)
N	360	300	338	390
R²		0.927	0.916	0.917
F		558.7	559.6	648.9
, , and ** respectively indicate significant at the significance level of 10%, 5%, and 1%, the content inside () represents the t-statistic.

Prior to the endogeneity test, the validity of the selected instrument is examined. The Anderson LM statistic, Cragg-Donald Wald statistic, and Sargan statistic all confirm the appropriateness of the instrument variable. The 2SLS regression shows that the coefficient of Dtra remains positive and statistically significant at the 1% level. This indicates that digital trade development effectively promotes the upgrading of China's export technological complexity. These results further support Hypothesis 1, confirming that digital trade has a positive and significant effect on export technological complexity. The findings also demonstrate the robustness of the conclusions. Additionally, the results suggest that digital trade exhibits a certain lag effect, yet this does not alter its overall positive influence on export technological complexity.

Other robustness tests

Remove samples from special periods

The COVID-19 pandemic caused unprecedented disruptions to global economic and social activities. China's pandemic containment policies, including widespread isolation measures, significantly negatively affected its export trade. To more accurately assess the long-term relationships among variables and enhance the robustness and credibility of the model conclusions, data from the pandemic period is excluded. This approach helps examine whether the model results are influenced by this exogenous shock and ensures the robustness of the empirical findings. Accordingly, data from 2020 to 2022 is removed, and the regression is re-run. The results are presented in Column (2) of Table 5. The coefficient on digital trade remains positive and statistically significant at the 1% level. This result is consistent with earlier findings, providing further support for the robustness of Hypothesis H1.

Eliminate special individual samples
The four municipalities directly under the Central Government, namely Shanghai, Beijing, Tianjin, and Chongqing, have larger economic scales, greater policy flexibility, and disproportionately abundant resources. These distinctive characteristics may introduce bias into the empirical estimates. To mitigate potential distortion, these cities are excluded, and the regression is re-run^[43]. As presented in Column (3) of Table 5, the coefficient on digital trade remains positive and statistically significant, with a magnitude comparable to the baseline full-sample estimate. This confirms that the positive effect of digital trade on export technological complexity is robust even after removing regions with exceptional policy advantages. The result demonstrates that the technological upgrading effect of digital trade is not driven primarily by unique policy benefits available in municipalities. Rather, it operates through general mechanisms such as reducing trade costs, accelerating technology diffusion, and improving resource allocation efficiency. These mechanisms appear universally applicable across other provinces. Moreover, both the goodness-of-fit and the F-statistic remain largely unchanged compared to the baseline model. This indicates that the adjusted sample does not diminish the model's explanatory power, further supporting the statistical reliability of the findings.

Adjust the calculation method of variables
To further test the robustness of the conclusions, an alternative measurement approach is adopted. Specifically, the method proposed by Verma et al.^[44]is referred to, and the digital trade index is recalculated using principal component analysis instead of the original method. Firstly, standardize the 15 items of digital trade, calculate the covariance matrix, solve the eigenvalues and eigenvectors, and sort them. Then, the principal components are selected based on the cumulative variance contribution rate, and the data is mapped through the principal component matrix to form a comprehensive digital trade indicator.

Subsequently, the model is re-estimated using a fixed effects specification. The results remain consistent, thereby strongly supporting the reliability of core findings. As shown in column (4) of Table 5, even after recalculating the composite index using PCA, the impact of digital trade on the technical complexity of China's exports remains statistically significant. Specifically, the revised estimate shows that a one-unit increase in the digitally recalculated trade index is associated with a 0.651 unit increase in export technological complexity, significant at the 1% level. This confirms that the positive effect of digital trade on export technological complexity is robust to alternative measurement approaches. The result provides strong additional support for the high robustness of Hypothesis H1.

Mediating effect test

The mediating effect of industrial structure is tested between digital trade and export technological complexity. The two-step method was adopted to test the mediating effect, and a regression analysis was conducted using models (5) and (6). Column (1) of Table 6 shows a coefficient of 0.692 for digital trade on industrial structure. It is significant at the 1% level. That is, for every 1% increase in the digital trade index, the industrial structure will rise by 0.692%. This indicates that digital trade significantly promotes industrial upgrading. Therefore, Hypothesis H2 is supported.

Table 6. Regression results of mediating effects.

	(1) IND	(2) Expy
Dtra	0.692***	0.473***
	(7.300)	(5.209)
IND		0.335***
		(7.063)
HC	0.175***	0.346***
	(3.372)	(7.352)
LNPgdp	0.001	0.768***
	(0.007)	(11.705)
SCL	0.112***	−0.083***
	(3.769)	(−3.061)
IL	−0.468***	0.288***
	(−11.827)	(6.898)
OPEN	0.010	−0.233***
	(0.160)	(−3.979)
N	390	390
R²	0.730	0.930
F	159.1	665.4
, , ** indicates significant at the significance level of 10%, 5%, and 1% respectively, the content inside () represents the t-statistic.

Column (2) of Table 6 further reveals the mediating role of industrial structure. The coefficient for industrial structure is 0.335 and is significant. This shows that industrial upgrading itself enhances export technological complexity. Advanced industrial structure directly increases the technological content of exports. After adding industrial structure as a mediator, the direct effect of digital trade drops from 0.704 to 0.473. However, it remains positive and significant. This confirms a partial mediating effect. Digital trade improves export technological complexity both directly and indirectly through industrial upgrading. Control variables show consistent results. Per capita GDP and human capital have positive effects. Openness to trade has a negative effect. This stability supports model robustness. Overall, the industrial structure is a key transmission channel. These results validate Hypothesis H3. They provide evidence that digital trade boosts export technological complexity by promoting industrial transformation.

Heterogeneity analysis

Due to variations in infrastructure, economic development levels, and human resources across regions, the impact of digital trade on total factor productivity in the service industry may differ among provinces. In China, coastal areas generally possess more advanced economic foundations, better information technology infrastructure, and superior transportation conditions. These advantages facilitate greater access to international markets and provide inherent benefits for digital trade development. In contrast, inland regions face relative isolation and encounter more obstacles in developing digital trade. To account for these regional disparities, the 30 provinces in the sample are classified into coastal and inland regions. This classification follows official guidelines issued by the National Development and Reform Commission and the National Bureau of Statistics. Specifically, 12 coastal provinces and 18 inland provinces were identified. The detailed regional division is presented in Table 7.

Table 7. Regional division.

Region	Specific province
Coastal cities	Liaoning Province, Hebei Province, Beijing Municipality, Tianjin Municipality, Shandong Province, Jiangsu Province, Shanghai Municipality, Zhejiang Province, Fujian Province, Guangdong Province, Guangxi Zhuang Autonomous Region, Hainan Province
Inland cities	Shanxi Province, Jilin Province, Heilongjiang Province, Anhui Province, Jiangxi Province, Henan Province, Hubei Province, Hunan Province, Sichuan Province, Guizhou Province, Yunnan Province, Shaanxi Province, Gansu Province, Qinghai Province, Inner Mongolia Autonomous Region, Ningxia Hui Autonomous Region, Chongqing Municipality
Data source: Compiled based on the documents of the National Development, and Reform Commission and the National Bureau of Statistics of China.

The regional heterogeneity regression results are shown in Table 8. The regression results show that digital trade has a coefficient of 0.941 on export technological complexity. This is significant at the 1% level. This value is notably higher than that of inland regions. It indicates that digital trade has a stronger promoting effect in coastal areas. This advantage stems from better digital infrastructure and more advanced digital trade foundations. Coastal regions benefit from superior geographical locations and policy support. They integrate earlier into global digital trade networks. They also achieve deeper fusion between digital technology and trade. Thus, digital trade more effectively boosts the technological upgrading of exports. Among control variables, human capital and economic development show strongly positive and significant effects. This reflects the role of higher human capital accumulation and economic development in coastal regions. These factors strongly support the technological upgrading effect of digital trade. Industrialization has a positive but small coefficient. This may be because coastal regions already have mature industrialization. Thus, its marginal contribution to export technological complexity is limited. Social consumption and trade openness are not significant. This may indicate a better alignment between consumption and export structures in coastal areas. It may also reflect that trade openness has moved beyond low-end lock-in, reducing its inhibitory effect.

Table 8. Regional heterogeneity regression results.

	Coastal cities expy	Inland cities expy
Dtra	0.941***	0.280***
	(6.353)	(3.172)
HC	0.440***	0.389***
	(5.634)	(6.365)
LNPgdp	0.774***	0.720***
	(6.539)	(11.079)
SCL	0.030	−0.031
	(0.672)	(−0.876)
LI	0.166*	0.080***
	(1.746)	(2.767)
OPEN	−0.071	−0.056
	(−0.817)	(−1.400)
N	156	234
R²	0.925	0.932
F	285.7	477.2
, , ** indicates significant at the significance level of 10%, 5%, and 1% respectively, the content inside () represents the t-statistic.

The regression results show a coefficient of 0.280 for digital trade. It is statistically significant at the 1% level. Although smaller than the coastal estimate, it confirms that digital trade still promotes export technological complexity inland. This weaker effect arises because inland regions started digital trade development later. Their digital infrastructure and supporting industries remain less developed. Among control variables, human capital and per capita GDP show positive and significant effects. This aligns with coastal findings. It confirms that human capital and economic development are fundamental to export upgrading across all regions. Industrialization also has a positive and significant coefficient. This suggests inland regions are in a phase of accelerated industrialization. In addition, social consumption and trade openness are not significant. This may reflect relatively limited consumer markets and lower trade openness in inland areas. Their potential influence on export technological complexity has not yet fully materialized.

Therefore, the impact of digital trade on the technical complexity of exports shows significant regional heterogeneity, which provides an empirical basis for formulating differentiated regional digital trade development policies.

[1]	Qi JY, Qiang HJ. 2021. Do digital service trade barriers affect service export complexity? An empirical analysis based on the OECD-DSTRI database. International Business (Journal of University of International Business and Economics) 4:1−18 (in Chinese) doi: 10.13509/j.cnki.ib.2021.04.001 CrossRef Google Scholar
[2]	Sheng B, Mao QL. 2017. Does import trade liberalization affect the technological complexity of Chinese manufacturing exports? World Economy 40:52−75 (in Chinese) doi: 10.19985/j.cnki.cassjwe.2017.12.004 CrossRef Google Scholar
[3]	Zheng ZP, Yue S. 2020. Institutional quality, population structure, and export technological complexity. Journal of Beijing University of Technology (Social Sciences Edition) 22:70−78 (in Chinese) doi: 10.15918/j.jbitss1009-3370.2020.2220 CrossRef Google Scholar
[4]	Gao X, Yuan K. 2020. Clean production environmental regulations and enterprise export technological complexity: micro evidence and influence mechanism. International Trade Issues 2:93−109 (in Chinese) doi: 10.13510/j.cnki.jit.2020.02.007 CrossRef Google Scholar
[5]	Yang S. 2016. Regional heterogeneity of the impact of environmental regulations on export technological complexity. Journal of Dalian University of Technology (Social Sciences Edition) 37:45−51 (in Chinese) doi: 10.19525/j.issn1008-407x.2016.04.009 CrossRef Google Scholar
[6]	Wu K, Tang Z, Zhang L. 2022. Population aging, industrial intelligence, and export technology complexity. Sustainability 14:13600 doi: 10.3390/su142013600 CrossRef Google Scholar
[7]	Liu Y, Hu W, Luo K, Guo Y, Wang Z. 2023. How does digital trade promote and reallocate the export technology complexity of the manufacturing industry? Evidence from 30 Chinese provinces, 2011–2020. PLoS One 18:e0291464 doi: 10.1371/journal.pone.0291464 CrossRef Google Scholar
[8]	Gong X. 2024. An empirical analysis of the impact of digital trade on export technological complexity. China Business Review 3:42−45 (in Chinese) doi: 10.19699/j.cnki.issn2096-0298.2024.03.042 CrossRef Google Scholar
[9]	Ouyang R, Liang J. 2023. Endogenous growth mechanism and Chinese strategies for the development of digital trade. Journal of University of Electronic Science and Technology (Social Sciences Edition) 25:1−16 (in Chinese) doi: 10.14071/j.1008-8105(2023)-1053 CrossRef Google Scholar
[10]	Jiang T, Hu Y, Haleem F, Zeng S. 2023. Do digital trade rules matter? Empirical evidence from TAPED. Sustainability 15:9074 doi: 10.3390/su15119074 CrossRef Google Scholar
[11]	Dai K. 2018. The impact of technological market development on export technological complexity and its mechanism. China Industrial Economics 7:117−135 (in Chinese) doi: 10.19581/j.cnki.ciejournal.2018.07.006 CrossRef Google Scholar
[12]	Hausmann R, Hwang J, Rodrik D. 2007. What you export matters. Journal of Economic Growth 12:1−25 doi: 10.1007/s10887-006-9009-4 CrossRef Google Scholar
[13]	Liu Y, Xia J. 2021. How does the platform economy facilitate smooth service consumption circulation: mechanism and policy design. Reform 11:19−29 (in Chinese) Google Scholar
[14]	Kerner R, Kitsing M. 2023. Small is beautiful and important: economies and firms trading in digital services. Journal of Business Economics and Management 24:93−111 doi: 10.3846/jbem.2023.18585 CrossRef Google Scholar
[15]	Malatji M. 2024. Accelerating the African continental free trade area through optimization of digital supply chains. Engineering Reports 6:e12711 doi: 10.1002/eng2.12711 CrossRef Google Scholar
[16]	Lendle A, Olarreaga M, Schropp S, Vézina PL. 2016. There goes gravity: eBay and the death of distance. The Economic Journal 126:406−441 doi: 10.1111/ecoj.12286 CrossRef Google Scholar
[17]	Li J, Wang F. 2024. A study on the competitiveness and influencing factors of the digital service trade. Sustainability 16:3116 doi: 10.3390/su16083116 CrossRef Google Scholar
[18]	Zhong S. 1998. Several concepts on industrial structure. Social Sciences Dynamics 3:47−49 (in Chinese) doi: 10.19354/j.cnki.42-1616/f.1998.03.016 CrossRef Google Scholar
[19]	Xu S, Liu Q, Lu X. 2021. Analysis of unbalanced spatial evolution characteristics and influencing effects of regional development in China: from the perspective of industrial structure transformation and upgrading. Finance and Trade Research 32:14−26 (in Chinese) doi: 10.19337/j.cnki.34-1093/f.2021.10.002 CrossRef Google Scholar
[20]	Li S, Gao L, Han C, Gupta B, Alhalabi W, et al. 2023. Exploring the effect of digital transformation on firms’ innovation performance. Journal of Innovation & Knowledge 8:100317 doi: 10.1016/j.jik.2023.100317 CrossRef Google Scholar
[21]	Wang C. 2022. Theoretical, mechanism, and strategy of the integrated development of digital economy and traditional industries. Business and Economic Research 8:190−192 (in Chinese) doi: 10.3969/j.issn.1002-5863.2022.08.048 CrossRef Google Scholar
[22]	Bas M, Strauss-Kahn V. 2015. Input-trade liberalization, export prices, and quality upgrading. Journal of International Economics 95:250−262 doi: 10.1016/j.jinteco.2014.12.005 CrossRef Google Scholar
[23]	Nham NTH, Bao NKQ, Ha LT. 2023. Nonlinear effects of digitalization on export activities: an empirical investigation in European countries. Technological and Economic Development of Economy 29:1041−1079 doi: 10.3846/tede.2023.17061 CrossRef Google Scholar
[24]	He DX, Yao ZQ. 2008. Effects, optimization goals, and policy measures of industrial structure adjustment in China. China Industrial Economics 5:46−56 (in Chinese) doi: 10.19581/j.cnki.ciejournal.2008.05.005 CrossRef Google Scholar
[25]	Sakas DP, Giannakopoulos NT, Terzi MC, Kanellos N, Liontakis A. 2023. Digital transformation management of supply chain firms based on big data from DeFi social media profiles. Electronics, 12:4219 doi: 10.3390/electronics12204219 CrossRef Google Scholar
[26]	Choi C. 2010. The effect of the Internet on service trade. Economics Letters 109:102−104 doi: 10.1016/j.econlet.2010.08.005 CrossRef Google Scholar
[27]	De Benedictis L, Tajoli L. 2007. Economic integration and similarity in trade structures. Empirical 34:117−137 doi: 10.1007/s10663-006-9024-x CrossRef Google Scholar
[28]	Chang X, Yang Z, Abdullah. 2024. Digital economy, innovation factor allocation, and industrial structure transformation—a case study of the Yangtze River Delta city cluster in China. PLoS One 19:e0300788 doi: 10.1371/journal.pone.0300788 CrossRef Google Scholar
[29]	Yu D, Li Y. 2021. Industrial organizational innovation in the era of the digital economy: a case study of the digital technology-driven industrial chain ecosystem. Reform 7:24−43 (in Chinese) Google Scholar
[30]	Yang Y, Chen N, Chen H. 2023. The digital platform, enterprise digital transformation, and enterprise performance of cross-border e-commerce—from the perspective of digital transformation and data elements. Journal of Theoretical and Applied Electronic Commerce Research 18:777−794 doi: 10.3390/jtaer18020040 CrossRef Google Scholar
[31]	Zhou M, Lu Y, Fu DH. 2016. Trade liberalization and industrial upgrading in China: facts and mechanisms. World Economy 39:78−102 (in Chinese) doi: 10.19985/j.cnki.cassjwe.2016.10.005 CrossRef Google Scholar
[32]	Jiang LD, Chen HX, Lu Y, Zhang GF. 2024. Upgrading of consumption structure and improvement of export product quality. Quantitative and Technical Economics Research 41:5−26 (in Chinese) doi: 10.13653/j.cnki.jqte.20240319.007 CrossRef Google Scholar
[33]	Zhang L, Pan A, Feng, S, Qin Y. 2022. Digital economy, technological progress, and city export trade. PLoS One 17:e0269314 doi: 10.1371/journal.pone.0269314 CrossRef Google Scholar
[34]	Han M, Zhang X. 2023. Empowering the manufacturing industry's export product upgrading through digital trade: a study from the perspective of technological complexity. Industrial Technology and Economics 42:106−114 (in Chinese) doi: 10.3969/j.issn.1004-910X.2023.02.011 CrossRef Google Scholar
[35]	Li W, Li T. 2023. Development of China's digital trade and comprehensive evaluation of provincial digital trade development: a quantitative analysis of Fujian Province. Price Theory and Practice 12:171−175 (in Chinese) doi: 10.19851/j.cnki.CN11-1010/F.2023.12.469 CrossRef Google Scholar
[36]	Zhang K, Liu DY, Yu SH. 2023. The impact of digital trade on export technological complexity: a test of the moderating mediating effect model based on financial development. Exploration of Economic Issues 2:144−159 (in Chinese) Google Scholar
[37]	Gan CH, Zheng RG, Yu DF. 2011. The impact of changes in China's industrial structure on economic growth and fluctuations. Economic Research 46:4−16, 31 (in Chinese) Google Scholar
[38]	Wang Z, Han C, Zhu W. 2022. The impact of digital financial development on export technological complexity. World Economic Research 8:26−42, 135−136 (in Chinese) doi: 10.13516/j.cnki.wes.2022.08.009 CrossRef Google Scholar
[39]	Hu X, Liu F. 2023. Research on the influencing factors and improvement paths of digital trade development in the Yangtze River Economic Belt: based on entropy-weighting TOPSIS and fsQCA method. PLoS One 18:e0284519 doi: 10.1371/journal.pone.0284519 CrossRef Google Scholar
[40]	Zhang WK. 2024. The impact of digital trade development on the quality of China's export products: the mediating effect of openness. Business and Economic Research 2:138−141 (in Chinese) Google Scholar
[41]	Fan HJ, Wu T, He SJ. 2022. How does the "Internet + Government Service" platform optimize the urban business environment? A perspective of interactive governance. Management World 38:126−153 (in Chinese) doi: 10.19744/j.cnki.11-1235/f.2022.0149 CrossRef Google Scholar
[42]	Wang SY, Zheng LK. 2019. Does servitization of manufacturing promote the upgrading of export products—based on the dual perspective of export product quality and export technology complexity. Journal of International Trade 11:45−60 (in Chinese) doi: 10.13510/j.cnki.jit.2019.11.004 CrossRef Google Scholar
[43]	Shi Biao. 2025. Research on the impact of fiscal and financial coordination on new-quality productivity and its path. Development Finance Research 2:3−14 (in Chinese) doi: 10.16556/j.cnki.kfxjr.2025.02.007 CrossRef Google Scholar
[44]	Verma A, Dandgawhal PS, Giri AK. 2023. Impact of ICT diffusion and financial development on economic growth in developing countries. Journal of Economics, Finance and Administrative Science 28:27−43 doi: 10.1108/JEFAS-09-2021-0185 CrossRef Google Scholar

{{lists.name}}

Impact of digital trade on the technological complexity of China's exports

Abstract

Rights and permissions

References

About this article

Cite this article

Article Metrics

Access History

Other Articles By Authors