Fostering Domestic Demand Through Digital–Real Economy Integration: Evidence from Household Consumption in China

Abstract

This paper examines the relationship between digital–real economy integration (DREI) and household consumption using panel data from 30 Chinese provinces over the period 2014–2024. By developing an entropy-weighted modified coupling coordination model, the level of DREI is quantitatively measured, and the mechanisms of its impact and its heterogeneity at the household consumption level are explored. The empirical results show that strengthening DREI has a significant positive impact on household consumption, particularly in China’s central region and in the goods sector. Additionally, this paper identifies three primary channels through which DREI can promote household consumption: optimizing the business environment to minimize transaction costs, developing logistics to improve the efficiency of resource allocation, and promoting financial inclusion to boost household consumption potential. The findings in this paper have significant policy implications for leveraging DREI to transform China’s economic growth pattern towards high-quality domestic demand in the era of artificial intelligence and sustainable development.

1. Introduction

This paper is meant to assess the influence of digital–real economy integration (DREI) on household consumption. In this regard, Chawla [1] highlights that artificial intelligence can be regarded as a fundamental driver of productivity and economic growth, significantly improving decision-making through advanced data analytics [2,3,4]. The coordination of digital economy policies and the integration of the digital and real economies can offer a promising and efficient way to advance the digital economy and modernize traditional industries [5,6]. The modern global economy is undergoing a paradigm shift in which boundaries between digital and physical realms of economic activity are increasingly becoming integrated and blurred [7]. This phenomenon is not merely a technological phenomenon; rather, it is an indication of a fundamental shift in the structure of economic activity, the creation of economic value, and the consumption of that value [8,9]. In an era of a rapidly advancing digital revolution in the global economy, digital–real economy integration and its influence on consumption are of paramount importance to policymakers, researchers, and industry players. This issue may be considered of paramount importance, both theoretically and practically.

China is an environment particularly conducive to investigating the impact of the integration of the digital and real economies on sustainable domestic demand. On the other hand, as a nation with the second-largest economy in the world [10], China has actively sought to create a paradigm shift in its export-oriented growth model to a dual circulation economy with an expansion in domestic demand as a major priority [11,12]. The Chinese government has made digital–real economy integration a key national policy priority, with a plethora of policies, including the Digital China program, as well as a massive equipment renewal program and a trade-in program for consumer goods, which have created a unique environment for observing the effects of integration-induced policy changes [5,7]. On the other hand, as a country with unparalleled digital infrastructure and a massive, diversified consumer base, China provides an environment for interesting exploration of the effects of digital integration on traditional sectors and on domestic demand patterns [13]. The widespread adoption of artificial intelligence, big data, and mobile internet across its vast territory, including rural areas, has created a unique environment for exploring the effects of digital integration on traditional sectors and on domestic demand patterns [2]. Thus, understanding the context of China is critical to understanding the country’s economic transformation and can serve as a case study for other emerging countries undergoing similar economic changes through the use of digital technology.

Recent studies focus on the relationship between digital and real economies. Li et al. [14] discuss the major forms and implications of deep integration between digital and real economies. In this study, digital transformation is seen to drive innovation and development in the real economy. Wang and Li [15] investigate the dynamic interaction and coupling coordination between digital and real economies. They also identified mismatches in sectoral definitions arising from the interplay and convergence between the digital and real economies. Wang et al. [16] investigate sector integration and the driving forces for convergence. In their study, digital transformation is seen as driving convergence across sectors and collaborative evolution in real industries [17].

The consumption structure has undergone a fundamental transformation due to the influence of digital technology. Zhou [18] provides empirical evidence of the positive influence of the digital economy on the upgrading of the consumption structure among urban residents, as emphasized by Wang and Li [15] and Wang [19]. Maleha et al. [20] have also documented the influence of changing consumer behavior in the digital age on the global economy. The significance of personal data management and transparency has been emphasized. There has been increased focus on the influence of integrating the digital and real economies on the consumption structure. Huo and Dong [5] have emphasized that the construction of a unified national market has significantly enhanced consumption growth, a trend further strengthened by the integration of the digital and real economies. The importance of integrating the digital and real economies in enhancing residents’ consumption has been emphasized by Zhao et al. [8], who also explained the mechanism as a synergy between supply and demand potential.

Nevertheless, there are some prominent gaps in the currently available literature (as shown in

Table 1. The summary of existing studies.

Primary Research Category	Core Research Topic	Author (Year)	Key Findings	This Study’s Contributions
Digital–Real Economy Integration Research	Forms and implications of deep digital–real integration	Li et al. [14]	Defines core integration patterns and framework	Develops an entropy-weighted modified DREI index with 2014–2024 China provincial panel data, improving integration measurement accuracy
	Coupling coordination and sectoral adaptation of digital and real economies	Wang and Li [15]	Significant dynamic linkage; identifies sectoral definition mismatch
	Cross-sector integration status and convergence drivers	Wang et al. [16], Wang et al. [17]	Clarifies sectoral integration heterogeneity and core drivers
Digital Economy and Consumption Research	Digital economy’s impact on urban consumption upgrading	Zhou [18], Wang and Li [15], Wang [19]	Significant positive effect on consumption structure optimization	Empirically tests three transmission channels; conducts refined heterogeneity analysis
	Global economic impact of digital-era consumer behavior changes	Maleha et al. [20]	Behavior reshapes the global economy; data management is critical
	Integration, unified national market and consumption growth.	Huo and Dong [5]	Integration amplifies the consumption effect of the unified market
	Mechanism of integration of household consumption	Zhao et al. [8]	Drives consumption via supply-demand two-way synergy

). Firstly, most of the available literature has focused on the unilateral impact of the digital economy, without adequately exploring the synergistic effects of integrating it with the real economy. More specifically, there is a lack of accurate measurements using coupling coordination models that adequately reflect the quality of interaction between the digital and real economies in China. Secondly, there is a lack of research on the transmission mechanisms through which the integration of the digital economy with the real economy impacts household consumption. There is a scarcity of systematic testing of the specific transmission mechanisms by which the integration of the digital economy with the real economy affects household consumption. In other words, the black box of how integrating the digital economy with the real economy affects household consumption has not been adequately explored. Thirdly, there is insufficient research on regional heterogeneity that would guide the development of differentiated policy recommendations. Additionally, most of the available research has not adequately considered consumption as a whole, nor has it explored the heterogeneous impact of integrating the digital economy with the real economy on different categories of consumption.

In order to solve the problems highlighted above, this study will examine the following research questions: (1) How does digital–real economy integration (DREI) generally influence household consumption in China? (2) By what mechanism (environmental optimization, logistics improvement, or financial deepening) does DREI influence household consumption? (3) How does the impact of this relationship differ based on region and type of consumption? This study aims to present an overview of the effects of digital–real economy integration on household consumption. To examine the research questions, the author uses a two-way fixed-effects panel data model, covering provincial-level data from 2014 to 2024.

This study introduces three innovations to address the aforementioned gaps. Firstly, we develop a comprehensive, entropy-weighted modified coupling coordination index for measuring digital–real economy integration (DREI) at the provincial level, addressing the limitations of unilateral digital economy indicators in the prior literature. This approach enables more accurate measurement of the depth and quality of interaction between the digital and real sectors, using provincial panel data from 2014 to 2024 in China. Secondly, we conduct a comprehensive study of the transmission mechanisms using empirical verification of three specific channels: business environment optimization, logistics scale expansion, and financial development deepening. This study provides a more comprehensive framework for linking the integration of the digital and real economies to household consumption, filling a gap in the existing literature. Thirdly, we conduct a more specific study on the heterogeneity issue, which has not been explored in the previous literature. Our findings reveal that the promotional role of DREI is more pronounced in central China and in the goods sector. These findings provide a solid ground for formulating differentiated and targeted policy measures according to specific regions and structures.

The structure of the rest of the paper is as follows. Section 2 develops the theoretical framework and proposes the research hypotheses. Section 3 describes the methodology and data sources. Section 4 reports the empirical results and analysis. Section 5 concludes the study with a summary of the findings.

2. Theoretical Analysis and Research Hypotheses

2.1. Direct Mechanisms of DREI in Reshaping Consumption Dynamics

The theoretical basis for which DREI facilitates household consumption occurs along two direct dimensions, as presented in Figure 1.

Supply-side empowerment underpins the fundamental logic of DREI’s impact on household consumption. By fully integrating digital technologies into production, DREI has driven the evolution of traditional manufacturing towards customization, intelligence, and green sustainability [21,22,23]. This has, in turn, increased the quality and variety of products available for consumption. Guedes et al. [24] have empirically confirmed that digital transformation has driven product upgrading by optimizing digital factor inputs. The improvement in supply quality not only satisfies existing demand but also generates new consumption patterns, thus stimulating latent demand through the supply-creating-its-own-demand mechanism.

Demand-side precision forms the fundamental channel through which DREI releases consumption potential. The market-oriented allocation of data elements enables enterprises to accurately capture changes in consumer demand, thereby matching supply with demand through user profiling and analysis [19]. Specifically, digital platforms have used big data and artificial intelligence technology to significantly reduce information asymmetry between supply and demand [8], thus reducing search costs and transaction frictions for households. With the aid of algorithmic recommendation systems, enterprises have been able to provide hyper-personalized products that connect with consumers’ utility functions, thereby converting consumer intent into actual behavior [5]. DREI has also driven the emergence of immersive consumption scenarios, including live-streaming commerce, which has the potential to increase consumer engagement and trust, especially in lower-tier cities, where physical retail infrastructure is relatively underdeveloped.

DREI’s direct effect on household consumption can be explained by existing theories of consumption in the literature. From the perspective of the permanent income hypothesis and life-cycle models, DREI stimulates household consumption by generating better expectations about future income due to efficiency gains in the real sector. Improved supply conditions lead to better quality and greater diversity of available goods, thus augmenting the budget constraint faced by consumers. On the demand side, DREI helps reduce search and transaction costs, as behavioral consumption theories based on information and decision friction suggest.

Based on the direct effects outlined above, we propose the primary hypothesis:

Hypothesis 1:

Digital–real economy integration has a significant positive impact on household consumption.

2.2. Indirect Transmission Pathways of DREI on Household Consumption

DREI has an indirect impact on household consumption by optimizing the macro environment, as presented in Figure 2.

2.2.1. Business Environment Optimization as a Driver of Consumption Prosperity

DREI contributes to prosperity in consumption by fundamentally enhancing the business environment. The in-depth application of digital governance enables more transparent rules, efficient administrative services, and accessible rights protection [14]. The in-depth application of digital governance can significantly lower transaction costs in the business environment. The information search costs before consumption, bargaining costs in transactions, and enforcement costs after purchase can be significantly lowered [6]. A more transparent and efficient business environment can boost market vitality and entrepreneur confidence [9], which can create jobs and provide income stability, a necessary condition for consumption prosperity.

2.2.2. Logistics Development Scale Expansion Facilitating Market Access

DREI can play an important role in the circulation of consumption by promoting the scale expansion and intelligent transformation of the logistics industry. As a bridge connecting production and consumption in physical form, a robust logistics system is vital to market prosperity [7,25,26]. In the DREI system, the integration of various digital technologies can promote the development of logistics systems into large-scale, intelligent, and collaborative ones [27]. In addition to improving the efficiency of logistics operations, expanding the geographic scope of logistics systems can reduce last-mile delivery costs, thus enabling more goods to reach consumers affordably and serving as an important mediator between digital integration and consumption realization.

2.2.3. Financial Development Deepening, Enhancing Household Capacity

The consumption potential released by DREI can be attributed to its ability to deepen financial development, particularly in digital financial inclusion. Wu and Chang [28] show that digital inclusive finance can significantly enhance household tourism consumption by mitigating liquidity constraints. Additionally, Sun et al. [29] demonstrate their ability to promote cultural consumption by bridging the gap in urban and rural consumption. Moreover, Fu et al. (2024) demonstrate that their approach promotes productive investment among rural households [30], thereby improving their long-run income and consumption capacity (Liu and Zhang, 2025) [31]. By mitigating information asymmetry and improving access to credit [29,32], DREI can enhance the financial backbone for sustainable consumption growth.

The indirect channels of transmission align with institutional and structural explanations of economic growth. The business environment channel aligns with institutional economics, which emphasizes that economic success depends critically on transaction costs, which are influenced by institutional conditions. The logistics channel relies on spatial economics and international trade theory. According to it, the degree of market integration and price convergence depends on access to markets and transportation infrastructure. The financial development channel draws on financial development theory and recent contributions from the field of digital financial inclusion.

Synthesizing these indirect pathways, we propose the secondary hypothesis:

Hypothesis 2:

Digital–real economy integration promotes household consumption improvement through three mediating pathways: optimizing the business environment, expanding the scale of logistics development, and deepening the level of financial development.

3. Methodology

3.1. Data Sources and Sample

This study utilizes comprehensive panel data for 30 provincial-level administrative regions in China (excluding Tibet, Hong Kong, Macau, and Taiwan) spanning 2014 to 2024. All data are primarily sourced from the China Statistical Yearbook, provincial statistical yearbooks, and annual statistical bulletins published on official websites of national and provincial statistical bureaus to ensure data reliability, consistency, and timeliness. The sample period captures the rapid development of China’s digital economy and provides sufficient variation for identifying causal relationships. Regarding the imputation of missing values, the following approach is followed. For the period from 2014 to 2024, if an indicator shows intermittent missing values accounting for less than 5% of the entire dataset, linear interpolation will be used between years.

3.2. Variable Definitions

3.2.1. Dependent Variable

The dependent variable in this paper is household consumption level (CL). Building on the findings of Cai et al. [33] and Wang et al. [34], this paper employs the natural logarithm of per capita residential consumption expenditure (calculated as total residential consumption expenditure divided by total permanent population) as the core indicator for household consumption. This indicator is an important official statistical measure that represents the exact expenditure by local residents on final consumption, both tangible goods and services. This measure is widely recognized as the most precise indicator in the current literature.

3.2.2. Independent Variable

The methodological steps below are outlined to ensure the transparency and replicability of the DREI index construction process. As the different indicators are not measured on the same scale or unit, each indicator should be standardized with min-max normalization before applying entropy weight. If the indicator is a positive one, the formula for the normalization is as follows: X_ij′ = (X_ij − min(X))/(max(X) − min(X)). However, if the indicator is negative, the normalization formula is reversed. The addition of a small constant of 0.0001 is required to avoid zeros in entropy calculations.

Then, the entropy weight is calculated in the following three stages. In the first step, we calculate the percentage of every standardized criterion: p_ij = X_ij′/Σ_jX_ij′. The second step includes calculating the information entropy for every criterion as e_i = −k × Σ_j (p_ij × ln(p_ij)) where k = 1/ln(n) and n is the number of provinces. In the third step, we get the objective weight for each criterion, which is equal to w_i = (1 − e_i)/Σ_i (1 − e_i). The higher the discriminatory power of an indicator across different provinces, the greater its entropy weight.

The core of digital–real economy integration (DREI) lies in collaborative development, for which a coupling evaluation model is usually adopted to measure the degree of integration between the two. As shown in

Table 2. Digital–real economy integration indicator system.

Subsystem	Criterion Layer	Secondary Indicator	Tertiary Indicator	Weight
Digital Economy	Digital Infrastructure	Digital Facility Construction	Optical cable length/Land area (10 k km/10 k km2)	0.0275
			Mobile phone base stations (10 k)	0.0196
		Digital Network Construction	Telephone penetration rate (sets/100 people)	0.0155
			Internet users/Total population (%)	0.0154
		Digital Popularization	Mobile internet users (10 k households)	0.0182
	Digital Industrialization	Digital Industry Construction	Software business revenue (10 k yuan)	0.0329
			Information service industry output (100 m yuan)	0.0291
			Telecom business volume (100 m yuan)	0.0251
		Digital Industry Scale	Technology contract transaction volume (10 k yuan)	0.0345
		Digital Industry Personnel	Information service employees/Total employment (%)	0.0311
	Industrial Digitalization	Digital Finance Coverage	Digital finance coverage breadth	0.0635
		Digital Finance Usage	Digital finance usage depth	0.0152
		Digitalization Degree	Digital finance digitalization degree	0.0155
		Digital Industry Construction	Mobile online payment level (%)	0.0543
			E-commerce sales (100 m yuan)	0.0280
	Digital Technology Environment	Technology R&D Environment	Three types of patent applications (items)	0.0273
			Industrial enterprise R&D projects (items)	0.0281
			Industrial enterprise R&D personnel (person-years)	0.0269
			Industrial enterprise R&D expenditure (10 k yuan)	0.0261
		Government Digital Governance	Government S&T culture expenditure/General budget (%)	0.0138
			Digital economy keywords in government reports (count)	0.0177
	Digital Technology Talent	University Talent Support	University enrollment (persons)	0.0174
			University full-time teachers (persons)	0.0170
		Government Talent Support	State-owned enterprise researchers (persons)	0.0177
		Cultural Environment	Public libraries (count)	0.0154
			Book and literature lending (1000 person-times)	0.0204
Real Economy	Agriculture	Agricultural Scale	Total agricultural output value (100 m yuan)	0.0141
		Agricultural Potential	Agricultural added value (100 m yuan)	0.0143
		Agricultural Modernization	Total agricultural machinery power (10 k kW)	0.0158
	Industry	Industrial Scale	Industrial enterprises (count)	0.0186
			Industrial enterprise total assets (100 m yuan)	0.0170
		Industrial Efficiency	Industrial enterprise main business income (100 m yuan)	0.0146
		Industrial Potential	Industrial added value (100 m yuan)	0.0160
	Construction	Construction Scale	Construction enterprises (count)	0.0153
			Construction enterprise total assets (10 k yuan)	0.0143
		Construction Efficiency	Construction total output value (10 k yuan)	0.0163
		Construction Potential	Construction added value (100 m yuan)	0.0167
	Transportation and Post	Transportation Scale	Highway, inland waterway and railway mileage (km)	0.0119
			Transportation, warehousing and postal employees (persons)	0.0139
		Transportation Potential	Transportation, warehousing and postal added value (100 m yuan)	0.0121
	Wholesale and Retail	Wholesale Scale	Wholesale enterprises (count)	0.0169
			Wholesale and retail employees (10 k persons)	0.0178
		Wholesale Efficiency	Wholesale and retail total sales (100 m yuan)	0.0216
		Wholesale Potential	Wholesale and retail added value (100 m yuan)	0.0151
	Accommodation and Catering	Accommodation Scale	Accommodation and catering enterprises (count)	0.0152
			Accommodation and catering employees (10 k persons)	0.0160
		Accommodation Efficiency	Accommodation and catering revenue (100 m yuan)	0.0168
		Accommodation Potential	Accommodation and catering added value (100 m yuan)	0.0165

, this paper constructs the DREI level indicator system following the methodology of An et al. [7] and Zhang et al. [35]. We used coupling coordination analysis of comprehensive indicators of digital economy and real economy, where the coupling degree is expressed as FusionDeg, and the coordination level is set as T (T = α × DigEco + β × ReaEco), with weights α and β taking the average contribution rates of digital economy and real economy from 2014 to 2024. The final DREI level index is obtained as follows:

$$\mathit{FusionDeg} = 2\times \sqrt{{\displaystyle \frac{\mathit{DigEco} \times \mathit{ReaEco}}{(\mathit{DigEco} + \mathit{ReaEco}{)}^2}}}$$

(1)

$$\mathit{DREI} = \sqrt{\mathit{FusionDeg} \times T}$$

(2)

3.2.3. Mechanism Variables

To test the mechanisms through which DREI affects household consumption by improving the business environment, expanding logistics development scale, and deepening financial development, this paper selects the business environment index (BEI), logistics scale index (LSI), and financial development (FD) as mechanism variables. BEI refers to the sum of external conditions faced by enterprises in their production and business activities, calculated using the entropy method, following recent literature [36]. LSI refers to the overall volume of logistics activities generated during commodity circulation and service delivery, measured as the natural logarithm of the region’s annual express delivery volume [18]. FD is measured by the ratio of total deposits and loans of financial institutions at year-end to regional gross domestic product (GDP), following Liu and Zhang [31].

3.2.4. Control Variables

Drawing on recent research including Wang and Li [15] and Zhao et al. [8], this paper selects the following control variables: urbanization level (UL), economic development level (EDL), degree of economic openness (DEO), industrial structure (IS), population size (PS), industrialization level (IL), informatization level (InL), and research and development (R&D) level (RDL). Specifically, we use the urban permanent population divided by total permanent population to reflect UL; the natural logarithm of per capita GDP to measure EDL; the ratio of total import and export volume to GDP to measure DEO; the ratio of tertiary industry output to secondary industry output to reflect IS; the natural logarithm of regional population to measure PS; the industrial added value to GDP ratio to measure InL; and enterprise R&D expenditure logarithm as a proxy variable for RDL. Among them, the baseline model uses a parsimonious set of the first five controls to avoid overfitting and ensure the stability of core estimation results, while the latter three variables are incorporated into the regression framework as supplementary controls in robustness tests to mitigate potential omitted variable bias further.

3.3. Econometric Models

To empirically test the hypotheses proposed above, this paper establishes the following econometric models. The baseline regression model is specified as follows:

$${\mathit{CL}}_{\mathit{it}} = \gamma + {\beta }_0\mathit{DRE}{I}_{\mathit{it}} + {\beta }_{1}U{L}_{\mathit{it}} + {\beta }_2\mathit{ED}{L}_{\mathit{it}} + {\beta }_3{\mathit{IS}}_{\mathit{it}} + {\beta }_4\mathit{DE}{O}_{\mathit{it}} + {\beta }_{5}P{S}_{\mathit{it}} + {\mu }_i + {\lambda }_t + {\epsilon }_{\mathit{it}}$$

(3)

where i represents the region, t represents the year, β₀ is the coefficient of the core explanatory variable of interest in this study, γ is the constant term, β₁ to β₅ are coefficients of explanatory variables, ${\mu }_i$ and ${\lambda }_{\mathrm{t}}$ represent regional fixed effects and time fixed effects respectively, and ${\epsilon }_{\mathit{it}}$ is the random disturbance term.

To deeply analyze the mechanism of DREI’s impact on household consumption, this study constructs an analysis model with business environment index, logistics development scale, and financial development level as mediating variables as follows:

$${\mathit{Mediator}}_{\mathit{it}} = {\gamma }^{’} + {\theta }_0\mathit{DRE}{I}_{\mathit{it}} + {\theta }_{1}U{L}_{\mathit{it} }+ {\theta }_2\mathit{ED}{L}_{\mathit{it} }+ {\theta }_3{\mathit{IS}}_{\mathit{it}} + {\theta }_4{\mathit{DEO}}_{\mathit{it}} + {\theta }_{5}P{S}_{\mathit{it}} + {\mu }_i + {\lambda }_t+{\epsilon }_{\mathit{it}}$$

(4)

4. Empirical Results Analysis

4.1. Baseline Regression Analysis

Based on annual panel data from 30 provincial-level administrative regions across China from 2014 to 2024, this study employs a two-way fixed effects model for empirical analysis; the results are shown in

Table 3. Impact of digital–real economy integration on household consumption.

Variables	(1) CL	(2) CL
DREI	0.7941 * (0.4090)	0.7140 *** (0.2611)
UL		1.9083 *** (0.3392)
EDL		0.2008 *** (0.0729)
IS		0.1329 *** (0.0291)
DEO		0.0972 (0.0640)
PS		0.0078 (0.0870)
Constant	9.1894 *** (0.0937)	5.8208 *** (0.7867)
Regional FE	Yes	Yes
Time FE	Yes	Yes
R-squared	0.9650	0.9833
Observations	330	330

Notes: Robust standard errors in parentheses; *** and * indicate significance at 1% and 10% levels, respectively.

.

From Table 3, we can observe that DREI has a significant positive impact on household consumption level, consistent with findings by Wang and Li [15], Zhou [18] and Zhao et al. [8]. Before and after controlling for various variables, the magnitude and direction of the regression coefficient do not change significantly; moreover, the significance level improves from 10% to 1%, further confirming the reliability of the regression conclusion. Hypothesis 1 is thus verified. DREI improves the real economy through digital technologies, optimizes the efficiency of consumption scenarios, and provides environmental guarantees for household consumption. Accordingly, DREI leverages digital capabilities to enhance productivity in the real economy, thereby promoting household consumption by improving product quality. Especially during periods of weak consumption demand, DREI can alleviate the pressure of consumption contraction and serve as a key path to stimulating consumption recovery, as emphasized by Su and Wu [37] in their analysis of sustainable development.

Based on the regression results for the control variables, the coefficients for urbanization rate, economic development, and industrial structure are all positive and significant. These findings are consistent with conventional theoretical and empirical studies [5,6]. Firstly, it can be assumed that higher disposable income in a region tends to be accompanied by higher urbanization. In this regard, agglomeration effects are theoretically conducive to production activities [38]. Hence, with a larger number of rural migrants moving to cities, not only does the agglomeration effect increase, but so does their productivity [21]. This leads to an increase in social output and, subsequently, in household disposable income, thereby boosting consumption demand and facilitating the fulfillment of aspirations to improve their living standards [34]. Secondly, optimizing the industrial structure plays a pivotal role in promoting the steady, long-term growth of an economy [14,39]. In this respect, through industrial innovation, supply levels are enhanced, and the supply system is optimized to meet residents’ consumption demands better, thereby boosting consumption levels. In addition, an increase in economic development levels directly impacts residents’ consumption levels.

In terms of economic openness and population size, although both factors have positive coefficients, neither is statistically significant. With respect to economic openness, this can be attributed to the measure, as represented by the total volume of trade as a proportion of GDP, focusing primarily on production-oriented aspects rather than those directly affecting consumption. Moreover, the lack of significance may be attributed to the relatively small sample size, which limits the ability to detect the effects of structural factors. With respect to population size, a larger population can be assumed to provide a larger labor supply and, as a consequence, a greater demand. However, the effect of this factor, as a primary determinant, can be considered indirect, as represented through income generation and urbanization, which have already been factored into the model.

4.2. Robustness Tests

To verify the reliability of baseline regression results, this paper conducts robustness tests using three methods: replacing the dependent variable, changing the estimation method to the Generalized Method of Moments (GMM), and adding control variables. The results are shown in

Table 4. Robustness test results.

Variables	(1) Replace Dependent Variable	(2) System GMM	(3) Add Controls
DREI	0.5752 *** (0.1602)	0.3776 * (0.2260)	0.5261 ** (0.2598)
UL	−0.0329 (0.1183)	1.1782 *** (0.1674)	1.8057 *** (0.3500)
EDL	0.0001 (0.0356)	0.4151 *** (0.0470)	0.1956 *** (0.0590)
IS	0.0122 (0.0113)	0.0315 *** (0.0099)	0.1322 *** (0.0287)
DEO	0.0699 ** (0.0296)	−0.4797 *** (0.0613)	0.1172 * (0.0630)
PS	−0.0224 (0.0590)	0.0388 ** (0.0196)	−0.0096 (0.0802)
IL			0.2929 * (0.1511)
InL			−0.0042 (0.0054)
RDL			0.0363 (0.0275)
AR(1) p-value		0.0192
AR(2) p-value		0.3814
Hansen J p-value		0.2471
Constant	0.1551 (0.6229)	2.2867 *** (0.5251)	5.8212 *** (0.7437)
Regional FE	Yes	No	Yes
Time FE	Yes	No	Yes
R-squared	0.2119	-	0.9295
Observations	330	270	330

Notes: Robust standard errors in parentheses; ***, **, and * indicate significance at 1%, 5%, and 10% levels respectively.

. Specifically, first, the dependent variable is replaced with the growth rate of the total retail sales of consumer goods to ensure that a particular measure of consumption does not influence the results. Second, the approach is changed from the baseline model to the system GMM. This is an effective solution to the problem of endogeneity that may arise from reverse causality between consumption and its determinants, while controlling for dynamic panel effects. Third, to address omitted variable bias, three key aspects are used as control variables: IL, InL, and RDL.

The empirical results in Table 4 further confirm the positive effect of DREI on CL. The regression coefficients of the core explanatory variables remain stable in direction, magnitude, and significance, consistent with the basic findings. This proves that the promotional impact of DREI on household consumption is robust and unaffected by changes in model specifications, estimation methods, or control variables. It provides substantial support for the findings of the research.

4.3. Mechanism Analysis

To further explore how DREI promotes household consumption, this study examines the business environment index, logistics development scale, and financial development level as mechanism variables of DREI’s impact on the consumption market. The mechanism analysis results are shown in

Table 5. Mechanism analysis results.

Variables	(1) BEI	(2) LSI	(3) FD	(4) CL	(5) CL	(6) CL
DREI	2.378 *** (0.381)	5.077 *** (1.838)	7.250 *** (1.474)	-	-	-
BEI	-	-	-	0.146 ** (0.066)	-	-
LSI	-	-	-	-	0.061 *** (0.017)	-
FD	-	-	-	-	-	0.044 *** (0.017)
UL	1.468 ** (0.675)	6.507 * (3.482)	3.515 (2.288)	1.615 *** (0.319)	1.435 *** (0.263)	1.681 *** (0.316)
EDL	−0.003 (0.148)	−0.862 (0.567)	−2.973 *** (0.513)	0.243 *** (0.061)	0.299 *** (0.058)	0.376 *** (0.084)
IS	−0.020 (0.033)	−0.034 (0.154)	0.453 ** (0.213)	0.132 *** (0.033)	0.139 *** (0.030)	0.117 *** (0.029)
DEO	−0.039 (0.097)	0.413 (0.483)	0.128 (0.432)	0.092 (0.072)	0.063 (0.064)	0.083 (0.074)
PS	−0.027 (0.138)	−1.786 ** (0.834)	−1.710 ** (0.749)	0.070 (0.078)	0.184 *** (0.060)	0.148 ** (0.068)
Constant	1.995 (1.822)	27.862 *** (5.641)	44.246 *** (9.476)	4.740 *** (0.518)	3.218 *** (0.680)	2.995 *** (0.780)
Regional FE	Yes	Yes	Yes	Yes	Yes	Yes
Time FE	Yes	Yes	Yes	Yes	Yes	Yes
R-squared	0.9523	0.9145	0.9324	0.9782	0.9689	0.9574
Observations	325	330	330	325	330	330

Notes: Robust standard errors in parentheses; ***, **, and * indicate significance at 1%, 5%, and 10% levels respectively.

.

As shown in Table 5, DREI has a significant positive effect on household consumption through improvements in the business environment, the scale of logistics development, and the depth of financial development. Hypothesis 2 is verified.

In the business environment channel, DREI plays an important role in improving the quality of the institutional environment. By applying digital technologies to optimize the administrative approval process and increase government transparency, DREI can reduce institutional transaction costs for enterprises. Such optimization of the institutional environment is beneficial for the vitality of the market ecosystem, encouraging entrepreneurship and reducing the operational difficulties for small and medium-sized enterprises [9]. As noted by Li et al. [40], the reduced institutional transaction costs not only contribute to the vitality of the market ecosystem but also create more job opportunities and income stability for residents. Thus, increased business confidence and rising income expectations among households directly translate into higher consumption propensities.

In logistics development, DREI plays an important role as a catalyst for the intelligent transformation and efficiency enhancement of modern circulation systems. By employing big data, Internet of Things, and artificial intelligence technologies in the logistics system, DREI can optimize the supply chain, reduce inventory costs and delivery times, and expand market reach to lower-tier cities [24]. As discussed in Khalil et al. [27] and Sonar et al. [41], digital empowerment provided by DREI can remove geographical barriers to trade, thereby ensuring that a wider variety of high-quality products is available for consumption at relatively lower prices. Therefore, DREI can remove supply-side constraints and enhance the matching efficiency between supply and demand, thereby unleashing latent consumption potential suppressed by logistical constraints.

In financial development, DREI can promote innovation in financial services by removing information asymmetry between lenders and borrowers [29,32]. This allows for financial inclusion to expand to underserved households and micro-enterprises through alternative data in credit scoring. This result aligns with Fu et al. [30] and Liu and Zhang [31] in that financial inclusion improves capital allocation efficiency and eases financial barriers to accessing credit. This allows households to better smooth consumption over time and purchase durable goods or upgrade to better services. In addition, as digital payment systems develop in the real economy, transactions become more streamlined, further boosting consumption.

4.4. Heterogeneity Analysis

4.4.1. Regional Heterogeneity

According to standard geographical division criteria, the sample in this study is divided into eastern, central, and western regional subsamples, and the results of the heterogeneity test are shown in

Table 6. Regional heterogeneity results.

Variables	(1) Eastern	(2) Central	(3) Western
DREI	0.1144 (0.2937)	2.5737 *** (0.5590)	−0.0074 (0.5656)
UL	1.7216 *** (0.2189)	−1.5059 *** (0.5271)	1.4820 * (0.8277)
EDL	−0.1016 (0.0807)	0.1043 (0.1349)	0.5356 *** (0.1726)
IS	0.0737 ** (0.0371)	0.0809 *** (0.0291)	0.3008 *** (0.0570)
DEO	0.0157 (0.0511)	0.4489 ** (0.1797)	0.0903 (0.1574)
PS	0.4579 *** (0.0911)	0.1917 * (0.1057)	−0.1264 (0.1715)
Constant	5.7110 *** (1.1028)	6.5759 *** (1.8270)	3.6722 ** (1.7654)
Regional FE	Yes	Yes	Yes
Time FE	Yes	Yes	Yes
R-squared	0.9866	0.9934	0.9868
Observations	121	88	121

Notes: Robust standard errors in parentheses; ***, **, and * indicate significance at 1%, 5%, and 10% levels respectively.

.

As indicated in Table 6, DREI has a significant positive effect on household consumption in the central region, but not in the eastern and western regions. This shows that the promoting effect of DREI on household consumption is mainly reflected in the central region. This conclusion is consistent with the distribution dynamics analysis by Zhang et al. [42].

In the eastern region, the insignificant impact of DREI on household consumption is primarily attributed to the law of diminishing marginal returns. This is because the eastern region is the most economically developed in China, meaning the digital and real economies are already highly integrated, with well-developed infrastructure and saturated digital consumption. Hence, any attempts to further deepen the integration would have little effect on the region’s overall consumption power, as it would already have experienced the initial high-impact effects of the digital dividends.

In the western region, the negative coefficient of DREI on consumption power is not statistically significant, indicating a structural mismatch and a short-run crowding-out effect. The rapid progress of digital–real economy integration in the western region may be outpacing its fundamental capabilities, making it inefficient at the moment. The higher costs of implementing digital infrastructure, as well as the lack of complementary skills among the population, could force local agents to divert resources from consumption to costly but initially non-productive digital changes. Besides, if digital integration is seen as beneficial mainly to capital-intensive sectors or to outside investors without effectively spilling over into local residents’ incomes, it could reinforce income inequalities or displace traditional livelihoods, thereby reducing local consumption in the short run.

In contrast, the central region exhibits a significant positive effect, attributed to its optimal readiness and substantial growth potential. In contrast to the eastern region, where saturation effects are a major problem, there is still considerable room for improving efficiency in the central region. In contrast to the western region, this region also benefits from well-developed transportation networks and a strong industrial base that can immediately leverage digital technologies to reduce transaction costs and increase income in agriculture and industry. In this case, DREI can successfully bridge production and consumption, enabling households to benefit from a wider and more diverse set of markets without the severe bottlenecks found in the West. The effectiveness of this translation of digital–real economy integration can be considered a potent and direct catalyst for increasing household consumption levels, thereby making the central region a major beneficiary of current DREI policy initiatives.

4.4.2. Consumption Type Heterogeneity Analysis

According to current consumption type classification standards, the sample in this paper is divided into two subcategories: goods consumption and service consumption. The corresponding heterogeneity test results are shown in

Table 7. Consumption type heterogeneity results.

Variables	(1) Goods Consumption	(2) Service Consumption
DREI	0.8282 *** (0.2139)	−0.1285 (0.6810)
UL	1.5910 *** (0.2983)	3.1917 *** (0.8037)
EDL	0.2125 *** (0.0642)	0.2819 (0.2063)
IS	0.1347 *** (0.0237)	0.1671 *** (0.0611)
DEO	−0.0047 (0.0534)	0.5845 *** (0.1485)
PS	0.0090 (0.0941)	−0.0397 (0.1483)
Constant	6.3141 *** (0.7859)	3.6617 * (2.1561)
Regional FE	Yes	Yes
Time FE	Yes	Yes
R-squared	0.9858	0.9934
Observations	330	330

Notes: Robust standard errors in parentheses; *** and * indicate significance at 1% and 10% levels, respectively.

.

From Table 7, we observe significant structural differences in the impact of DREI on household consumption. The goods consumption sector benefits significantly, which may stem from the fact that goods consumption has standardized and warehouse-able characteristics, making it easier to reduce costs and improve efficiency through digitalization. This finding aligns with research by Zhou [18] on consumption structure upgrading and Suali et al. [43] on e-commerce supply chains. However, this effect is not significant in the service consumption sector, possibly constrained by factors such as service consumption’s dependence on scenario experience, the non-visualization of service quality, and a lack of trust mechanisms, which make it difficult to activate the service consumption market effectively.

4.5. Discussion

The empirical results offer robust and theoretically coherent support for a causal connection between digital–real economy integration (DREI) and household consumption in China. The estimated regression coefficients, signs, and statistical significance levels align perfectly with the a priori hypotheses put forward earlier and make a useful contribution to the extant literature through novel empirical results. Specifically, the significantly positive estimate of the DREI coefficient in the baseline estimation indicates not only statistical robustness but also economic significance, as a one-standard-deviation change in provincial DREI leads to a 0.714% increase in per capita household consumption expenditure, controlling for two-way fixed effects and all other control variables. The positive sign of the coefficient supports Hypothesis 1 and compensates for the lack of focus on the synergistic interaction between the digital and real economies in extant research on consumption growth. The core empirical result obtains strong support after various robustness checks. These checks consist of alternative specifications that involve substituting the dependent variable, applying system-GMM to control for endogeneity, and adding additional control variables. In all these alternative models, the effect of DREI remains positive and highly statistically significant, with little change in the magnitude of the estimate.

Hypothesis 2 finds support from the results of the mediation analysis as DREI has significantly positive effects on all three suggested mediators, namely business environment, logistics scale, and financial development, each of which, in turn, exerts a positive and statistically significant effect on household consumption expenditure. This reveals institutional, infrastructural, and financial mechanisms underlying the mediating effect that are not found in the previously reviewed literature. With respect to heterogeneity, the considerably larger positive coefficient obtained for central China, compared with the estimates in the east and west regions, follows the logical pattern of decreasing marginal returns to scale and readiness for structural transformation. Also, a greater effect on goods than on services is expected, since standardized goods are generally more compatible with digitization.

5. Conclusions and Policy Implications

5.1. Summary of Findings

This study utilizes comprehensive panel data from 30 provinces, municipalities, and autonomous regions across China from 2014 to 2024, employing a two-way fixed-effects model to analyze the impact of DREI on household consumption and to deeply explore the transmission mechanisms of the business environment, logistics scale, and financial development. We find that DREI has a significant promoting effect on the consumption market. The improvement of the DREI level also effectively promotes consumption through three pathways: improving the business environment, expanding the scale of logistics development, and deepening the level of financial development. In terms of heterogeneity test results, the consumption-promotion effect of DREI development is more pronounced in the central region and the goods consumption sector.

5.2. Policy Recommendations

Based on the empirical results from this research, which demonstrate that DREI boosts household consumption through the transmission channels of the business environment, logistics scale, and financial development, with a more significant effect in the central region and the goods consumption sector, the following policy suggestions are made.

Firstly, speed up the development of an inclusive digital infrastructure to strengthen the logistical base for domestic consumption. This suggestion is directly supported by empirical evidence showing that logistics scale is one of the main transmission mechanisms through which DREI boosts household consumption. Further improvement in the development of digital infrastructure, such as constructing data centers, laying fiber-optic networks, and building 5G base stations, should focus on reducing the urban–rural gap in digital infrastructure. As part of this policy proposal, the development of technology-empowerment services offered by digital enterprises, particularly in less-developed regions, should be pursued.

Secondly, use digital networks to improve the business environment and integrate markets; doing this is directly correlated with the empirical finding that DREI boosts consumption due to the improved business environment and its more significant impact in the central region. Leveraging digital networks to reduce information asymmetry and transaction costs in the institutional environment, as the core mechanisms for improving the business environment found through the empirical analysis, is proposed. On matters related to regional coordination, it is recommended that attention be paid to fostering interregional connectivity and collaboration to avoid homogeneous industrial competition via digital methods. Instead of devising complicated coordination measures, practical approaches will suffice to maximize the positive externalities of DREI’s greater impact in the central region and boost household consumption in other regions.

Thirdly, pursue deeper reform in the factor market and address the imbalance between goods and service consumption, which aligns with the empirically proven fact that DREI has a more significant impact on goods consumption. To address this concern, the following measures are proposed: on the one hand, improve financial development by lowering the entry cost in the fintech market, promoting inclusive digital finance, and ensuring balanced distribution of resources and talents between regions, which can help to promote the financial development transmission mechanism; on the other hand, take measures to stimulate service consumption through digital governance of service quality and online–offline integration of service consumption scenes.

5.3. Limitations and Future Research Directions

There are several limitations of this study. Firstly, the issue of measurement arises, since DREI relies on the entropy-weighted coupling coordination among variables; thus, it is unclear whether the qualitative features of integration, such as organizational culture transformation or the strategic alignment of digital and traditional economic sectors, can be assessed using this approach. It would be worth testing alternative ways of measuring the variable, such as principal component analysis, to evaluate its sensitivity to methodological assumptions.

Secondly, the data is analyzed at the provincial level, omitting heterogeneity within provinces. Consumption, as well as DREI scores, are expected to vary quite significantly from one city to another within a certain province. Hence, aggregating at the provincial level may lead to an underestimation of actual relationships and a decrease in their policy relevance for sub-provincial units of analysis. Future studies would explore city- or firm-level data to provide more granular insights.

Thirdly, this research lacks external validity beyond the Chinese context, since the unique characteristics of the institutional setting, the development of digital infrastructure, and the country’s policy priorities make it an outlier from a cross-country perspective. The effectiveness of DREI in fostering consumption might depend on institutional preconditions specific to China. Cross-country comparative studies are needed to identify which aspects of the DREI–consumption relationship are context-specific and which may be universal.