Exploring the Influence of the Digital Economy on Energy, Economic, and Environmental Resilience: A Multinational Study across Varied Carbon Emission Groups

Abstract

Rapid advancements in digital technologies have accelerated global change, underscoring the critical role of resilience in addressing the escalating energy, economic, and environmental challenges. This paper investigates the effects and mechanisms of the digital economy on energy, economic, and environmental resilience within the context of these challenges. By utilizing panel data from 66 countries spanning the period from 2000 to 2020, this analysis employs robust panel data models and incorporates tests such as the Hausman and Leamer tests, and exploratory factor analysis. The results reveal a notable positive impact of the digital economy on resilience across various countries and time periods. However, when it comes to carbon emissions, a more intricate pattern emerges, suggesting a negative influence on resilience in environmental, energy, and economic domains. Interestingly, countries with below-average carbon emissions show more positive effects on economic resilience due to the digital economy. On the other hand, the effect of the digital economy on energy resilience is less prominent in below-average carbon-emitting nations, while carbon emissions have a more significant impact within this subgroup. Above-average carbon-emitting countries experience limited effects of the digital economy on environmental resilience, while below-average carbon-emitting countries face challenges with significant carbon emissions impacting their environmental resilience.

1. Introduction

The transformative impact of digital technologies, including Blockchain, Artificial Intelligence (AI), and renewable energy technologies, on society, business, and consumption patterns is undeniable [1,2,3]. As exemplified by the remarkable growth of the Chinese digital economy, these technological advancements have not only reshaped global economic landscapes but have also introduced significant changes in energy consumption and generation patterns [4,5]. In light of escalating economic, energy, and environmental threats, the concept of resilience has emerged as a crucial consideration for maintaining a high quality of life and responding effectively to adverse events [6]. The exploration of this nexus is imperative for understanding the opportunities and challenges posed by the digital economy in advancing sustainability within a rapidly evolving world. The more we enhance the urban systems’ resilience, the more it intersects with how the digital economy is evolving [7,8,9].

Resilience, highlighted for its significant role in global development agendas like the Sustainable Development Goals and initiatives such as the “How to Make Cities More Resilient” campaign, holds a central position in policymaking [10,11,12]. This importance is further underscored by documents such as the World Bank’s 2013 “Guide to Building Resilience in East Asia” [13], along with the integration of resilient city building in China’s 14th Five-Year Plan for National Economic and Social Development and Long-Range Objectives for 2035 [14]. Recognizing its crucial role in enhancing the well-being of vulnerable populations, resilience is a key strategy for addressing the multifaceted dimensions of vulnerability [15]. Achieving resilience involves implementing effective strategies to tackle vulnerabilities and contribute to sustainable development across various domains, including energy [16], economy [17], and environment [18].

In the realm of energy [16], studies emphasize the importance of resilience in ensuring consistent service delivery despite operational challenges, with a lack of resilience potentially leading to the loss of valuable energy system services. Research has explored the trade-offs between energy efficiency and heat resilience [19,20] and the length of survival during power outages [21]. Digital technologies play a crucial role in enhancing energy resilience by enabling improved monitoring, control, and optimization of energy systems, leading to increased efficiency and reliability [22]. Concerning the impact of the digital economy on the sustainability of energy sources, traditional fossil fuels have historically been pivotal in providing a stable energy supply for human civilization. However, in our shift towards carbon neutrality, it is crucial to harmonize energy security and resilience with the imperative of reducing carbon emissions [23]. Energy resilience, defined as the capacity of an energy system to consistently deliver satisfactory service levels despite diverse operational challenges, becomes paramount. A lack of resilience may lead to the loss of valuable energy system services, prompting abrupt transitions or shifts towards different situations and configurations [24].

Digital technologies play a key role in enhancing the monitoring, control, and optimization of energy systems, leading to increased efficiency and reliability. Gatto and Drago (2020) [15], for example, delve into the Global Energy Resilience Index (GERI), considering factors such as approved electrification plans with digital technologies, information availability to consumers through digital platforms, incentives for renewable energy, and the integration of carbon pricing and monitoring mechanisms with digital systems. Figure 1 illustrates the percentage of electricity access in certain countries as an indicator within the GERI index based on World Bank data. It emphasizes the practical application of the GERI index, showcasing how it considers real-world data to evaluate and compare the energy resilience of different nations. Moreover, proposals for cross-border trading and innovative energy networks have emerged to bolster regional energy security [25]. In summary, cultivating a comprehensive understanding of energy resilience is essential to confront the challenges posed by the digital economy and ensure a sustainable and resilient energy future.

Regarding the economy [17], the digital economy, recognized by the United Nations as a major force for sustained economic development, plays a vital role in innovation, structural change, and economic growth [26]. Digital finance, a fundamental pillar of the digital economy, positively affects regional economic resilience by breaking down geographical boundaries and generating knowledge spillovers [27,28,29]. Although the digital economy may pose short-term financial risks, it contributes to sustainable economic development in the long run [30]. The transformation of economic structure and competitive order through digital technologies leads to financial disintermediation, altering the traditional roles of financial intermediaries like banks or financial institutions [31].

In the context of the environment [18], vulnerability to environmental disasters is influenced by social factors such as social class, ethnicity, age, and nationality [32,33]. Poor communities, particularly in developing countries, are more vulnerable to climate change and other hazards due to inadequate management policies [34,35]. Concerning the impact of the digital economy on environmental resilience, countries worldwide grapple with a spectrum of natural and anthropogenic hazards that influence their vulnerability levels. This susceptibility is shaped by factors such as geographical location, topography, climate, and management strategies [36]. The concept of resilience in the face of natural hazards, originally defined by [37], emphasizes a community’s ability to recover using its own resources. This framework has become integral in assessing hazard impacts, extending its application to diverse fields, including ecology [38], social systems [39], and human–environment systems [40,41]. Therefore, bolstering environmental resilience becomes paramount for mitigating damages, with resilience patterns exhibiting geographic variability among nations [18].

Moreover, the digital economy plays a pivotal role in contributing to sustainability goals by curbing carbon emissions through optimized industrial output [42], real-time energy consumption monitoring, and support for eco-friendly sectors like smart and shared transportation [43]. Consequently, delving into the relationship between carbon emissions and the digital economy proves indispensable for achieving sustainable development and carbon reduction, aligning with global sustainability objectives [44]. The assessment and comparison of resilience across different countries and groups serve as valuable tools in informing emergency responses, recovery strategies, and mitigation efforts during critical circumstances.

According to the concepts of three types of resilience and the digital economy, this article raises three questions:

• Are there variations in digital economy’s effects on economic resilience among different countries with different levels of carbon emission?
• Does the digital economy affect energy resilience in a similar manner between below-average and above-average carbon-emitting countries?
• Do countries with above-average and below-average CO2 emissions respond similarly to the digital economy’s impacts on environmental resilience in relation to their carbon emissions?

Based on these questions, the following hypotheses are formulated:

H₀. 

The digital economy has no discernible impact on the specific components of resilience (economic, energy, or environmental) within different countries with varying levels of carbon emissions.

H₁. 

The digital economy positively influences economic, energy, and environmental resilience.

In order to investigate the questions above, using 66 countries’ annual data from 2000 to 2020, we used the panel data regression model with fixed and random effects. If H₁ is accepted, it implies that the digital economy plays a role in enhancing economic, energy, and environmental resilience in the studied countries. On the other hand, if the statistical analysis does not provide sufficient evidence to reject the hypothesis of H₀, it suggests that there is no significant impact of the digital economy on economic, energy, and environmental resilience. The analysis included random effects models, with the Hausman test indicating the significant selection of the panel model (p-value ≤ 0.10). Collinearity, normality, and correlation tests were conducted to examine the relationships among variables. The main benefit of the panel data regression approach is that it makes use of a dataset with many observations of different people over a long period of time [45]. This allows for a comprehensive analysis of the influences of the digital economy on 3E resilience (Energy, Economy, and Environmental), considering the variations across different countries in terms of their economy, energy, and environment.

This paper is organized as follows: The associated literature is briefly summarized in Section 2. In Section 3, the specifics of the econometric technique are described together with an interpretation of the data sources and variables. The findings and analysis are covered in Section 4. The study is concluded in Section 5 along with any pertinent policy implications.

2. Literature Review

With the surge in population, energy demand, urbanization, and environmental impact accompanying the onset of the digital era, a pressing need arises to prioritize the development of resilience solutions and the establishment of sustainable global energy infrastructure [45]. In alignment with this imperative, the concept of the “digital economy” has evolved, considering digitized knowledge and information as a pivotal production factor, contemporary information networks, and the effective utilization of ICT (Information and Communication Technology) for productivity growth and economic optimization [46]. ICT encompasses technologies related to information processing, communication, and the convergence of various technologies for handling information [47]. While previous research has explored the characteristics and effects of the digital economy, there remains a gap in thorough research on its connection with global energy, economic, and environmental resilience. The integration of the digital economy, carbon emissions, and resilience in a unified framework, as well as a deeper exploration of the impact mechanism, is still limited, with some studies focusing on theoretical impact and mediating mechanisms.

The primary objective of this paper is to investigate how the digital economy influences economic, energy, and environmental (3E) resilience. Regarding the connection between economic resilience and the digital economy, existing research primarily discusses influencing factors and measurement methods. Saputro and Suwito (2022) [48] emphasize policy interventions to advance the digital economy in resource-oriented cities. Tang et al. (2022) [49] highlight the efficient application of digital finance in credit financing, while He et al. (2023) [50] analyze the connection between urban economic resilience and digital finance in China, finding a significant positive impact. Daud and Ahmad (2022) [51] suggest that digital finance has the potential to foster knowledge spillover and offer more efficient capital support to the economy. However, there is a lack of global studies investigating the mechanism of the connection between the digital economy and enduring economic growth.

The relationship between the internet-based economy and energy resilience has been insufficiently explored in existing studies. Although some research has focused on energy resilience, a notable gap exists in considering the digital economy in conjunction with this concept. Gatto and Drago (2020) [15] introduced the GERI index, ranking and defining energy resilience for both OECD and non-OECD nations. Li et al. (2020) [52] conducted research on 277 cities in China, establishing a digital economy index and a green economy efficiency index, revealing that the internet-based economy significantly enhanced the efficiency of the green economy. However, studies exploring the impacts of a technology-based economy on energy resilience across different countries are limited.

In terms of the influences of the digital economy on environmental resilience, it is crucial to attain sustainability objectives, particularly by reducing carbon emissions, indicating potential resilience [44]. Research by Chadha et al. (2007) [53] highlights the importance of disaster recovery plans, while Zhai and Yue (2022) [54] emphasize the role of the environment in shaping economic adaptability. Moghim and Garna (2019) [18] investigate the vulnerability of 141 countries to natural and anthropogenic hazards, revealing variations in resilience across different locations. However, there is a lack of research addressing this relationship in different countries with varying carbon emission levels.

Despite some studies exploring certain aspects of the relationship between the digital economy and resilience, there is insufficient research addressing the mechanisms through which the digital economy affects energy, economy, and environmental resilience on a global scale. Given the escalating risks faced by various regions due to climate change and other factors, understanding the effects of the digital economy on different forms of resilience becomes crucial. To fill this gap, this paper aims to investigate the connection between the digital economy and diverse types of resilience in different countries, considering their varying economic, energy, and environmental conditions. By comparing the resilience of different groups, countries, governments, and decision-makers, we can gain insights into relative resilience and determine whether there is a need to learn from more resilient nations when it comes to emergency responses, recovery, and mitigation in critical circumstances.

3. Method and Data Description

3.1. Model

In recent studies, panel data regression has been employed to estimate the impact of the digital economy on consumer spending and economic resiliency, revealing positive effects on urban economic resilience [55]. The digital economy was found to be positively impacted by elements including population quality and industrial structure. Similarly, panel data regression has been utilized to examine the relationship between Information and Communication Technology (ICT) and energy use, indicating that fostering ICT development can lead to energy efficiency gains and economic expansion in South Asian countries, even if there has been a rise in energy usage [56]. The panel data model offers advantages such as accounting for omitted variables through fixed and random effect models [57], capturing non-observable time and region-specific intercepts [58], and addressing issues related to relevant omitted variables’ impact on the dependent variable [59].

In our study, the selection of an appropriate panel model was a crucial step in analyzing the relationships between observed variables and identifying underlying factors or latent variables. To determine the most suitable model, we employed the Hausman test. This statistical test allows us to compare the random effects and the fixed effects models and assess their relative performance. By evaluating the test results, we were able to identify the model that best captures the dynamics of the data and provides reliable estimates. In addition to selecting the appropriate panel model, we conducted exploratory factor analysis (EFA) as part of our analytical approach [60]. EFA is a statistical technique used to uncover the fundamental elements or latent variables that contribute to the observed patterns in the data. By examining the interrelationships between variables, EFA helps to identify common factors that may explain the variation in the observed data. This analysis provided us with valuable insights into the complex relationships between the observed variables and allowed us to understand the underlying factors that drive the dynamics of the digital economy’s impact on resilience. By combining the Hausman test and exploratory factor analysis, we were able to employ a robust analytical framework that accounted for the complexities and interdependencies of the variables in our panel data model. These analytical techniques enhanced the validity and reliability of our findings, providing an expanded comprehension of the connections between the digital economy, observed variables, and different forms of resilience. We formulated a panel data model to investigate the connection between various forms of resilience and the digital economy, as follows:

$${CRES}_{it}={\beta }_0+{\beta }_1{DIG}_{it}+{\beta }_2{Cr}_{it}+{\beta }_3{Pd}_{it}+{\beta }_4{Eo}_{it}+{\beta }_5{Fs}_{it}+{ϵ}_{it}$$

(1)

$${GRES}_{it}={\beta }_0+{\beta }_1{DIG}_{it}+{\beta }_2{Cr}_{it}+{\beta }_3{Pd}_{it}+{\beta }_4{Eo}_{it}+{\beta }_5{Fs}_{it}+{ϵ}_{it}$$

(2)

$${VRES}_{it}={\beta }_0+{\beta }_1{DIG}_{it}+{\beta }_2{Cr}_{it}+{\beta }_3{Pd}_{it}+{\beta }_4{Eo}_{it}+{\beta }_5{Fs}_{it}+{ϵ}_{it}$$

(3)

The panel data models presented in Equations (1)–(3) illustrate the relationships between various variables in our study. The subscript “i” denotes different countries, while the subscript “t” represents different time periods. According to the literature review, the economic resilience index ($CRES$) in Equation (1) is defined as the dependent variable, which is influenced by the digital economy ($DIG$), carbon emissions ($Cr$), population density ($Pd$), economic openness ($Eo$), and food security ($Fs$), along with an error term ($\epsilon$). Similarly, Equation (2) represents the energy resilience index ($GRES$) as the dependent variable, influenced by the same set of independent variables. Equation (3) captures the environmental resilience index ($VRES$), influenced by the aforementioned variables.

$CRES$ is derived from a combination of subindices, including labor productivity, discretionary income, reliance on foreign trade, gross domestic product, innovation index, education expenditure, and research costs, as cited in [60]. $GRES$ incorporates variables related to energy access, energy efficiency, renewable energy, and hydrogen-based renewable energy, as described in [15]. $VRES$ encompasses variables such as environmental risk, energy consumption, access to clean water, sanitation, air pollution, and greenhouse gas emissions, as outlined in [18]. The digital economy variable ($DIG$) specifically captures the impact of the digital economy on resilience, while carbon dioxide emissions ($Cr$), population density ($Pd$), economic openness ($Eo$), and food security ($Fs$) serve as control variables. By considering these factors and their interplay within the panel data models, our study provides a comprehensive framework for analyzing and understanding the connection between different forms of resilience and the digital economy.

3.2. Database and Variables

This study explores the influence of the digital economy on economic, environmental, and energy resilience using panel data spanning from 2000 to 2020. In

Table 1. Descriptive statistics of the variables.

Variable	Abbreviation	Obs	Mean	Std	Min	Max
Dependent variable
Economic resilience	CRES	1428	7.099	2.895	2.011	11.985
Energy resilience	GRES	1428	3.052	1.720	0.004	5.995
Environmental resilience	VRES	1428	7.034	2.297	3.016	10.996
Independent variable
Digital Economy	DIG	1428	19.911	5.735	10.015	29.941
Control variables
Carbon emissions	Cr	1320	6.460	4.201	0.168	25.61
Economic openness	Eo	1428	98.853	67.109	15.683	442.62
Food security	Fs	1428	614.584	2526.239	2.477	21,594.8
Population density	Pd	1428	94.691	14.211	46.300	247.57

, we present details on key study factors, including the total number of observations (Obs), average (Mean), deviation (Deviation), and minimum and maximum values (Min and Max). The study establishes a digital economy index system, building upon previous research on digital infrastructure and digital industrialization [5], as outlined in Table 1. To assess economic resilience, guided by Martin’s regional economic resilience definition, we categorize the explanatory variable into three dimensions: resistance and recovery, adaptation and adjustment, and innovation and transformation [61]. The environmental resilience index follows Moghim’s research [18], while the energy resilience variable is derived from Gatto’s study [15].

The data for analyzing how the digital economy has changed the economic, energy, and environmental resilience is obtained from various sources, as shown in Appendix A. The digital economy variables are sourced from the International Telecommunication Union (ITU) [62], the United Nations (UN) [63], and the World Bank [64,65]. These variables include fixed broadband, fixed phone, mobile cellular, and other services, telecommunication infrastructure index, people who use the Internet [66], e-participation index, medium and high-tech manufacturing value added, and online service index. The social support variable is measured using the value added per person in the service sector. For economic resilience analysis, the data is sourced from the World Bank [64] and includes variables such as reliance on international commerce, per-person disposable income, overall societal labor productivity, GDP per-person, fiscal self-sufficiency rate, per-person local fiscal spending, and per-person fixed asset investment. The Organization for Economic Cooperation and Development (OECD) [67] is the source of the innovation and transformation capability variables, which also include the regional innovation and entrepreneurship index, per capita fiscal expenditure on education, and employment index for the scientific research industry. Energy resilience variables are sourced from Regulatory Indicators for Sustainable Energy (RISE) [68], the US Energy Information Administration (EIA) [69], and the Global Energy Statistical Yearbook [70]. These variables include energy access (access to electricity), efficiency in energy use (primary energy’s intensity level), the contribution of renewable energy to final energy consumption, and hydrogen-based renewable energy (total renewable electricity consumption) [71]. Environmental resilience data is obtained from various sources. The environmental risk variable is sourced from the International Disaster Database (EMDAT) [72] and measures the level of exposure or vulnerability to natural disasters. Air pollution data is sourced from the World Health Organization (WHO) [73] and includes mean levels of pollutants. Access to clean water, reduced greenhouse gas emissions, and improved sanitation variables are also sourced from the WHO and the United Nations Climate Change database [56,74]. The description of each variable is presented in Appendix A.

4. Results and Discussion

4.1. The Effects of Digital Economy on Economic Resilience

The panel data estimation outcomes are presented in

Table 2. The estimation results of the influence of the digital economy on economic resilience (Model 1).

Variable	Model (1)
	Coefficient	p-Value
Dig	0.022	0.020
Cr	−0.093	0.013
Eo	0.059	0.125
Pd	−0.003	0.005
Fs	0.092	0.017
Constant	8.010	0.000
Observation	1320
R2		27.06
F Test	3.27	0.000
Normality of Residual		0.870
Leamer Test	8.280	0.000
Hausman Test	11.75	0.038
Autocorrelation Test	0.347	0.558
Heteroskedasticity Test	0.84	0.359
Ramsey Reset test	0.45	0.715

. Diagnostic tests were conducted to determine the most appropriate estimation method. One of these tests is the Leamer test, which evaluates the validity of the random effects assumption in panel data analysis. The test helps in choosing the suitable panel data model [75]. According to the Leamer test, the H0 hypothesis that assumes no effect from the digital economy on different resilience, is rejected with a 99% confidence level. This rejection suggests that the panel data method should be used for estimation.

Furthermore, the Hausman test, which favors the random effects method [76], is rejected with a 95% confidence level. Consequently, it is recommended to employ the fixed effects panel data method to estimate the first model in Equation (1).

According to the estimation results, the intercept of the first model is estimated to be 8.010, which, with 99% confidence, is statistically significant. The variable’s coefficient of “Dig” is estimated to be 0.022, with a p-value of 0.020, which is statistically significant at a 95% confidence level. Therefore, the digital economy improves economic resilience. One possible explanation for this relationship could be attributed to the fact that the modernization of industrial structures is made possible by the digital economy, leads to increased employment opportunities, and enables the restructuring of employment patterns [77]. Some studies support the findings regarding the advantages of the digital economy for economic stability. Shahadat et al. (2023) [78] conducted research that revealed a positive effect of the digital economy on supply chain efficiency and agility, ultimately contributing to improved economic resilience. Additionally, another study demonstrated that the adoption of digital technologies within industries enhances productivity, fosters innovation, and increases competitiveness [79]. This is consistent with a study by Feng et al. from 2023 [80], which emphasizes the importance of industry structure and urban size structure as determining factors that impact how regional integration affects economic resilience. Collectively, these results support the idea that the digital economy is vital to enhancing economic resilience.

Among the control variables, carbon emissions negatively affect economic resilience at a 95% confidence level. One possible reason may be that the environmental costs associated with carbon emissions, such as pollution and climate change, pose risks to economic stability and sustainability. Studies have shown that the negative externalities of carbon emissions can lead to increased resource scarcity, regulatory burdens, and potential disruptions to industries and supply chains [81,82,83]. Furthermore, moving toward a low-carbon economy and the adoption of sustainable practices have been identified as crucial factors for enhancing economic resilience in the face of environmental challenges. By reducing carbon emissions and promoting sustainable practices, economies can mitigate environmental risks, enhance resource efficiency, and foster long-term economic stability and adaptability [84]. The coefficient of the variable “Cr” is estimated to be −0.930. Economic openness (Eo) does not significantly affect how resilient the economies of the research countries are. This suggests that the degree of economic openness, such as trade and foreign direct investment, has little bearing on the ability of the analyzed countries to weather economic downturns. Factors other than economic openness may have a greater influence on their economic resilience, such as domestic policies [85], institutional frameworks [86], or specific economic conditions [87]. The coefficient of the variable “Population Density” (Pd) is estimated to be −0.003, with a p-value of 0.005. The p-value represents the statistical significance of the coefficient, indicating the probability of observing such a relationship by chance. In this case, the p-value of 0.005 shows that, with 99% confidence, the coefficient is statistically significant. A negative coefficient indicates an inverse relationship between population density and economic resilience. In other words, as population density increases, economic resilience tends to decrease. This finding suggests that higher population density may pose challenges to economic resilience, possibly due to factors such as increased competition for resources, strained infrastructure, or reduced flexibility in response to shocks or disruptions [88]. Contrarily, the variable “Food Security” has an effect on economic resilience that is both positive and statistically significant. The coefficient of the variable “Food Security” (Fs) is estimated to be 0.092, which has a 95% level of confidence that is statistically significant. This implies that countries or regions with higher levels of food security demonstrate greater economic resilience [89]. Food security ensures stable access to nutritious food, reduces vulnerability to shocks, and supports overall economic stability. These findings highlight the importance of considering population density and food security as influential factors in understanding and enhancing economic resilience.

The normality of the error terms has been tested using the Kolmogorov–Smirnov test [90]. According to the results, the p-value obtained from this test was 0.870, indicating that the residuals follow a normal distribution. Consequently, conventional tests such as chi-square tests, t-tests, F tests, and Z tests can be applied.

To investigate serial correlation, the Durbin–Watson test was employed. The H0 hypothesis of no serial correlation was not rejected based on the test results, with a p-value greater than the significance level. Therefore, there is no evidence of serial correlation in the error terms of the first model. The heteroscedasticity of the variance in the first model was examined using the Breusch–Pagan–Cook–Weisberg test. The p-value obtained for this test was 0.359, indicating that the H0 hypothesis of homoscedastic variance was confirmed. Thus, the residuals of the first model exhibit homoscedastic variance.

Furthermore, the model’s specification was evaluated using the Ramsey Reset test. The p-value obtained for this test was 0.715, indicating that the H0 hypothesis was not rejected. Therefore, there is evidence to support the model’s specification, indicating that the digital economy has improved economic resilience.

The statistical tests performed on the initial model validate its reliability and establish the relationships between the variables. Specifically, Model 1, corresponding to Equation (1), delves into the effects of economic resilience on the digital economy. The results affirm the model’s validity and indicate a positive impact, providing evidence that the digital economy contributes significantly to economic resilience.

4.2. The Effects of Digital Economy on Energy Resilience

The findings of the estimation for panel data for the second model (Equation (2)) are presented in

Table 3. The estimation findings of the digital economy’s effects on energy resilience (Model 2).

Variable	Model (2)
	Coefficient	p-Value
Dig	0.005	0.009
Cr	−0.032	0.000
Eo	−0.001	0.349
Pd	−0.024	0.018
Fs	−0.008	0.010
Constant	4.187	0.000
Observation	1320
R2		24.16
Wald Test	14.69	0.012
Normality of Residual		0.206
Leamer Test	2.50	0.000
Hausman Test	1.75	0.186
Autocorrelation Test	0.646	0.425
Heteroskedasticity Test	1.75	0.185
Ramsey Reset test	0.67	0.570

. Before conducting the estimation, diagnostic tests were carried out to select the appropriate estimation method. The Leamer test rejected the hypothesis that the data are not combined at a 99% confidence level, indicating the need for the panel data method. Furthermore, the Hausman test, which determines the selection of the random effect method, was not rejected. Therefore, unlike the first model, the random effects panel data method should be utilized to estimate the second model. According to the estimation outcomes, the intercept of the second model is estimated to be 4.187, at a 99% confidence level, which is statistically significant. The coefficient for the variable “Dig” is estimated to be 0.050, also 99% confidently significant statistically. Although the coefficient may be small, its statistical significance suggests that even a slight increase in digital economy activities can contribute to improving energy resilience. This shows that energy resilience is favorably impacted by the digital economy. One possible explanation for this relationship is that energy efficiency is promoted by the digital economy, technological advancements, and the integration of renewable energy sources, which collectively enhance overall energy resilience [91]. Furthermore, digital technologies enable better monitoring, management, and optimization of energy systems, leading to more reliable and responsive energy infrastructure [92].

When it comes to control variables, the coefficient for the variable “Cr” (carbon emissions) in the second model is estimated to be −0.032, indicating a negative impact on energy resilience. The p-value of 0.000 indicates that this negative effect is statistically significant at a 99% level of confidence. This suggests that there is strong evidence to support the finding that carbon emissions have a detrimental effect on energy resilience. One possible explanation for this relationship is that carbon emissions contribute to environmental degradation, which can increase vulnerability and reduce the capacity to cope with energy disruptions and shocks. Environmental degradation, including air pollution and climate change induced by carbon emissions, can disrupt energy systems, hinder energy production and distribution, and pose challenges to energy infrastructure and supply chains [93,94]. These disruptions can undermine the resilience of energy systems and limit their ability to withstand and recover from disturbances. This result aligns with the research by Talebian et al. (2023) [94], which underscores the detrimental impacts of carbon emissions on energy resilience and the imperative of prioritizing sustainability [95]. Another contributing factor may be the reliance on fossil fuels and inefficient energy systems, which can impede both economic and energy resilience. The heavy dependence on fossil fuels exposes economies to price volatility, supply disruptions, and geopolitical risks, all of which can negatively impact both economic and energy resilience [96]. In contrast, transitioning to cleaner and more sustainable energy sources, such as renewable energy, can enhance energy resilience by reducing reliance on finite and polluting resources [97]. On the one hand, this transition can contribute to a more diversified and robust energy portfolio, enabling greater adaptability and resilience in the face of disruptions and shocks [98]. On the other hand, the more energy resilience and government support there is for energy supply, the higher the energy output achieved through renewable energy compared to non-renewable energy [99]. This highlights the positive relationship between energy resilience and the utilization of renewable energy sources. It also emphasizes the importance of prioritizing energy resilience as a necessity for achieving sustainable development and enhancing overall resilience, as supported by Esfandi et al. (2022) [88].

Regarding other variables, Economic openness (Eo) does not have a significant effect on energy resilience in the countries under study. The coefficient of the variable “Population Density” (Pd) is estimated to be −0.024, which is statistically significant at a 99% confidence level. Therefore, population density has a negative impact on energy resilience. Similarly, the estimate for the coefficient of the variable “Food Security” (Fs) is −0.008, which is statistically significant at a 99% confidence level, indicating that food security also negatively and significantly impacts energy resilience. One possible reason is that insufficient food security can lead to increased vulnerability in energy systems, particularly in terms of energy production, distribution, and access [100]. When a population lacks access to an adequate and reliable food supply, it can affect their overall well-being and ability to meet their energy needs. This, in turn, can undermine energy resilience as the availability and reliability of energy sources may be compromised in such circumstances.

According to the results of the Kolmogorov–Smirnov test, the residuals of the second model also follow a normal distribution. The H₀ hypothesis of the Wooldridge serial correlation test, which suggests the absence of serial correlation, is not rejected. Therefore, the disturbance terms of the second model exhibit no serial correlation. Considering the outcomes presented in Table 3, the H₀ hypothesis of the test for heteroscedasticity is confirmed, indicating that the residuals of the second model have homoscedastic variance. Therefore, the model specification has been correctly formulated, and energy resilience is positively impacted by the digital economy.

According to the results, the normality of the residuals in the second model was assessed using the Kolmogorov–Smirnov test. The p-value obtained of 0.206 suggests that the residuals follow a normal distribution. Therefore, there is no proof to disprove the hypothesis that the residuals of the second model exhibit normality. To investigate the presence of serial correlation in the disturbance terms of the second model, the Wooldridge serial correlation test was conducted. The test’s outcomes indicate a p-value of 0.425, demonstrating that the disturbance terms of the second model do not exhibit any meaningful serial association. Moreover, the test for heteroscedasticity in the residuals of the second model was performed, and the p-value obtained was 0.185. This confirms that the residuals have homoscedastic variance. Therefore, Model 2, corresponding to the Equation (2) specification, has been correctly formulated, and the residuals of the second model exhibit consistent variance across the range of the independent variables.

4.3. Environmental Resilience and the Impact of the Digital Economy

Panel regression estimation was employed to investigate how the digital economy specifically affects environmental resilience, as depicted in

Table 4. The estimation results of environmental resilience and the influence of the digital economy.

Variable	Model (3)
	Coefficient	p-Value
Dig	0.083	0.032
Cr	−0.105	0.027
Eo	0.218	0.204
Pd	−0.061	0.000
Fs	−0.046	0.071
Constant	6.960	0.000
Obs	1320
R2		38.58
Wald Test	69.18	0.000
Normality of Resid		0.445
Leamer Test	12.00	0.000
Hausman Test	5.06	0.409
Autocorrelation Test	0.161	0.690
Heteroskedasticity Test	0.25	0.617
Ramsey Reset test	0.94	0.421

based on Equation (3). In order to determine the most appropriate estimation method, diagnostic tests were conducted, and their results are presented in the lower part of Table 4. These tests provide valuable insights into selecting the suitable estimation approach. The findings from the Breusch–Pagan test indicate that there is a 99% probability that the homoscedasticity hypothesis is false. This suggests that the panel data method is well-suited for the analysis. Furthermore, the Hausman test does not reject the random effects model selection, thus affirming the appropriateness of employing the panel random effects method to estimate the third model.

According to the estimation results, the intercept of the third model is estimated to be 960.6, which, with 99% confidence, is statistically significant. At a 95% confidence level, the calculated coefficient for the variable “Digital Economy” (Dig) is 0.083, which is statistically significant. Therefore, the digital economy improves environmental resilience. One of the possible reasons why the digital economy improves environmental resilience is its potential to enable more efficient resource management. Digital technologies can facilitate real-time monitoring and data collection, allowing for better understanding and management of natural resources such as water, energy, and land. This enhanced visibility and control enable more effective resource allocation and utilization, reducing waste and promoting sustainability. Moreover, the digital economy fosters innovation in clean technologies and renewable energy [101]. Digital solutions can optimize energy systems, improve energy efficiency, and integrate renewable energy sources into the grid. The digital economy facilitates the development and implementation of smart grids, energy management systems, and sustainable energy solutions by utilizing sophisticated analytics, artificial intelligence, and Internet of Things (IoT) technology [102,103]. Additionally, the digital economy promotes sustainable production and consumption patterns. E-commerce platforms and digital marketplaces enable more efficient supply chains, reducing the need for physical transportation and associated carbon emissions. Digital platforms also facilitate sharing economy models, collaborative consumption, and circular economy practices, which minimize waste generation and promote resource efficiency [104].

When considering the control variables, carbon emissions show a negative coefficient of −0.105 (p-value = 0.027) at a 95% confidence level, indicating that carbon emissions significantly reduce environmental resilience. One possible reason is that this negative impact can be attributed to their contribution to climate change and associated environmental hazards [36]. Carbon emissions, primarily originating from burning fossil fuels, are a significant source of greenhouse gas emissions that cause climate change and global warming. Extreme weather events, including storms, heatwaves, droughts, and floods, may occur more frequently and with greater ferocity as a result of these changes in climate patterns, which pose significant challenges to ecosystems and natural resources. Additionally, carbon emissions contribute to air pollution, adversely affecting both human health and the environment. These factors undermine the resilience of ecosystems and their ability to withstand and recover from disturbances [105]. This is in line with Eldesoky et al. (2022) [106], who mentioned that the more cities that are exposed to extreme and frequent heatwaves, the more crucial it becomes to understand the combined influence of environmental dimensions on urban resilience.

Among other variables, economic openness (Eo) does not demonstrate a significant effect on environmental resilience in the countries under study, with a coefficient of 0.218 (p-value = 0.204). On the other hand, population density (Pd) exhibits a statistically significant negative impact on environmental resilience, with a coefficient of −0.061 (p-value < 0.001) at a 99% confidence level. Similarly, the variable for food security (Fs) shows a significant negative effect, with a coefficient of −0.046 (p-value = 0.071) at a 90% confidence level. High population density can give rise to various environmental challenges, including increased resource consumption, habitat fragmentation, and pollution. The pressure exerted by dense populations on natural resources and ecosystems leads to environmental degradation and a decline in resilience. Additionally, limited access to food and inadequate food security exacerbate these challenges, resulting in unsustainable agricultural practices, deforestation, and land degradation [107].

According to the Kolmogorov–Smirnov test, the p-value for the normality of residuals is 0.445, indicating that the residuals follow a normal distribution. The Durbin–Watson test, which examines serial correlation, yields a p-value of 0.690, suggesting that the disturbance terms in the third model do not exhibit serial correlation. The p-value for the heteroskedasticity test in terms of heteroskedasticity is 0.617, confirming that the residuals of the third model, corresponding to Equation (3), have homogeneous variance. Additionally, the p-value for the Ramsey Reset test is 0.421, implying that the model specification has been correctly formulated.

To assess multicollinearity, the Variance Inflation Factor (VIF) test was conducted for all research variables in the three models, as shown in

Table 5. Results of co-integration tests for research models.

Variable	VIF	1/VIF
Dig	1.00	0.998
Cr	1.11	0.903
Eo	1.61	0.622
Pd	1.49	0.673
Fs	1.01	0.998

. The calculated VIF values range from 1.00 to 1.61, and their corresponding 1/VIF values range from 0.622 to 0.998. Since all VIF values are below the threshold of 5, it can be concluded that the research models do not suffer from multicollinearity. The threshold of 5 is often used as a guideline in the Variance Inflation Factor (VIF) test to identify potential issues of multicollinearity. When the VIF value of a variable exceeds 5, it suggests that there may be a high correlation between those variables and other independent variables in the model, indicating the presence of multicollinearity. In such cases, the interpretation of the individual variable’s impact on the dependent variable becomes less reliable [108].

4.4. Economic Resilience in Above-Average and Below-Average Carbon-Emitting Country Groups as a Result of the Digital Economy

Economic resilience and the digital economy’s effects have been assessed for two groups of countries, divided into below-average and above-average carbon emissions regions based on average CO2 emissions. (In Appendix B, the list of countries is displayed).

Table 6. Estimation of economic resilience model by different carbon emission country groups.

Variable	Above-Average Carbon-Emitting Countries		Below-Average Carbon-Emitting Countries
	Coefficient	p-Value	Coefficient	p-Value
Dig	0.023	0.099	0.002	0.013
Cr	−0.040	0.076	−0.050	0.006
Eo	−0.008	0.333	0.008	0.428
Pd	−0.003	0.006	−0.048	0.001
Fs	0.025	0.000	0.086	0.000
Constant	7.589	0.000	9.935	0.000
Observation	660		660
F test	2.06	0.055	2.18	0.025
R2 Adj.		10.79		27.21

compares the findings of the estimations for how the digital economy will affect these two groups’ economic resilience. According to the results, the economic resilience of the below-average carbon-emitting country group is higher than that of the above-average carbon-emitting country group, and at a 99% confidence level, this difference is statistically significant (p < 0.01). Additionally, only the high-emitting country group’s coefficient for the digital economy variable is statistically significant at a 90% confidence level (p < 0.1), but for the low-emitting country group, it is significant at a 99% confidence level (p < 0.01). Therefore, compared to nations with high emissions, those with low emissions are more affected by the digital economy in terms of economic resilience. It is important to note that the effect of carbon emissions on the group of countries with below-average carbon emissions’ economic resilience is particularly important (p < 0.01), underscoring the crucial role of carbon emission reduction in enhancing economic resilience within this group. The more we reduce carbon emissions and leverage the potential of the digital economy, the more we can promote sustainable economic growth and resilience. This finding is consistent with the study by Dong et al. (2022) [44], which emphasizes the beneficial contribution of the digital economy to the reduction of carbon emission intensity.

4.5. The Impact of the Digital Economy on Energy Resilience in Countries with High and Low Carbon Emissions

The estimation findings for energy resilience in above-average and below-average carbon-emitting nations are shown in

Table 7. Energy resilience estimation findings for above-average and below-average carbon-emitting nations.

Variable	Above-Average Carbon-Emitting Countries		Below-Average Carbon-Emitting Countries
	Coefficient	p-Value	Coefficient	p-Value
Dig	0.002	0.000	0.027	0.027
Cr	−0.059	0.048	−0.047	0.002
Eo	−0.001	0.297	−0.001	0.654
Pd	−0.017	0.001	−0.087	0.000
Fs	−0.008	0.096	−0.007	0.051
Constant	4.786	0.000	3.463	0.000
Observation	660		660
Wald test	26.82	0.000	54.33	0.000
R2 Adj.		20.44		21.12

, which also demonstrates the statistical significance of the variables. For the above-average carbon-emitting country group, the digital economy variable’s coefficient (Dig) is 99% confident that it is statistically significant (p < 0.01), indicating its significant impact on energy resilience. Similarly, at a 90% confidence level, the carbon emission variable (Cr) is statistically significant (p < 0.1), highlighting the importance of carbon emissions in energy resilience within this group. In contrast, for the below-average carbon-emitting country group, at a 90% confidence level, the digital economy variable’s coefficient is significant, while carbon emissions (Cr) do not exhibit statistical significance. These results imply that in above-average carbon-emitting nations, the digital economy has a greater impact on energy resilience, while carbon emissions play a more significant role in energy resilience within the below-average carbon-emitting group. This finding is consistent with a recent study by Zou et al. (2023) [109], which also demonstrates that the comprehensive analysis of energy resilience in a hot–humid area of China reveals a projected rise in foreseeable energy needs for the urban housing market. These results underscore the significance of accounting for climate change in the design of efficient energy systems in cities. Siuta et al. (2022) [110] also highlight the transformative impact of digitalization on energy systems and emphasize the crucial role of reducing carbon emissions for achieving sustainable energy transitions.

4.6. The Impact of the Digital Economy on Environmental Sustainability in Groups of Above-Average and Below-Average Carbon-Emitting Countries

Table 8. Estimation of environmental resilience model by different carbon emission country groups.

Variable	Above-Average Carbon-Emitting Countries		Below-Average Carbon-Emitting Countries
	Coefficient	p-Value	Coefficient	p-Value
Dig	0.011	0.147	0.062	0.000
Cr	−0.029	0.000	−0.036	0.002
Eo	0.001	0.786	−0.001	0.644
Pd	−0.099	0.032	0.001	0.004
Fs	−0.044	0.088	−0.096	0.000
Constant	6.887	0.000	7.001	0.000
Observation	660		660
Wald test	18.44	0.072	14.25	0.0145
R2 Adj.		13.27		30.13

presents the estimation results for the connection between environmental resilience and the digital economy in above-average carbon-emitting and below-average carbon-emitting countries. The p-values associated with the digital economy’s correlation reveal that it is not statistically significant in above-average carbon-emitting countries (p = 0.147), suggesting limited influence on environmental resilience. However, at a 99% confidence level, the coefficient of the digital economy variable is statistically significant in nations with below-average carbon emissions (p < 0.01), indicating a stronger impact on environmental resilience. Furthermore, considering the coefficient of the Cr variable representing carbon emissions, it is observed to be statistically significant in both above-average carbon-emitting countries (p < 0.001) and below-average carbon-emitting countries (p = 0.002). This suggests that carbon emissions have a significant influence on environmental resilience in both groups, with a larger impact observed in countries with lower carbon emissions. These findings underscore the importance of improving environmental resilience by taking into account carbon emissions and the digital economy, particularly in above-average carbon-emitting countries that may have implemented stringent environmental regulations and sustainable practices.

4.7. The Interrelationships between the Digital Economy and Three Different Resilience

To examine the interrelationships among three variables—economic resilience, energy, and the environment—and the potential intersection effects between the digital economy and other variables, three different tests were employed. In

Table 9. Simultaneous Estimation of Correlations in Three Models Using the SUR Method.

Variable	CRES (Model 1)		GRES (Model 2)		VRES (Model 3)
	Coefficient	p-Value	Coefficient	p-Value	Coefficient	p-Value
Dig	0.011	0.000	0.060	0.009	0.087	0.000
Cr	−0.090	0.000	−0.024	0.015	−0.108	0.016
Eo	−0.002	0.263	−0.001	0.402	0.0001	0.934
Pd	−0.023	0.002	−0.118	0.014	−0.033	0.000
Fs	0.021	0.019	−0.008	0.026	−0.001	0.896
Constant	7.382	0.000	4.194	0.000	6.951	0.000
Observation	1320		1320		1320
Wald test	2093.42	0.000	2442.34	0.0000	1890.74	0.0000
R2 Adj.	29.85		25.93		23.64

, three research models were simultaneously fitted using the seemingly unrelated regressions (SUR) method. It is evident that in this approach, the digital economy is confirmed to be a significant factor in enhancing economic resilience, energy, and environmental resilience at a confidence level of 99%. Based on the correlation matrix results, there is no significant correlation or dependence among the three variables of economic resilience, energy, and the environment. Therefore, there is no evidence of a relationship between them.

Furthermore, the endogeneity of the main variable, the digital economy, was also examined in all three models using the Durbin–Wu–Hausman test. The null hypothesis of this test posits that this variable is exogenous. According to

Table 10. Results of the Endogeneity Test for the Digital Economy Variable.

p-Value	Value	Model
0.582	0.302	Model 1
0.562	0.336	Model 2
0.472	0.518	Model 3

, the computed statistic for the first model is 0.302, for the second model is 0.336, and for the third model is 0.518. Therefore, in all three models, the mentioned variable is considered exogenous. According to the results of this test, there is no substantial collinearity between this variable and the other model variables, and there is no issue for establishing classical assumptions in the model.

5. Conclusions and Policy Implications

This study examines the influences of the digital economy on energy, economic, and environmental resilience across 66 countries from 2000 to 2020. The findings demonstrate that the digital economy significantly enhances resilience in the environmental, energy, and economic domains for various countries and time periods. The digital economy, however, has a detrimental impact on resilience in each of the three domains when carbon emissions are taken into account. The research also shows that below-average carbon-emitting nations’ economic resilience is more affected by the digital economy, highlighting the crucial role of reducing carbon emissions in enhancing their economic resilience. In contrast, energy resilience in the below-average carbon-emitting group is less influenced by the internet-based economy, with carbon dioxide emissions exerting a greater impact on the resilience of the energy system within this group. Additionally, the digital economy has limited influence on environmental resilience in countries with above-average carbon-emitting, while lower carbon-emitting countries experience larger carbon emissions, making their environmental resilience more susceptible to such emissions. These findings underscore the complex relationship between the global resilience landscape, carbon emissions, and many aspects of the digital economy.

Considering this study’s empirical findings, the following recommendation can be made to further enhance the understanding and application of the findings. The complex connection between resilience, carbon emissions, and the digital economy requires a collaborative approach among countries. Governments, international organizations, and stakeholders should work together to share best practices, exchange knowledge, and develop joint initiatives aimed at promoting sustainable and resilient digital economies. International collaboration can help identify common challenges and develop innovative solutions that address global issues related to energy, economic, and environmental resilience. Moreover, recognizing the potential negative impact of the digital economy on resilience through increased carbon emissions, efforts should be made to promote sustainable digital practices. This could involve encouraging the adoption of energy-efficient technologies, promoting renewable energy sources to power digital infrastructure, and raising awareness among businesses and individuals about the carbon footprint associated with digital activities. Emphasizing sustainability in the digital economy can contribute to a more resilient and environmentally conscious approach.

While this study provides valuable insights into the complex relationship between the digital economy, energy, economic, and environmental resilience, there are several limitations that should be acknowledged. Firstly, the study relies on aggregated country-level data, and the generalization of findings to specific regions or localities within countries may overlook nuances in resilience dynamics at more localized levels. The heterogeneity within countries could impact the accuracy and applicability of the results at finer geographical scales. Secondly, while the study examines the relationship between the digital economy and resilience, it may not account for all potential influencing factors. Other socio-economic, political, or cultural variables that were not included in this analysis could play a role in shaping resilience outcomes. Additionally, future research could explore additional environmental indicators to provide a more comprehensive understanding of the environmental implications of the digital economy.