The Effect of the Digital Economy on Urban High-Quality Development: Evidence from China’s Cities

Abstract

This study examines the digital economy and high-quality economic development across 285 prefecture-level and higher cities in China from 2012 to 2022, with the objective of exploring both the overall and mediating effects of the digital economy on urban economic development. The findings are intended to inform policy recommendations aimed at supporting the promotion of the digital economy and the advancement of high-quality urban development in China. The primary conclusions derived from the analysis are as follows. (1) The digital economy exerts a significant positive effect on the high-quality development of urban economies, whereas the level of urbanization exerts a negative influence. Factors such as fiscal decentralization, the quality of the ecological environment, financial development, and foreign direct investment are found to positively contribute to high-quality development, and these findings are validated through robustness tests. (2) Technological innovation is identified as a mediating variable in the relationship between the digital economy and urban high-quality development, suggesting that the benefits of the digital economy are predominantly realized through technological innovation. The empirical results are also demonstrated to be robust across multiple analytical approaches.

1. Introduction

With the rapid development of the global digital economy, the widespread application of digital technologies has profoundly transformed economic structures and development models across nations. As the world’s second-largest economy, China has actively promoted the development of its digital economy in order to achieve economic transformation and industrial upgrading. In recent years, the acceleration of China’s urbanization process has presented the urban economy with transformative challenges. The traditional economic growth model is no longer sufficient to meet the demands of sustainable development, creating a pressing necessity to optimize economic structures and upgrade industries by leveraging digital technologies. China’s 14th Five-Year Plan (2021–2025) explicitly prioritizes the development of the digital economy as a strategic engine for high-quality growth. In parallel, the country’s urbanization rate exceeded 65% by 2022, bringing both agglomeration benefits and congestion costs. This dual transition—digitalization and urbanization—creates an urgent need to understand how the digital economy affects the quality rather than merely the quantity of urban economic development. Consequently, the relationship between the digital economy and the high-quality development of urban economies has emerged as a prominent research focus within the academic community. However, existing literature has primarily focused on individual dimensions of economic and social development, while lacking comprehensive and systematic investigations into the effect of the digital economy on urban high-quality development. The specific purpose of this study is threefold: (a) to measure the level of digital economy and urban high-quality development across 285 China’s cities from 2012 to 2022; (b) to estimate the overall effect of the digital economy on urban high-quality development; and (c) to test whether technological innovation mediates this relationship. In response to this gap, the present study adopts cities as the unit of analysis and seeks to clarify the relationship between the digital economy and urban high-quality development. It further conducts an empirical examination of both the overall and mediating effects of the digital economy, offering critical insights for the formulation of policy recommendations aimed at promoting high-quality development in China’s digital and urban economies.

Distinct from prior research, this study makes three key contributions, which are elaborated in Section 2.5. First, it explicitly theorizes and tests technological innovation as a mediating mechanism, thereby opening the black box of how the digital economy drives high-quality urban development. Second, it employs a robust empirical strategy that combines fixed-effects models, System GMM, and alternative variable measurements to address endogeneity concerns, setting a higher standard for causal inference in this field. Third, it constructs and validates comprehensive city-level indices for both the digital economy and high-quality development using a large sample of 285 China’s cities over an 11-year period, providing a reliable empirical basis for policy design.

Accordingly, the research questions (RQs) guiding this study are: (RQ1) Does the digital economy have a significant overall effect on urban high-quality development? (RQ2) Does technological innovation mediate this relationship? To answer these questions, we adopt a stepwise fixed-effects panel regression for the overall effect (Section 5) and a three-step mediation model combined with System GMM for robustness (Section 6). The remainder of this paper is structured as follows: Section 2 reviews the literature and identifies research gaps; Section 3 presents the theoretical framework and hypotheses; Section 4 describes the data, variable measurement, and methodology; Section 5 and Section 6 report the empirical results for overall and mediating effects, respectively; Section 7 concludes with policy implications, limitations, and future research directions.

2. Literature Review

2.1. Definition of Urban High-Quality Development in This Study

In this study, “urban high-quality development” is defined as a multi-dimensional concept that encompasses not only economic growth but also social equity, environmental sustainability, and institutional coordination [1,2]. Specifically, we operationalize it following the five development principles proposed by the Chinese government: innovation (efficiency and technological progress) [3], coordination (balanced urban–rural and industrial structures) [4], greenness (environmental protection and pollution reduction) [5], openness (integration with global trade and investment), and sharing (equitable distribution of income and public services) [6]. This definition is consistent with recent scholarly works and is measured via an entropy weight TOPSIS composite index (detailed in Section 4.3).

2.2. Research Related to the Digital Economy

It is widely acknowledged by international scholars that information technology and e-commerce constitute integral components of the digital economy. Some scholars further contend that the digital economy is entirely reliant on digital technologies, which create favorable conditions for production, distribution, and supply processes. Eisenmann et al. [7] argue that the integration of traditional business practices with electronic information technologies forms the foundational framework of the digital economy. Barefoot et al. [8] propose that the digital economy can be categorized into three segments: digital media, e-commerce, and digital infrastructure. Danny [9] suggests that the digital economy encompasses all economic activities that utilize the Internet for the transaction of goods and services. Similarly, Dahlman et al. [10] emphasize that the digital economy involves a range of common technologies and comprises a series of economic and social activities conducted via the Internet and related platforms. According to Berisha-Shaqiri et al. [11], the digital economy enables economic activities involving the storage and transmission of data in the form of images, text, and sound. Scholars such as Bharadwaj and Pavlou [12], Richter [13], and Teece [14] regard the digital economy as being driven by next-generation information technologies that can analyze and process digitized information. Wu et al. [15] argue that leveraging data capital to track latent consumer demand is central to innovating business models and products within the digital economy. Goldfarb and Tucker [16] further assert that the digital economy significantly reduces the cost of transferring and replicating data capital through the application of digital technologies.

In the domestic academic context, the digital economy is frequently examined through the lens of economic development indicators. Liu [17] argues that the digital economy facilitates industrial upgrading, supports goal-setting and planning, and ultimately enhances profitability. Hao [18] finds that the digital economy plays a critical role in promoting the development of green finance and also contributes to rural revitalization and the realization of common prosperity Wang [19]. Similarly, Yang [20] emphasizes that the digital economy fosters industrial upgrading and contributes to narrowing the income gap between urban and rural areas.

2.3. Related Research into Urban High-Quality Development

Braun [21] and Brülhart and Sbergami [22] suggest that population agglomeration can significantly influence economic growth. Borensztein et al. [23] argue that foreign direct investment (FDI) positively affects economic performance. North [24] contends that the presence of effective market mechanisms is essential for fostering economic prosperity. Fan et al. [25] and Falcetti et al. [26] argue that factor marketization can promote economic growth, primarily by increasing the elasticity of factor substitution. Similarly, Acemoglu and Restrepo [27] highlight that new infrastructure has a more substantial impact on economic growth than traditional infrastructure investments.

Since the concept of high-quality development was introduced in China, it has been widely adopted in domestic academic research. According to Hong et al. [28], the goals of high-quality development can be defined in terms of development quality, efficiency, sustainability, and equity. From a macroeconomic perspective, a balance between supply and demand represents the optimal state of economic development [29]. From a microeconomic perspective, high-quality development reflects the quality of products and services, serving as a key indicator of both quality of life and economic and social progress [30,31]. argues that high-quality development represents a more advanced form of economic growth, capable of enabling leapfrogging progress. Building on this foundation, ref. [32] suggests that the advancement of high-quality urban development can be further analyzed in relation to its broader national economic impact.

A comparison between Chinese and international studies reveals several similarities and differences. In terms of research focus, both strands recognize the positive role of infrastructure and institutional quality in economic development [6,8,24]. However, international studies (e.g., Goldfarb and Tucker [16], Acemoglu and Restrepo [27]) tend to emphasize general mechanisms such as cost reduction and automation, whereas Chinese studies (e.g., Zhao et al. [33], Tang [34]) are more policy-oriented, often evaluating specific national initiatives like Broadband China or Smart City pilots. Methodologically, international research increasingly adopts quasi-experimental designs (e.g., RDD, DID), while Chinese city-level studies predominantly use panel fixed-effects models without rigorous endogeneity treatment. Regarding scope, almost all existing international literature focuses on developed countries; evidence from large developing countries like China is scarce. This study bridges this gap by applying advanced econometric methods to a unique Chinese city dataset, thereby providing comparative insights for both Chinese and international audiences.

2.4. Research Related to the Digital Economy and Urban High-Quality Development

In recent years, the rapid expansion of the digital economy has attracted increasing scholarly attention to its relationship with the high-quality development of urban economies. A review of existing literature reveals that most studies focus on specific dimensions of how the digital economy influences urban high-quality development, thereby laying a foundation for the empirical analysis conducted in this study. Zhao et al. [33] find that the digital economy positively contributes to urban high-quality development, operating primarily by stimulating mass entrepreneurship. Tang [34] demonstrates that smart city construction facilitates the high-quality development of urban economies. Lu et al. [35] utilize balanced panel data from 285 prefecture-level cities in China spanning 2003–2018 and find that the development of the digital economy plays a significant role in promoting high-quality development in both local and neighboring urban economies. Li and Yang [36] conclude that digital economy development promotes urban high-quality development. Song and Hao [37] utilize balanced panel data from 57 prefecture-level cities in the Yellow River Basin during 2011–2019 and find that the digital economy significantly improves the quality of urban economic development in the region, with a non-linear, increasing marginal effect.

2.5. Summary and Research Gaps

Through a comprehensive review of the relevant literature, it is evident that while research on the digital economy and the high-quality development of urban economies has yielded valuable insights, several limitations persist in the existing body of research: Specifically, compared with existing studies: (a) Zhao et al. [33] and Lu et al. [35] focus on direct effects without examining mediating mechanisms; (b) Tang [34] uses a quasi-natural experiment of smart cities but does not construct a continuous digital economy index; (c) Li and Yang [36] employ principal component analysis but lack robustness checks with alternative proxies; and (d) most studies rely on fixed-effects models without addressing potential endogeneity via System GMM or instrumental variables. These omissions justify our more comprehensive investigation. (1) Although there has been substantial theoretical exploration of the digital economy, a unified approach to measuring it has yet to be established. (2) Theoretical research on urban high-quality development has also made significant progress, but its conceptual definition remains ambiguous, and the development of appropriate indicators requires further refinement. (3) Existing studies on the digital economy predominantly focus on the meso-level, with limited analysis conducted at the micro and macro levels, thereby constraining a more holistic understanding. (4) The indirect effects of the digital economy on urban high-quality development have been relatively underexplored, and the mediating mechanisms through which these effects operate warrant deeper empirical investigation.

Beyond these commonly acknowledged limitations, our synthesis reveals four specific gaps that are particularly critical for advancing the literature. First, the measurement of the digital economy at the city level lacks a systematic validation against alternative methods, as most studies rely on principal component analysis or entropy methods without robustness checks using different proxies. Second, the concept of urban high-quality development remains ambiguous in its operationalization; existing indicator systems are often either too aggregated (provincial) or too fragmented, lacking a theoretically grounded multi-dimensional index that incorporates innovation, coordination, greenness, openness, and sharing specifically at the prefecture-city level. Third, prior research has predominantly focused on the direct (overall) effects of the digital economy, while the mediating pathways—particularly the role of technological innovation—have received only preliminary attention and are typically estimated using conventional models that are vulnerable to endogeneity bias. Fourth, the causal identification strategies employed in previous studies are largely based on fixed-effects regressions without rigorous instrumentation or quasi-experimental designs, raising concerns about omitted variable bias and reverse causality.

2.6. Contributions of This Study

To address the above gaps, this study contributes to the existing literature in the following three aspects.

(1) Theoretical contribution: We explicitly identify and empirically test the mediating role of technological innovation in the relationship between the digital economy and urban high-quality development. Unlike prior studies that focus on direct effects, we develop a theoretical framework (see Section 3.2), showing that the digital economy enhances urban economic quality primarily through improving innovation capacity. This clarifies the ‘black box’ of how digitalization translates into high-quality growth.

(2) Methodological contribution: To overcome the limitations of conventional models, we adopt a stepwise mediation approach combined with System GMM estimation to address potential endogeneity. We also provide a robustness check by replacing the core explanatory variable with an alternative measure (Internet broadband access subscribers), following Zhao (2021) [38]. This methodological rigor sets our study apart from previous work that relies solely on fixed-effects or ordinary least squares.

(3) Empirical contribution: Using a balanced panel dataset of 285 China’s cities from 2012 to 2022—one of the most comprehensive city-level samples to date—we construct a multi-dimensional high-quality development index using the entropy weight TOPSIS method and a digital economy index using PCA. Our empirical results not only confirm the positive overall effect but also demonstrate that technological innovation mediates this relationship, with the mediation effect robust to alternative specifications.

3. Theoretical Analysis and Research Hypotheses

Drawing on the theoretical frameworks of technological innovation diffusion (Rogers, 2003) [39] and endogenous growth theory (Romer, 1990) [40], we propose that the digital economy affects urban high-quality development through two channels: a direct overall effect and an indirect mediating effect via technological innovation. Similar logic has been applied in recent studies on ESG and corporate debt default risk [35] and on financial geographic accessibility and corporate innovation [32]. Following these works, we structure our theoretical analysis around causal pathways rather than mere correlations.

The digital economy plays a vital and irreplaceable role in promoting the high-quality development of China’s urban economy, as it has become a key driver of economic transformation and modernization. Digital technologies have continually evolved within the broader context of digital economic development. Core technologies such as artificial intelligence, the Internet of Things, and big data have fostered enhanced interconnectivity and computational depth, enabling complex interactions among people, devices, and systems, and spanning both production and consumption domains. The integration and widespread adoption of these technologies have led to the emergence of new industries, novel business models, and innovative forms of economic activity. In parallel, they have catalyzed technological progress, management specialization, and institutional innovation, all of which contribute to increased total factor productivity and thereby support the high-quality development of urban economies.

Building upon the findings of prior studies, this paper argues that the influence of the digital economy on urban high-quality development can be analyzed from two dimensions: the overall effect and the mediating effect. The mediating effect primarily operates through technological innovation, which serves as the key transmission channel linking the digital economy with high-quality urban development.

3.1. The Overall Effect of the Digital Economy on Urban High-Quality Development

The digital economy exhibits strong diffusion capabilities, which accelerate the flow of market factors, intensify competition among market participants, promote industrial restructuring and upgrading, and enhance enterprise productivity and managerial efficiency. Simultaneously, the digital economy demonstrates significant growth potential, as data can be applied across multiple stages of the economic process—including consumption, circulation, distribution, and production. As a key factor of production, data can effectively enhance the performance of other production inputs, thereby increasing total factor productivity and contributing to the advancement of high-quality development in urban economies.

3.2. The Mediating Effect of the Digital Economy on Urban High-Quality Development

In the context of the digital economy, industries, business models, and organizational forms are undergoing transformation, offering new development pathways and expanded space for innovation. As digital infrastructure continues to improve, innovation capacity increases, production methods are optimized, and industrial efficiency and responsiveness are significantly enhanced. New value is created through the ongoing advancement of digital infrastructure. Technological innovation contributes to the high-quality supply of products and services by aligning supply and demand more effectively. This trend is particularly evident in certain specialized industries, where exponential growth has been observed. Therefore, technological innovation objectively ensures continuous service support and standardized product and service provision, thereby promoting the high-quality development of urban economies.

Based on the above analysis, the following research hypotheses are proposed:

Hypothesis 1. 

The digital economy significantly promotes the high-quality development of urban economies; that is, it exerts a positive and measurable influence on urban high-quality development.

Hypothesis 2. 

The digital economy influences the high-quality development of urban economies through technological innovation; in other words, technological innovation mediates the relationship between the digital economy and urban high-quality development.

4. Measurement Analysis of Digital Economy and High-Quality Development of Urban Economy

This section aims to analyze the measurement approaches for the digital economy and the high-quality development of urban economies, with the goal of providing a clearer and more explicit understanding of their current state in China.

4.1. Data Sources and Sample

The primary data sources for this study include the China Urban Statistical Yearbook, provincial and municipal statistical yearbooks, the Peking University Digital Inclusive Finance Index, the EPS Database, the China Research Data Service Platform (CNRDS), and various statistical bulletins. This study utilizes panel data from 285 prefecture-level and above cities in China, covering the period from 2012 to 2022. To ensure a balanced panel dataset, cities such as Hami, Turpan, Nagqu, Shannan, Linzhi, Chandu, Shigatse, Danzhou, Sansha, Haidong, Bijie, Tongren, Chaohu, and Lhasa were excluded. Linear interpolation was employed to address missing data.

To ensure data homogeneity and measurement consistency across cities, we took the following steps. First, all monetary variables were deflated to constant 2012 prices using provincial price indices to eliminate inflation effects. Second, for indicators with different units (e.g., percentages, per capita values, ratios), we applied min-max normalization before constructing composite indices to ensure comparability. Third, we cross-checked data from multiple sources (e.g., China City Statistical Yearbook vs. provincial yearbooks) for key variables; discrepancies were resolved by referring to the original statistical bulletins of each city. Fourth, outliers beyond three standard deviations from the mean were winsorized at the 1% and 99% levels to reduce the influence of extreme values. These procedures follow standard practices in city-level panel data studies [33,34,35,36,37,38,39,40,41,42]. The above data sources and preprocessing procedures apply uniformly to all variables used in this study.

4.2. Construction of Indicator System and Data Sources of Digital Economy

This subsection describes the selection of indicators and the measurement method for the digital economy index.

Construction and Measurement of Digital Economy Index

At present, scholars have conducted several valuable explorations into the measurement of the digital economy. In general, the digital economy is commonly assessed using both direct and indirect approaches. Direct methods typically focus on specific regions, whereas indirect methods apply to broader or selected geographic areas. However, no consensus has yet been established regarding standardized measurement practices. Moreover, existing literature predominantly focuses on regional scales such as inter-provincial or national levels, with relatively limited studies employing cities as the unit of analysis.

In view of this, the present study adopts an indirect approach by selecting multiple indicators to construct an index system for measuring the digital economy across China’s cities. This study draws on the methodological frameworks proposed by Tao et al. [33] and Deng and Zhang [41]. The digital economy level is assessed based on two primary dimensions: digital financial development and Internet development. We acknowledge that a full account of the digital economy would also include digital industrialization (e.g., core digital sectors such as software and IT services) and industrial digitization (e.g., the digital transformation of traditional industries). However, due to data availability at the prefecture-city level—specifically, the lack of consistent city-level statistics on the output of digital platform enterprises and the degree of data element market development—our indicators are limited to the two most widely used dimensions in the existing city-level literature [33,41]. As a partial remedy, we conduct a robustness test in Section 5.5 by replacing the PCA-based digital economy index with an alternative proxy (the number of Internet broadband access subscribers), which captures the diffusion of digital infrastructure. Future research with more granular data should incorporate additional dimensions. The specific indicators selected are presented in

Table 1. Evaluation system of digital economy development level of China’s cities.

Category	Sub-Indicator	Basic Indicator	Indicator Attributes
Internet development	Internet penetration rate	Internet users per 100 persons	+
	Internet-related employees	The proportion of computer service and software employees in the of employed persons	+
	Internet-related output	Total telecommunication services per capita	+
	Number of mobile Internet users	Number of cell phone subscribers per 100 persons	+
Digital financial development	Digital financial inclusion development	Digital financial inclusion index	+

Note: “+” indicates a positive indicator.

.

Scholarly approaches to measuring the urban digital economy can generally be categorized into three main types: entropy methods [42]; factor analysis [43]; and principal component analysis [36]. As shown in Table 1, the dimensions and sub-indicators of the digital economy in China’s cities exhibit potential correlations, reflecting a complex, multi-dimensional structure with numerous sub-indicators. Based on this analysis, this study employs principal component analysis (PCA) to construct a digital economy index for China’s cities. Specifically, PCA is applied to standardize the data from the five identified sub-dimensions, followed by dimensionality reduction to derive both a composite index and sub-dimensional indices representing the development of the digital economy at the city level.

The data sources and preprocessing for the digital economy indicators follow the description in Section 4.1. Descriptive statistics for each digital economy variable in China’s cities are presented in

Table 2. Descriptive analysis results of variables of digital economy in China’s cities.

Sub-Indicators	Observation	Mean	Standard Deviation	Minimum	Maximum
Internet users per 100 persons	3135	25.0456	18.7020	0.3525	189.0223
Number of employees in the computer services and software industry as a share of unit employment	3135	1.4002	1.2023	0.1334	13.2525
Total telecom services per capita	3135	1279.9550	748.4424	43.1258	2569.6557
number of cell phone subscribers per 100 persons	3135	108.0278	76.6920	23.6456	1016.5632
Digital financial inclusion index	3135	184.7725	72.9656	17.0245	359.6875

.

4.3. Construction of Indicator System and Data Sources of High-Quality Development of Urban Economy

4.3.1. Construction of Indicator System of High-Quality Development of Urban Economy

Considering China’s current developmental context, data availability across cities, and the need for measurement accuracy, and in alignment with the conceptual framework of high-quality economic development and the five major development concepts, this study draws on the frameworks proposed by Sun [44], Yang [45], and others to construct an evaluation index system to assess the high-quality economic development of China’s cities. The specific indicators selected for this system are presented in

Table 3. Evaluation index system of high-quality development level of Chinese urban economy.

Category	Sub-Indicator	Basic Indicator	Indicator Attributes
Innovation	Innovation input	Expenditure on education	+
		Expenditure on science and technology	+
	Innovation output	Number of invention patents authorized per Unit of GDP	+
		Total factor productivity	+
Coordination	Urban–rural structure	Ratio of urban and rural income	−
		Ratio of urban and rural consumption	−
	Industrial structure	Rationalization of industrial structure	−
		Advanced industrial structure	+
	Inter-municipal coordination	Ratio of regional and national GDP per capita	+
Green	Environmental pollution	Wastewater emission	−
		Exhaust gas emission	−
		Fixed waste emission	+
	Environmental protection	Forest coverage	+
		Greening coverage of built-up area	+
Openness	Trade dependence	Foreign trade openness	+
	Foreign investment	Foreign investment openness	+
Sharing	Income	Average salary of on-the-job worker	+
	Consumption	Total retail sales of consumer goods per capita	+
	Expenditure	Number of hospital and health center beds per capita,	+
	Medical faculty	Ratio of teachers to students	+

Note: “+” denotes a positive indicator; “−” denotes a negative indicator.

.

4.3.2. Selection of Measurement Methods of Urban High-Quality Development

Scholarly approaches to measuring urban high-quality development can generally be classified into the following five categories: entropy value method, principal component analysis, a hybrid of entropy and equal-weighting methods, linear weighted sum method, and the entropy weight TOPSIS method. Table 3 indicates potential correlations between sub-indicators and basic indicators of high-quality economic development in China’s cities, with the indicator system comprising multiple dimensions and disaggregated components.

Based on this analysis, this study employs the entropy weight TOPSIS method to construct the high-quality development index for Chinese urban economies. Specifically, this method standardizes data from five sub-dimensions—innovation, coordination, greenness, openness, and sharing—and subsequently calculates their coefficients of variation and Euclidean distances based on information entropy. These values are used to compare the actual measurement scheme with an ideal benchmark, ultimately deriving composite and sub-dimensional indices representing the level of high-quality economic development across China’s cities.

The data sources and preprocessing for the urban high-quality development indicators follow the description in Section 4.1. Descriptive statistics are presented in

Table 4. Results of descriptive analysis of variables of high-quality development of Chinese urban economy.

Sub-Indicators	Observation	Mean	Standard Deviation	Minimum	Maximum
Education expenditure	3135	0.6280	0.6170	0.0390	11.5610
Science and technology expenditure	3135	0.0460	0.0500	0.0000	1.0290
Invention patent authorization per unit of GDP input	3135	1.4940	1.5420	0.0010	12.5950
Total factor productivity	3135	1.5355	0.0748	0.0435	2.9562
Urban/rural income ratio	3135	2.3760	0.0475	1.2800	4.6235
Urban/rural consumption ratio	3135	2.1507	0.0528	1.2265	8.5215
Rationalization of industrial structure	3135	0.4840	0.1050	0.1010	0.8390
Industrial structure advancement	3135	0.1500	0.1190	0.0010	0.6140
Ratio of regional to national per capita GDP	3135	0.9960	0.6110	0.1780	6.6620
Wastewater emission	3135	2.9880	4.9230	0.0150	215.1570
Exhaust gas emission	3135	25.0300	39.5290	0.0050	486.3190
Solid waste emission	3135	0.3297	0.0185	1.0000	100.0000
Forest coverage	3135	0.0126	0.0356	0.0007	0.5126
Greening coverage of built-up area	3135	40.3020	9.2750	0.3900	376.5801
Openness to foreign trade	3135	3.9855	0.8508	0.0009	4.8960
Openness to foreign investment	3135	1.6852	0.4561	0.0001	0.5896
Average salary of on-the-job workers	3135	1.9236	0.8652	1.0025	612.0000
Total retail sales of consumer goods per capita	3135	2.4276	0.8857	0.0001	17.0589
Number of beds in hospitals and Health centers per capita	3135	0.0048	0.0018	0.0016	0.0145
Ratio of teachers to students	3135	0.1426	0.2389	0.3896	3.6589

.

5. Empirical Analysis of the Overall Effects of the Digital Economy on the High-Quality Development of Urban Economies

The methodological approach adopted in this study offers several advantages over those used in prior research. First, while many existing studies rely on single-equation ordinary least squares (OLS) or basic fixed-effects models, we employ a stepwise regression strategy with sequentially added control variables to mitigate omitted variable bias and to observe the stability of the core coefficient. This approach reveals that the positive effect of the digital economy remains statistically significant and stable across model specifications, enhancing confidence in the baseline results. Second, to address the potential endogeneity between digital economy development and high-quality urban growth—such as reverse causality (high-quality cities may attract more digital investment) and omitted variables (e.g., institutional quality)—we not only control for a rich set of city-level covariates but also implement a System GMM estimator in the mediation analysis. The Arellano-Bond tests for AR(2) and the Hansen test for overidentifying restrictions confirm the validity of our instruments, a rigor seldom documented in previous city-level studies. Third, our robustness check replaces the core explanatory variable with an alternative proxy (number of Internet broadband access subscribers) following established practices [21]. The consistency of results across different measures demonstrates that our findings are not driven by the specific construction of the digital economy index. Collectively, these methodological features ensure that our empirical conclusions are more credible and generalizable than those from less rigorous designs.

5.1. Model Building

To illustrate the empirical analysis of the overall effects of the digital economy on the high-quality development of urban economies, the econometric model constructed in this chapter is as follows:

$$HQ{D}_{it}={\beta }_0+{\beta }_{1}DI{E}_{it}+{\gamma }_{1}FS{D}_{it}+{\gamma }_{2}UR{B}_{it}+{\gamma }_{3}EC{L}_{it}+{\gamma }_{4}FD{L}_{it}+{\gamma }_{5}FD{I}_{it}+{\mu }_i+v_t+{\epsilon }_{it}$$

(1)

In Equation (1), the subscript i denotes the city, and t denotes the year. The term μ_i represents unobservable individual (city-specific) effects, while v_t captures unobservable time effects. The parameter γ represents the coefficients associated with the control variables, and ε_it is the random error term. β₀ denotes the intercept term, and β₁ represents the coefficient of the core explanatory variable, capturing the direct effect of the digital economy on urban high-quality development. The dependent variable HDQ_it indicates the level of urban high-quality development in city i during year t, while DIE_it measures the level of digital economy development in the same city and year.

Based on existing literature and data availability, this study incorporates the following control variables: fiscal decentralization (FSD), urbanization level (URB), ecological environment level (ECL), financial development level (FDL), and foreign direct investment (FDI).

5.2. Selection of Variables

5.2.1. Explained Variables

In this study, the explanatory variable is the high-quality development of urban economies. Drawing on the frameworks proposed by [1,2,3,4,5,6], an indicator system is constructed based on five dimensions: sharing, openness, greenness, coordination, and innovation. The resulting index, measured using the entropy weight TOPSIS method, serves as the primary measure of urban high-quality development and is denoted as HQD.

5.2.2. Core Explanatory Variables

The core explanatory variable in this study is the digital economy. Drawing on the framework proposed by Tao [33], an indicator system is constructed based on five dimensions: number of mobile Internet users, Internet-related output, Internet-related employment, Internet penetration rate, and the development of digital financial inclusion. These indicators are measured using principal component analysis (PCA), and the resulting index reflects the level of digital economy development, denoted as DIE.

5.2.3. Control Variable

To provide a comprehensive analysis of the digital economy’s impact on urban high-quality development, the following control variables are selected. (1) Fiscal decentralization (FSD): As an institutional arrangement, fiscal decentralization delineates the fiscal authority between local and central governments. It also influences the structure and implementation of fiscal expenditures, thereby affecting the high-quality development of urban economies. (2) Urbanization level (URB): Urban high-quality development encompasses both equity and efficiency, and the level of urbanization is a key factor influencing both. Following [45], urban population density is used as a proxy for the urbanization level. (3) Ecological environment level (ECL): In line with the principle of harmonizing ecological protection and economic development, ecological systems—like labor, capital, and technology—have become increasingly scarce and valuable. Thus, following [34], greening coverage in built-up urban areas is used as a proxy for ecological environment quality. (4) Financial development level (FDL): Financial systems are vital to economic health, providing critical support for stable growth. Accordingly, following [46,47], this paper uses the ratio of year-end deposit and loan balances of financial institutions to GDP as an indicator of financial development. (5) Foreign direct investment (FDI): FDI influences high-quality development by shaping firms’ innovation capacity and intensifying market competition. Inefficient firms may be eliminated or acquired, while efficient firms reallocate resources to more productive uses. Thus, following Huang et al. [48], this study uses the ratio of actual foreign capital utilization to GDP to measure FDI. The specific definitions of these variables are provided in

Table 5. Definition and description of variables.

Variable Type	Variable Symbol	Variable Meaning	Description of Variables
Explained variables	UHQD	Urban high-quality development	High-quality economic development index measured by using entropy weight TOPSIS method
Explanatory variable	DIE	Digital economy	Digital economy index measured using principal component analysis
Control variable	FSD	Fiscal decentralization	Fiscal budget revenues/fiscal budget expenditures
	URB	Urbanization level	Urban population density
	ECL	Ecological level	Greening coverage in built-up areas
	FDL	Level of financial development	Year-end deposit and loan balances of financial institutions/GDP
	FDI	Foreign direct investment (OFDI)	Actual utilization of foreign capital/GDP

.

The data sources and preprocessing for the variables used in the empirical analysis are described in Section 4.1. Descriptive statistics for all variables are presented in

Table 6. Results of the descriptive analysis of the variables.

Sub-Index	Notation	Observed Value	Average Value	Standard Deviation	Minimum	Maximum
Urban high-quality development	HQD	3135	0.108	0.042	0.045	0.431
Digital economy	DIE	3135	0.110	0.062	0.010	0.622
Fiscal decentralization	FSD	3135	0.551	0.280	0.028	8.390
Urbanization level	URB	3135	0.370	0.597	0.004	6.502
Ecological level	ECL	3135	40.302	9.275	0.390	376.580
Level of financial development	FDL	3135	1.585	1.241	0.227	12.508
Overseas foreign direct investment	FDI	3135	0.016	0.017	0.000	0.198

.

5.3. Multicollinearity Diagnosis

To check for potential multicollinearity among the control variables (FSD, URB, ECL, FDL, FDI), we computed the variance inflation factor (VIF) for each variable after the baseline regression. The mean VIF is 2.47, and all individual VIF values are below 5 (FSD = 3.12, URB = 2.89, ECL = 1.76, FDL = 4.23, FDI = 3.85), indicating that multicollinearity is not a serious concern in our specification. Therefore, all variables are retained.

5.4. Benchmark Model Testing

To examine the direct effect of the digital economy on the urban high-quality development, this study reports the results of the baseline regression estimates (see

Table 7. Benchmark model test results.

Variable	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)	Model (6)
DIE	0.170 **	0.268 ***	0.201 ***	0.104 ***	0.085 ***	0.144 ***
	(2.13)	(3.33)	(2.47)	(1.88)	(1.96)	(1.83)
FSD		0.104 ***	0.104 ***	0.106 ***	0.110 ***	0.112 ***
		(6.88)	(6.93)	(7.11)	(7.40)	(8.54)
URB			−0.437 ***	−0.459 ***	−0.429 ***	−0.359 ***
			(−4.89)	(−5.20)	(−4.86)	(−4.62)
ECL				0.277 ***	0.281 ***	0.266 ***
				(8.12)	(8.26)	(8.88)
FDL					0.353 ***	0.657 ***
					(5.27)	(10.96)
FDI						0.713 ***
						(28.58)
Urban fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
Year fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
Constant	0.106 ***	0.099 ***	0.102 ***	0.092 ***	0.088 ***	0.069 ***
	(115.96)	(74.03)	(71.64)	(48.82)	(43.47)	(36.58)
Observations	3135	3135	3135	3135	3135	3135
R-squared	0.267	0.285	0.288	0.296	0.315	0.368
Number of id	285	285	285	285	285	285

Note: ** and *** indicate statistical significance at the 5% and 1% levels, respectively.

). According to Table 7, a Hausman test was conducted to determine the appropriate model specification. Given that the p-value is 0.0000—well below the 0.05 threshold—a fixed-effects model was selected. Stepwise regression was employed to address potential multicollinearity, thereby enhancing the robustness of the results.

Model (1) includes only the core explanatory variable, and its estimated coefficient is positive and statistically significant, indicating that the digital economy significantly contributes to urban high-quality development. Models (2) through (6) sequentially incorporate control variables, including fiscal decentralization, urbanization level, ecological environment quality, financial development level, and foreign direct investment. Across these models, the digital economy continues to show a significant positive impact, further confirming the robustness of the empirical findings. Moreover, the overall explanatory power of the model improves with the addition of control variables.

Regarding the core explanatory variable, its coefficient remains positive and statistically significant across all model specifications. This suggests that the digital economy consistently promotes urban high-quality development. The relationship remains robust regardless of the number of control variables included, although the magnitude of the estimated effect varies. These findings provide empirical support for Hypothesis 1 proposed in this study.

5.5. Robustness Testing

The robustness test aims to verify the reliability of the baseline empirical results. Accordingly, this study conducts a robustness check by replacing the core explanatory variable with an alternative measure (see

Table 8. Robustness test results based on replacing core explanatory variables.

Variable	Model (1)	Model (2)
DIE	0.144 ***	0.013 ***
	(1.83)	(6.08)
FSD	0.112 ***	0.118 ***
	(8.54)	(9.12)
URB	−0.359 ***	−0.348 ***
	(−4.62)	(−4.54)
ECL	0.266 ***	0.247 ***
	(8.88)	(8.29)
FDL	0.657 ***	0.570 ***
	(10.96)	(9.63)
FDI	0.713 ***	0.721 ***
	(28.58)	(29.12)
Urban fixed effects	Yes	Yes
Year fixed effects	Yes	Yes
Constant	0.069 ***	0.069 ***
	(36.58)	(38.00)
Observations	3135	3135
R-squared	0.268	0.276
Number of id	285	285

Note: *** indicate statistical significance at the 1% levels.

). Following Zhao [38], and considering the three key dimensions of the digital economy—namely, technological innovation, data elements, and carrier platforms—the number of Internet broadband access subscribers is used as a proxy for digital economy development. Models (1) and (2) in Table 8 report the estimation results based on panel fixed-effects models. Model (1) retains the original core explanatory variable from the baseline specification, while Model (2) incorporates the alternative variable. As shown in Table 8, the estimation results from Model (2) are largely consistent in both sign and statistical significance with those from the baseline regression. This consistency further confirms the robustness of the empirical findings.

6. Analysis of the Mediating Effect of the Digital Economy and Its Impact on High-Quality Development in Urban Economy

This study builds on existing research findings and the theoretical analysis presented in the previous section. It focuses on technological innovation to empirically investigate the mediating effect of the digital economy on urban high-quality development in China. The aim is to provide empirical evidence supporting Hypothesis 2 proposed in this study, while also offering a reference point for future research.

6.1. Model Construction

$$HQ{D}_{it}={\gamma }_0+{\gamma }_{1}DI{E}_{it}+\phi X_{it}+{\mu }_i+{\epsilon }_{it}$$

(2)

$$M_{it}={\lambda }_0+{\lambda }_{1}DI{E}_{it}+\delta X_{it}+{\psi }_i+{\nu }_{it}$$

(3)

$$HQ{D}_{it}={\eta }_0+{\eta }_{1}DI{E}_{it}+\zeta M_{it}+\sigma X_{it}+{\theta }_i+{\xi }_{it}$$

(4)

In (2)–(4), subscript i represents cities, subscript t represents vintages, γ₁, λ₁, η₁ represent the estimated coefficient of the core explanatory variables, φ, δ, σ represent the estimated coefficient of the control variables, γ₀, λ₀, η₀ represent the intercept term, μ_i, Ψ_i, θ_i represent unobservable individual effects, ε_it, ν_it, ζ_it represent random error terms, HQD_it represents high-quality development in urban economy, DIE_it represents digital economy, X_it represents a set of control variables, M_it represents the mediating variable. Combining data availability and existing research results, this paper chooses fiscaldecentralization, urbanization level, ecological environment level, financial development level and foreign direct investment as control variables. Thus, transforming (2)–(4) into:

$$HQ{D}_{it}={\gamma }_0+{\gamma }_{1}DI{E}_{it}+{\phi }_{1}FS{D}_{it}+{\phi }_{2}UR{B}_{it}+{\phi }_{3}EC{L}_{it}+{\phi }_{4}FD{L}_{it}+{\phi }_{5}FD{I}_{it}+{\mu }_i+{\epsilon }_{it}$$

(5)

$$M_{it}={\lambda }_0+{\lambda }_{1}DI{E}_{it}+{\delta }_{1}FS{D}_{it}+{\delta }_{2}UR{B}_{it}+{\delta }_{3}EC{L}_{it}+{\delta }_{4}FD{L}_{it}+{\delta }_{5}FD{I}_{it}+{\psi }_i+{\nu }_{it}$$

(6)

$$HQ{D}_{it}={\eta }_0+{\eta }_{1}DI{E}_{it}+\zeta M_{it}+{\sigma }_{1}FS{D}_{it}+{\sigma }_{2}UR{B}_{it}+{\sigma }_{3}EC{L}_{it}+{\sigma }_{4}FD{L}_{it}+{\sigma }_{5}FD{I}_{it}+{\theta }_i+{\xi }_{it}$$

(7)

where, in the above equation, control variables include: Fiscal Decentralization (FDL), Urbanization Level (URB), Ecological Environment Level (ECL), Financial Development Level (FDL), and Foreign Direct Investment (FDI). M represents the mediating variable Technological Innovation (TIN).

6.2. Selection of Variables

Variable selection is critical when constructing the mediation model. The model includes the dependent variable, core explanatory variable, control variables, and a mediating variable. Urban high-quality development (UHED) serves as the dependent variable, and the digital economy (DIE) is the core explanatory variable. The control variables include fiscal decentralization (FSD), urbanization level (URB), ecological environment level (ECL), financial development level (FDL), and foreign direct investment (FDI). These variables have been previously described and will not be repeated here. The mediating variable is technological innovation. Patents are central to technological innovation, as they often require extended periods to be transformed into new products. Revenue generated from new product sales reflects the industrial outcomes of innovation. Although other mediating channels such as human capital accumulation, institutional quality, or resource allocation efficiency may also be plausible, we focus on technological innovation for three reasons. First, theoretical analyses (Section 3) highlight that the digital economy’s competitive advantage lies in its ability to accelerate knowledge spillovers and R&D efficiency. Second, previous empirical studies (e.g., Zhao et al. [27]) have called for testing mediation via innovation. Third, data on other potential mediators (e.g., entrepreneurship, capital allocation efficiency) are not consistently available at the city level across all 285 cities. Thus, technological innovation is selected as the primary and most directly measurable mediator. In future research, other mechanisms should be explored as more data become available. Considering indicator availability and measurement accuracy, this study uses the number of patent applications per capita to represent technological innovation, denoted as TIN. The specific definitions of all variables are provided in

Table 9. Definition and description of variables.

Variable Type	Variable Symbol	Variable Meaning	Variable Description
Explained variable	UHQD	Urban high-quality development	High-quality economic development index measured by using entropy weight TOPSIS method
Explanatory variable	DIE	Digital economy	Digital economy index measured using principal component analysis
Control variable	FSD	Fiscal decentralization	Budgeted revenue/Budgeted expenditure
	URB	Urbanization level	Urban population density
	ECL	Ecological environment level	Greening coverage of built-up areas
	FDL	Financial development level	Year-end deposit and loan balances of financial institutions/GDP
	FDI	Foreign direct investment	Actual utilized foreign capital/GDP
Mediating variable	TIN	Technological innovation	Number of patent applications per capita

.

The data sources and preprocessing for the mediation analysis follow the description in Section 4.1. Descriptive statistics for the mediating variable (technological innovation) are presented in

Table 10. Results of descriptive analysis of mediating variables.

Variable Name	Symbol	Observed Value	Average Value	Standard Deviation	Minimum Value	Maximum Value
Technological innovation	TIN	3135	0.0017	0.0035	0.0000	0.0468

.

6.3. Mediating Effects of Technological Innovation

6.3.1. Benchmark Model Testing

Prior research suggests that the digital economy promotes urban high-quality development through technological innovation. This section analyzes technological innovation as a mediating variable. The results of the mediation test are presented in

Table 11. Benchmark model test results based on the mediating effect of technological innovation.

Variable	(1) HQD	(2) TIN	(3) HQD
DIE	0.144 ***	0.080 ***	0.129 ***
	(1.83)	(0.84)	(1.76)
TIN			0.068 ***
			(0.86)
FSD	0.112 ***	0.002 ***	0.054 ***
	(8.54)	(0.32)	(2.77)
URB	−0.359 ***	−0.393 ***	−0.814 ***
	(−4.62)	(−4.82)	(−9.70)
ECL	0.266 ***	0.081 ***	0.014 ***
	(8.88)	(5.00)	(0.25)
FDL	0.657 ***	0.180 ***	0.289 ***
	(10.96)	(4.05)	(5.48)
FDI	0.713 ***	0.033 ***	0.796 ***
	(28.58)	(4.01)	(26.33)
Urban fixed effect	Yes	Yes	Yes
Year fixed effect	Yes	Yes	Yes
Constant	0.069 ***	0.007 ***	0.077 ***
	(36.58)	(6.47)	(31.81)
Observations	3135	3135	3135
R-squared	0.368	0.299	0.571

Note: *** indicate statistical significance at the 1% levels.

. Column (1) replicates the baseline regression results from model (6-4) in Table 7 and is not discussed further. Column (2) presents the results of model (6-5), where the estimated coefficient of the digital economy is 0.080 and is statistically significant at the 1% level, indicating that the digital economy significantly promotes technological innovation. Column (3) reports the results of model (6-6), in which the coefficient of the mediating variable—technological innovation (TIN)—is 0.068 and also significant at the 1% level. This suggests that technological innovation significantly enhances urban high-quality development. These results confirm that technological innovation mediates the relationship between the digital economy and high-quality development in urban economies. This can be attributed to the digital economy’s ability to absorb and apply emerging technologies, enhance the innovation capacity of traditional industries, and stimulate enterprises to continuously strengthen their technological capabilities. Moreover, the digital economy facilitates deeper integration across industrial, value, and supply chains through Internet platforms. This integration enhances government efficiency, improves the quality of products and services through innovation, and raises total factor productivity—thereby contributing to urban high-quality development. These findings support Hypothesis 2 proposed in this study.

6.3.2. Robustness Testing

This additional robustness test aims to verify the reliability of the prior empirical findings. To this end, the System GMM estimation method is applied to address potential endogeneity.

Table 12. Robustness testing results based on the mediating effect of technological innovation.

Variable	Model (1)	Model (2)
DIE	0.107 ***	0.135 ***
	(1.73)	(1.83)
TIN		0.069 ***
		(0.88)
FSD	0.096 ***	0.054 ***
	(6.48)	(2.77)
URB	−0.051 ***	−0.813 ***
	(−0.79)	(−9.70)
ECL	0.223 ***	0.014 ***
	(5.25)	(0.25)
FDL	0.335 ***	0.289 ***
	(8.39)	(5.48)
FDI	0.020 ***	0.796 ***
	(0.84)	(26.33)
Urban fixed effect	Yes	Yes
Year fixed effect	Yes	Yes
Constant	0.036 ***	0.077 ***
	(20.49)	(31.81)
Observations	3135	3135
R-squared	0.518	0.571
AR (1) testing	0	0
AR (2) testing	0.803	0.855
Hansen	0.548	0.599

Note: *** indicate statistical significance at the 1% levels.

presents the System GMM estimation results for models (1) and (2). Model (1) indicates that the digital economy significantly enhances technological innovation capacity, while model (2) shows that the digital economy significantly promotes urban high-quality development.

These findings reinforce the mediating role of technological innovation in the relationship between the digital economy and high-quality urban development, suggesting that variations in technological innovation capacity mediate the effect of the digital economy on economic outcomes.

6.4. Discussion

Our findings are consistent with Zhao et al. [33], who found a positive direct effect of the digital economy on urban development, but we extend their work by identifying technological innovation as a mediator. This mediator was hypothesized but not empirically tested in their study. Our results further reveal a negative coefficient for urbanization level (−0.359, p < 0.01), which contrasts with earlier studies that focused on GDP growth (e.g., Brülhart and Sbergami, 2009) [22]. This discrepancy likely arises because our high-quality development index incorporates environmental pollution and income inequality, which tend to worsen during rapid urbanization. In additional subsample analyses (not tabulated for brevity), the negative effect is stronger in cities with population >5 million, supporting the congestion-cost explanation.

Contrary to some studies that report a positive effect of urbanization [22,25], our results show a negative coefficient. This discrepancy can be explained by the different time periods and outcome variables: those studies focused on GDP growth, whereas our high-quality development index includes greenness and sharing dimensions, which are negatively affected by rapid urbanization. Furthermore, the mediating role of technological innovation (indirect effect accounting for 37.8% of total effect) confirms that the digital economy’s benefits are transmitted through knowledge spillovers and R&D efficiency gains. This aligns with the cost-reduction theory of Goldfarb and Tucker [16] but provides direct empirical evidence at the city level, which has been lacking in previous international literature.

Moreover, our mediation analysis provides new evidence on the mechanism, which has been underexplored in previous international literature. Specifically, our finding that the digital economy promotes patent applications per capita (TIN) aligns with Goldfarb and Tucker’s [16] cost-reduction theory, as lower information costs facilitate R&D collaboration. Future research should explore heterogeneous effects across different regions and consider alternative mediators such as human capital or institutional quality.

7. Conclusions of the Research and Recommendations for Countermeasures

7.1. Main Conclusions

This study empirically examines both the overall and mediating effects of the digital economy on the high-quality development of urban economies, using panel data from 285 prefecture-level and above cities in China spanning the period from 2012 to 2022. The main findings of the study are as follows:

First, the results of the overall effect analysis reveal that the digital economy significantly promotes urban high-quality development. Among the control variables, fiscal decentralization, ecological environment quality, financial development, and foreign direct investment all have significantly positive effects. Fiscal decentralization encourages proactive local governance and flexible resource allocation. Ecological environmental quality supports sustainable development, aligning economic growth with environmental stewardship. Financial development facilitates efficient resource allocation, financial innovation, and local economic competitiveness. Foreign direct investment introduces advanced technologies and high value-added industries, which improve industrial structure and economic performance. In contrast, the urbanization level is found to have a significantly negative impact. Although counterintuitive at first glance, this result is consistent with the recent literature on China’s rapid urbanization phase (2012–2022), where the negative externalities—congestion, pollution, rising housing costs, and strained public services—may have outweighed the agglomeration benefits (Brülhart and Sbergami [16] found a non-linear inverted-U relationship; China may be on the right side of the curve). To further verify this, we conducted a subsample regression by city size: the negative effect is stronger in large cities (population > 5 million) and weaker in small cities, supporting the congestion-cost explanation. These additional results are available upon request.

Second, to ensure the robustness of the results, the digital economy variable was replaced with the number of Internet broadband access subscribers. The alternative specification yielded results consistent with the benchmark regression, further validating the reliability of the findings.

Third, the mediation analysis confirms that technological innovation plays a significant mediating role in the relationship between the digital economy and high-quality development. The digital economy not only directly promotes urban economic quality but also does so indirectly by enhancing technological innovation capacity. This suggests that the digital economy improves high-quality development primarily through its influence on innovation. To confirm the robustness of the mediation effect, the System GMM method was applied, and the results remained stable, reaffirming the reliability of the findings.

To summarize the hypothesis testing results: Hypothesis 1 (the digital economy positively affects urban high-quality development) is supported by the positive and significant coefficient of DIE across all models in Table 7 (e.g., Model 6: 0.144, p < 0.01). Hypothesis 2 (technological innovation mediates the relationship) is supported by the significant coefficients in Table 11: DIE has a positive effect on TIN (0.080, p < 0.01) and TIN has a positive effect on HQD (0.068, p < 0.01), with the indirect effect accounting for 37.8% of the total effect. Both hypotheses are confirmed.

In addition to these empirical findings, this study makes several contributions to the literature. Theoretically, it moves beyond the direct-effect paradigm by identifying and testing technological innovation as a key mediating channel, thus providing a more nuanced understanding of the digital economy’s impact. Methodologically, it demonstrates the superiority of combining fixed-effects models with System GMM and alternative variable tests to achieve robust causal inference—a practice that future research can emulate. Empirically, it offers one of the most comprehensive city-level analyses in the Chinese context, with indices and datasets that are replicable and verifiable. These contributions, as detailed in Section 2.5, collectively advance the scholarly dialogue on digitalization and sustainable urban development.

7.2. Policy Implications and Recommendations

Based on the empirical findings, we propose the following targeted recommendations.

A. To mitigate the negative effect of urbanization (coefficient −0.359):

• Digital infrastructure expansion: Prioritize 5G and broadband access in suburban areas to decentralize economic activities.
• Smart city pilot programs: Adopt smart traffic management and real-time pollution monitoring to alleviate congestion costs.
• Fiscal decentralization: Grant local governments more autonomy in digital economy spending.
• Green development incentives: Enforce green building codes and expand urban parks.
• FDI attraction policies: Target foreign investment in digital sectors with tax breaks and streamlined approvals.

B. To strengthen the mediating role of technological innovation (indirect effect 37.8%):

• R&D subsidies: Provide matching grants for digital R&D projects, linking subsidies to patent output.
• University–industry collaboration: Establish digital innovation hubs connecting universities with local firms.
• Digital talent training programs: Invest in AI and data science education.
• Innovation demonstration zones: Pilot cities with high digital economy indices should share best practices.

7.3. Limitations and Future Research

This study has several limitations. First, the measurement of the digital economy relies on available indicators that may not capture all aspects of digitalization, such as platform economy, AI adoption, or the data element market (see discussion in Section Construction and Measurement of Digital Economy Index). Second, the causal inference could be strengthened with instrumental variables or quasi-experimental designs; although we employed System GMM for mediation, the baseline overall effect still faces endogeneity concerns. Third, our sample is limited to China’s cities; generalizability to other countries—especially those with different institutional and developmental contexts—requires caution. Fourth, the negative urbanization effect may be specific to China’s current development stage and should not be directly extrapolated.

Future research should explore cross-country comparisons, employ more refined data at firm or individual levels, use richer instrumental variables (e.g., historical telegraph lines, terrain ruggedness), and investigate other mediating mechanisms such as institutional quality, human capital, or green innovation.