A Study on the Relationship Between Rural E-Commerce Development and Farmers’ Income Growth

Abstract

The development of rural e-commerce is conducive to expanding channels for increasing farmers’ income. This paper analyzes the relationship between rural e-commerce development and farmers’ income growth in 31 provinces (municipalities/autonomous regions) across China. First, based on a review of relevant literature, an index system for rural e-commerce development was constructed, and the entropy weight method was used to measure the comprehensive index of national rural e-commerce development. Next, a panel data regression model was established to analyze the relationship between rural e-commerce development and farmers’ income growth, followed by regional heterogeneity analysis and robustness tests. The analysis found that during the study period, China’s rural e-commerce development level increased year by year, but there were regional differences; the development of rural e-commerce in China has a significant positive impact on farmers’ income growth, but the degree of impact varies across different regions. Finally, suggestions are put forward to promote the further development of rural e-commerce in China.

1. Introduction

The Party Central Committee and the State Council attach great importance to the development of e-commerce. Since 2020, General Secretary Xi Jinping has given important instructions on the development of e-commerce on multiple occasions in different settings. The Government Work Report has mentioned e-commerce several times and highly affirmed its significant role in economic development. As an emerging business format that integrates the development of issues related to agriculture, rural areas, and rural residents (the “Three Rural Issues”), rural e-commerce, in contrast to the mature development of urban e-commerce, is permeating throughout the entire agricultural industry chain and rapidly transforming the business models, economic development patterns, and the production and living styles of farmers in rural areas of China [1]. Through its integration with industries such as cultural tourism and finance, it has not only solved the problem of agricultural product sales but also driven the development of services such as rural tourism and finance, forming a complete rural industrial chain and enhancing the overall economic benefits of rural areas.

With the rapid development and widespread adoption of digital technologies, the digital divide between urban and rural areas has been continuously narrowing. Rural e-commerce has broken through temporal–spatial limitations and information technology barriers, creating broader market opportunities for rural regions, unlocking new income growth channels for farmers, and activating the potential for rural revitalization. In 2023, China’s annual rural online retail sales reached CNY 2.49 trillion, marking a 12.9% year-on-year increase and nearly a 13-fold growth compared to 2014 [2]. National agricultural product online retail sales amounted to CNY 587.03 billion, growing by 12.5% year-on-year. Meanwhile, the per capita disposable income of farmers approached CNY 22,000. According to data from the National Bureau of Statistics, rural residents’ per capita disposable income reached CNY 6596 in the first quarter of 2024, representing a real growth of 7.7% after the deduction of price factors—2.4 percentage points higher than the growth rate of urban residents’ per capita disposable income [3]. Rural e-commerce demonstrates enormous development potential and promising prospects. Its expansion will further stimulate employment and entrepreneurship in rural areas, drive rural economic development, narrow urban–rural disparities, and contribute to the construction of beautiful villages and the implementation of rural revitalization strategies. Therefore, investigating the relationship between rural e-commerce development and farmers’ income growth in China is of significant necessity.

This study focuses on the relationship between the development of rural e-commerce in China and the increase in farmers’ income. It aims to construct a research mechanism and an empirical analysis framework between the two from a macro perspective and to reveal the internal logic of the relationship between them. The research objects of this paper are the 31 provinces (municipalities directly under the Central Government/autonomous regions) in China, and the research period spans from 2011 to 2023.

Firstly, by reviewing relevant domestic and international research findings, the influence mechanism through which rural e-commerce affects the increase inof farmers’ income is clarified. Subsequently, an evaluation index system for the development of rural e-commerce in China was constructed, and the entropy method [4,5] was used to calculate the comprehensive index of the development of rural e-commerce in each province (municipality directly under the Central Government or autonomous region). We then established a panel data regression model [6], taking the per capita disposable income of rural residents as the dependent variable and the rural e-commerce development index as the core explanatory variable, to analyze the relationship between rural e-commerce development and rural household income. Given the regional disparities in rural e-commerce development across China, we conducted regional heterogeneity tests for the eastern, central, western, and northeastern regions to examine how rural e-commerce impacts farmers’ income differently across these areas. Finally, to ensure the scientific rigor and robustness of our regression results, we performed robustness analyses by replacing both the dependent and explanatory variables, verifying the consistency of our findings across alternative specifications.

2. Literature Review

2.1. Rural E-Commerce

The concept of rural e-commerce lacks a universally agreed-upon definition.

Grandon et al. (2004) defined e-commerce as the process of buying or selling products or services through electronic data transmission over the Internet and the World Wide Web based on the research of prior scholars. They also elaborated on the multitude of benefits that e-commerce offers to both buyers and sellers. This definition emphasizes the core mechanism of e-commerce while highlighting its value-adding aspects for market participants [7]. Joshi et al. (2017) defined e-commerce as activities that conduct or facilitate commercial transactions over the Internet. They emphasized that e-commerce extends beyond mere online buying and selling; any commercial transaction conducted electronically falls within its scope [8]. Jain, V. et al. (2021) define e-commerce as the transaction of goods and services through electronic media and the Internet, which requires businesses to adopt Internet connectivity and relevant information technologies. They specify that this involves directly selling goods or services to users through the platforms of Internet providers. The authors subsequently discuss the development trends of e-commerce in India [9]. Chatterjee (2019) distinguishes rural e-commerce from mainstream e-commerce, defining it as a model that establishes digital platforms for rural products to reach broader consumer bases. She emphasizes that rural e-commerce transcends mere transactional mechanisms, fundamentally enhancing the integration of rural areas with regional and national markets. This process not only expands market access for rural producers but also serves as a catalyst for improving income levels in rural communities [10].

An Linli (2018) conceptualizes rural e-commerce as an emerging industrial chain that connects agricultural production, consumption, and circulation through digital platforms, facilitating efficient rural resource utilization, transforming traditional agriculture, and boosting farmer income [11]. Sun, F., and Li, J. (2022) employed grounded theory methodology from the perspective of common interest orientation to examine rural e-commerce, positing that it represents a tangible implementation within the framework of rural revitalization and shared prosperity [12].

Drawing from existing literature and scholarly insights, this paper defines rural e-commerce as a digital transaction process that leverages modern information technologies—including the Internet, computers, and multimedia—to enable agricultural producers and operators to conduct online sales, purchases, and electronic payments for products and services.

2.2. Rural E-Commerce Development Models

Kaplan S E (2003) believed that e-commerce enterprises had broken the traditional transaction methods and accelerated the development speed of the industry. He also held the view that the e-commerce industry should have different development models in the initial stage, the early development stage, and the mature stage [13]. Eric Ng (2005) conducted a survey of Australian agricultural-related enterprises, gaining an in-depth understanding of their practices in the process of selecting B2B e-commerce models. He then developed a preliminary implementation framework for Australian agribusiness organizations seeking to select business-to-business (B2B) e-commerce models [14]. Wen, W. (2007) analyzed an intelligent online marketing system for agricultural products, arguing that e-commerce could be integrated with rural e-commerce. He proposed a Knowledge-based Intelligent E-commerce System (KIES) that utilized electronic maps integrated with the Global Positioning System (GPS) [15]. Jalali, A. A., et al. (2011) employed a mixed-method approach integrating quantitative and qualitative methodologies to examine the development trends and practical experiences of e-commerce infrastructure in rural Iran. The study proposed a tailored practical model for e-commerce implementation specifically designed for rural Iranian contexts [16]. Taniya Ferreira (2017) found that the rapid development of the Internet can improve users’ work efficiency and meet people’s living needs through their research. They particularly pointed out that modern and efficient e-commerce business models can be directly applied to rural e-commerce [17]. Sun, M., et al. (2021) employed correlation, cluster, and regression analyses to demonstrate that both economic and environmental factors are critical to e-commerce business models. The study further noted that the sustainability of such models enables companies to gain a leading position within the framework of global sustainable development [18].

2.3. Influencing Factors of Rural E-Commerce Development

Existing research has explored multiple dimensions of factors influencing rural e-commerce development, encompassing both analyses of hardware conditions such as infrastructure and technological applications and examinations of institutional environments and stakeholder behaviors.

Nishida (2014) employed a method combining qualitative and quantitative approaches to analyze the influencing factors of the development of rural e-commerce in Japan’s prefectures. The results showed that the rural population, the degree of perfection of infrastructure, the local economic development status, and the level of human capital all have a certain impact on the development of rural e-commerce, and these are all significantly positive relationships [19]. Rahayu, R., and Day, J. (2015) examined the development of rural e-commerce in developing countries using Indonesia as a case study, finding that technical infrastructure, organizational environment, and individual-level factors significantly influenced the advancement of rural e-commerce [20]. Maribel P (2016) found that the advancement of artificial intelligence technology can not only make the business processes of rural e-commerce efficient but also significantly improve the application level of rural e-commerce [21]. Vyas, A., et al. (2017) analyzed the challenges confronting India’s e-commerce industry, arguing that advancing rural e-commerce requires not only a focus on improving product quality but also proactive efforts to gather market intelligence, solicit customer feedback, and address client needs [22]. Cano, J. A., et al. (2022) analyzed 99 articles published in Scopus from 2021 to 2022 that focused on logistics, e-commerce, and sustainability. They identified urban logistics, urban transportation, and environmental impacts as key factors influencing e-commerce development [23]. Zennaro, I., et al. (2022) argued that e-commerce, increasingly prevalent as a sales channel in global markets, has transformed the role of logistics within supply chains. They analyzed how logistics influences e-commerce development through aspects such as supply chain network design (SCND), outbound logistics (OL), warehousing (WR), and IT and data management (E-IT) [24].

Guo, N., and Chen, H. (2022) defined the connotation of high-quality development for rural e-commerce in China first and then constructed an evaluation index system for such development. Employing the entropy method and an obstacle factor diagnosis model, they analyzed the level of high-quality development and influencing factors of China’s rural e-commerce under the new development philosophy. Their findings indicate that the overall development quality of China’s rural e-commerce shows an upward trend, though coordination and innovation remain the primary obstacles to its high-quality advancement [25]. Wang, Y., and Zhang, Z. (2023) analyzed the willingness to pursue digital transformation in rural e-commerce and its influencing factors using data from a survey on digital upgrading of rural e-commerce in Zhejiang Province. Employing the Theory of Planned Behavior (TPB) and Structural Equation Modeling (SEM), their study revealed that behavioral attitude is the primary determinant of such willingness [26]. Wang, L., et al. (2024) developed a measurement system for rural e-commerce development capabilities centered on readiness, utilization, and impact. Using a Panel Vector Autoregression (PVAR) model, they first identified key influencing factors and then analyzed the spatiotemporal dynamic characteristics of rural e-commerce development capabilities and inter-provincial interaction relationships through Exploratory Spatiotemporal Data Analysis (ESTDA) [27].

2.4. The Relationship Between Rural E-Commerce and Farmer Income Growth

International studies have highlighted the income-enhancing effects of e-commerce in rural contexts. Wilson, P. (2000) pointed out that geographical factors pose bottlenecks to the economic development in rural areas. However, the introduction of rural e-commerce can create a new sales approach, thereby widening the channels for income growth, increasing farmers’ opportunities to earn more income, and consequently driving the rural economy to reach a new height [28]. Jensen, R. (2007) investigated the impact of digital technologies on fishermen in southern India, arguing that e-commerce adoption enables farmers to seize greater market arbitrage opportunities. The study further noted that this occurs not only by expanding access to price-disparity markets but also by mitigating costs and risks inherent in agricultural trade, thereby driving income growth for rural producers [29]. Goyal, A. (2008) conducted market research in India, revealing that the widespread adoption of the Internet accelerated information dissemination. As a result, farmers were able to sell soybeans at higher prices, thereby increasing their incomes [30]. Aker, J. C. (2010) investigated the relationship between mobile phone use and agricultural markets in Niger, finding that Internet access enables farmers to obtain abundant and actionable market information. This, in turn, facilitates the adoption of new agricultural technologies, enhances productivity, and boosts farmers’ incomes [31]. Courtois, P., and Subervie, J. (2015) investigated farm-level agricultural sales in sub-Saharan African countries, developing a bargaining model between farmers and traders at the farm gate. Their study revealed that selling produce through e-commerce platforms can increase farmers’ income [32]. Shimamoto, D., et al. (2015) examined the factors influencing rice sales prices in Cambodia, arguing that the widespread adoption of the Internet and mobile phone usage can consistently enhance market sales and prices of agricultural products [33]. Surjeet Singh Dhaka (2017) noted that rapid Internet development allows rural e-commerce users to capture greater benefits from digital transactions [34]. M Hafiz Yusoff (2019), analyzing 1060 Malaysian respondents, confirmed a positive correlation between rural e-commerce platform use and income growth [35].

Chinese scholars have conducted extensive research on the impact of e-commerce on farmers’ income, thoughtfully integrating findings with China’s unique policy practices. Gao, W. et al. (2010) conducted a study and discovered that e-commerce platforms provided farmers with channels to access price information and production demands. This enabled farmers to reduce their cost expenditures, thereby leading to an increase in their incomes [36]. Qin, Z., et al. (2019) examined the impact of rural e-commerce development on household income using difference-in-differences (DID) and fixed-effect models. Their findings show that e-commerce poverty alleviation policies have significant positive effects on households’ average pure income, average agricultural operating income, and average non-farm income. The study also reveals that these policies can transform rural households’ income structure through poverty alleviation mechanisms in the short term [37]. Tang, W., and Zhu, J. (2020) examined the impact of e-commerce in developing countries using China’s “Taobao Villages” as a case study. They argued that e-commerce, increasingly prevalent in these contexts, has exerted profound impacts on economic growth and daily life, and highlighted “Taobao Villages” as an effective means of rural revitalization and narrowing urban–rural disparities [38]. Qiu, H., et al. (2024) examined the impact and mechanisms of e-commerce operations on farmers’ income using ordinary least squares (OLS) and mediating-effect models, based on data from the 2020 China Rural Revitalization Survey. Their findings show that e-commerce operations promote income growth by enhancing information access, reducing operational costs, and improving financial support for farmers. The heterogeneity analysis revealed that the effect of e-commerce on farmers’ income varies across income compositions, education levels, and regional distributions [39]. Guan, X., et al. (2024) examined the effects of rural e-commerce on farmers’ income and intra-rural income inequality using quantile regression methods, based on data from the 2021 China Rural Revitalization Survey (CRRS). Their findings indicate that rural e-commerce effectively enhances farmers’ income levels, with particularly pronounced income growth effects for low-income households [40]. Liu, M., et al. (2021) analyzed the factors influencing farmers’ decisions to adopt e-commerce using an endogenous switching regression (ESR) model, based on rural household data from Shandong, Henan, and Shaanxi provinces in China in 2019. The study argued that farmers derive benefits from e-commerce adoption, with a heterogeneity analysis revealing that rural households located closer to towns gain significantly more returns from e-commerce than those farther away [41]. Sun, X., et al. (2021) posited that the integration of e-commerce with rural industries expands the market reach and sales channels for agricultural products, extends the original industrial chain of rural products, enhances their added value, and thereby promotes sustainable economic development in rural areas [42]. Li, G., and Qin, J. (2022) conducted an empirical analysis of panel data from 57 counties in Zhejiang Province with Taobao Villages (TB-villages) between 2010 and 2018. Using the continuous difference-in-differences (DID) method, they examined the impact of TB-villages on rural residents’ income and its underlying mechanisms. The study found that TB-villages have a significant positive effect on rural residents’ income: specifically, a 1% increase in the proportion of TB-villages is associated with a 3.6% growth in rural residents’ average pure income [43]. Zhu, C., and Luo, W. (2024) analyzed the spatiotemporal evolution and driving factors of coupling coordination between rural revitalization and rural e-commerce in Hunan Province using observational data from its 14 prefecture-level cities spanning 2013–2021. Employing a coupling coordination model and spatial econometric models, their study revealed that the degree of coordinated development between rural revitalization and e-commerce has gradually increased, though regional disparities in this coupling coordination persist [44]. Li, X., et al. (2025) used fuzzy-set qualitative comparative analysis (fsQCA) to investigate how the length, breadth, and depth of modern agricultural industrial systems affect farmers’ wage income, operating income, property income, and transfer income. Their findings reveal that farmers’ income levels are influenced by the combined effects of the length, breadth, and depth of these industrial systems [45]. Wang, Y., and Wu, Y. (2025) utilized panel data from 30 Chinese provinces spanning 2011–2020 to examine how rural e-commerce development affects rural employment quality in the context of the digital economy. Their findings indicate that rural e-commerce enhances employment quality by improving industrial structure, and they propose starting with farmer education to boost literacy levels, thereby amplifying the positive effects of e-commerce on employment. The study also highlights that rural e-commerce has a more pronounced impact on central and western regions compared to eastern areas [46].

Overall, scholars both domestically and internationally have conducted extensive research on how rural e-commerce development boosts farmers’ income, deepening our understanding of this topic. However, most studies in the Chinese context rely on micro-level data and are geographically concentrated in eastern China, where rural e-commerce is more developed, limiting the generalizability of their findings. This paper addresses this gap by adopting a macroscopic perspective to analyze national-level rural e-commerce development, investigate regional disparities in income-increasing effects, and identify actionable strategies to improve e-commerce infrastructure in lagging areas, thereby contributing to the reduction in regional economic inequalities. Representative studies are presented here, as shown in

Table 1. Summary of existing representative studies.

Author (Year)	Research Objectives	Research Methods	Research Indicators	Research Conclusions
Goyal (2008) [30]	Impact of Internet penetration on agricultural product prices	Market Research (India)	Speed of Information Dissemination, Selling Prices of Agricultural Products	Internet penetration accelerates information flow, enabling farmers to sell soybeans at higher prices.
Aker (2010) [31]	Relationship between mobile phone usage and farmers’ income	Empirical Study (a Survey of Farmers in Niger)	Market Information Access, Technology Adoption Rate, Productivity	Mobile phone usage enhances farmers’ access to market information, thereby boosting productivity and increasing income.
Courtois and Subervie (2015) [32]	Relationship between e-commerce platform usage and farmers’ income growth	Bargaining Model (Sub-Saharan Africa)	Bargaining Power, Selling Prices	Selling agricultural products through e-commerce platforms enhances farmers’ income.
M Hafiz Yusoff (2019) [35]	Association between e-commerce platform usage and income growth	Survey Research (1060 Respondents in Malaysia)	Platform Usage Frequency, Income Growth Rate	Rural e-commerce platform usage exhibits a significant positive correlation with income growth.
Qiu et al. (2024) [39]	Mechanisms of e-commerce-driven income growth and heterogeneous effects	Ordinary Least Squares (OLS) Regression and Mediation-Effect Model	Information Access, Operational Costs, Financial Support	E-commerce promotes income growth via three pathways—information access, cost reduction, and financial support—but its effects vary with education levels and regional disparities.
Guan et al. (2024) [40]	The impact of e-commerce development on farmers’ income	Quantile Regression Analysis	Income Levels, Gini Coefficient	E-commerce significantly boosts the income of low-income farmers.
Liu et al. (2021) [41]	Determinants of farmers’ e-commerce adoption and its impact on income	Endogenous Switching Regression (ESR) Analysis	Geographic Location (Distance to Towns), Adoption Decisions	Farmers located closer to towns benefit more from e-commerce, with geographic location emerging as a key determinant of income disparities among adopters.
Li and Qin (2022) [43]	The impact of ‘Taobao Villages’ on farmers’ income	Continuous Difference-in-Differences (DID) Analysis	Proportion of Taobao Villages, Growth Rate of Farmers’ Net Income	A 1% increase in the proportion of Taobao Villages leads to a 3.6% growth in farmers’ net income.
Li (2022) [4]	Analysis of rural e-commerce development levels and regional disparities	Analytic Hierarchy Process (AHP) and Cluster Analysis	Infrastructure, Talent, Digital Governance	China’s rural e-commerce exhibits a regional divide characterized by stronger development in southern and eastern regions compared to northern and western areas.
Li et al. (2025) [45]	The impact of modern agricultural industrial systems on farmers’ multidimensional income	Fuzzy-Set Qualitative Comparative Analysis (fsQCA)	Wage Income, Operating Income, Property Income	Farmers’ income is influenced by the combined effects of the length, breadth, and depth of agricultural industrial systems, necessitating multidimensional coordinated optimization.

.

3. Research Methodology

3.1. Literature Research Method

We systematically collected domestic and international research literature on the relationship between rural e-commerce and farmers’ income growth, covering research perspectives such as the conceptualization of rural e-commerce, development models, influencing factors, evaluation indicators, and the nexus with farmers’ income. After critically reviewing existing studies, we identified the research direction of this paper: examining the role of rural e-commerce development in boosting rural residents’ income at a national macro level, conducting an in-depth analysis of how rural e-commerce impacts income growth across different regions in China, and exploring strategies to enhance e-commerce development in relatively lagging areas. This study addresses gaps in previous research, thereby offering a more comprehensive understanding of the topic.

3.2. Empirical Analysis Method

This paper took 31 provinces (municipalities directly under the Central Government and autonomous regions) across the country as the research objects, and the research period was from 2011 to 2023. The number of sample observations was 403 (31 × 13 = 403). Since the time span is shorter than the number of research objects, and each individual was observed the same number of times with the same observation time interval, the data used in this study are short balanced panel data [47]. Therefore, we first constructed an evaluation index system for the development of rural e-commerce and used the entropy method to calculate the comprehensive index of the development of rural e-commerce in each province (municipality directly under the Central Government and autonomous region), so as to measure the development level of rural e-commerce in China. In the regression analysis stage, in order to ensure the stationarity of the data and avoid spurious regression caused by non-stationary data, we first conducted a data stationarity test using the Harris–Tzavalis test (HT test). Immediately afterward, in order to identify the unstable factors existing in the model and ensure the reliability of the analysis results, we used the Variance Inflation Factor test (VIF test) to check whether there is serious multicollinearity among various variables. On this basis, we carried out the F-test and the Hausman test to select an appropriate regression model for the analysis in this paper. According to the test results, we chose the fixed-effect model to analyze the relationship between the development of rural e-commerce and the income of rural residents [48]. In addition, considering the differences in the development of e-commerce in different rural regions of China, we also conducted a regional heterogeneity test on the development of rural e-commerce and farmers’ income in the eastern, central, western, and northeastern regions of China. Finally, in order to ensure the scientificity and persuasiveness of the regression analysis results, we carried out a robustness analysis by replacing the explained variables and explanatory variables.

4. Analysis of the Influencing Mechanism and Research Hypotheses

The rapid advancement of digital technologies has provided a catalyst for rural e-commerce development, enabling agricultural product transactions to transcend geographical constraints, diversify transaction methods, and create new entrepreneurial and employment opportunities for farmers. Moreover, rural e-commerce has spurred the rise of related industries such as tourism and logistics, significantly boosting rural household incomes [49,50], as shown in Figure 1.

First, rural e-commerce development expands agricultural product distribution channels. By overcoming traditional spatial limitations, rural e-commerce has emerged as a critical sales channel, enabling previously unsold local produce to access national and international markets. This transformation has created tangible economic opportunities through entrepreneurship and employment for rural residents.

Second, rural e-commerce drives industrial transformation and upgrading. Historically dominated by traditional farming and primary processing with low-value-added products, rural industries now benefit from e-commerce-driven innovation. High-value-added and processed agricultural products have become increasingly prevalent, enhancing product quality and market competitiveness while generating substantial economic returns for farmers.

Third, rural e-commerce fosters economic diversification. The growth of online agricultural sales has increased demand for logistics services, creating income-generating opportunities. Concurrently, advancements in digital financial services and rural tourism—facilitated by e-commerce platforms—have diversified income sources. Additionally, rural e-commerce has attracted educated youth and migrant workers back to their hometowns, leveraging their skills to promote local agricultural products, build regional brands, and inject vitality into rural economies through innovative business models.

After conducting the analysis of the mechanisms through which rural e-commerce development promotes farmers’ income growth, and drawing on previous studies exploring the relationship between the two, we hypothesize that rural e-commerce development positively influences farmers’ income growth in China [51]. However, given regional disparities in infrastructure and resource endowments—where eastern regions enjoy superior development conditions compared to central, western, and northeastern regions—we further hypothesize that the income-increasing effects of rural e-commerce will exhibit significant regional heterogeneity [48,52].

5. Model Specification and Variable Definitions

5.1. Model Specification

Our research focuses on the relationship between the development of rural e-commerce and the increase in farmers’ income. We also analyze the heterogeneity of the impact of rural e-commerce development on farmers’ income growth in different regions. When examining the influence of rural e-commerce development on the income of rural residents, regression analysis is a commonly used and effective method. This is mainly attributed to the advantages of this model in revealing the relationships between variables, controlling confounding factors, quantifying the degree of influence, and evaluating policies [53].

Firstly, the relationship between the development of rural e-commerce and the income of rural residents is complex, with potentially multiple factors interacting. Regression analysis can effectively uncover this relationship. By setting the development of rural e-commerce as the independent variable and the income of rural residents as the dependent variable and constructing a regression model, the correlation between the two can be clearly presented.

Secondly, numerous factors affect the income of rural residents. Besides the development level of rural e-commerce, differences in factors such as residents’ education levels, policy support, and the construction status of village infrastructure also play roles. Regression analysis allows these control variables to be incorporated into the study, thus effectively isolating the net impact of rural e-commerce development on residents’ income. In practical analysis, evaluation factors like residents’ educational attainment, the number of family laborers, and the status of village infrastructure are included as control variables in the regression model. This way, when analyzing the impact of rural e-commerce development on residents’ income, the interference of other factors can be eliminated, and the true effect of rural e-commerce development can be accurately evaluated.

Thirdly, regression analysis can not only determine the direction of the relationship between variables but also precisely quantify the extent to which the development of rural e-commerce impacts the income of rural residents. By examining the magnitude and significance of regression coefficients, one can intuitively understand the degree to which changes in rural e-commerce development will cause changes in rural residents’ income. In specific regression results, if the regression coefficient is positive and significant, it indicates that the development of rural e-commerce has a positive and promoting effect on the income of rural residents. The specific value of the coefficient reflects the intensity of this promoting effect. Conversely, if the coefficient is negative and significant, it implies an inhibitory effect. This provides specific data support for policymakers and researchers, facilitating the formulation of more precise and effective policy measures.

Based on the impact mechanism and research hypotheses mentioned above, and referring to the research methods of previous scholars [54,55], we first conduct a baseline regression analysis to explore the relationship between the development of rural e-commerce and the income growth of rural residents without the influence of control variables. Next, we examine this relationship after incorporating a multitude of control variables. Subsequently, we carry out a heterogeneity analysis of the relationship between the development of rural e-commerce and farmers’ income growth in different regions. Finally, we replace the explained variable and the core explanatory variable to conduct a robustness analysis of the relationship between the development of rural e-commerce and farmers’ income growth. The models are as follows:

ln_pinc_it = α₀ + α₁ecsd_it + ε_it

(1)

ln_pinc_it= β₀ + β₁ecsd_it + β₂control_it + ε_it

(2)

ln_pcc_it= λ₀ + λ₁ecsd_it + λ₂control_it + ε_it

(3)

ln_pinc_it = φ₀ + φ₁cln_ecsv_it + φ₂control_it + ε_it

(4)

where the following definitions hold:

ln_pinc_it represents the natural logarithm of per capita disposable income of rural residents in province “i” in year “t”;

ecsd_it is the comprehensive index of rural e-commerce development;

control_it includes a set of control variables;

ln_pcc_it is the logarithm of per capita consumption expenditure as an alternative dependent variable;

ln_ecsv_it is the logarithm of e-commerce transaction volume as an alternative explanatory variable;

α, β, λ, φ are coefficients to be estimated;

ε_it is the error term.

5.2. Variable Definitions

Rural Residents’ Income (ln_pinc_it): Natural logarithm of per capita disposable income of rural residents in province “i” in year “t”. We referred to the approaches adopted by previous scholars and took the logarithm of the data to measure the income of rural residents [54].

Rural E-commerce Development Index (ecsd_it): Composite index reflecting the comprehensive development level of rural e-commerce in province “i” in year “t”. Although the Alibaba Research Institute published its “E-commerce Development Index” based on big data for 2013–2017, this study constructs a multidimensional evaluation system using 12 indicators and calculates the index via the entropy weight method to ensure data consistency and timeliness.

Control Variables (control_it):

Human Capital Endowment: Proxied by the proportion of rural labor with junior high school education or above.

Government Fiscal Support: Measured by local government expenditure on agriculture, forestry, and water conservancy.

Regional Economic Development: GDP per capita (logarithmized).

Urbanization Level: Proportion of urban population in total population.

Primary Industry Value Added: Share of agriculture, forestry, animal husbandry, and fisheries in regional GDP.

Financial Development: Rural household loans per capita from banking institutions.

These variables are included based on theoretical mechanisms where e-commerce interacts with human capital, policy support, and macroeconomic conditions to affect income growth.

Alternative Dependent Variable (ln_pccit): Natural logarithm of per capita consumption expenditure, serving as a robustness check for income measurement.

Alternative Explanatory Variable (ln_ecsvit): Natural logarithm of core e-commerce transaction volume (excluding tourism and service sectors), addressing potential endogeneity in the comprehensive index.

The evaluation index system for rural e-commerce development and farmers’ income growth involved in the analysis of this manuscript is shown in

Table 2. Evaluation index system for rural e-commerce development and farmers’ income growth.

System	Evaluation	Index Measurement Method	Unit
Rural e-commerce subsystem	Fixed Investment in E-commerce Services	Fixed asset investment in transportation, warehousing, and postal services across the society	CNY 100 million
	Fixed Investment in E-commerce Industry	Fixed asset investment in information transmission, software, and information technology services across the society	CNY 100 million
	Internet Penetration Rate	Ratio of the number of rural netizens to the rural population	%
	Mobile Phone Ownership	Number of mobile phones per 100 rural households at year-end	unit
	Computer Ownership	Number of computers per 100 rural households at year-end	unit
	Rural Delivery Routes	Length of routes delivering to rural households on postal sections	km
	Proportion of Post-Serviced Administrative Villages	Ratio of administrative villages with postal service to all administrative villages	%
	Optical Fiber Cable Coverage	Length of optical fiber cable lines per square kilometer	km
	E-commerce Sales Volume	Total transaction value of goods and services sold via online orders	CNY 100 million
	Number of Taobao Villages	Number of villages where active online stores account for ≥10% of local households and annual e-commerce transactions reach CNY ≥10 million	unit
	Postal Service Volume	Monetary value of total services/products provided by postal departments	CNY 100 million
	Express Delivery Volume	Total quantity of express services handled by courier companies	100 million units
Farmers’ income increase subsystem	Per Capita Disposable Income of Rural Residents	Sum of income for final consumption and savings, measuring rural residents’ living standards and purchasing power	CNY
	Per Capita Consumption Expenditure of Rural Residents	Total expenditure on rural residents’ personal and household living consumption, and collective consumption for individuals	CNY

.

6. Data Description and Descriptive Statistics

This paper selects relevant data from 31 provincial-level administrative regions in China (excluding Hong Kong, Macau, and Taiwan) during 2011–2023 for analysis. The data are sourced from the National Bureau of Statistics website, Annual Reports on the Development of Rural E-commerce in China, China Rural Statistical Yearbooks of respective years, Alibaba Research Institute, Annual Research Reports on Taobao Villages in China, and statistical bulletins on the annual national economic and social development of each provincial-level administrative region. For partially missing data, the interpolation method is adopted for calculation. Given the inconsistent measurement units and differences in magnitude across datasets, direct analysis would produce meaningless results. Thus, to ensure scientific and accurate measurement when calculating the rural e-commerce development level, all indicator data undergo standardization to eliminate dimensional effects. Additionally, since standardized original data may yield zero values—affecting subsequent calculations and analyses—this paper follows the approaches of Chen Yuhong, Xiang Fulin, etc. [56,57], applying a translation operation to standardized data. Specifically, 0.0001 is added to all standardized values. Logarithmic transformations are applied to the per capita disposable income (pinc) for measuring rural residents’ income growth, per capita GDP (edv) for regional economic development, the alternative dependent variable of rural residents’ per capita consumption expenditure (pcc), and the alternative explanatory variable of rural e-commerce transaction volume (ecsv). The rural human capital level (hrc) is measured by farmers’ average education years. Government financial support (fe) is measured by the ratio of local general budget expenditure to regional GDP. The urbanization level (ur) is measured by the ratio of the permanent urban population to the total permanent population of the region. Primary industry development (ppi) is measured by the ratio of primary industry value added to regional GDP. Financial institution support (lvr) is measured by the ratio of financial institution loans to regional GDP.

Table 3. Descriptive statistics of variables.

Variable	Obs	Mean	Std. Dev.	Min	Max
ln_pinc	403	9.4841	0.4431	8.3612	10.6687
ln_pcc	403	9.2792	0.4212	8.0539	10.3347
ecsd	403	0.1849	0.1166	0.0292	0.8304
ln_ecsv	403	7.4278	1.5907	2.1041	10.6536
hrc	403	7.7335	0.8147	3.8039	9.9552
fe	403	0.2888	0.2044	0.1050	1.3538
ln_edv	403	10.8915	0.4715	9.6818	12.2075
ur	403	0.5971	0.1292	0.2281	0.8983
ppi	403	1.3966	0.5213	0.2000	2.58000
lvr	403	1.5561	0.4717	0.6689	2.9975

shows the descriptive statistics results for each variable.

From the perspective of the dependent variable, after taking the logarithm of the per capita disposable income of rural residents, the maximum and minimum values are 10.6687 and 8.3612 respectively, indicating certain disparities in the income levels of rural residents across China. Regarding the core explanatory variable, the rural e-commerce development index reaches a maximum of 0.8304 (observed in Guangdong Province) and a minimum of 0.0292 (observed in Tibet Autonomous Region), with an average value of 0.1849. This demonstrates unbalanced rural e-commerce development in China, suggesting a need to strengthen support for e-commerce development in relatively underdeveloped regions.

In terms of control variables, the standard deviation of urbanization development level is 0.1292, and that of government financial support is 0.2044. The standard deviations of the regional economic development level, financial institution support, primary industry development level, and human resource development level are 0.4715, 0.4717, 0.5213, and 0.8147, respectively. These figures indicate relatively small variation ranges in control variables, ensuring the accuracy of subsequent regression analysis results.

7. Empirical and Regression Result Analysis

7.1. Benchmark Regression Analysis

7.1.1. Harris–Tzavalis (HT) Unit Root Test (for Data Stationarity Test)

This paper uses Stata 18 to conduct an empirical analysis of the relationship between the development of rural e-commerce and rural residents’ income nationwide, based on the panel data of 31 provincial-level administrative regions in China from 2011 to 2023. Since the analysis involves short balanced panel data, to ensure data stationarity, facilitate subsequent analysis and modeling, and avoid spurious regression caused by non-stationary data, a Harris–Tzavalis test (HT test) is first applied for data stationarity testing. The test results are shown in

Table 4. Results of HT unit root test.

Variable	ln_pinc	ln_pcc	ecsd	ln_ecsv	hrc
Statistic	0.5498	0.2913	0.2358	−0.0437	−0.0891
p-Value	0.0000	0.0000	0.0000	0.0000	0.0000
Result	stable	stable	stable	stable	stable
Variable	fe	ln_edv	ur	ppi	lvr
Statistic	0.1683	0.4085	0.3422	0.1193	0.2321
p-Value	0.0003	0.0062	0.0000	0.0000	0.0000
Result	stable	stable	stable	stable	stable

.

The results show that the p-values of all indicator data are below 0.05. Therefore, the null hypothesis is rejected at the 5% significance level, indicating that the data are stationary—justifying further model construction and relevant analyses.

7.1.2. Variance Inflation Factor (VIF) Test for Multicollinearity

To identify unstable factors in the model, ensure the reliability of analysis results, and enhance the explanatory power of the model, a Variance Inflation Factor (VIF) test is performed to examine whether severe multicollinearity exists among variables. The test results are presented in

Table 5. Results of VIF test for multicollinearity.

Variable	VIF	1/VIF
ur	8.14	0.122855
ln_edv	6.48	0.154287
fe	4.47	0.223580
hrc	3.87	0.258313
lvr	2.80	0.387754
ecsd	2.58	0.387754
ppi	1.90	0.526450
Mean VIF	4.32

.

Generally, if the Variance Inflation Factor (VIF) exceeds 10 or falls below 1, it indicates a high correlation among independent variables. In such cases, measures should be taken to reduce multicollinearity to ensure model stability and valid inferences. However, as shown in Table 4, the mean VIF value of this study is 4.32, and all variable VIF values are below 10. This indicates that there is no severe multicollinearity among variables, satisfying the model assumptions for reliable regression analysis.

7.1.3. F-Test

Panel data regression analysis typically involves three models: pooled OLS, the random-effect model, and the fixed-effect model. It is necessary to perform the F-test and Hausman test to select an appropriate model. The F-test results are presented in

Table 6. Results of F-test.

Fixed-effect (within) regression			Number of obs = 403
Group variable: year			Number of groups = 13
R-squared:			Obs per group:
Within = 0.8513			min = 31
Between = 0.9961			avg = 31.0
Overall = 0.8925			max = 31
			F(7, 383) = 313.17
corr(u_i, Xb) = 0.6160			Prob > F = 0.0000
ln_pinc	Coefficient	Std. err.	t	p > |t|	[95% conf. interval]
ecsd	0.5616161	0.081709	6.87	0.000	0.4009616	0.7222705
hrc	0.0255134	0.0140599	1.81	0.070	−0.0021309	0.0531577
fe	0.0317401	0.0605104	0.52	0.600	−0.0872341	0.1507143
ln_edv	0.541813	0.0403838	13.42	0.000	0.4624114	0.6212146
ur	0.3790538	0.1431655	2.65	0.008	0.0975651	0.6605425
ppi	0.0058702	0.0017902	3.28	0.001	0.0023504	0.00939
lvr	0.0422408	0.0229421	1.84	0.066	−0.0028675	0.0873491
_cons	2.923868	0.4107824	7.12	0.000	2.116197	3.731539
sigma_u	0.12013644
sigma_e	0.11601267
rho	0.51745724 (fraction of variance due to u_i)
F-test that all u_i = 0: F(12, 383) = 10.79			Prob > F = 0.0000

.

From the table data, it can be seen that the p-value is 0.0000, which is less than 0.05. Therefore, we reject the null hypothesis that “all fixed-effect coefficients are zero,” confirming the presence of fixed effects and invalidating the pooled regression model.

7.1.4. Hausman Test

Next, the Hausman test is conducted to determine whether fixed effects or random effects should be employed for subsequent analysis [58]. The test results are presented in

Table 7. Results of Hausman test.

Hausman FE RE, Constant Sigmamore
	Coefficients
	(b)	(B)	(b − B)	Sqrt (diag(V_b − V_B))
	FE	RE	Difference	Std. Err.
ecsd	0.5616161	0.7273939	−0.1657778	0.0206537
hrc	0.0255134	0.0384127	−0.0128993	0.0017627
fe	0.0317401	−0.045787	0.0775271	0.0099367
ln_edv	0.541813	0.8176957	−0.2758827	0.0291442
ur	0.3790538	−0.306655	0.6857088	0.0735395
ppi	0.0058702	0.0158202	−0.00995	0.001059
lvr	0.0422408	0.1544903	−0.1122495	0.0116456
_cons	2.923868	−0.0503436	2.974212	0.3135371

b = consistent under H0 and Ha; obtained from xtreg. B = inconsistent under Ha, efficient under H0; obtained from xtreg. Test of H0: difference in coefficients not systematic. chi2(7) = (b − B)’[(V_b − V_B)^(−1)](b − B) = 99.46. Prob > chi2 = 0.0000.

.

From the test results, it can be seen that the p-value is 0.000, leading to the rejection of the null hypothesis and acceptance of the alternative hypothesis that fixed effects are preferred over random effects. This indicates that the fixed-effect model should be adopted for further analysis in this study.

7.1.5. Fixed-Effect Regression Analysis

As described previously, Stata 18 is employed to analyze the relationship between rural e-commerce development and farmers’ income. The regression results are presented in

Table 8. Regression results of fixed-effect model.

	(1)	(2)
Variables	ln_pinc	ln_pinc
ecsd	1.972 ***	0.218 ***
	(0.120)	(0.0487)
hrc		0.0541 ***
		(0.0151)
fe		0.337 ***
		(0.0871)
ln_edv		0.892 ***
		(0.0332)
ur		0.865 ***
		(0.167)
ppi		0.00508 *
		(0.00262)
lvr		0.0792 ***
		(0.0144)
Constant	9.120 ***	−1.479 ***
	(0.0250)	(0.265)
Observations	403	403
Number of id	13	31
R-squared	0.409	0.984

Note: Standard errors in parentheses, *** p < 0.01, * p < 0.1.

.

Column (1) presents the impact of rural e-commerce development on farmers’ income without control variables. The regression coefficient of 1.972 (p < 0.01) indicates a significant positive effect, confirming that rural e-commerce development directly promotes income growth. Column (2) incorporates control variables including human capital, government fiscal support, and financial development. The coefficient of 0.218 (p < 0.01) suggests that a one-unit increment in the development level of rural e-commerce within China will lead to a 0.218% elevation in farmers’ income. Evidently, the development of rural e-commerce in China is capable of effectively promoting an upward trend in farmers’ income levels.

The regression analysis incorporating control variables further demonstrates that six key determinants (human capital development, government fiscal support, regional economic development, urbanization rate, primary industry value added, and financial institution lending) all exhibit statistically significant positive effects on rural household income at the 1% significance level.

The underlying mechanisms operate as follows: Enhanced human capital development, proxied by educational attainment, strengthens rural residents’ cognitive capacities and skill acquisition efficiency, thereby accelerating their technological adoption and practical application capabilities. This establishes a self-reinforcing cycle between human capital investment and income growth. Government interventions manifest through dual channels: direct fiscal allocations for rural infrastructure modernization and agricultural subsidies improve production fundamentals, while policy-guided private capital inflows stimulate the emergence of specialized agro-industries and anchor enterprises, generating substantive employment opportunities. Regional economic development emerges as the most influential driver through its dual-sector transformation effect—it not only optimizes agricultural production efficiency through technological spillovers but also induces secondary processing industries and tertiary service sectors (particularly agro-tourism), thereby creating diversified income streams. Urbanization facilitates labor market integration through rural–urban migration while simultaneously expanding agricultural product demand via urban consumption growth, effectively bridging production–supply chains. The expansion of primary industry value added reflects structural improvements in agricultural productivity, driving scale economies, vertical industry chain integration, and value-added product differentiation. Financial institutional innovations address systemic constraints through dual mechanisms: diversified financing instruments alleviate liquidity constraints, while risk-hedging products (e.g., crop insurance, futures contracts) mitigate vulnerability to climatic shocks and price volatility, thereby stabilizing income expectations.

7.2. Heterogeneity Analysis

Given regional disparities in rural e-commerce development conditions and income gaps among residents, this paper conducts regional heterogeneity analysis on the impact of rural e-commerce development on farmers’ income across China’s four major economic regions—the East, Central, West, and Northeast—based on China’s economic regional classification [59]:

Northeastern Region: Liaoning, Jilin, Heilongjiang;

Eastern Region: Beijing, Tianjin, Hebei, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong, Hainan;

Central Region: Shanxi, Anhui, Jiangxi, Henan, Hubei, Hunan;

Western Region: Inner Mongolia Autonomous Region, Guangxi Zhuang Autonomous Region, Chongqing, Sichuan, Guizhou, Yunnan, Tibet Autonomous Region, Shaanxi, Gansu, Qinghai, Ningxia Hui Autonomous Region, Xinjiang Uygur Autonomous Region.

The analysis results are presented in

Table 9. Results of heterogeneity analysis.

	East	Middle	West	Northeast
Variables	ln_pinc	ln_pinc	ln_pinc	ln_pinc
ecsd	0.0830	0.788 ***	0.219	3.243 ***
	(0.0518)	(0.231)	(0.246)	(0.302)
hrc	0.00429	0.0596 **	0.0102	0.0196
	(0.0169)	(0.0272)	(0.0225)	(0.0251)
fe	1.022 ***	1.537 ***	0.0102	0.446 ***
	(0.151)	(0.316)	(0.115)	(0.136)
ln_edv	1.130 ***	0.546 ***	0.743 ***	0.464 ***
	(0.0325)	(0.0574)	(0.0698)	(0.0862)
ur	0.248	1.990 ***	1.723 ***	1.705 **
	(0.239)	(0.387)	(0.344)	(0.804)
ppi	0.0122 *	0.0115 **	0.0111 ***	0.00952 ***
	(0.00719)	(0.00457)	(0.00423)	(0.00207)
lvr	0.153 ***	0.0526	0.0448 ***	0.115 ***
	(0.0237)	(0.0437)	(0.0164)	(0.0326)
Constant	3.653 ***	1.351 ***	0.379	2.955 ***
	(0.355)	(0.485)	(0.620)	(0.615)
R-squared	0.992	0.995	0.988	0.998

Note: Standard errors in parentheses, *** p < 0.01, ** p < 0.05, * p < 0.1.

.

The heterogeneity regression results demonstrate that the estimated coefficients for the eastern, northeastern, central, and western regions are 0.083, 3.243, 0.788, and 0.219, respectively (Table 9), indicating that digital economy penetration and e-commerce maturation—through infrastructure enhancement and industrial linkage effects—create diversified income-generation opportunities.

Notably, the northeastern region exhibits exceptional responsiveness, with each unit increase in e-commerce development corresponding to a 3.243-unit income growth. This outsized effect stems from two synergistic mechanisms:

(1)

Effective leveraging of unique agricultural endowments (premium rice, soybeans, wild fungi, and mountain herbs) through specialized e-commerce platforms;

(2)

Emergence of complementary industries including cold-chain logistics, packaging services, live-stream commerce, and influencer-driven rural tourism.

The analysis reveals a converging trend in e-commerce’s income effects across regions, despite current disparities. While eastern China’s early-mover advantage in e-commerce development shows diminishing marginal returns (attributable to market saturation), the western and northeastern regions demonstrate accelerating growth dynamics.

This convergence is driven by the following:

• Targeted infrastructure investments improving last-mile connectivity and reducing rural–urban digital divides;
• Reverse migration of skilled labor enhancing local e-commerce capabilities;
• Innovative marketing paradigms (fieldside live-streaming, geo-tagged specialty products) overcoming traditional distribution bottlenecks.

Conversely, eastern China’s mature e-commerce ecosystem faces structural constraints:

• Declining marginal utility of additional platform penetration;
• Intensified competition eroding seller profit margins;
• Market saturation in high-value agricultural segments.

This regional rebalancing effect underscores the success of China’s coordinated rural revitalization strategy, particularly its emphasis on bridging digital infrastructure gaps and fostering place-based e-commerce models.

7.3. Robustness Analysis

7.3.1. Overall Robustness Analysis

To ensure the robustness of the research findings, this paper conducts robustness tests on regression results by adopting the method of substituting dependent and independent variables, following established literature. Specifically, the robustness tests are as follows:

Dependent Variable Substitution: Per capita rural household consumption expenditure is used as an alternative measure for rural income, given its strong correlation with income levels.

Independent Variable Substitution: Rural e-commerce transaction volume—reflecting market activity and scale—is employed as a proxy for the rural e-commerce development index, after logarithmic transformation.

The robustness test results using these substituted variables are presented in

Table 10. Results of overall robustness analysis.

Variables	Replace Explained Variable	Replace Explanatory Variables
ecsd	0.315 ***
	(0.0638)
hrc	0.0384 *	0.0551 ***
	(0.0198)	(0.0144)
fe	0.613 ***	0.324 ***
	(0.114)	(0.0825)
ln_edv	0.840 ***	0.858 ***
	(0.0435)	(0.0311)
ur	1.040 ***	0.717 ***
	(0.218)	(0.154)
ppi	0.00190	0.00732 ***
	(0.00342)	(0.00252)
ln_ecsv		0.0459 ***
		(0.00600)
lvr	0.110 ***	0.0744 ***
	(0.0188)	(0.0137)
Constant	−1.180 ***	−1.341 ***
	(0.347)	(0.235)
Observations	403	403
R-squared	0.973	0.985
Number of id	31	31

Note: Standard errors in parentheses, *** p < 0.01, * p < 0.1.

.

After the substitution of the independent variable (rural e-commerce transaction volume), the regression coefficient for rural e-commerce’s impact on farmers’ income is 0.0459 (p < 0.01). Similarly, substituting the dependent variable (per capita rural consumption expenditure) yields a coefficient of 0.315 (p < 0.01), maintaining the significant positive relationship. Furthermore, including control variables such as human capital development, government fiscal support, and regional economic development does not alter the positive sign of the coefficients, aligning with benchmark regression results. These consistent findings across alternative specifications validate the robustness of our empirical conclusions.

7.3.2. Regional Robustness Analysis

As can be seen from the previous analysis, when using the fixed-effect model to analyze the promotion effect of rural e-commerce development on farmers’ income growth in China, the regression coefficients are significant. This result indicates that at the national level, there is a clear and significant correlation between rural e-commerce development and farmers’ income growth. In other words, the core relationship that rural e-commerce development in China promotes farmers’ income growth holds at the macro level.

However, the results of the regional heterogeneity analysis, considering the different conditions and policies of rural e-commerce development in the four major economic regions, show that the impact of rural e-commerce development on farmers’ income growth in the eastern and western regions is not significant. In the eastern region, due to its relatively developed economy and diverse industrial structure, farmers have a wide range of income sources. The marginal contribution of e-commerce to farmers’ income growth is relatively small, and its impact may be masked by other more important income channels, resulting in non-significant regression results. In the western region, the relatively backward transportation, logistics, and communication infrastructure affects farmers’ convenience and enthusiasm for participating in e-commerce, hindering the role of e-commerce in promoting farmers’ income growth.

After the replacement of the explanatory and explained variables, the regression results are significant. This result enhances the reliability and robustness of the research conclusion, indicating that the promotion of farmers’ income growth by rural e-commerce development in China has a certain degree of universality and stability.

To better reflect the influencing factors and mechanisms of rural E-commerce development and farmers’ income growth in the four major economic regions of China and to differentiate policies for promoting rural E-commerce development and farmers’ income growth, we replaced the explanatory and explained variables of rural E-commerce development and farmers’ income growth in each region and conducted a regional robustness analysis. The results are shown in

Table 11. Results of regional robustness analysis.

	Replace Explained Variable (ln_pcc)				Replace Explanatory Variables (ln_pinc)
Variables	East	Middle	West	Northeast	East	Middle	West	Northeast
ecsd	0.192 **	1.250 ***	0.915 ***	3.264 ***
	(0.0899)	(0.283)	(0.273)	(0.830)
hrc	0.0474 *	0.0167 *	0.0193 *	0.0223 *	0.0121 *	0.0375 *	0.0273 *	0.0563 *
	(0.0294)	(0.0334)	(0.0249)	(0.0692)	(0.0130)	(0.0293)	(0.0208)	(0.0576)
fe	1.657 ***	1.199 ***	0.197 *	0.164 *	0.679 ***	0.625 **	0.0132 *	0.645 **
	(0.263)	(0.387)	(0.127)	(0.373)	(0.119)	(0.309)	(0.104)	(0.305)
ln_edv	1.043 ***	0.457 ***	0.604 ***	0.234	0.926 ***	0.496 ***	0.706 ***	0.517 **
	(0.0563)	(0.0703)	(0.0773)	(0.237)	(0.0292)	(0.0708)	(0.0595)	(0.217)
ur	0.705 *	0.785 *	1.314 ***	0.734 *	0.570 ***	1.312 ***	1.602 ***	1.868 ***
	(0.415)	(0.474)	(0.381)	(2.213)	(0.186)	(0.394)	(0.313)	(1.638)
ppi	0.0412 ***	0.00889 *	0.00035 *	0.00183 *	0.0162 ***	0.00456	0.00415 *	0.00648 *
	(0.0125)	(0.00560)	(0.00468)	(0.00570)	(0.00516)	(0.00512)	(0.00407)	(0.00484)
lvr	0.139 ***	0.103 *	0.0941 ***	0.260 ***	0.102 ***	0.164 ***	0.0291 *	0.00842
	(0.0411)	(0.0535)	(0.0182)	(0.0899)	(0.0173)	(0.0371)	(0.0151)	(0.0722)
ln_ecsv					0.0780 ***	0.0241 *	0.0396 ***	0.0111 *
					(0.00855)	(0.0123)	(0.00745)	(0.0172)
Constant	−1.683 ***	3.074 ***	1.408 **	5.724 ***	−2.014 ***	1.961 ***	0.432 *	0.711 *
	(0.616)	(0.594)	(0.686)	(1.694)	(0.241)	(0.649)	(0.503)	(1.406)
Observations	130	78	156	39	130	78	156	39
R-squared	0.977	0.994	0.985	0.988	0.996	0.995	0.990	0.991
Number of id	10	6	12	3	10	6	12	3

Note: Standard errors in parentheses, *** p < 0.01, ** p < 0.05, * p < 0.1.

.

From the results of the regional robustness test, after the replacement of the explained variable, the variable of the level of e-commerce development is significant in the eastern region with a coefficient of 0.192. It is also significant in the central, western, and northeastern regions, and there are large differences in the coefficient values. This indicates that the development of rural e-commerce has different degrees and levels of significance of influence in different regions. Therefore, regional differences must be taken into account when formulating relevant policies. The variable of e-commerce human resources is significant at the 10% level in all four major regions in China. The coefficient values are small but all positive, suggesting that this variable has the same direction of influence in different regions, yet the degree of influence is relatively weak. The financial support variable is highly significant in the eastern and central regions and is significant at the 10% level in the western and northeastern regions. The coefficient values vary significantly across regions, reflecting different degrees of impact on each region. Educational attainment, urbanization, and other variables also show different degrees of significance in different regions, with certain differences in their coefficients. This demonstrates that differences in the regional economic and social environment lead to different effects of the same variable.

After the replacement of the explanatory variables, judging from the coefficients and significance levels of each variable, the impact and significance of the same explanatory variable on the explained variable vary significantly across different regions. This reflects substantial differences among China’s four major economic regions in terms of economic and social environments, industrial structures, and resource endowments, resulting in different mechanisms and degrees of influence of the same factors on economic outcomes. Some variables are highly significant in some regions, moderately significant in others, significant at the 10% level in some, and non-significant in others. This further demonstrates the characteristics of regional heterogeneity.

This result indicates that different regions have their own unique economic and social environments as well as development characteristics. It further demonstrates that regional heterogeneity is a crucial factor that cannot be ignored in the research of this manuscript. Such results provide a basis for formulating region-differentiated rural e-commerce development policies. Different regions should formulate targeted e-commerce development strategies to promote farmers’ income growth according to their own resource endowments, industrial characteristics, and development stages.

8. Conclusions and Recommendations

8.1. Conclusions

The development of rural e-commerce serves as a critical driver for accelerating agricultural modernization, advancing rural revitalization strategies, fostering rural entrepreneurship, and promoting income growth for farmers. Using China’s 31 provinces (municipalities/autonomous regions) as the research sample, this study first constructs an evaluation index system for rural e-commerce development. The entropy method is then applied to calculate comprehensive development indices for each province. Panel data regression models are subsequently established, with per capita disposable income of rural residents as the dependent variable and the rural e-commerce development index as the core independent variable, to analyze the relationship between rural e-commerce and farmers’ income. Considering regional disparities in e-commerce adoption, regional heterogeneity tests are conducted across China’s East, Central, West, and Northeast regions. Finally, robustness analyses are performed by substituting dependent and independent variables to ensure scientific rigor and validity. Key findings are detailed in the following subsections.

8.1.1. Regional Disparities in Rural E-Commerce Development

Longitudinal analysis reveals significant provincial disparities in rural e-commerce development across China from 2011 to 2023. The eastern region maintains technological leadership with a composite development index of 0.8303 in Guangdong Province (2023), contrasting sharply with Tibet’s 0.1454. This 5.7-fold differential underscores substantial regional gaps influenced by geographical constraints, transportation infrastructure, and digital penetration levels.

Notably, coastal provinces (Zhejiang, Jiangsu, Guangdong) demonstrate sustained growth trajectories, while Hainan lags at 0.1787 despite incremental progress. Central China exhibits relative equilibrium, with Henan Province emerging as a regional leader (margin advantage < 5%). Western China presents divergent patterns—Sichuan and Chongqing show accelerated growth rates, whereas broader western regions face developmental constraints from complex topography and sparse populations. In northeastern China, Liaoning Province leads industrial transformation efforts, leveraging historical manufacturing advantages.

8.1.2. Income Enhancement Effects of Rural E-Commerce

The development of rural e-commerce in China has a positive impact on the growth of farmers’ income to a certain extent. Through regression results, when there are only core explanatory variables, the regression coefficient is 1.972 and is significant at the 1% confidence level. After adding control variables such as the level of human resources, the intensity of government financial support, the level of regional economic development, the level of urbanization, and the loan support from financial institutions, the regression coefficient is 0.218, still significant at the 1% confidence level. In the regional heterogeneity analysis of the eastern, central, western, and northeastern regions, the positive impact of rural e-commerce development on farmers’ income increase still exists. This indicates that the development of rural e-commerce in China can indeed increase farmers’ income.

8.1.3. Regional Heterogeneity Mechanisms

The development of rural e-commerce in China has different degrees of impact on the income of rural residents in different regions. The pulling effect in the northeastern region is higher than that in the central region, and the pulling effect in the central region is higher than that in the eastern region. This is because the disposable income levels of rural residents in the northeastern and central regions were originally at a below-average level, so the pulling effect of rural e-commerce development is relatively obvious. Although both the western region and the northeastern region are underdeveloped areas, the conditions for the development of rural e-commerce in these two regions are different. The northeastern region has convenient transportation. As a former heavy-industrial base, it has a relatively strong industrial foundation and still has certain development strength and potential as a whole. However, the western region has a vast territory, complex geology, inconvenient transportation, a sparse population, and a backward economy. The development of e-commerce started late, and the conditions for developing e-commerce are relatively weak, resulting in a low level of rural e-commerce development and an insignificant pulling effect on the increase in rural residents’ income.

8.2. Recommendations

8.2.1. Formulate Differentiated Policies for Rural E-Commerce Development

For provinces in eastern China with relatively advanced rural e-commerce development—such as Guangdong, Zhejiang, and Jiangsu—policies should encourage innovation in business models and provide financial support, tax incentives, and other policy measures to promote the high-end and internationalized development of rural e-commerce.

For central, western, and northeastern regions—especially those with relatively lagging development—the policy focus should prioritize infrastructure construction to comprehensively address shortcomings in rural e-commerce development. In western regions, increase investment in transportation, network communication, and other infrastructure to improve network coverage and broadband speed. In remote areas like Tibet and Qinghai, accelerate the construction of 5G base stations and fiber-optic networks to create foundational conditions for rural e-commerce. In northeastern regions, address the challenges of winter (low-temperature environments) by promoting the construction of temperature-controlled systems and building competitive modern agricultural supply chain systems. These efforts will ensure agricultural product quality and enhance logistics efficiency, thereby supporting sustainable income growth for farmers.

8.2.2. Innovate E-Commerce Application Scenarios

Building on local characteristics and resource endowments, rural areas should leverage and exploit their rich cultural heritage, folk customs, and specialty products to distill core elements for brand building. Examples include Tibet’s unique religious beliefs and mysterious culture, Qinghai’s long history and rich customs, Hebei’s scenic mountains and rivers, and Xinjiang’s snow-capped mountains and vast grasslands. In northeastern regions, leveraging the advantage of their old industrial base, efforts should be made to integrate deep processing of agricultural products with e-commerce, creating iconic brands such as Northeast black soil rice.

In western regions with rich rural tourism resources—such as Guizhou—combining agricultural product sales with folk cultural experiences can create new models for integrated rural industrial development. These initiatives inject fresh vitality into the rural economy, broaden channels for rural residents to increase their income and prosper, and narrow regional development gaps.

8.2.3. Promote Digital Empowerment of Rural E-Commerce Development Promptly

In light of the current development of the digital economy, efforts should be intensified to promote and enhance the application of smart agricultural equipment. This will accelerate the digital upgrading of the entire agricultural industry chain. Additionally, it is crucial to facilitate the integration of technologies such as artificial intelligence, big data, mobile payment, and blockchain with e-commerce and vigorously develop rural e-commerce. Actively cultivate the e-commerce development model of “agricultural base + live-streaming with goods”. Build a number of live-streaming bases for special agricultural products to comprehensively promote and sell the special agricultural products in rural areas, drive the development of related industries, and solve the employment problems in rural areas.

8.2.4. Build Rural E-Commerce Human Capital

Regarding the impact of human resources factors on the development of rural e-commerce, a combination of training and talent introduction should be adhered to. Training in e-commerce application skills should be strengthened to cultivate a professional e-commerce-skilled workforce. Provide rural residents with training related to agricultural knowledge and e-commerce, aiming to improve farmers’ cultural qualities and e-commerce operation skills, and foster e-commerce-skilled talents.

Establish e-commerce demonstration stores to encourage more market players to utilize e-commerce platforms, and promote the in-depth application of e-commerce technologies in all aspects of agricultural production, agricultural product sales, and services, thereby enhancing agricultural production efficiency.

Guide college students and migrant workers to return to their hometowns for entrepreneurship and employment, which can contribute to the stable development of rural e-commerce and increase farmers’ income levels.

9. Research Limitations and Future Prospects

9.1. Research Limitations

9.1.1. Data Collection and Processing

The data used in this paper mainly come from the official website of the National Bureau of Statistics, relevant reports, and yearbooks. The breadth and depth of the data may be limited. For some missing data, the interpolation method was employed. Although this ensures data integrity to a certain extent, it may introduce errors and affect the accuracy of the research results. Moreover, this paper analyzes the data from rural areas spanning from 2011 to 2023. The relatively short time span makes it difficult to capture the long-term development trends of rural e-commerce.

9.1.2. Selection of Evaluation Indicators

When constructing the index system for rural e-commerce development, 12 indicators were selected, and the entropy weight method was used to calculate the comprehensive index. However, it is possible that not all key factors influencing rural e-commerce development were comprehensively covered, resulting in certain limitations in the explanatory power of the model.

9.1.3. Empirical Analysis

When establishing the regression model to analyze the relationship between rural e-commerce and farmers’ income growth, although variables such as the level of human resources, government financial support, and financial support were controlled, there may still be omitted variables, leading to biased estimation results. Additionally, in the regional heterogeneity analysis, although the country was divided into four major regions (the East, the Central, the West, and the Northeast), there may be significant differences among different provinces within the same region. As a result, it may not be possible to accurately understand the specific situation of rural e-commerce development and farmers’ income growth in different areas.

9.2. Research Prospects

9.2.1. Expanding Data Dimensions

We propose expanding data collection channels by incorporating primary data sources such as surveys and field investigations to obtain richer, more detailed grassroots information. Additionally, extending the time span of the research period will facilitate tracking the long-term trajectory of rural e-commerce development, analyzing its enduring influencing factors, and forecasting potential trends. This broader temporal scope will enhance the study’s ability to capture dynamic, long-term impacts on farmers’ income.

9.2.2. Optimizing Models and Indicator Systems

We will further explore the potential factors that influence the development of rural e-commerce and the increase in farmers’ income. We will improve the indicator system by incorporating evaluation indices such as farmers’ willingness to start businesses and tax incentives. By doing so, we aim to enhance the explanatory power of the model and make the research findings more reliable.

Meanwhile, we will take into account the differences among different provinces and counties within the same region to analyze the heterogeneity of the relationship between the development of rural e-commerce and the increase in farmers’ income. Based on this analysis, we will formulate more targeted regional development strategies to better promote the balanced development of rural e-commerce in various regions and the sustained growth of farmers’ income.