Digital Economy and Carbon Emission Intensity: Evidence from ASEAN

Abstract

After the carbon neutrality goal was proposed, the question of how to achieve low-carbon and green development has become the focus of ASEAN countries. The digital economy will play an important role in promoting carbon emission reduction in ASEAN countries. This paper measures the digital economy index of ASEAN countries and explores the impact of the digital economy on carbon emission intensity across 10 ASEAN countries from 2011 to 2020. The results show that the digital economy can reduce the intensity of carbon emissions in ASEAN countries. In addition, the impact of the digital economy on carbon emission intensity will vary depending on the level of economic development, resource endowment, and government financial support capacity. The mechanism analysis shows that the digital economy can ultimately reduce the intensity of carbon emissions by promoting industrial structure upgrading, stimulating technological innovation, and promoting the transformation of energy consumption structure. Based on the above conclusions, specific recommendations are proposed for the digital economy to realize carbon reduction in ASEAN countries.

1. Introduction

The long-standing reliance on an extensive development model has resulted in consistently high levels of energy use and carbon emissions, causing serious environmental pollution (Cao et al., 2021 [1]; Deng and Zhang, 2021 [2]). This phenomenon is particularly serious in ASEAN countries.

After entering globalization, the aggregate carbon emissions produced by ASEAN countries have witnessed exponential growth, reaching 1.651 billion tons in 2020—a figure representing a more than twenty-two-fold increase over the 1965 level. Consequently, their share of the worldwide carbon footprint, which stood at 0.65% initially, climbed to 4.74% in 2020. This expansion was propelled by a mean increase of 5.8% per year, dramatically outpacing the 2.1% international average. (as shown in Figure 1). Especially after entering the 21st century, industrial development and urbanization-driven sectors like construction, transportation, and residential electricity use have significantly raised energy consumption, which in turn has escalated carbon emissions. IEA statistics reveal a continuing dominance of fossil fuels in ASEAN’s energy mix. Notably, around 75% of its power supply is generated from fossil fuels, and coal constitutes the largest share at over 50%. Furthermore, it is projected that 75% of the additional energy demand by 2030 will be fulfilled by these conventional energy sources.

Given the responsibility for carbon reduction and sustainable development, in response to climate responsibilities and sustainable development goals, each ASEAN member state has presented its nationally determined climate pledges, outlining specific targets for cutting greenhouse gas output. Following the conclusion of the 2021 UN Climate Change Conference (COP26), a shared commitment to carbon neutrality has been formally adopted by all ASEAN members, with the Philippines being the sole exception at this stage. Long-term strategies outlining pathways to carbon neutrality have been submitted to the United Nations by several ASEAN members, including Singapore, Indonesia, Cambodia, and Thailand, demonstrating ASEAN’s determination to reduce carbon emissions. It is noteworthy that these countries all have underscored the critical role played by the digital economy in their carbon neutrality strategies (ESCAP, 2023 [3]). Such confidence is anchored in the potential of digitalization to drive structural change and decarbonize traditional industries. In the new situation, leveraging the digital economy to advance carbon reduction and eco-friendly growth has become an important development task for ASEAN countries.

Current studies examining the drivers behind carbon emissions in ASEAN nations have centered predominantly on economic growth and foreign direct investment,(Chandran and Tang, 2013 [4]; Heidari et al., 2015 [5]; Zhu et al., 2016 [6]; Tebourbi et al., 2023 [7]; Pata et al., 2023 [8]), trade (Atici, 2012 [9]; Salman et al., 2019 [10]), technology industry (Haini, 2021 [11]), urbanization (Wang et al., 2016 [12]), energy transition (Liu et al., 2017 [13]; Shafiq et al., 2020 [14]), and so on. There is no exploration of the digital economy factor yet. In fact, after joining globalization and undertaking industrial transfer, the carbon reduction pressure on ASEAN countries has gradually increased. In order to avoid sacrificing the environment for economic development, it seems more important for ASEAN countries to develop a digital economy.

This study addresses a clear gap in the literature by examining the ASEAN community as a dynamic force within the Asia-Pacific area, analyzing how advances in its digital economic activities shape outcomes related to lowering carbon output. The key contributions of this work include: (1) The impact of the digital economy on carbon emissions in the ASEAN region is systematically examined. Existing literature has predominantly focused on the influence of traditional economic factors on carbon emissions, while largely overlooking the role of the digital economy as an emerging driving force. This study fills this gap by revealing the emission reduction role of the digital economy in this important and representative alliance of emerging economies. (2) Second, in terms of the impact mechanism, it fully elucidates the core transmission framework of “industrial structure upgrading—technological innovation—energy structure improvement.” This study not only verifies the independent mediating effects of the three pathways but also clarifies the systematic emission reduction mechanism they constitute, thereby deepening the understanding of the “black box” between the digital economy and its environmental impacts.

The remainder of this study is structured to flow logically from theory to evidence. Commencing with a review of the literature and hypothesis development in Section 2, the study then outlines its research model and data in Section 3. Section 4 presents the statistical findings and their interpretation. The final section synthesizes the core conclusions and puts forward practical implications.

2. Literature Review and Hypothesis

2.1. Digital Economy and Carbon Emission

Existing literature has developed two competing theoretical perspectives on the relationship between the digital economy and carbon emissions. Research supporting the “enabling effect” argues that digital technologies can systematically drive carbon reduction at both macro and micro levels. At the macro level, smart technology clusters not only enhance monitoring capabilities for carbon emissions and carbon sinks (Tortorella and Fettermann, 2018 [15]; Ghobakhloo, 2020 [16]; Fei et al., 2022 [17]), but also provide technical support for pathway planning toward carbon peak and carbon neutrality goals (Fei et al., 2022 [17]). Big data technologies offer foundational methodologies for carbon emission governance (Hampton et al., 2013 [18]; Kwon et al., 2014 [19]; Shin and Choi, 2015 [20]), while digital tools optimize carbon market mechanisms to achieve efficient allocation of emission allowances (Xie, 2022 [21]). At the micro level, enterprises utilize digital solutions to precisely monitor energy consumption patterns, significantly enhancing energy efficiency and green total factor productivity(Xu et al., 2021 [22]).

However, proponents of the “environmental cost” argument present a contrasting perspective, suggesting that digital technologies may generate substantial carbon footprints. Zhou et al., 2019 [23] highlight the high energy consumption of digital infrastructure and its reliance on high-carbon intermediate goods, which could exacerbate carbon emissions. Notably, Yu and Zhu, 2023 [24], found a pronounced inverted U-shaped relationship between digital economic development and carbon emissions during its early stages, indicating that the environmental impacts of digital technologies exhibit stage-specific characteristics.

Although existing research provides valuable theoretical frameworks and empirical evidence for understanding the relationship between the digital economy and carbon emissions, significant limitations remain. Existing empirical research on the relationship between the digital economy and carbon emissions predominantly focuses on digitally mature economies such as China, Europe, and the United States, exhibiting significant sample selection bias. This has resulted in a critical knowledge gap regarding the ASEAN region—a typical alliance of emerging economies. This gap raises dual concerns: theoretically, ASEAN nations are undergoing a unique phase of concurrent industrialization and digitalization, characterized by systemic differences in industrial structure, energy dependency, and digital policies compared to larger mature economies. Overlooking this contextual specificity may lead to theoretical predictions failing in practice. From a policy perspective, the absence of region-specific evidence makes it difficult for ASEAN members to formulate precise low-carbon digitalization strategies, potentially underestimating emission reduction potential or overlooking localized carbon costs. Therefore, clarifying the specific carbon emission effects of the digital economy in ASEAN is not only an academic necessity for expanding the geographical scope of research but also an urgent policy requirement for supporting the region’s green transition. Therefore, this paper aims to explore the relationship between ASEAN’s digital economic development and carbon emissions.

Given the explosive growth of ASEAN’s digital economy in recent years—marked by rapid expansion of internet users and a doubling of its total digital economy size between 2019 and 2022 (e-Conomy SEA, 2022 [25])—there is reason to believe the region may have collectively moved beyond the initial phase of digital infrastructure development. The “enabling effect” of digital technologies is thus expected to dominate. Based on this, this study proposes the following hypothesis:

Hence, according to the above literature review, we propose the following hypothesis.

Hypothesis 1.

ASEAN’s digital economy development reduces carbon emission intensity.

2.2. The Underlying Mechanism

Scholars have largely identified the principal channels through which the digital economy affects carbon emission levels, primarily involving three key transmission mechanisms: advancement in industrial sophistication, enhancement of technological capabilities, and optimization of energy systems.

2.2.1. Upgrading Industrial Structure

The digital sector actively drives industrial modernization and structural advancement. This transformation supports emission reduction via updated operational methods and improved resource utilization (Li et al., 2019 [26]; Tian et al., 2019 [27]). The industrialization of digital technology has directly given birth to numerous emerging industries driven mainly by digital technology as the production factor. The direct substitution of traditional industries by these emerging industries will change the transformation of economic growth models and industrial structures toward greater ecological sustainability. Concurrently, these emerging industries promote efficient allocation and circulation of labor, capital, and other factors within and between industries. This process further supports operational optimization and lowers the carbon impact per unit of output (Shahbaz et al., 2020 [28]). From the perspective of industrial digitization, digital technology is extensively merged with sectors characterized by high carbon emissions, including power generation, manufacturing, transport, building, and energy infrastructure. Digital technology ultimately reduces carbon emission intensity by improving the efficiency of the entire workshop process in production, logistics, inventory management, and equipment maintenance (Fei et al., 2022 [17]), and upgrading production technology. In addition, digital consumption, online shopping, and digital office activities have directly reduced the frequency and scope of offline activities by transforming people’s lifestyles, thereby reducing the transportation energy consumption and pollution of the entire city.

2.2.2. Technological Innovation

Technological innovation is a key factor through which the digital economy affects carbon emission intensity, principally via three channels. (1) Technological upgrading effect. The industrial pollution control strategy and energy management model have achieved system upgrades under the application of digital technology, and the interconnection between control departments and manufacturing equipment has been achieved through digital technology (Kong et al., 2022 [29]); (2) Technology sharing effect. The digital economy has the effects of sharing, penetration, and spillover. Digital technology promotes the connection, sharing, and collaboration among innovative entities by breaking monopolies (Paunov and Rollo, 2016 [30]), ultimately improving the overall innovation efficiency of the supply chain and greatly reducing carbon emission intensity (Xing et al., 2019 [31]). (3) Market feedback effect. Digital technology mines and analyzes big data on green product production and consumer preferences, prompting manufacturers and enterprises to increase investment in green technology to accurately meet consumer demand, thereby reducing carbon emission intensity in the production process.

2.2.3. Energy Structure Transformation

The digital economy impacts how energy is used through several key channels. (1) Industrial transformation. Digital technology improves the approval efficiency of green loans for enterprises, reduces the financing difficulties within the clean energy domain, and thus promotes the development of new energy enterprises (Baloch et al., 2021 [32]). (2) Lifestyle transformation. By dynamically simulating changes in energy supply and demand through big data, carbon footprint can be optimized, thereby guiding regions with concentrated carbon footprints to transform towards low-carbon lifestyles in a targeted manner (Han and Xie, 2017 [33]). (3) Regional coordination. Digital technology is also conducive to establishing a renewable energy consumption market and environmental regulatory system for inter-regional cooperation. Digital technology enhances coordination between online and offline energy systems, leading to improved utilization of renewable resources and better supply-demand balance across the region.

In terms of its impact mechanism can be summarized as follows: it first catalyzes the emergence of low-energy-consumption industries (industrial structure upgrading), directly replacing high-carbon activities and optimizing factor allocation. Subsequently, the operation and competition of new business models inherently drive innovation in green technologies and processes (technological innovation). Ultimately, the simultaneous leap in industrial structure and technological sophistication lays a solid foundation for the clean and efficient transformation of the energy system (improved energy structure). These three elements are interlinked, collectively forming the core mechanism driving carbon emission reductions. The mechanism through which the digital economy impacts carbon emissions is illustrated in Figure 2. Accordingly, the research hypothesis is formulated:

Hypothesis 2.

The development of ASEAN’s digital economy reduces carbon emission intensity by promoting industrial structure upgrading, strengthening technological innovation, and transforming energy structure.

2.3. Evidence from ASEAN Countries

On the one hand, the digital economy has reshaped behavioral patterns across society by giving birth to numerous low-carbon and emerging industries. E-commerce and digital services have gradually become the leading engines of economic growth, fostering an economic rebalancing from industry-led to service-oriented growth, ultimately achieving low-carbon development.

Digital economy promotes economic growth. The triennium marked a pivotal phase in the expansion of ASEAN’s digital economy. 100 million (over 20%) out of 460 million internet users in ASEAN countries are first-time internet users. The total output value of ASEAN’s digital economy reached $200 billion in 2022, doubling from 2019. Jobs originating from the digital economy reached 160,000 in specialized fields across ASEAN, with an additional workforce of approximately 30 million supported by the sector’s expansion. (e-Conomy SEA, 2022 [25]).

Digital economy promotes low-carbon lifestyle. E-commerce and digital services are gradually becoming the dominant engines of economic growth. As of 2021, the penetration rate of e-commerce in the six ASEAN countries has reached 80%, which is the main driving force for the growth of the digital economy. It is expected that the total volume of commodity transactions in the e-commerce industry will reach 233 billion US dollars by 2025, corresponding to a compound annual growth rate of 35% (e-Conomy SEA, 2021 [34]). Delivery services, ride hailing services, e-commerce retail, and electronic office are all developing rapidly. Industry projections for the delivery service market are converging on a forecast of USD 12 billion by 2025, while the ride-hailing market is forecast to reach USD 19 billion. (China Institute of Inclusive Finance, 2022 [35]).

On the other hand, in alignment with sustainable development objectives, industrial digitalization has been elevated to a strategic national priority across ASEAN. Member states are prioritizing the integration of digital solutions into conventional sectors such as manufacturing, mining, agriculture, logistics, and energy infrastructure. The deployment of these technologies is simultaneously accelerating the ongoing shift toward clean and alternative energy solutions in both industrial operations and household consumption. Strategic policy direction is essential to transition away from the prevailing heavy dependence on fossil fuels and to facilitate a fundamental shift in the energy mix.

Within their carbon neutrality frameworks, ASEAN nations have explicitly outlined goals to modernize energy systems and drive urban development through technological adoption. (refer to

Table 1. Content related to “Digital Economy Promoting Emission Reduction” in ASEAN Countries’ Carbon Neutrality Strategies.

Country	Related Content in Carbon Neutrality Strategy 2
Cambodia	Build modern power operation system, including digital generator operation, digital power dispatch, and standardized digital codes for the power grid.
Thailand	(1) Establish digital renewable energy control center to manage power facilities; (2) Establish digital ecosystem service system for biodiversity and natural resource use detection and paid management; (3) Establish sustainable development financial center to promote digital low-carbon products.
Laos	Propose the “Industry 4.0” initiative, focusing on the digital transformation of energy infrastructure construction, agriculture and biotechnology, healthcare, intelligent equipment, robotics technology, automation, online office, digital industry, Internet of Things, 5G, etc.
Singapore	(1) Launch digital carbon emission recording tools to help businesses record, detect, and reduce their carbon footprint; (2) Promote the construction of smart cities and promote comprehensive decarbonization in areas such as life, transportation, and construction; (3) The Smart State and Future Economic Initiative Committee was established, and in November 2018, the Service and Digital Economy Technology Roadmap (SDE-TRM) was launched, with Service 4.0 identified as the main engine of Singapore’s digital economy growth.
Malaysia	(1) Launch the world’s first digital free trade zone (in cooperation with Alibaba); (2) Propose the “2020 Go Digital Vision”, deploy smart grids that can provide stable electricity supply, and build smart microgrids in remote areas; (3) Develop the ICT industry, and popularize digital urban development models such as remote online work, and reducing transportation load.

² Data source: The author sorted it out from the carbon neutral strategies of ASEAN countries. https://eciu.net/netzerotracker and https://asiapacificenergy.org/apef/index.html#main/lang/en/time/[1990,2023]/geo/[BRN]/matrix (accessed on 13 December 2023).

). Furthermore, regional strategies advocate for a comprehensive energy transition, emphasizing the scaled utilization of renewables and the accelerated electrification of urban transport networks. These initiatives inherently rely on support from digital and intelligent technological platforms. The adoption of such technologies can catalyze the overhaul of conventional industrial production paradigms. ASEAN’s environmental transition gains accelerated momentum from the rapid spread of digital economic activities.

3. Methodology and Data

3.1. Economic Model

An econometric model is developed as following to investigate the connection between two key variables in ASEAN nations—digital economic development and carbon emission intensity.

$${\mathrm{C}\mathrm{a}\mathrm{r}\mathrm{b}\mathrm{o}\mathrm{n}}_{\mathrm{i},\mathrm{t}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{D}\mathrm{i}\mathrm{g}\mathrm{i}\mathrm{t}\mathrm{a}\mathrm{l}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\alpha }}_2{\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i},\mathrm{t}}+{\mathrm{u}}_{\mathrm{i}}+{\mathrm{v}}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$$

(1)

${\mathsf{\alpha }}_0$ is the baseline intercept, representing the starting value of carbon emissions when all independent variables are zero. $\mathrm{i}$ and $\mathrm{t}$ represent country and year, respectively; ${\mathrm{C}\mathrm{a}\mathrm{r}\mathrm{b}\mathrm{o}\mathrm{n}}_{\mathrm{i},\mathrm{t}}$ is the dependent variable, which is carbon emission intensity; ${\mathrm{D}\mathrm{i}\mathrm{g}\mathrm{i}\mathrm{t}\mathrm{a}\mathrm{l}}_{\mathrm{i},\mathrm{t}}$ is the core explanatory variable, which is the digital economy; ${\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i},\mathrm{t}}$ encompasses all controlled factors influencing carbon emission intensity; ${\mathsf{\alpha }}_1$ is the impact coefficient of the digital economy, ${\mathsf{\alpha }}_2$ is the influence coefficient matrix of the control variable set; The model includes country-fixed effects (${\mathrm{u}}_{\mathrm{i}}$) and time-fixed effects (${\mathrm{v}}_{\mathrm{t}}$), with ${\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$ as the stochastic error term.

A mechanism analysis is formally evaluated through an extended model as following that introduced mediating variables to the core framework.

$${\mathrm{M}}_{\mathrm{i},\mathrm{t}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{D}\mathrm{i}\mathrm{g}\mathrm{i}\mathrm{t}\mathrm{a}\mathrm{l}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }}_2{\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i},\mathrm{t}}+{\mathrm{u}}_{\mathrm{i}}+{\mathrm{v}}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$$

(2)

$${\mathrm{C}\mathrm{a}\mathrm{r}\mathrm{b}\mathrm{o}\mathrm{n}}_{\mathrm{i},\mathrm{t}}={\mathsf{\rho }}_0+{\mathsf{\rho }}_1{\mathrm{D}\mathrm{i}\mathrm{g}\mathrm{i}\mathrm{t}\mathrm{a}\mathrm{l}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\rho }}_2{\mathrm{M}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\rho }}_3{\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i},\mathrm{t}}+{\mathrm{u}}_{\mathrm{i}}+{\mathrm{v}}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$$

(3)

${\mathsf{\beta }}_0$ and the ${\mathsf{\rho }}_0$ are the intercept of this model. The two intercept terms both represent the theoretical level when all influencing factors in the model specification are zero. The model incorporates ${\mathrm{M}}_{\mathrm{i},\mathrm{t}}$ as the intermediary factor and ${\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i},\mathrm{t}}$ as the set of control variables, which deliberately excludes the mediator itself. To control for unobserved cross-country and temporal heterogeneity, ${\mathrm{u}}_{\mathrm{i}}$ (country fixed effects) and ${\mathrm{v}}_{\mathrm{t}}$ (time fixed effects) are included, leaving ${\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$ as the model’s stochastic disturbance. Diagnosing the mediation effect is a multi-stage process. The entire analysis is premised on a significant ${\mathsf{\alpha }}_1$. Once this premise is established, the significance of the pathway coefficients ${\mathsf{\beta }}_1$ and ${\mathsf{\rho }}_2$ is examined. The diagnostic outcome is then determined by the status of ${\mathsf{\rho }}_1$: its significance leads to a conclusion of partial mediation, and its non-significance points to a complete mediation effect.

3.2. Variable Selection

3.2.1. Explained Variable

Carbon emission intensity. Due to the differences in economic size, land area, population size, and energy structure among ASEAN countries, simply using the absolute value of total carbon emissions cannot fully compare the carbon emission efficiency between countries. Therefore, referring to Xie (2022) [21], the carbon emission intensity for each country is represented by its carbon emissions-to-GDP ratio. which can also horizontally compare the carbon emission efficiency of each country.

3.2.2. Explanatory Variable

Digital economy. The measurement system for digital economy indicators is quite mature, mainly including digital infrastructure indicator group, industrial digitalization indicator group, and digital industrialization indicator group. The empirical foundation for this analysis rests on the data from ASEAN members that are both accessible and meet the required quality standards. An indicator system is constructed to measure the digital economy of ASEAN countries, and use the entropy weighting method commonly used in academia for comprehensive measurement of the digital economy (measurement methods shows in Appendix A). The specific composition of the indicator system is shown in

Table 2. ASEAN Digital Economy Indicator System.

	Indicators	Unit	Index Attribute
Digital Economy	Subscription and usage of online services	People/per hundred people	+
	Number of mobile phone users	People/per hundred people	+
	Percentage of Individuals using the Internet	Percent	+
	Fixed broadband service subscriptions	People/per hundred people	+
	Number of ATMs	per 100,000 people	+
	Proportion of telecommunications, computer, and information services trade to total trade	Percent	+

.

3.2.3. Other Variables

Control variables specifically include population density, per capita electricity consumption, methane emissions, and industrial value added. The mediating variables specifically include industrial structure upgrading, technological innovation, and energy structure transformation. The specific variable explanations are shown in

Table 3. Variable definitions.

	Variables	Explanation	Source
Explained variable	Carbon Emission Intensity	carbon emissions-to-gross domestic product (GDP) ratio	Our World in Data database World Bank database
Explanatory variable	Digital Economy	measure through the entropy method	ASEAN Statistical Yearbook World Bank database
Control variables	Population Density	the number of people per kilometer of land area	World Bank database
	Per capita Electricity Consumption	per capita electricity consumption	IEA database IndexMundi database
	Rural Emission	methane emissions	World Bank database
	Industrial Development	industrial added value	World Bank database
Mediating variables	Industrial Structure Upgrading	the ratio of the tertiary sector’s share to the secondary sector’s share in GDP	Asian Development Bank (ADB) database
	Technological Innovation	the added value of medium and high-tech manufacturing	World Bank database
	Energy Structure Upgrading	obtaining clean fuels (as a percentage of the population)	World Bank database
Other variables	Industrial Robot	number of robots installed	International Federation of Robotics (IFR) database
	Fixed telephone subscriptions	Fixed telephone subscriptions per 100 people in 2000	World Bank database

.

3.3. Data Source

Due to the scattered data statistics of ASEAN countries, the sources of data in this paper are relatively diverse. Each data indicator is unified at the national and annual levels, so it is suitable for empirical testing, and there will be no experimental estimation bias caused by inconsistent data sources in different countries or years. The empirical data sources for each indicator used in this paper are shown in Table 3.

3.4. Descriptive Statistics

The descriptive profile of the variables is provided in

Table 4. Descriptive Statistics.

	Variables	Average	SD	Min Value	Max Value
Explained variable	Carbon Emission Intensity	0.058	0.030	0.011	0.180
Explanatory variable	Digital Economy	0.405	0.253	0.001	0.882
Control variables	Population Density	12,515.78	18,632.34	880.31	66,859.3
	Per capita Electricity Consumption	36.335	10.940	22.142	73.673
	Rural Emission	62,932	76,532.780	1560	281,000
	Industrial Development	3.080	3.435	0.160	12.100
Mediating variables	Industrial Structure Upgrading	905.475	2299.812	27.800	7965.878
	Technological Innovation	1.459	0.611	0.371	3.037
	Energy Structure Upgrading	30.134	23.290	0.260	83.727
Other variables	Industrial Robot 3	59.676	33.855	4.100	100
	Fixed telephone subscriptions	12.201	16.188	0.248	48.2

³ International Federation of Robotics (IFR) database only includes robot data from Indonesia, Malaysia, the Philippines, Singapore, Thailand, and Vietnam. Due to the fact that these countries are among the ASEAN countries with more obvious characteristics of digital economy and carbon emission development, sample selection bias can be avoided. Therefore, this article uses this data for robustness testing.

. These results indicate that the values are distributed with adequate uniformity, meeting the essential prerequisites for further analysis.

Among ASEAN members, Singapore leads in the scale and maturity of its digital economy. Singapore also ranks first among ASEAN countries in the World Bank’s ranking of business environment convenience. Singapore has strong capabilities in communication infrastructure, technological development, population quality, import and export of high-tech products, and government financial support, which leads to a highly sophisticated digital economy. Established industrial economies such as Malaysia and Thailand demonstrate robust digital development. As a developed nation with a favorable business climate and advanced economy, Brunei similarly hosts a thriving digital sector. Furthermore, both Cambodia and Vietnam are making significant progress in this area. In addition, Cambodia and Vietnam have developed rapidly in the past five years, with a trend of surpassing other countries. Laos and Myanmar have relatively low levels of digital economy development (See Figure 3).

4. Results and Discussion

4.1. Benchmark Estimation Results

Table 5. Baseline regression results.

Variables	Carbon Emission Intensity
	(1)	(2)	(3)	(4)	(5)	(6)
Digital Economy	−0.094 **	−0.116 ***	−0.199 ***	−0.213 ***	−0.211 ***	−0.227 ***
	(−2.432)	(−2.935)	(−5.084)	(−5.229)	(−5.158)	(−5.461)
Per Capita GDP		−0.000 **	−0.000 ***	−0.000 ***	−0.000 ***	−0.000 ***
		(−2.018)	(−4.815)	(−4.956)	(−4.853)	(−4.007)
Industrial Development			0.004 ***	0.004 ***	0.004 ***	0.004 ***
			(4.723)	(4.871)	(4.669)	(4.386)
Rural Emission				−0.000	−0.000	−0.000
				(−1.202)	(−1.323)	(−1.498)
Per Capita Electricity Consumption					−0.005	−0.004
					(−0.910)	(−0.829)
Population Density						−0.000 *
						(−1.667)
Constant	0.093 ***	0.149 ***	0.011	0.011	0.062	0.063
	(5.456)	(4.570)	(0.261)	(0.262)	(0.888)	(0.915)
Country Fixed Effects	Yes	Yes	Yes	Yes	Yes	Yes
Year Fixed Effects	Yes	Yes	Yes	Yes	Yes	Yes
R-squared	0.708	0.722	0.784	0.788	0.790	0.798

Notes: t-statistics in parentheses. *** means p < 0.01, ** means p < 0.05, and * means p < 0.1.

summarizes the regression outcomes examining how the digital economy influences carbon emission levels across ASEAN member states. Statistically significant, the findings demonstrate an inverse relationship between the progression of the digital economy and carbon emission intensity. After adding control variables, the coefficient of the impact of the digital economy on carbon emission intensity remains significantly negative at the 1% level. Taking Model (6) as an example, the regression coefficient is −0.227. This indicates that, holding other factors constant, a one-unit increase in the digital economy development level of ASEAN countries leads to an average decrease of 0.227 units in their carbon emission intensity. This result provides strong support for the research hypothesis, confirming that the digital economy can indeed effectively enhance the green efficiency of regional economic development. The digital economy of ASEAN countries has reduced carbon emission intensity by reducing communication costs between people, reducing information asymmetry, promoting production innovation, and transforming lifestyles.

Among the control variables: (1) The coefficient for the impact of economic development level on carbon emission intensity is significantly negative at the 1% level, indicating that higher economic development correlates with lower carbon emission intensity. This provides empirical support for the applicability of the Environmental Kuznets Curve in the ASEAN region. (2) The coefficient for the impact of industrial intensity on carbon emission intensity is significantly positive at the 1% level, indicating that greater industrial value-added correlates with higher carbon emission intensity. This reflects the reality that coal and oil remain the primary energy sources for industrial development in most ASEAN countries, both currently and in the foreseeable future. (3) The coefficient for population density shows a significant positive impact on carbon emission intensity at the 5% level, indicating that countries with higher population densities exhibit greater carbon emission intensity. Among ASEAN nations, Singapore has the highest population density. However, Singapore stands as a unique case, actively practicing low-carbon environmental principles, prioritizing carbon emission control, and serving as a model garden city. The next countries in the ranking are the Philippines, Vietnam, and Indonesia. These three nations are typical industrial-dominated economies heavily reliant on fossil fuels for development. The concentration of large populations has led to significant production and lifestyle pollution, substantially increasing their carbon emission intensity. Conversely, Laos and Myanmar have the lowest population densities in ASEAN. Low population density implies reduced industrial activity and urban pollution. Additionally, both nations have lower economic development levels with fewer large industrial enterprises, and small-scale operations and individual activities contribute minimally to carbon emissions. Moreover, both nations possess substantial hydropower capacities. In 2021, hydropower accounted for 73.9% of Laos’ total electricity generation and 54% of Myanmar’s. Consequently, the reality of substantial renewable energy meeting the development needs of smaller populations results in lower carbon emission intensities compared to other countries.

4.2. Robustness Testing

4.2.1. Alternative Measures of Digital Economy

Industrial robots serve as a pivotal carrier of digital technologies in the real economy, deeply integrating core digital elements such as automation, artificial intelligence, and the Internet of Things. They effectively represent the level of industrial digitization, making them an ideal proxy variable for measuring the development of the digital economy (Shahbaz et al., 2020 [28]). Accordingly, to establish an empirical connection, the analysis utilizes data on installed industrial robotics, treating this metric as a proxy measure to assess its relationship with carbon emission intensity. The statistical evidence presented in

Table 6. Robustness Test.

Variables	(1)	(2)	(3)	(4)	(5)
	Carbon Emission Intensity	Annual Growth Rate of Carbon Emissions	OLS	SYS-GMM	IV Method
Digital Economy		−0.908 *	−0.060 *	−0.017 *
		(−1.920)	(−1.668)	(−1.696)
Industrial Robot	−0.007 *
	(−1.915)
Carbon Emission Intensity (−1)			0.866 ***	1.046 ***
			(9.059)	(4.668)
Fixed telephone subscriptions × internet user					−0.372 **
					(2.123)
Constant	−0.044	0.321	−0.004	0.000	−0.004
	(−0.420)	(0.410)	(−0.076)	(.)	(−0.076)
Control Variables	Yes	Yes	Yes	Yes	Yes
Country fixed effects	Yes	Yes	Yes	Yes	Yes
Year fixed effects	Yes	Yes	Yes	Yes	Yes
AR(1)				0.000
AR(2)				0.694
Sargan test				0.659
R-squared	0.978	0.344	0.920		0.782

Notes: t-statistics in parentheses, *** means p < 0.01, ** means p < 0.05, * means p < 0.1.

reveals an inverse association between industrial automation and the outcome variable (p < 0.10), as shown in the first specification, suggesting that their deployment in ASEAN nations is associated with a reduction in carbon intensity.

4.2.2. Alternative Measures of Carbon Emission Intensity

In this study, the annual growth rate of carbon emissions is adopted as an alternative metric to represent emission intensity in empirical analysis. According to the outcomes displayed in column (2) of Table 6, the digital economy exhibits a statistically significant negative coefficient at the 10% level. This finding implies that the advancement of digital economic activities is associated with a deceleration in the yearly increase of carbon emissions.

4.2.3. Endogeneity Test

Due to the temporal and inertial characteristics of carbon emissions. Its determinants are dual in nature, encompassing both short-term economic drivers and the persistent influence of historical carbon emission patterns. If the model lacks dynamic influencing factors of carbon emission intensity, it may lead to endogeneity problems. Based on this, we use a dynamic panel model for estimation, which adds a one-period lagged term of carbon emission intensity to the control variables. The formal representation appears below:

$${\mathrm{C}\mathrm{a}\mathrm{r}\mathrm{b}\mathrm{o}\mathrm{n}}_{\mathrm{i},\mathrm{t}}=\mathsf{\alpha }+{\mathsf{\beta }}_1{\mathrm{D}\mathrm{i}\mathrm{g}\mathrm{i}\mathrm{t}\mathrm{a}\mathrm{l}}_{\mathrm{i},\mathrm{t}}+{\mathsf{\beta }}_2{\mathrm{C}\mathrm{a}\mathrm{r}\mathrm{b}\mathrm{o}\mathrm{n}}_{\mathrm{i},\mathrm{t}-1}+\mathsf{\gamma }{\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}_{\mathrm{i},\mathrm{t}}+{\mathrm{u}}_{\mathrm{i}}+{\mathrm{v}}_{\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i},\mathrm{t}}$$

(4)

The regression outputs reported in columns (3) and (4) of Table 6 demonstrate that a country’s carbon emission intensity from the previous period has a clear positive effect on its current level. This confirms the presence of a persistent inertial tendency in the emission profiles of ASEAN economies. Meanwhile, an inverse relationship for the digital economy remains evident, with results attaining the 1% significance level, affirming that even after accounting for this inertial dynamic, the digital economy maintains a statistically substantial suppressing influence on emission levels.

4.2.4. Instrumental Variable Method

To further address estimation biases arising from endogeneity, this paper draws upon the methodology of Wang and Zhang (2023) [36]. It employs the interaction term between the 2000 fixed telephone subscriptions (number of fixed-line telephone main lines per 100 people) and the previous year’s national internet user count to form a new instrumental variable for the digital economy. The rationale for this selection is twofold: First, a region’s early telecommunications infrastructure level is highly correlated with its subsequent capacity and willingness to develop modern digital infrastructure such as the internet and mobile communications. This telecommunications legacy path-dependently influences digital economic development, satisfying the principle of relevance. Second, the 2000 fixed-line telephone penetration rate is a historical variable unlikely to directly affect contemporary carbon emission intensity unless it influences subsequent digital economic development trajectories. This satisfies the exogeneity requirement. The regression results, as shown in column (5) of Table 6, remain statistically significant, confirming the robustness of the conclusion.

4.3. Heterogeneity Analysis

4.3.1. Heterogeneity of Economic Development Level

Referring to She and Wu (2022) [37], countries are classified into three categories based on their per capita carbon emission growth rate and GDP: developed countries (Singapore, Brunei), emerging industrial countries (Indonesia, Malaysia, Philippines, Thailand, Vietnam), and underdeveloped countries (Cambodia, Myanmar, Laos). Referring to Su et al. (2018) [38], the regression outputs in column (1) of

Table 7. Heterogeneity analysis.

Variables	Carbon Emission Intensity
	(1)	(2)	(3)
Digital Economy	−0.163 **	−0.229 ***	−0.234 ***
	(−2.326)	(−4.272)	(−4.826)
Digital Economy × emerging industrial countries	−0.106 *
	(−1.816)
Digital Economy × emerging industrial countries	0.049
	(0.673)
Digital Economy × non resource endowment countries		0.003
		(0.069)
Digital Economy × countries with middle levels of fiscal expenditure			0.007
			(0.154)
Digital Economy× countries with lower levels of fiscal expenditure			0.054
			(1.265)
Control Variables	Yes	Yes	Yes
Constant	0.175 *	0.065	0.106
	(1.921)	(0.872)	(1.202)
Country fixed effects	Yes	Yes	Yes
Year fixed effects	Yes	Yes	Yes
R-squared	0.830	0.798	0.805

Notes: t-statistics in parentheses, *** means p < 0.01, ** means p < 0.05, * means p < 0.1.

present the following findings: the digital economy variable, along with its interaction term with the dummy variable for emerging industrial nations, both show statistically significant negative coefficients at the 5% level. The interaction between the digital economy and the dummy for underdeveloped economies is insignificant. These results imply that the carbon intensity reduction effect attributable to the digital economy is stronger in both developed and emerging industrial countries, and its effect on underdeveloped countries is still not prominent. This may be because developed and emerging industrialized countries in ASEAN entered globalization earlier (Saboori and Sulaiman, 2013 [39]), and through the introduction of technology, capital, and talent, digitalization and information technology were improved, and carbon emission intensity was reduced relatively faster and more efficiently through digitalization technology (Zhu et al., 2016 [6]). However, the digital economy in Myanmar, Cambodia, and Laos has a small size and slow progress. Furthermore, these nations are currently undergoing a phase of accelerated industrial expansion, which is associated with elevated levels of carbon emissions per unit of economic output. Consequently, the efficacy of digital economic initiatives in mitigating emission intensity remains limited under such conditions.

4.3.2. Heterogeneity of Resource Endowment

Referring to Guo et al. (2022) [40], we divide countries into resource endowment countries (net energy exporting countries, including Indonesia, Malaysia, Brunei, and Myanmar) and non-resource endowment countries (net energy importing countries, including Singapore, the Philippines, Thailand, Vietnam, Cambodia, and Laos) based on their net energy imports and exports. The estimation results presented in column (2) of Table 7 reveal a statistically significant negative coefficient for the digital economy at the 5% level. In contrast, the coefficient of the interaction term between the digital economy and non-resource-endowed countries lacks statistical significance. These findings suggest that the effect of the digital economy in lowering the intensity of carbon emissions is more substantial in resource-rich nations, while its effect on non-resource endowment countries is still not prominent. The reasons may include two aspects. On the one hand, ASEAN resource endowment countries heavily rely on fossil fuels (Lean and Smyth, 2010 [41]), so the impact of the digital economy on diminishing carbon emission intensity is still in a marginal increasing state. Non resource endowment countries passively rely on imports for resources, and the digital economy cannot create more energy substitutes in the short term, so the carbon reduction effect of the digital economy is relatively weak. On the other hand, international trade often requires higher quality product supply. In order to import products with higher added value, energy importing countries may choose to send technology, capital, talent and other factors to energy exporting countries to improve product quality (Atici, 2012 [9]; Salman et al., 2019 [10]), which enables energy exporting countries to achieve low-carbon transformation of their industries by introducing digital technology.

4.3.3. Heterogeneity of Government Financial Support Capacity

In 2020, the proportion of government expenditure in ASEAN countries to GDP was 29% (Brunei); 26.03% (Singapore); 22.14% (Thailand); 19.56% (Philippines); 18.48% (Malaysia); 18.38% (Cambodia); 16.55% (Indonesia) (Trading Economics [42]), respectively. We classify countries into countries with higher levels of fiscal expenditure (including Singapore, Thailand, the Philippines and Brunei); countries with middle levels of fiscal expenditure (including Malaysia, Cambodia, and Indonesia); countries with lower levels of fiscal expenditure (including Laos, Myanmar, and Vietnam). The results of Table 7 column (3) show that the carbon reduction effect brought about by the digital economy is more evident in countries where fiscal expenditures are substantial, but its effect on countries with lower fiscal expenditure levels is still not prominent. This may be because the more government fiscal expenditure there is, the higher the level of support allocated to the digital economy, the greater its potential to achieve carbon emission reductions.

4.4. Mediation Effect Analysis

4.4.1. Industrial Structure Upgrading

Results in the first specification of

Table 8. Mediation effect analysis.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
	Industrial Structure Upgrading	Carbon Emission Intensity	Technological Innovation	Carbon Emission Intensity	Energy Structure Upgrading	Carbon Emission Intensity
Digital Economy	0.302 ***	−0.212 ***	7.617 **	−0.223 ***	50.245 ***	−0.109 **
	(4.547)	(−5.323)	(2.106)	(−5.335)	(7.623)	(−2.163)
Industrial Structure Upgrading		−0.080 ***
		(−2.949)
Technological Innovation				−0.000
				(−0.843)
Energy Structure Upgrading						−0.001 *
						(−1.914)
Control Variables	Yes	Yes	Yes	Yes	Yes	Yes
Constant	2.993 ***	0.423 ***	−26.701 *	0.067	75.637 ***	0.187 **
	(9.721)	(4.264)	(−1.841)	(0.871)	(6.355)	(2.042)
Country fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
Year fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
R-squared	0.991	0.819	0.981	0.799	0.996	0.766

Notes: t-statistics in parentheses, *** means p < 0.01, ** means p < 0.05, * means p < 0.1.

establish a significant positive association between digital economy and manufacturing transformation, demonstrating its capacity to accelerate industrial structure. The second set of results deepens this finding, revealing a notable mitigating relationship between digital economic activity and emission levels. The effect size is estimated at −0.212. Compared to the benchmark regression coefficient of −0.227, this suggests that the progression toward more advanced industrial forms is a key mechanism underpinning the digital economy’s contribution to lower carbon emission intensity. Digital technologies are fostering operational optimization and energy transition across ASEAN. By improving industrial processes, reshaping consumption behaviors, and diminishing dependence on traditional energy sources, they actively mitigate carbon emission intensity.

4.4.2. Technological Innovation

The estimates in the second column in Table 8 confirm with greater precision the suppressive role of the digital economy on carbon emissions. The measured estimate is −0.212. Meanwhile, column (4) demonstrates a notable negative influence of the digital economy on carbon emission intensity. The estimated coefficient of –0.223, when compared to the benchmark value of –0.227, suggests that technological innovation serves as a mediating channel through which the digital economy contributes to lower carbon intensity. This mediating pathway is observed to be especially pronounced in Malaysia, Thailand, and Vietnam. This may be because Malaysia, Thailand, and Vietnam have stronger broadband connectivity capabilities (Raja, 2019 [43]), and digital technology has upgraded production technology and improved energy allocation efficiency through developed networks (Zhi et al., 2024 [44]). The intermediary effect of other countries is relatively insignificant. Singapore’s economic development is highly dependent on oil imports, which leads to less significant innovation effects (Munir et al., 2020 [45]). Indonesia and the Philippines have a high dependence on fossil fuels (as shown in Figure 4). The elimination of existing energy infrastructure may lead to resource waste and economic development turbulence. So, the positive feedback effect of technological innovation in the short term is still not significant enough. Laos and Cambodia have relatively underdeveloped technology, so the intermediary effect is weak.

4.4.3. Energy Structure Upgrading

As evidenced in Table 8, column (5), the digital economy exhibits a significantly positive correlation with the shift in energy consumption patterns. Furthermore, the results in column (6) reveal a persistently negative influence of the digital economy on carbon emission intensity. The coefficient of −0.109 suggests that its mitigating effect is partially channeled through modernizing the energy mix, a conclusion supported by its larger absolute value compared to the benchmark coefficient of −0.227. The industrial development of ASEAN countries heavily relies on fossil fuels, causing serious environmental pollution. So ASEAN countries urgently need to achieve low-carbon development through energy structure transformation (Handayani et al., 2022 [46]). The intelligence- and information technology-driven digital economy emerges as a vital tool. It enables the development of renewables via digital methods and advocates for their use to curb carbon emissions (Shafiq et al., 2020 [14]).

5. Conclusions and Policy Implications

This research develops a Digital Economy Development Index for the ASEAN region and employs a ten-year panel dataset (2011–2020) covering all ten member states to assess its environmental implications, supplemented by mechanism analysis. The key conclusions can be summarized as follows: (1) The progression of digitalization in the ASEAN is associated with a decline in regional emission intensity. (2) The extent of this influence is moderated by national economic conditions, resource characteristics, and fiscal capabilities. (3) Three underlying mechanisms are identified—structural transformation in industry, advancement in technical capabilities, and shifts in energy use patterns.

This study systematically investigates the interplay between digital economy and carbon emission across ASEAN member states, identifying significant structural barriers that hinder effective transformation. To address these challenges, we propose the following strategic recommendations:

First, ASEAN countries should promote energy transformation through digital technology to reduce carbon emissions. (1) As most ASEAN nations are still undergoing industrialization, applying digital technologies in industrial processes can enhance energy efficiency. Specifically, accurately calculating and estimating energy usage through big data and improving production accuracy through artificial intelligence are two areas of interest. (2) ASEAN countries should strengthen the exploration, measurement, development, and utilization of renewable energy through digital technology. In addition to promoting large-scale digital energy infrastructure in big cities, it is also necessary to popularize digital technologies supporting small-scale renewable energy generation facilities in rural and remote areas, gradually establish a digital power grid with internal and external connections, and explore the connection mode of smart grids within and between regions. (3) ASEAN countries export more high-carbon products. ASEAN countries can achieve product upgrades through digital technology and green management, thereby reducing the implicit carbon emissions in trade. At present, most ASEAN countries have established high-tech industrial parks and export trade zones, and have attracted foreign investment to establish low-carbon green factories and ICT industries in such parks through preferential policies like simplifying land use processes. Examples of this include the Jurong Industrial Park in Singapore and the Kuantan Industrial Park in Malaysia. ASEAN countries should further consider the construction of supporting facilities in industrial parks, such as establishing digital industrial parks and e-commerce service parks, and leveraging the positive role of digital finance and e-commerce. (4) ASEAN countries should cultivate low-carbon consumption habits across society as a whole; actively promote services such as online payment, remote office work, online shopping; and promote green transportation to transform the way people consume.

Second, ASEAN countries should apply digital technology according to their actual situations. For emerging industrialized countries, due to the significant locking effect of high-carbon energy structure, their reform is more difficult. So, they should give priority to applying digital solutions to upgrade legacy industries and reform the urbanization process through, for example, research and development related to clean coal technology, digital substitution of industrial processes, digital office work, etc., gradually achieving a smooth transition towards digitalization throughout the entire system. In many less-developed nations, the digital economy is still in its early stages. As a result, its role in reducing carbon emissions has not yet become substantial. However, this also provides opportunities for developing countries to directly move beyond digital transformation and enter digital innovation. These countries can introduce foreign investment and advanced technology, as well as use digital means to build new urban buildings, energy, and communication facilities. In addition, a new “wild goose mode” within ASEAN should be formed, with Singapore as the first tier; Malaysia, Thailand, Indonesia, Vietnam, and the Philippines as the second tier; and other countries as the third tier, forming an alliance operation mechanism that combines digital industry cooperation, intelligent technology sharing, and digital talent training, ultimately achieving the transfer of digital elements within the region.

Finally, the experiences of the ASEAN nations offer a replicable model for other developing nations facing similar challenges in balancing economic growth with carbon reduction. To enhance the policy relevance and global impact of our findings, we propose the following generalized recommendations. (1) Integrate Digital and Green Infrastructures from the Outset: Developing countries should prioritize the synchronous development of digital and green infrastructures. By embedding low-carbon standards into the planning of 5G networks, data centers, and other digital utilities, nations can avoid high-carbon lock-in effects from the very beginning, achieving synergistic advancement in digital capabilities and energy structure optimization. (2) Activate Mechanism-Driven Decarbonization Pathways: The core insight is to consciously leverage the digital economy through its key transmission mechanisms. Policymakers should tailor strategies to national conditions—countries with a solid industrial base can focus on “industrial digitization” to enhance energy efficiency, while those with strong service sector potential can accelerate “digital industrialization” to foster new, low-emission growth engines. The fundamental goal is to deliberately unclog the three core channels identified in this study—industrial structure upgrading, technological innovation, and energy structure transformation—to ensure digital investments translate into tangible emission reductions.

6. Limitations and Prospects

This study has several limitations. Firstly, while the entropy weight method provides an objective approach to measuring the digital economy, it has inherent drawbacks. The method determines indicator weights solely based on data dispersion, which may not align with their theoretical importance. This could lead to underestimating crucial but statistically concentrated aspects of digital development, such as AI integration or big data capabilities. Moreover, the entropy method is sensitive to extreme values and changes in sample selection, potentially affecting the stability and comparability of results across different contexts. Second, the concurrent examination of three mediation mechanisms introduces model complexity that may obscure the distinct effects of each pathway. The interconnected nature of these channels poses challenges in accurately attributing emission reductions to specific mechanisms. Additionally, the exclusion of other variables that may influence carbon emissions from the framework analysis may constitute a shortcoming of the paper.

Hence, future research could develop more nuanced digital economy metrics that capture its core features beyond infrastructure. Additionally, employing research designs that can disentangle the unique contributions of each mechanism would provide deeper insights into their relative importance for emission reduction.