The nonlinear relationship between digital economy and environmental pollution: evidence from 287 cities in China

In the context of global digital transformation and the transition toward green development, the environmental impact of the digital economy requires further systematic investigation. Utilizing panel data from 287 Chinese cities (2013–2022) and leveraging big data techniques, this study examines the nonlinear effects of the digital economy on environmental pollution by employing panel threshold regression and machine learning. Our results reveal a significant inverted U-shaped relationship, alongside pronounced technological heterogeneity. Specifically, green digital technologies monotonically suppress pollution without a threshold effect, while non-green digital technologies exhibit a more marked inverted U-shaped pattern. Significant regional heterogeneity is also evident, with a lower threshold value in more developed regions compared to less developed ones. Mechanism analysis shows that the initial phase of digital infrastructure expansion aggravates pollution, whereas in later stages, technology-driven industrial upgrades facilitate energy conservation and emission reduction. These findings provide empirical support for formulating staged and targeted policies to steer the digital economy toward sustainable development.

1 Introduction

In recent years, China’s economy has maintained rapid growth, driven in part by accelerated urbanization. However, this process has also accelerated the overexploitation of resources, leading to numerous environmental pollution issues (Adams and Klobodu, 2017; Liang et al., 2019; Zhang et al., 2024). China’s economy has transitioned from a phase of high-speed growth to one of high-quality development. Under constraints of natural resources, transforming the economic growth model is an inevitable choice for addressing the challenge of low ecological quality in sustainable development (Qu and Hao, 2022). With the rapid rise of new-generation information technologies, the digital economy has become a new engine for China’s high-quality economic development (Li, 2019; Jiao and Sun, 2021). While effectively enhancing production efficiency, the development of the digital economy also consumes substantial energy and generates pollutant emissions. However, a “green paradox of digitalization” emerges: while effectively enhancing production efficiency, the development of the digital economy also consumes substantial energy and generates pollutant emissions, potentially becoming a significant source of high energy consumption and high emissions itself.

The global climate crisis and the digital technology revolution coincide to create new avenues for greening the digital economy. On one hand, technologies such as the internet, artificial intelligence, and blockchain contribute to energy conservation, efficiency gains, and environmental monitoring, providing new channels for reducing environmental pollution (Huang et al., 2023). On the other hand, the high energy consumption of digital infrastructure itself and the issue of electronic waste may exacerbate environmental pressures. This duality suggests that the relationship between the digital economy and environmental quality is likely not a simple linear one, but rather involves complex nonlinear dynamics.

Confronted with the accelerating green economic transition and the opportunities presented by digital technology, exploring the mechanisms between the two is crucial (Jiang and Deng, 2022). The existing literature features three perspectives on their relationship: “promotion theory,” “inhibition theory,” and “nonlinear theory.” However, several limitations remain: most studies rely on single indicators, failing to distinguish the heterogeneous effects of green versus non-green technologies; moreover, most statistical methods are based on linear assumptions, which contradict the nonlinear characteristics of the real world and offer limited capacity for investigating complex nonlinear relationships (Ma et al., 2020). There is a need to explore how to maximize the positive impact of the digital economy on green development.

To address existing limitations, this study makes three key contributions. First, it constructs a multi-dimensional digital economy index at the city level and innovatively distinguishes between green and non-green digital technologies to examine their heterogeneous environmental impacts. Second, it integrates panel threshold models and machine learning methods to systematically capture nonlinear and threshold effects, moving beyond linear assumptions. Third, it proposes and validates a dynamic mechanism in which the digital economy influences pollution through an “infrastructure energy consumption effect “followed by a “green innovation effect,” deepening understanding of the digital-green paradox. These approaches provide new empirical and theoretical insights for designing targeted digital green policies.

Against this backdrop, this study aims to address the following questions: (1) Does the digital economy significantly impact environmental pollution? (2) What are the pathways through which the development of the digital economy affects local environmental issues? (3) Do these pathways exhibit threshold effects or nonlinear effects? If such effects exist, how can threshold values be determined for different stages? (4) Are these pathways constrained by factors such as technology and geography?

The remainder of this study is structured as follows: Section 2 reviews the literature. Section 3 proposes the theoretical hypotheses. Section 4 discusses the data and methodology. Section 5 presents the empirical findings. Section 6 provides a further analysis of the results. Section 7 concludes with policy recommendations.

2 Literature review

2.1 The promoting effect of the digital economy on green development

Multiple studies confirm that the development of the digital economy contributes to improved urban environmental quality and reduced pollution and carbon emission intensity. Quasi-experimental evidence from studies leveraging the “Broadband China” demonstration city policy as an exogenous shock shows that digital economy advancement significantly lowers PM2.5 concentrations and the Air Quality Index (Tian et al., 2024), while also reducing both the total volume and intensity of CO₂ emissions (Feng et al., 2023; Liu et al., 2023). This negative correlation between the digital economy and environmental pollution remains robust across broader empirical settings (Yuan et al., 2024). For instance, Zhang (2023), constructing a digital economy index via principal component analysis, demonstrates that it significantly reduces urban environmental pollution—proxied by industrial sulfur dioxide emissions—through mechanisms of human capital enhancement and industrial agglomeration.

Digitalization spurs green technological innovation, which in turn suppresses pollution and carbon emissions. This mechanism has been consistently validated within spatial econometric, threshold, and Difference-in-Differences mediation frameworks. The digital economy fosters green research and development through “innovation spillovers,” an enabling effect particularly salient in specific digital contexts such as smart cities (Yang et al., 2022).

Furthermore, digitalization drives the structural transformation from secondary to tertiary industries, promotes servitization of production, and facilitates industrial upgrading, thereby significantly enhancing environmental performance. Research by Liu et al. (2025) indicates that the digital economy substantially boosts energy efficiency through intermediary pathways like industrial structure optimization.

Digital technologies and data governance also enhance energy efficiency and promote the substitution of clean energy. Studies by Zhang et al. (2022), Shahbaz et al. (2022), and Xu et al. (2022a) show that digitalization considerably improves corporate energy savings and accelerates the energy transition. At the industrial level, technologies like the Industrial Internet and big data analytics significantly reduce energy consumption per unit of output (Huang et al., 2023).

Finally, digitalization improves environmental quality by enabling cross-regional resource allocation and strengthening government environmental governance—encompassing monitoring, enforcement, and institutional capacity. Huang and Lei (2021) posit that environmental regulation can stimulate advanced industrial structures and low-carbon energy consumption via innovation compensation and investment screening effects. In environmental governance, technologies like blockchain aid in tracking transaction-based carbon emissions, providing a technical foundation for precise regulation.

2.2 The inhibitory effect of the digital economy on green development

In the early stages of digitalization or under inadequate supporting conditions, the expansion of the digital economy may temporarily increase pollution or emissions. Threshold regression analyses identify a “rise-then-fall” pattern, where digital expansion below the threshold level is accompanied by rising emissions (Lan and Ma, 2025). This phenomenon can be attributed to the substantial energy and resource consumption inherent in digital economic activities. For example, Jiang et al. (2021) estimate that Bitcoin blockchain operations in China would generate massive carbon emissions. From a derived-effects perspective, consumption upgrading fostered by the digital economy - such as cross-border e-commerce - increases logistics and transportation volume, thereby driving up energy consumption.

Economic development levels and institutional/market environments serve as “amplifiers” for the pollution-reduction effects of the digital economy. At lower developmental or marketization levels, the environmental benefits of digital economy development are significantly weakened, with risks of negligible or even temporarily adverse impacts. As development levels improve, these pollution-reduction effects become substantially stronger (Yang et al., 2022).

Panel threshold models focusing on CO₂ emissions show that when the digital economy index remains below the threshold, digitalization progress correlates with increased emissions. This upward pressure can only be partially counteracted through technological advancement channels. Statistical evidence does not support the existence of a dual-threshold model (Sun et al., 2024).

During early development stages, multiple factors contribute to short-term environmental pressures: rising energy consumption from data centers and communication networks (Zhao et al., 2023), increased logistics intensity driven by e-commerce and instant delivery, and rebound effects in consumption and travel resulting from digitalization. As technological efficiency improves, energy structures transform, and governance capacity strengthens, these adverse effects are gradually offset and eventually converted into net environmental improvements. This explanation aligns with the segmented evidence presented by Lan and Ma (2025). Overall, the “inhibitory effect” predominantly emerges during low digital maturity stages or below established thresholds, but can be reversed through green technological progress and optimization of industrial and energy structures.

2.3 The nonlinear relationship between the digital economy and green development

Early research on the relationship between the digital economy and environmental pollution primarily focused on whether its role was “promoting” or “inhibiting,” leading to seemingly contradictory conclusions. To reconcile these differences, the “nonlinear theory” has gradually emerged as the prevailing consensus. Regarding environmental pollution or carbon emissions, the panel threshold analysis consistently identifies a single threshold, indicating that the marginal environmental effects of the digital economy undergo structural changes across different developmental stages. When marketization crosses the threshold (e.g., the single marketization threshold reported by Yuan et al. (2024)), the pollution-reducing effects of the digital economy become more pronounced. Threshold tests for CO₂ emissions reveal a significant single threshold but no significant dual threshold (at the 10% level), suggesting a more likely “single inflection point” structure within the sample period (Sun et al., 2024). Empirical evidence also supports an inverted U-shaped relationship: digital expansion initially increases pollution but significantly suppresses it upon reaching maturity, accompanied by spatial spillovers and the activation of intermediary channels (Lan and Ma, 2025). Chen and Fu (2023) further revealed the complexity of this relationship in their study of the Yangtze River Economic Belt, identifying innovation levels as a key threshold variable influencing the digital economy’s environmental impact. Similarly, He et al. (2019) also identified a significant nonlinear threshold effect in their analysis of green credit, providing theoretical support for this framework.

The “segmented” effect observed in regional scale grouping aligns with threshold conclusions: higher-development groups are more likely to exhibit emission reductions and environmental improvements (Li and Yang, 2024). The digitalization process of neighboring cities alters the nonlinear pathway of local environmental performance, carrying implications for regional coordination policies (Huang et al., 2023). Recent studies have found that the digital economy exerts negative spillover effects on pollution intensity in neighboring regions (Yang et al., 2024), further enriching the spatial dimension of the relationship between the digital economy and the environment.

Chen et al. (2023) revealed the mediating mechanisms through which the digital economy influences green development via three pathways: “technological innovation,” “industrial structure optimization,” and “environmental regulation.” The effectiveness of these mechanisms may also be modulated by developmental stage thresholds. At low maturity stages, inhibitory mechanisms (e.g., infrastructure energy consumption) dominate; while after threshold crossing, facilitating mechanisms (e.g., industrial green transformation, targeted regulation) take the lead, thereby achieving an overall shift in environmental effects.

The aforementioned literature lays the foundation for this study, yet certain limitations remain. First, most studies employ single or composite indicators to measure the digital economy, failing to effectively separate and distinguish the heterogeneous impacts of “green” versus “non-green” digital technologies. This makes it difficult to answer the question of “what kind of digital economy should be developed” to more effectively drive green transformation. Second, extensive research relies on linear models for inference, struggling to capture the complex nonlinear realities of the digital economy’s environmental impacts. While a few studies employ threshold models, they often focus solely on macro-level threshold variables, failing to fully reveal the micro-level mechanisms driving causal inflection points. Third, existing mechanism tests primarily validate single transmission pathways, lacking systematic analysis of the dynamic interactions among multiple pathways—particularly conflicting ones—during developmental stages.

To address these gaps, this paper contributes as follows: By innovatively categorizing city-level digital economic development into green and non-green digital technologies, it establishes an empirical foundation for examining the impact of technological heterogeneity on the environment. Second, it integrates cutting-edge econometric and machine learning methods to capture complex nonlinear relationships. By combining panel threshold regression models with the XGBoost-SHAP machine learning framework, it systematically identifies threshold effects, nonlinear patterns, and key feature contributions in the environmental impacts of the digital economy, overcoming the limitations of linear assumptions. Third, it proposes and tests the “dual-path dynamic game” theoretical mechanism. This study integrates the Environmental Kuznets Curve (EKC) with the technological innovation life cycle theory, proposing that the digital economy influences environmental pollution through two dynamically interacting pathways: the “infrastructure energy consumption effect” (dominant in the early stage) and the “green innovation effect” (dominant in the later stage). This theory provides micro-mechanism support for phased policy design. The innovation of this integrated framework lies in its dual capability: not only does it predict the existence of nonlinear relationships, but it also identifies the core drivers of transition points—namely, the dynamic interaction between the “infrastructure energy consumption effect” and the “green innovation effect” at different developmental stages. This lays the theoretical foundation for subsequently distinguishing the heterogeneous impacts of technological attributes (H3) and regional contexts (H4).

3 Research hypothesis

Existing research (Li, 2023) also suggests a nonlinear relationship between the digital economy and carbon emissions, with industrial structure and marketization levels potentially serving as critical threshold variables. This provides direct empirical evidence supporting the hypothesis framework of this paper. To systematically address the four research questions posed in the introduction, this paper follows the logical sequence of “testing overall effects—revealing mechanism pathways—analyzing conditions for heterogeneity” and proposes the following research hypotheses:

H1:

H1:. The digital economy exhibits an inverted U-shaped nonlinear relationship with regional environmental pollution.

H2a:

H2a:. The Digital Economy Exerts a Nonlinear Mediating Effect on Environmental Pollution Through Energy Consumption.

H2b:

H2b:. The digital economy exerts a restraining mediating effect on environmental pollution by promoting green technological innovation.

H3:

H3:. Green digital technologies exhibit a monotonically inhibitory effect on environmental pollution.

H4:

H4:. The environmental promoting effects of the digital economy exhibit lower threshold values in developed regions.

4 Research methods and data

4.1 Variable description

4.1.1 Explained variable

This study employs industrial SO₂ emissions, CO₂ emissions, PM_2.5 concentration, and industrial wastewater discharge to evaluate the degree of environmental pollution. The Entropy Weighting Method is used to synthesize these four indicators into a comprehensive environmental pollution index (EP). A higher index value indicates greater environmental pollution in the region. The entropy weighting method is an objective weighting approach whose core principle lies in assigning weights based on the variability of indicator values: the greater the variability of an indicator value, the lower its information entropy, and the higher the weight assigned. This more effectively reflects the differential contributions of different pollution dimensions to the composite index (Shannon, 1948). This method avoids the biases inherent in subjective weighting, enhancing the scientific rigor and comparability of composite indicators. To mitigate the issue of heteroscedasticity, the comprehensive environmental pollution index is treated logarithmically in this research.

4.1.2 Explanatory variable

Drawing on the approach of Liu et al. (2020) of using internet development as the core metric and incorporating digital transactions into the framework for constructing an indicator system, this paper measures the comprehensive development level of the digital economy from two dimensions—internet development and digital financial inclusion—based on the availability of relevant data at the city level. Following the methodology of Huang et al. (2019), for measuring internet development at the city level, four indicators are employed: internet penetration rate, relevant workforce statistics, related output metrics, and mobile phone penetration rate. The corresponding actual content for these four indicators is: number of broadband internet users per 100 people, proportion of employees in computer services and software industries relative to total urban unit employees, per capita telecommunications service volume, and number of mobile phone users per 100 people. The raw data for these indicators can be obtained from the China prefecture-level City Statistical Annual Report. For digital finance development, the China Digital Inclusive Finance Index is employed, jointly compiled by Peking University Digital Finance Center and Ant Group. Through principal component analysis (PCA), the data from the five indicators are standardized and reduced in dimension to derive the comprehensive digital economy development index, denoted as DE. Table 1 presents the evaluation framework for the DE index.

Table 1

Table 1. Descriptive statistical results.

4.1.3 Control variables

To better explore the relationship between the digital economy and environmental pollution, we incorporate the following four control variables into our regression model:

1. Foreign Direct Investment (FDI): FDI can introduce advanced technologies, generating technology spillover effects that reduce pollution (Sapkota and Bastola, 2017). However, some scholars argue that FDI facilitates the development of pollution-intensive industries in host countries (Wang, 2021a). This study evaluates FDI using the ratio of FDI to GDP.

2. Environmental Regulation (reg): Environmental regulation is estimated using the ratio of government investment in pollution control to GDP.

3. Secondary Industry Share (Indus): The development status of the secondary industry significantly impacts the economic structure and the environment. On the one hand, rational development of the secondary industry can promote economic growth. On the other hand, certain sectors within the secondary industry may exhibit high resource consumption and pollution emissions. This paper uses the share of secondary industry to assess its proportion within the economic structure.

4. R&D Expenditure as a Percentage of GDP (R&D): R&D investment is crucial for enhancing the innovation capacity and competitiveness of industrial enterprises. Higher R&D expenditure enables enterprises to develop new technologies and products, enhance production efficiency, and drive innovation. This paper uses the proportion of R&D expenditure by industrial enterprises above a designated size relative to GDP to measure the intensity of R&D investment.

4.1.4 Variable description

This study selected 287 Chinese cities from 2013 to 2022 as the research sample. The dataset’s limitation lies in the fact that the research data only extends to 2022, potentially failing to accurately reflect the current and future relationship between the digital economy and environmental pollution. The explanatory variable is the environmental pollution index calculated by the entropy weight method, while the core explanatory variable is the digital economy index computed via PCA. The environmental pollution index undergoes logarithmic transformation to address heteroscedasticity and balance data volatility. Table 2 presents descriptive statistics for key variables. The original data for this study were sourced from Peking University Digital Finance Center, China Industrial Statistical Yearbook, China City Yearbook, China Statistical Yearbook on Environment, National Intellectual Property Administration, municipal statistics bureaus, China Energy Statistical Yearbook, and China prefecture-level City Statistical Annual Report. PM2.5 concentrations were derived using a global ground-based PM2.5 concentration grid data (calibrated via the Global Wavenumber Raster model) combined with data from China’s national air quality monitoring stations, as outlined in the methodology presented in Liu et al. (2022). Partially missing values were imputed using regression-based imputation and linear interpolation.

Table 2

Table 2. Comprehensive index system of urban DE level.

4.2 Research methods

This study employs a three-tier methodological framework integrating bidirectional fixed-effects models, panel threshold models, and machine learning approaches: ‘fundamental testing—structural identification—multifactorial analysis.’ The two-way fixed effects model is employed to preliminarily validate the nonlinear relationship between the digital economy and environmental pollution. The panel threshold model further identifies the structural thresholds where the impact mechanism shifts. Machine learning methods, from a data-driven perspective, reveal the importance of key variables, their nonlinear contributions, and sample heterogeneity. These three approaches mutually corroborate and complement each other, collectively enhancing the robustness and explanatory power of the research conclusions.

4.2.1 Double-fixed-effect model

This study constructs the following model to test the inverted U-shaped relationship:

${lnEP}_{it}={\alpha }_0+{\alpha }_1{DE}_{it}+{\alpha }_2{DE}_{it}^2+{\displaystyle \sum _{k=1n}}{\gamma }_k+X_{kit}+{\mu }_i+{\lambda }_t+{\epsilon }_{it}(1)$

where i represents city and t represents time. ${lnEP}_{it}$ represents the logarithm of comprehensive environmental pollution index, ${DE}_{it}$ is the core explanatory variable (digital economy), ${DE}_{it}^2$ is its squared term, $X_{kit}$ denotes a set of control variables, ${\mu }_i$ is the individual fixed effect, ${\lambda }_t$ is the time fixed effect, and ${\epsilon }_{it}$ is the random error term. If ${\alpha }_1$ > 0 and ${\alpha }_2$ < 0, this supports the inverted U-shaped hypothesis.

The model’s assumptions primarily include: First, each individual and time point possesses unique characteristics that influence the explanatory variables. Second, no multicollinearity issues exist. Third, the error term should be independent and normally distributed with constant variance, free from autocorrelation and heterogeneity.

This model is suitable for panel data. When individual and temporal heterogeneity exists in the data, it controls for unobservable individual and time effects.

4.2.2 Panel threshold model

Using Hansen’s (1999) panel threshold model, this method has been widely applied to examine complex nonlinear relationships between environmental and economic variables. With the digital economy composite index as the threshold variable, the threshold values are identified in Formulas 2, 3:

Single-threshold model:

${lnEP}_{it}={\beta }_0+{\beta }_1{DE}_{it}I\left({DE}_{it}\le \gamma \right)+{\beta }_2{DE}_{it}I\left({DE}_{it}>\gamma \right)+{\beta X}_{it}+{\mu }_i+{\lambda }_t+{\epsilon }_{it}(2)$

Multi-threshold model:

$\begin{array}{l}{lnEP}_{it}={\beta }_0+{\beta }_1{DE}_{it}I\left({DE}_{it}\le {\gamma }_1\right)+{\beta }_2{DE}_{it}I\left({\gamma }_1{<DE}_{it}\le {\gamma }_2\right)\\ +{\beta }_3{DE}_{it}I\left({de}_{\text{it}}>{\gamma }_2\right){+\beta X}_{it}+{\mathrm{\mu }}_i+{\mathrm{\lambda }}_t+{\mathrm{\epsilon }}_{it}\end{array}(3)$

where γ denotes the threshold value and I(⋅) represents the indicator function. The significance of threshold effects was tested via Bootstrap sampling (1,000 iterations).

4.2.3 Machine learning framework

To identify key factors influencing environmental pollution and validate the central role of the digital economy among multiple control variables, this study employs machine learning methods for feature importance analysis and testing the nonlinear relationship. Specifically:

Using DE, Indus, reg, R&D, FDI, and pgdp as features and lnEP as the target variable, XGBoost calculates feature importance scores based on information gain to rank the relative importance of the digital economy among numerous influencing factors. TreeExplainer computes SHAP values for each feature, quantifying the average marginal contribution of the digital economy to environmental pollution predictions and revealing both global patterns and individual heterogeneity in feature impacts.

SHAP is a method inspired by Shapley values in game theory to interpret machine learning model predictions. It quantifies the contribution of each feature to the model’s output and reveals the relationship between features and predictions (Pelegrina et al., 2023). This study constructs an interpretable model of lnEP’s influencing factors in Formula 4 (Li, 2022):

${lnEP}_{it}={lnEP}_{base}+f\left(X_{i,DE}\right)+f\left(X_{i,Indus}\right)+f\left(X_{i,reg}\right)+f\left(X_{i,R\&D}\right)+f\left(X_{i,F\text{DI}}\right)+f\left(X_{i,pgdp}\right)(4)$

$X_{ij}$ denotes the j th feature of the i th sample, where i = 1, … , n and j∈{DE, Indus, reg, R&D, FDI, pgdp}. lnEPbase represents the average of lnEP across all samples, and f (x_ij) is the SHAP value for the j th feature of the i th sample.

5 The empirical result analysis

5.1 Benchmark regression results

Regression analysis was conducted using Formula 1 to preliminarily explore the relationship between the digital economy and environmental pollution. Table 3 presents the regression results for the pooled OLS model, time fixed-effects model, individual fixed-effects model, and dual fixed-effects model (time and individual). The findings reveal that the inverted U-shaped relationship between the digital economy and environmental pollution is gradually becoming more pronounced and stable.

Table 3

Table 3. Baseline regression results.

In the pooled OLS model, the coefficient for the squared term of the digital economy is significantly negative, while the linear term is insignificant, indicating an initial emergence of the inverted U-shaped relationship. After incorporating time fixed effects, the absolute value of the squared term coefficient increases to −7.8193, further strengthening the relationship. In the individual fixed effects model, the coefficient for the linear term of the digital economy was significantly positive, while the coefficient for the squared term was significantly negative, formally confirming the inverted U-shaped relationship. Finally, in the dual fixed effects model, both the linear and squared term coefficients were highly significant, making the inverted U-shaped relationship the most robust. The calculated inflection point value was approximately 0.243, indicating that the inhibitory effect of the digital economy on environmental pollution begins to manifest when the digital economy index exceeds 0.24.

Regarding control variables, significant differences emerged across models: the share of secondary industry was significantly positive in the mixed OLS and time-fixed effects models but became insignificant after controlling for individual effects, indicating its pollution effect primarily stems from inherent regional differences. R&D investment was significantly negative in the individual fixed effects model, suggesting technological innovation improves the environment in the long term. Environmental regulations were significantly positive under the individual fixed effects model, potentially due to the endogeneity of regulatory intensity. Foreign direct investment was significantly positive in the two-way fixed effects model, providing evidence for the “pollution haven” hypothesis.

5.2 Panel threshold model regression results

Table 4 reports the results of the threshold effect tests. The single-threshold F-statistic was 121.22 with a p-value of 0.008, rejecting the “no threshold” hypothesis. The dual-threshold F-statistic was −82.23 with a p-value of 0.385, accepting the “no second threshold” hypothesis. This indicates the presence of a single threshold with an approximate value of 0.25.

Table 4

Table 4. Threshold effect test results.

Table 5 shows the regression results for the single threshold model. When DE ≤ 0.25, the coefficient for the digital economy is significantly positive (3.098) and significant at the 1% level. This indicates that in the early stages of digital economic development, its growth is primarily manifested through the rapid expansion of high-energy-consuming digital infrastructure (e.g., 5G base stations, data centers). The negative environmental effects generated by this segment dominate, thereby exacerbating environmental pollution. When DE > 0.25, the coefficient is significantly negative with a high absolute value and statistical significance. This indicates that beyond the threshold, the synergistic effects of digital technology integration with industries emerge—industrial internet optimizes production processes, AI monitoring enables pollution source tracing, and the green benefits outweigh the energy consumption costs, exerting a suppressing effect on environmental pollution.

Table 5

Table 5. Panel threshold model regression results.

Regression results on both sides of the threshold are shown in Table 5. Regarding control variables: At the high digital economy stage, while the secondary industry share coefficient remains positive, it has decreased. The foreign direct investment coefficient is significantly negative, and the environmental regulation coefficient is significantly positive. This indicates that combining a digital economy with stringent environmental regulations can effectively curb the “pollution haven” effect of foreign investment.

Economically, a one-unit increase in the digital economy index leads to a 3.098-unit rise in environmental pollution before the threshold and a 1.963-unit decrease after the threshold, resulting in a net emission reduction effect of 5.061 units.

5.3 Heterogeneity analysis

5.3.1 Technology heterogeneity

Classifying digital technologies into “green digital technologies” and “non-green digital technologies”, this study measures the level of green digital technology and non-green digital technology by quantifying the number of green digital patents (GP) and the number of non-green digital technology patents (NGP). Specifically, city-level patent data was obtained from the National Intellectual Property Administration’s patent database. Using the “Green List of International Patent Classification” published by the World Intellectual Property Organisation (WIPO) and overlaying keywords related to “digital technologies” (such as IoT, big data, artificial intelligence, and blockchain), a dual screening and identification process was conducted. This ultimately yielded the annual number of green digital technology invention patents and utility model patents granted in each city, denoted as variable GP. The difference between the total number of digital technology patents and the GP value is denoted as the variable NGP. We separately examined their impacts on environmental pollution, with results shown in Table 6. The coefficients for green digital technologies were significantly negative and exhibited no threshold effect, indicating they directly serve pollution control and can reduce emissions without requiring scale accumulation. Robustness tests reveal that using a lagged green patent variable still yields a significant negative impact with stable coefficients, confirming the sustained environmental effects of green digital technologies. Non-green digital technologies exhibit a pronounced “U-shaped” relationship with a threshold of 4.24, indicating that they yield environmental improvements at moderate scales but exceed efficiency gains through increased energy consumption and environmental impacts beyond this threshold.

Table 6

Table 6. Regression results of technology heterogeneity.

This study’s finding of a “monotonically decreasing” relationship for green digital technologies aligns with Wang et al. (2021b), whose research indicates that internet development promotes green growth by driving industrial upgrading and corporate innovation—a more direct pathway potentially bypassing the traditional “pollute first, then clean up” process. Green digital technologies, both in their technical core and application orientation, directly serve to reduce emissions, conserve energy, and control pollution. Consequently, they generate environmental improvement effects from the very outset of technology adoption, bypassing the high-energy-consumption phase of traditional digital technologies that typically follow a pattern of “first scale expansion, then efficiency enhancement.” Notably, the U-shaped relationship observed for non-green digital technologies indicates that even traditionally “non-green” digital technologies can yield environmental benefits through efficiency gains and resource optimization within a moderate development scale. This reflects the inherent environmental positive externalities of digital technologies. However, beyond the optimal scale, environmental burdens such as energy consumption and electronic waste become apparent, suggesting the need for regulatory mechanisms to govern the moderate development of digital technologies.

It should be noted that this study measures green digital technology levels using the number of green digital patents, with results primarily reflecting the positive leadership role of innovation capacity and knowledge reserves in emission reductions. Simultaneously, the analysis of non-green digital technologies reveals the complexity of digital technology’s environmental effects. Even technologies classified as “non-green” exhibit significant variations in environmental impact depending on their development stage.

5.3.2 Regional development heterogeneity

This study uses the median per capita GDP in 2018 (approximately ¥50,500) as the grouping threshold to classify the sample into developed and underdeveloped regions for regional development heterogeneity analysis. The results are presented in Table 7. For developed regions, the threshold value for the digital economy’s impact on environmental pollution is 0.21, with a coefficient of −1.7885 after exceeding the threshold; while the threshold for less developed regions is 0.23, with a coefficient of −8.0923 beyond this threshold. The interaction term regression test indicates highly significant regional differences, and the coefficient difference test for each group further confirms the statistical significance of this heterogeneity. This disparity may be related to structural factors between regions: developed regions feature a lighter industrial structure and stringent environmental regulations, have jointly reduced the institutional costs and technological adaptation barriers for the green transformation of the digital economy, enabling it to cross the environmental inflection point earlier; Less developed regions rely on energy-intensive industries, where digital technologies are more readily applied to expand production scale, requiring higher thresholds to transition toward green applications. Jin et al. (2025) and Pan et al. (2022) independently demonstrated the regional unevenness of the environmental effects of the digital economy from the perspectives of carbon emissions and total factor productivity (TFP), respectively. Their findings indicate that regional resource endowments, policy implementation capacity, and innovation absorption capabilities serve as crucial contextual variables shaping its green performance.

Table 7

Table 7. Regression results of regional development heterogeneity.

5.4 Robustness test

5.4.1 Municipalities and provincial capitals are excluded

As regional economic hubs, municipalities directly under the central government and provincial capitals enjoy abundant resources and policy advantages. After excluding these cities and retaining only prefecture-level city samples for regression analysis, the results shown in Table 8 indicate an inflection point of approximately 0.24. The coefficient sign and significance level remain unchanged, confirming the exclusion of administrative tier interference.

Table 8

Table 8. The baseline regression results (Excluded municipalities and provincial capitals).

5.4.2 Tail reduction treatment

To mitigate the impact of outliers on regression results, a 1% two-tailed tail trimming was applied to the sample. All variables below the first percentile were replaced with the first percentile value. All variables above the 99th percentile were replaced with the 99th percentile value. As shown in Table 9, the signs and significance levels of all coefficients remained unchanged, consistent with the aforementioned results.

Table 9

Table 9. The baseline regression results (Tailed processing).

5.4.3 Endogeneity problem

Pollutant emissions may be influenced by other unmeasurable variables, such as public environmental awareness, heterogeneous characteristics of government officials, and the effects of control variables, thereby introducing an omitted variable issue. Simultaneously, a reverse causal relationship may exist between a city’s digital economy and pollutant emissions. From a spatial perspective, Xu et al. (2022b) concluded that the digital economy’s impact exhibits spatial spillover effects, further complicating causal identification. Consequently, we employ an instrumental variables approach to investigate the relationship between the digital economy and urban pollution emissions. Drawing on the methodology of Bai et al. (2022), we select the spherical distance between each city center and Hangzhou as the instrumental variable.

We select the logarithm of the spherical distance (log_distance) between each city center and Hangzhou as the instrumental variable for the digital economy. The underlying logic is that the development of digital finance, exemplified by Alipay, originated in Hangzhou, implying that Hangzhou should lead in digital economic development. In general, cities closer to the geographic center of Hangzhou exhibit stronger digital economic development. Additionally, beyond correlation with the endogenous variable, instrumental variables must satisfy the exclusion property. Given the natural exogeneity of spherical distance variables from each city to Hangzhou, the instrumental variable selected in our study essentially meets the exclusion requirement. Specifically, to address the issue of Hangzhou’s distance being 0, we applied appropriate technical adjustments to ensure the feasibility of logarithmic operations.

Table 10 reports the estimation results for the instrumental variables. The first-stage regression reveals that the instrumental variable, log_distance, exhibits a statistically significant negative correlation with the core explanatory variable, DE, at the 1% level. This aligns with the theoretical expectation that “the farther a city is from Hangzhou, the lower its digital economy development level,” thereby validating the correlation condition of the instrumental variables. The Anderson LM test statistic is significant at the 1% level, and the Cragg-Donald Wald F-statistic for the first-stage regression is 125.555, indicating no weak instrumentation problem. After addressing endogeneity, the second-stage regression results show that the coefficients for digital economy and environmental pollution remain significantly negative.

Table 10

Table 10. Estimation results of instrumental variable method.

5.4.4 Core explanatory variables are replaced

This paper re-calculates DE using the entropy weight method and performs the regression. Table 11 presents the results after replacing the core explanatory variables, with no significant changes in coefficient signs or significance levels.

Table 11

Table 11. The baseline regression results (replacing the core explanatory variable).

6 Further analysis

6.1 1Analysis the key factors influencing environmental pollution

To identify key factors influencing environmental pollution and validate the robustness of conclusions, this study employs the XGBoost machine learning model. To mitigate the risk of model overfitting, the dataset is randomly split into a 70% training set and a 30% testing set. The model optimization process combined grid search with cross-validation, ensuring model accuracy while strictly controlling overfitting. The optimal hyperparameter combination was determined as follows: n_estimators = 200, max_depth = 6, learning_rate = 0.1, subsample = 0.8, colsample_bytree = 0.8, random_state = 42. Comprehensive testing demonstrates that the XGBoost and SHAP model frameworks exhibit robust performance in feature selection, parameter tuning, and data variability, ensuring reliable interpretability. The final model achieves performance metrics of R² = 0.655, RMSE = 0.322, and MAE = 0.246 on the test set, indicating strong simulation and explanatory capabilities for environmental pollution.

XGBoost was employed to extract feature contributions for environmental pollution interpretation, generating a feature importance map (Figure 1) to reveal each feature’s contribution to model predictions. According to the XGBoost feature importance map, factors influencing environmental pollution are ranked in descending order: DE, R&D, pgdp, Indus, FDI, and reg. DE is the most significant factor in urban ecological environments, exerting a major impact on environmental pollution levels.

Figure 1

Figure 1. Feature importance chart.

To further explore the specific direction and magnitude of each feature’s impact on environmental pollution, SHAP values were calculated for each sample. Figure 2 illustrates the marginal contributions of influencing factors, with each point representing a distinct level. Blue and red sample points denote lower and higher true values of the influencing factors, respectively. The x-axis corresponds to the SHAP value of each sample, while the y-axis is sorted in descending order based on the feature’s average absolute SHAP value. Higher values indicate a greater influence of the feature on environmental pollution. Consequently, indicators with a positive impact on pollution exhibit a “blue left, red right” pattern, while those with a negative impact show a “red left, blue right” pattern. The contribution of individual features reveals that DE negatively affects pollution—higher DE values correlate with smaller SHAP values and corresponding pollution indices. Across the city, R&D and pgdp generally exert positive effects on environmental pollution, while Indus exhibits a potentially negative impact, reg and FDI demonstrate more complex patterns in their influence on environmental pollution.

Figure 2

Figure 2. SHAP summary plot.

SHAP value analysis reveals that DE ranks first among all influencing factors and exerts a significant negative impact on environmental pollution. This indicates that promoting digital economic development is a key strategy for improving environmental pollution.

To further explore the contribution patterns of DE and other control variables to environmental pollution, a SHAP dependence plot was generated (Figure 3). The x-axis represents feature values, while the y-axis displays the corresponding SHAP values for each feature. Each point represents the SHAP value of a specific feature in the data sample. The SHAP dependence plot reflects the specific contribution of individual variables to environmental pollution at different values.

Figure 3

Figure 3. SHAP dependence plot.

The subplots in the figure illustrate the univariate contributions of DE, Indus, reg, R&D, FDI, and pgdp to environmental pollution. Taking the DE subplot as an example, the SHAP values exhibit complex trends as DE itself varies. Overall, when DE is at lower levels, SHAP values are predominantly negative, indicating an adverse impact on environmental pollution. As DE levels increase, SHAP values gradually shift from negative to positive, indicating a transition from negative to positive effects on environmental pollution. Beyond a certain threshold, SHAP values exhibit significant growth, suggesting that DE’s positive contribution to environmental pollution intensifies with its own development, revealing nonlinear and threshold characteristics.

In the Indus subplot, Indus’s SHAP values show an overall upward trend. As Indus increases, SHAP values gradually shift from negative to positive and continue to grow. This indicates that the impact of industrial development on environmental pollution transitions from suppression to promotion, with the promotional effect strengthening as industrial levels rise.

In the reg subplot, overall, when reg is at a low level, the SHAP value is negative, indicating a negative impact on environmental pollution. As the reg level increases, the SHAP value gradually rises, revealing a positive impact on environmental pollution.

In the R&D subplot, SHAP values exhibit significant fluctuations and nonlinear characteristics. At different R&D levels, SHAP values can be either positive or negative, indicating that R&D’s impact on environmental pollution is complex. Its contribution direction and magnitude vary across different intervals, with multiple inflection points reflecting the intricate nonlinear relationship between R&D and environmental pollution.

In the FDI subplot, the SHAP value of FDI evolves from negative to positive as its level increases. At lower FDI levels, it inhibits environmental pollution. As FDI rises, this inhibitory effect weakens and gradually shifts to a promotional role. Beyond a certain threshold, the positive contribution becomes more pronounced.

In the pgdp subplot, the SHAP value exhibits an overall nonlinear relationship. At different pgdp levels, SHAP values are both positive and negative, indicating that the impact of per capita GDP on environmental pollution varies across different levels. The direction of contribution changes with pgdp levels, featuring multiple inflection points that reveal a complex interaction pattern between per capita GDP and environmental pollution.

Overall, these variables exhibit gradual changes or multi-inflection point patterns in their nonlinear effects on environmental pollution. Substantial positive reinforcement occurs only when variables reach critical intensity levels, further highlighting the complexity and nonlinearity of these impacts.

The machine learning analysis in this section complements and cross-validates the econometric findings on several key points. First, the top ranking of the digital economy (DE) in the XGBoost feature importance analysis corroborates its statistically significant and substantial impact in the benchmark regression. Second, the SHAP dependency plots illustrate that DE’s marginal effect on environmental pollution shifts from negative to positive as its value increases, which aligns with the single threshold and the inverted U-shaped relationship identified by the panel threshold model. Furthermore, the heterogeneity in SHAP values across samples aligns with the econometric results on regional and technological grouping, jointly highlighting the context-dependent and heterogeneous environmental effects of the digital economy. This consistency across methodologies enhances the robustness of our conclusions and provides a more solid empirical foundation for policy deliberation. Building on this, the following section delves into the mediating mechanisms through which the digital economy influences environmental pollution.

6.2 Analysis of the mechanisms affecting environmental pollution

The development of the digital economy initially promotes environmental pollution before subsequently inhibiting it. To further investigate this mechanism, we now conduct an in-depth analysis to identify the specific channels through which the digital economy influences environmental pollution. This study selects electricity consumption and green digital technology innovation as key mediating variables. The mechanism testing model is constructed in Formulas 5, 6:

${Med}_{it}={\upsilon }_0+{\upsilon }_1·{DE}_{it}+{\upsilon }_2·{DE}_{it}^2+{\mu }_i+{\lambda }_t+{\epsilon }_{it}(5)$

${lnEP}_{it}={\delta }_0+{\delta }_1·{DE}_{it}+{\delta }_2·{DE}_{it}^2+{\delta }_3·{med}_{it}+{\mu }_i+{\lambda }_t+{\epsilon }_{it}(6)$

In Formulas 5, Med represents the mediating variables, including power consumption (pow) and green technological innovation (GP), while other variables remain consistent with Formula 1. If υ₁, υ₂, δ₁, δ₂, δ₃ are statistically significant, it indicates that the mediating variables exert a mediating effect. This study employs a stepwise regression approach to address limitations in mediating effect statistical testing. The Bootstrap method is used to repeatedly sample 5,000 samples, enhancing statistical reliability.

The impact of the digital economy on environmental pollution transmission mechanisms primarily relies on dual restructuring of energy and technology. At the electricity consumption level, the explosive growth of the digital economy manifests directly as an exponential expansion of computing infrastructure. Such high-energy-consumption facilities inevitably drive up electricity usage, thereby exacerbating environmental pollution. However, from a long-term perspective, leveraging digital tools like smart grid dynamic scheduling and distributed energy management systems, the digital economy can transform this situation into a “digital emission reduction dividend,” mitigating environmental pollution by reducing electricity consumption. The intermediary role of green technological innovation manifests in two ways: first, the digital economy creates an iterative optimization environment for clean technology R&D; second, digital technologies themselves integrate directly into green production processes, ultimately achieving green manufacturing and reducing environmental pollution.

Based on the regression results in Table 12, the transmission mechanism of the digital economy’s impact on environmental pollution is analyzed as follows:

Table 12

Table 12. Results of mediation effect analysis.

Column (1) shows that the coefficient of the quadratic term for the digital economy is significantly negative at the 1% level, indicating an inverted U-shaped nonlinear effect where the digital economy first promotes and then inhibits environmental pollution. Column (2) reveals the mediating effect of electricity consumption: the coefficients for DE and DE2 are positive and negative, respectively, at the 1% significance level, suggesting the digital economy also exerts a nonlinear impact on electricity consumption, initially increasing it before reducing it. Column (3) shows that after incorporating electricity consumption as a mediating variable, the coefficient for DE2 remains significantly negative, confirming the inverted U-shaped relationship between the digital economy and environmental pollution. Notably, compared to Column (1), the coefficient for DE has decreased, indicating that electricity consumption plays a nonlinear mediating role in the digital economy’s impact on environmental pollution—specifically, amplifying the effect initially before mitigating it later. Column (4) reveals that the primary term coefficient of the digital economy exhibits a significant negative impact on green technological innovation at the 1% level, while the secondary term coefficient is significantly positive, indicating a U-shaped pattern in the digital economy’s influence on green technological innovation. Column (5) results show that after incorporating green technological innovation as a mediating variable, the coefficient of DE2 remains significantly negative, further confirming the inverted U-shaped relationship between the digital economy and environmental pollution. More importantly, compared to the benchmark model (1), the absolute value of the digital economy coefficient significantly increases, verifying that green technological innovation plays a crucial mediating role in emission reduction within the digital economy’s impact on environmental pollution.

Test results based on the Bootstrap method (5,000 repeated samples) reinforce the statistical validity of the aforementioned mediating effect. Collectively, these findings demonstrate that both electricity consumption and green technological innovation play significant transmission roles in the digital economy’s impact on environmental pollution.

Mechanism testing reveals that the inverted U-shaped trajectory of “power consumption”—rising initially before declining—illustrates the digital economy’s transition from an “infrastructure energy consumption effect” (dominant in the early phase), centered on capital deepening and energy expenditure, to a “green innovation effect” (dominant in the later phase), focused on intelligent scheduling and energy efficiency optimization. Meanwhile, the U-shaped growth trajectory of “green technological innovation” reflects how the digital economy, after overcoming initial cost constraints, leverages its general-purpose technology attributes and digital platform efficiency to systematically reshape the supply-side conditions for green innovation activities. This shift moves beyond passively inducing demand to actively empowering transformation, thereby becoming the core force driving the emergence of pollution inflection points. The transition of the “green innovation effect” from latent to dominant hinges critically on whether the scale of digital infrastructure and the cost of green technologies can surpass specific thresholds. Once this critical point is crossed, the general-purpose technology attributes of the digital economy will systematically reduce the marginal cost of green innovation, driving a dynamic shift in the dominant mechanism from the “infrastructure energy consumption effect” to the “green innovation effect.” These two complementary pathways collectively reveal the deep-seated logic underpinning the digital economy’s environmental impact. Yang et al. (2024) and Li (2023) highlight industrial upgrading and green technological progress as key mechanisms through which the digital economy influences environmental pollution, aligning with our study’s “green digital innovation” mediation pathway. However, while the “green digital innovation” pathway explains most of the moderating effect, other unobserved mechanisms may coexist, for instance, while supporting environmental regulations promote green industrial upgrading, they may also trigger spatial and sectoral restructuring in the labor market (Li et al., 2025). The impact of environmental legislation can amplify the attractiveness of foreign direct investment in renewable energy sectors under production reduction policies, thereby generating more positive spillover effects for renewable energy technologies (Lai and Wang, 2024). This study identifies a dominant core mechanism, revealing the intrinsic dynamic game process through which the digital economy impacts the environment.

7 Conclusions and policy recommendations

Based on previous research, this paper theoretically examines the nonlinear relationship between the digital economy and green development, proposing three research hypotheses. Empirical results validate these hypotheses. Specifically, utilizing data from 287 Chinese cities spanning 2013 to 2022, this study employs a combination of two-way fixed effects models, panel threshold models, and machine learning methods to empirically investigate the impact of the digital economy on green development and its underlying mechanisms. The conclusions are as follows: (1) The digital economy exerts a significant influence on environmental pollution, occupying a central position among all influencing factors. (2) The digital economy affects environmental pollution through a nonlinear pathway: first exacerbating energy consumption, then promoting green technological innovation. This pathway is dynamic and unfolds in stages. (3) The pathway exhibits a pronounced nonlinear effect, taking the form of an inverted U-shaped curve, and demonstrates threshold effects. (4) The pathway exhibits technological heterogeneity, with green technologies achieving emission reductions without threshold requirements. (5) Regional development heterogeneity is evident, with developed regions demonstrating lower thresholds and stronger emission reduction effects. Compared with existing research, this paper not only verifies the nonlinear relationship between the digital economy and environmental pollution but also further identifies its structural thresholds, technological heterogeneity mechanisms, and regional heterogeneity conditions. It provides more systematic empirical evidence and decision-making references for the question of ‘how to develop the digital economy to better achieve green transformation.’ Based on these findings, the following recommendations are proposed:

1. Implement phased, targeted regulation based on development thresholds. Research confirms that the impact of the digital economy on environmental pollution exhibits a single threshold (DE ≈ 0.25). For regions below this threshold, policy focus should shift to “green infrastructure constraints and upfront planning.” Strict energy intensity access requirements and clean energy allocation ratios should be imposed on new facilities such as data centers and 5G base stations. Special transformation funds should be established to support their energy efficiency upgrades, thereby mitigating environmental risks during the initial phase of digital expansion. For regions exceeding this threshold, efforts should concentrate on “unlocking digital emission reduction dividends.” Tools like tax incentives and green credit should motivate enterprises to adopt technologies such as industrial internet and artificial intelligence for energy-saving upgrades and targeted pollution control, transforming digital advantages into environmental governance effectiveness.

2. Implement differentiated guidance strategies based on technological attributes. Green digital technologies (e.g., environmental AI, carbon blockchain) exhibit linear emission reduction effects, while non-green digital technologies (e.g., certain high-energy-consuming data centers) demonstrate U-shaped environmental impacts. Therefore, green digital technologies should receive “dual incentives for R&D and practical application” through dedicated programs in national and local science initiatives, alongside promoting their demonstration and procurement in environmental monitoring and smart energy sectors. For non-green digital technologies, establish a “dual-control mechanism for efficiency and scale.” While encouraging energy efficiency improvements through technological upgrades, reference the U-shaped inflection point of their environmental impact to explore tiered resource pricing for operations exceeding reasonable scale, guiding their orderly low-carbon transition.

3. Design coordinated development pathways based on regional endowments. Research reveals significant regional heterogeneity in the environmental effects of the digital economy, with threshold values lower in developed regions (0.21) than in less developed regions (0.23). Developed regions (e.g., eastern urban clusters) should focus on “deep integration of digital technologies with green industries,” leveraging digital advantages to develop high-end green manufacturing and environmental services, while exploring the inclusion of digital emission reduction contributions into carbon market trading systems. Less developed regions (e.g., central and western cities) should adopt a “parallel green infrastructure and industrial upgrading” strategy, prioritizing coordinated deployment of digital infrastructure with local clean energy sources like wind and solar. Simultaneously, digital technologies should empower the green transformation of traditional industries to avoid falling into the “digital expansion-high carbon lock-in” trap.

4. Improve incentive-compatible institutional safeguards. It is recommended to incorporate a “Digital Economy Green Development Index”—integrating environmental quality and green economic indicators—into local government performance evaluations. Concurrently, provide certification facilitation and financing support to enterprises actively adopting green digital technologies for emissions reduction, while imposing corrective deadlines and differentiated electricity pricing on digital infrastructure operators failing to meet standards. This dual approach of incentives and constraints will guide multiple stakeholders toward a concerted effort in green digital transformation. Throughout this process, the boundaries of government intervention must be carefully calibrated. Notably, governments must rationally consider governance approaches. Wu et al. (2024b) found that air pollution directs capital to renewables. Therefore, policies should mandate corporate environmental disclosure to reduce information asymmetry, helping markets better identify and support true green innovators. Guo et al. (2025) found that the ZWCP policy significantly reduced energy consumption intensity in pilot regions, demonstrating its strong reference and replication value. While some governments promote environmental protection through market segmentation, this may reduce market efficiency, thereby hindering green technology development and application and ultimately impeding environmental goals (Lai et al., 2024). Therefore, the core of institutional safeguards should be maintaining fair competition within the unified national market, dismantling barriers to factor mobility, and providing a robust market foundation for the synergistic innovation of digital and green technologies.

This study has several limitations. First, the data cutoff is 2022, which fails to capture the latest environmental impacts of digital technologies. Second, the classification of green and non-green digital technologies remains relatively broad and could be further refined. Third, beyond the two core pathways identified, other mediating or moderating variables may exist. Future research could leverage more granular technology data and longer-span panel data, combined with case studies or experimental methods, to delve deeper into the multi-pathway dynamic mechanisms linking the digital economy and green development.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Author contributions

YZ: Writing – original draft, Writing – review and editing, Formal Analysis, Data curation, Investigation.

Funding

The author(s) declared that financial support was received for this work and/or its publication.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Adams, S., and Klobodu, E. (2017). Urbanization, democracy, bureaucratic quality, and environmental degradation. J. Policy Model. 39 (6), 1035–1051. doi:10.1016/j.jpolmod.2017.04.006

CrossRef Full Text | Google Scholar

Bai, F., Huang, Y., Shang, M., and Ahmad, M. (2022). Modeling the impact of digital economy on urban environmental pollution: empirical evidence from 277 prefecture-level cities in China. Front. Environ. Sci. 10, 991022. doi:10.3389/fenvs.2022.991022

CrossRef Full Text | Google Scholar

Chen, Z., and Fu, B. (2023). The impact of digital economy on environmental pollution from the perspective of innovation level: a case Study of 40 cities along the Yangtze river Economic belt. Resour. Environ. Yangtze Basin 33 (1), 126–138. Available online at: https://kns.cnki.net/kcms2/article/abstract?v=IU01-AKyIMB2QL7H42PNbDN27DzWnLu_lBBQai91LlwhwtoE3QkHjnP5tiXiXxHHHN2mgslYSAp48eF_6JzTfiCZTewDxvh_S_mBBOOcIjnL0OKt3sGbB-pCjSi96rpB-J1TL-6JOwrprLRKUfj77loZY8mfo9VuLcODeVQZ5kxIzV9NUG33dtSjH6pJq9Hq&uniplatform=NZKPT&language=CHS.

Google Scholar

Chen, S., Yang, Y., and Wu, T. (2023). Digital economy and green total factor productivity---based on the empirical research on the resource-based cities. Environ. Sci. Pollut. Res. 30, 47394–47407. doi:10.1007/s11356-023-25547-y

PubMed Abstract | CrossRef Full Text | Google Scholar

Feng, Y., Wang, H., Liu, J., and Nie, C. (2023). The effect of information infrastructure on total carbon emissions and intensity: evidence from a quasi-natural experiment in China. J. Environ. Plan. Manag. 67 (13), 3310–3338. doi:10.1080/09640568.2023.2220898

CrossRef Full Text | Google Scholar

Guo, B., Qian, Y., Guo, X., and Zhang, H. (2025). Impact of zero-waste city pilot policies on urban energy consumption intensity: causal inference based on double machine learning. Sustainability 17 (11), 5039. doi:10.3390/su17115039

CrossRef Full Text | Google Scholar

Hansen, B. E. (1999). Threshold effects in non-dynamic panels: estimation, testing, and inference. J. Econ. 93 (2), 345–368. doi:10.1016/s0304-4076(99)00025-1

CrossRef Full Text | Google Scholar

He, L., Zhang, L., Zhong, Z., Wang, D., and Wang, F. (2019). Green credit, renewable energy investment and green economy development: empirical analysis based on 150 listed companies of China. J. Clean. Prod. 208, 363–372. doi:10.1016/j.jclepro.2018.10.119

CrossRef Full Text | Google Scholar

Huang, L., and Lei, Z. (2021). How environmental regulation affects corporate green investment: evidence from China. J. Clean. Prod. 279, 123560. doi:10.1016/j.jclepro.2020.123560

CrossRef Full Text | Google Scholar

Huang, Q., Yu, Y., and Zhang, S. (2019). Internet development and productivity improvement in manufacturing: internal mechanism and China’s experience. China Ind. Econ. (8), 5–23. doi:10.19581/j.cnki.ciejournal.2019.08.001

CrossRef Full Text | Google Scholar

Huang, S., Han, F., and Chen, L. (2023). Can the digital economy promote the upgrading of urban environmental quality? Int. J. Environ. Res. Public Health 20 (3), 2243. doi:10.3390/ijerph20032243

PubMed Abstract | CrossRef Full Text | Google Scholar

Jiang, Y., and Deng, F. (2022). Multi-dimensional threshold effects of the digital economy on green economic growth? New evidence from China. Sustainability 14 (19), 12888. doi:10.3390/su141912888

CrossRef Full Text | Google Scholar

Jiang, S., Li, Y., Lu, Q., Hong, Y., Guan, D., Xiong, Y., et al. (2021). Policy assessments for the carbon emission flows and sustainability of bitcoin blockchain operation in China. Nat. Commun. 12, 1938. doi:10.1038/s41467-021-22256-3

PubMed Abstract | CrossRef Full Text | Google Scholar

Jiao, S., and Sun, Q. (2021). Digital economic development and its impact on economic growth in China: research based on the perspective of sustainability. Sustainability 13 (18), 10245. doi:10.3390/su131810245

CrossRef Full Text | Google Scholar

Jin, W., Wang, Y., Yan, Y., Zhou, H., Xu, L., Zhang, Y., et al. (2025). Digital economy, green finance, and carbon emissions: evidence from China. Sustainability 17 (12), 5625. doi:10.3390/su17125625

CrossRef Full Text | Google Scholar

Lai, A., and Wang, Q. (2024). How coal de-capacity policy affects renewable energy development efficiency? Evidence from China. Energy 286, 129515. doi:10.1016/j.energy.2023.129515

CrossRef Full Text | Google Scholar

Lai, A., Du, Q., and Wang, Q. (2024). One region's gain is another's loss: the impact of market segmentation on renewable energy development in China. Energy Policy 192, 114215. doi:10.1016/j.enpol.2024.114215

CrossRef Full Text | Google Scholar

Lan, T., and Ma, J. (2025). How does a city's digital economy development impact pollution emissions? Front. Environ. Sci. 13, 1538077. doi:10.3389/fenvs.2025.1538077

CrossRef Full Text | Google Scholar

Lei, X., Ma, Y., Ke, J., and Zhang, C. (2023). The non-linear impact of the digital economy on carbon emissions based on a mediated effects model. Sustainability 15 (9), 7438. doi:10.3390/su15097438

CrossRef Full Text | Google Scholar

Li, X. (2019). New features and the formation mechanism of new growth drivers of digital economy. Reform 11, 40–51. Available online at: https://link.cnki.net/urlid/50.1012.F.20191106.1540.002.

Google Scholar

Li, Z. (2022). Extracting spatial effects from machine learning model using local interpretation method: an example of SHAP and XGBoost. Comput. Environ. Urban Syst. 96, 101845. doi:10.1016/j.compenvurbsys.2022.101845

CrossRef Full Text | Google Scholar

Li, Z. (2023). “The impact of digital economy development on carbon emissions: a multi-dimensional study based on threshold effect,” in Proceedings of the 2nd international academic conference on blockchain, information technology and smart finance, 1229–1243.

Google Scholar

Li, H., and Yang, Z. (2024). Does digital economy development affect urban environment quality: evidence from 285 cities in China. PLoS ONE 19 (2), e0297503. doi:10.1371/journal.pone.0297503

PubMed Abstract | CrossRef Full Text | Google Scholar

Li, X., Liu, J., and Ni, P. (2021). The impact of the digital economy on CO2 emissions: a theoretical and empirical analysis. Sustainability 13 (13), 7267. doi:10.3390/su13137267

CrossRef Full Text | Google Scholar

Li, Z., Cao, Y., Lai, A., and Wang, Q. (2025). Geospatial labor reallocation: the impact of China’s clean air action on inter-city labor distribution. Environ. Resour. Econ. 88 (3), 1937–1969. doi:10.1007/s10640-025-00996-w

CrossRef Full Text | Google Scholar

Liang, L., Wang, Z., and Li, J. (2019). The effect of urbanization on environmental pollution in rapidly developing urban agglomerations. J. Clean. Prod. 237, 117649. doi:10.1016/j.jclepro.2019.117649

CrossRef Full Text | Google Scholar

Liu, B., and Fang, X. (2025). Analysis of the impact of digital economy on carbon emissions and its mediation effect. Front. Environ. Sci. 13, 1635153. doi:10.3389/fenvs.2025.1635153

CrossRef Full Text | Google Scholar

Liu, J., Yang, Y., and Zhang, S. (2020). Research on the measurement and driving factors of China’s digital economy. Shanghai J. Econ. (6), 81–96. doi:10.19626/j.cnki.cn31-1163/f.2020.06.008

CrossRef Full Text | Google Scholar

Liu, H., Liu, J., Li, M., Gou, P., and Cheng, Y. (2022). Assessing the evolution of PM2.5 and related health impacts resulting from air quality policies in China. Environ. Impact Assess. Rev. 93, 106727. doi:10.1016/j.eiar.2021.106727

CrossRef Full Text | Google Scholar

Liu, B., Li, Y., Tian, X., Sun, L., and Xiu, P. (2023). Can digital economy development contribute to the low-carbon transition? Evidence from the city level in China. Int. J. Environ. Res. Public Health 20 (3), 2733. doi:10.3390/ijerph20032733

PubMed Abstract | CrossRef Full Text | Google Scholar

Liu, F., Kumar, J., Sun, H., and Edziah, B. K. (2025). Harnessing the digital economy for sustainable energy efficiency: an empirical analysis of China's Yangtze river Delta. Economics 19, 20250136. doi:10.1515/econ-2025-0136

CrossRef Full Text | Google Scholar

Ma, J., Ding, Y., Cheng, J. C. P., Jiang, F., Tan, Y., Gan, V. J., et al. (2020). Identification of high impact factors of air quality on a national scale using big data and machine learning techniques. J. Clean. Prod. 244, 118955. doi:10.1016/j.jclepro.2019.118955

CrossRef Full Text | Google Scholar

Pan, W., Xie, T., Wang, Z., and Ma, L. (2022). Digital economy: an innovation driver for total factor productivity. J. Bus. Res. 139, 303–311. doi:10.1016/j.jbusres.2021.09.061

CrossRef Full Text | Google Scholar

Pelegrina, G. D., Duarte, L. T., and Grabisch, M. (2023). A k-additive Choquet integral-based approach to approximate the SHAP values for local interpretability in machine learning. Artif. Intell. 325, 104014. doi:10.1016/j.artint.2023.104014

CrossRef Full Text | Google Scholar

Ping, Y., and Li, Z. (2024). Is there a digital rebound in the process of urban green development? New empirical evidence using ensemble learning methods. Sustainability 16 (10), 4281. doi:10.3390/su16104281

CrossRef Full Text | Google Scholar

Qu, J., and Hao, X. (2022). Digital economy, financial development, and energy poverty based on mediating effects and a spatial autocorrelation model. Sustainability 14 (15), 9206. doi:10.3390/su14159206

CrossRef Full Text | Google Scholar

Sapkota, P., and Bastola, U. (2017). Foreign direct investment, income, and environmental pollution in developing countries: panel data analysis of Latin America. Energy Econ. 64, 206–212. doi:10.1016/j.eneco.2017.04.001

CrossRef Full Text | Google Scholar

Shahbaz, M., Wang, J., Dong, K., and Zhao, J. (2022). The impact of digital economy on energy transition across the globe: the mediating role of government governance. Renew. Sustain. Energy Rev. 166, 112620. doi:10.1016/j.rser.2022.112620

CrossRef Full Text | Google Scholar

Shannon, C. E. (1948). A mathematical theory of communication. Bell Syst. Tech. J. 27 (3), 379–423. doi:10.1002/j.1538-7305.1948.tb01338.x

CrossRef Full Text | Google Scholar

Sun, P., Nisar, U., Qiao, Z., Ahmad, S., Kathuria, K., Bahir, A. A., et al. (2024). Digital economy, technology, and urban carbon emissions nexus: an investigation using the threshold effects and mediation effects tests. Front. Environ. Sci. 12, 1454256. doi:10.3389/fenvs.2024.1454256

CrossRef Full Text | Google Scholar

Tian, L., Zhang, C., and Lei, X. (2024). Digital economy's role in environmental sustainability: air quality enhancement through the 'Broadband China' initiative. Pol. J. Environ. Stud. 34, 7411–7421. doi:10.15244/pjoes/193135

CrossRef Full Text | Google Scholar

Wang, Z., and Zhang, Z. (2023). The carbon emission reduction effect and mechanism analysis of digital economy: evidence of prefecture-level cities in China. Appl. Econ. 56 (53), 6803–6820. doi:10.1080/00036846.2023.2276087

CrossRef Full Text | Google Scholar

Wang, J., Wang, W., Ran, Q., Irfan, M., Ren, S., Yang, X., et al. (2021a). Analysis of the mechanism of the impact of internet development on green economic growth: evidence from 269 prefecture cities in China. Environ. Sci. Pollut. Res. 29, 9990–10004. doi:10.1007/s11356-021-16381-1

PubMed Abstract | CrossRef Full Text | Google Scholar

Wang, Y., Liao, M., Wang, Y., Xu, L., and Malik, A. (2021b). The impact of foreign direct investment on China's carbon emissions through energy intensity and emissions trading system. Energy Econ. 97, 105212. doi:10.1016/j.eneco.2021.105212

CrossRef Full Text | Google Scholar

Wu, J., Lai, A., Li, Z., and Wang, Q. (2024a). Investment efficiency of renewable energy enterprises when exposed to air pollution: evidence from China. Int. Rev. Econ. & Finance 96 (Part C), 103722. doi:10.1016/j.iref.2024.103722

CrossRef Full Text | Google Scholar

Wu, L., Gui, Y., Zhang, Q., and Yu, M. (2024b). “Digital economy, environmental regulation, and carbon emissions: based on a dynamic panel threshold model,” in Proceedings of the 2024 9th international conference on intelligent information processing, 169–178.

Google Scholar

Xu, Q., Zhong, M., and Li, X. (2022a). How does digitalization affect energy? International evidence. Energy Econ. 107, 105879. doi:10.1016/j.eneco.2022.105879

CrossRef Full Text | Google Scholar

Xu, S., Yang, C., Huang, Z., and Failler, P. (2022b). Interaction between digital economy and environmental pollution: new evidence from a spatial perspective. Int. J. Environ. Res. Public Health 19 (9), 5214. doi:10.3390/ijerph19095074

PubMed Abstract | CrossRef Full Text | Google Scholar

Yang, Z., Gao, W., Han, Q., Qi, L., Cui, Y., and Chen, Y. (2022). Digitalization and carbon emissions: how does digital city construction affect China's carbon emission reduction? Sustain. Cities Soc. 87, 104201. doi:10.1016/j.scs.2022.104201

CrossRef Full Text | Google Scholar

Yang, P., Lin, Z., Zhu, Z., and Ying, F. W. (2024). Re-evaluating the impact and mechanism of digital economy on regional pollution intensity from the perspective of spatial spillover. Environ. Sci. Pollut. Res. 31, 9062–9077. doi:10.1007/s11356-023-31794-w

PubMed Abstract | CrossRef Full Text | Google Scholar

Yuan, H., Liu, J., Li, X., and Zhong, S. (2024). The impact of digital economy on environmental pollution: evidence from 267 cities in China. PLoS ONE 19, e0297009. doi:10.1371/journal.pone.0297009

PubMed Abstract | CrossRef Full Text | Google Scholar

Zeng, S., Yang, P., and Xu, Y. (2019). Internet popularization and the servitization of industrial structure: on the development strategy of rural service industry under the background of rural revitalization. Ind. Econ. Rev. 10, 36–55.

Google Scholar

Zhang, C. (2023). The impact of digital economy development on environmental pollution: an empirical study based on urban panel data. Jiangxi, China: Jiangxi University of Finance and Economics.

Google Scholar

Zhang, R., Fu, W., and Kuang, Y. (2022). Can digital economy promote energy conservation and emission reduction in heavily polluting enterprises? Empirical evidence from China. Int. J. Environ. Res. Public Health 19 (16), 9812. doi:10.3390/ijerph19169812

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhang, J., Liang, F., and Gao, P. (2024). Can big data reduce urban environmental pollution? Evidence from China's digital technology experimental zone. PLoS ONE 19 (10), e0288429. doi:10.1371/journal.pone.0288429

PubMed Abstract | CrossRef Full Text | Google Scholar

Zhao, S., Peng, D., Wen, H., and Wu, Y. (2023). Nonlinear and spatial spillover effects of the digital economy on green total factor energy efficiency: evidence from 281 cities in China. Environ. Sci. Pollut. Res. 30, 81896–81916. doi:10.1007/s11356-022-22694-6

PubMed Abstract | CrossRef Full Text | Google Scholar