Integrating Geospatial Intelligence and Digital Transformation for Advancing Environmental Sustainability in Emerging EU Economies

Abstract

This study explores the interplay between digital transformation and environmental sustainability in four emerging European Union economies—Bulgaria, Hungary, Poland, and Romania—through the lens of geographic information science and technology (GIS&T). Leveraging panel data from 2010 to 2022, digitalization is quantified through infrastructure, affordability, and user adoption metrics, consolidated into a comprehensive digitalization index. Environmental sustainability is assessed via greenhouse gas emissions and the share of renewable energy in total energy consumption, capturing critical spatial and ecological dynamics. Applying a generalized linear model, the results reveal that enhanced digital infrastructure and affordability contribute to emissions reduction, whereas increased user adoption correlates with higher energy demand and emissions. Furthermore, digitalization fosters improved integration of renewable energy sources, highlighting the strategic role of GIS&T in monitoring and optimizing sustainability initiatives. These insights emphasize the necessity of embedding geospatial analytics into policy frameworks to align digital transformation with environmental and socioeconomic sustainability goals.

1. Introduction

In the era of rapid technological advancements, digital transformation has emerged as a pivotal force reshaping economies across the globe [1]. In this context, global initiatives such as the European Green Deal and the UN’s 2030 Agenda for Sustainable Development underscore the importance of integrating digital transformation with sustainable practices. As part of the EU’s efforts to achieve climate neutrality by 2050, these policies encourage member states, particularly emerging economies such as Bulgaria, Hungary, Poland, and Romania, to adopt digital technologies that not only foster economic growth but also contribute to environmental sustainability. For these countries, digitalization presents a unique opportunity to accelerate their development while addressing the pressing challenges of climate change and promoting green growth in line with EU priorities. While classifications of emerging markets vary, institutions such as the International Monetary Fund (IMF) continue to categorize Bulgaria, Hungary, Poland, and Romania as emerging economies based on factors such as income levels, market accessibility, and financial system maturity [2,3,4,5]. For these emerging economies within the European Union (EU), the integration of digital technologies presents both opportunities and challenges. While digitalization drives economic growth and enhances competitiveness, its environmental implications remain a critical area of investigation. The intersection of digital transformation and environmental sustainability has garnered significant attention, particularly in the context of mitigating climate change and promoting sustainable development [6].

The definition of digital transformation has not reached a consensus among academic researchers [7]. These definitions highlight various characteristics of digital transformation but do not provide concrete definitions [8]. However, Mergel et al. [9] define digital transformation as the necessity for companies to enhance their competitiveness in the internet era by providing goods and services both online and offline through the use of new technologies. Similarly, Fitzgerald et al. [10] describe digital transformation as the utilization of digital technologies to achieve significant improvements in business, including optimizing operational processes, enhancing customer experience, and developing new business models. To better understand the concept of digital technology, the literature defines this term as a set and new paradigm of intelligent and innovative technologies within the context of Industry 4.0, which includes big data analytics, the Internet of Things (IoT), and cloud computing, all of which facilitate connectivity, communication, and process automation [11,12].

In the context of this diversity of definitions and perspectives on digital transformation, it is important to emphasize that, regardless of specific definitions, the literature acknowledges the essential role of digital technologies in modernizing economies and enhancing competitiveness while also supporting the transition to more sustainable economic practices. Additionally, digital technology impacts traditional industries by providing smart equipment and production management methods that support environmentally friendly production techniques, thereby encouraging the modernization and transformation of traditional industries to reduce carbon emissions [13]. Furthermore, digital investments drive clean productivity growth through the development of sustainable production processes [14].

Within the European Union, this process is particularly significant for emerging countries such as Bulgaria, Hungary, Romania, and Poland, as these economies face substantial challenges related to sustainable development, including the need to reduce air emissions and increase the use of renewable energy. The United Nations has adopted a set of 17 Sustainable Development Goals, with the primary aim of eradicating poverty, protecting the environment, and ensuring equitable prosperity globally [15]. Consequently, approximately 200 countries and regions have reached a common consensus to develop clean, low-carbon energy through the Paris Agreement of 2015 [16].

Although the European Union has made considerable advancements in the adoption of renewable energy sources over the past two decades—achieving 18% of gross final energy consumption in 2018, nearly twice the level recorded in 2005—significant progress is still required to meet long-term environmental targets [17]. This expansion of renewable energy has contributed to the reduction of greenhouse gas emissions and a decreased dependence on fossil fuels; however, substantial challenges persist. Emerging economies within the European Union, such as Poland, Romania, Bulgaria, and Hungary, have also achieved notable advancements in the utilization of renewable energy, yet further efforts are necessary to realize these environmental objectives fully. In this context, the integration of digital technologies emerges as a critical enabler in accelerating the transition towards sustainable and environmentally responsible development, presenting new opportunities for resource optimization and emissions reduction. Specifically, a report by the International Energy Agency [18] suggests that digital technologies could reduce global energy consumption in buildings by up to 10% by 2040. These reductions are primarily driven by enhancements in energy efficiency through the use of smart thermostats, lighting, and appliances that dynamically adjust energy consumption based on real-time data and user behavior. Another study conducted by PwC [19] underscores the benefits of implementing Industry 4.0 technologies, such as IoT, AI, and big data analytics, in reducing industrial CO₂ emissions by up to 15% through process optimization and waste reduction. These findings illustrate the transformative potential of digitalization in significantly enhancing environmental performance, particularly in the domains of energy efficiency, renewable energy integration, and emissions reduction.

Moreover, the strategic implementation of these technologies within emerging economies is of particular significance. These economies, while grappling with unique challenges in their pathways to modernization and sustainable development, also present significant opportunities. According to research by Pinto et al. [20], within the European Union, Bulgaria, Poland, and Romania are recognized as “champions of digital transformation” due to their above-average performance across key digitalization indicators, reflecting strong commitments and successes in this area. Conversely, Hungary, identified within the “empowerment-driven disparities” cluster, exhibits above-average performance in digital empowerment but lags in business digitalization and broadband accessibility [20]. Therefore, digital transformation holds the potential to play a pivotal role in accelerating the transition of these economies towards more sustainable practices, thereby enhancing their global competitiveness and significantly contributing to environmental objectives.

In this context, this study is of special significance as it addresses the pressing need to clarify the role of digital transformation in enhancing environmental performance within emerging economies. These economies are at a crucial stage of development, where balancing economic expansion with environmental sustainability has become increasingly imperative. The primary objective of this research is to assess the extent to which the adoption of advanced digital technologies accelerates the transition towards sustainable economic practices and analyze the mechanisms through which these technologies can be strategically leveraged to support environmental objectives. By providing a comprehensive understanding of this dynamic, the study aims to offer valuable insights and recommendations for policymakers and stakeholders engaged in economic and environmental development, thereby contributing to the achievement of long-term sustainable development.

To achieve the objectives of this study, we have formulated a series of research questions aimed at clarifying the relationship between digital infrastructure and environmental sustainability in emerging economies within the European Union:

• Q1: To what extent does the improvement of digital infrastructure contribute to the reduction of greenhouse gas emissions in emerging economies?
• Q2: Is there a significant relationship between the affordability of digital technologies and the increase in the use of renewable energy sources?
• Q3: How does user adoption of digital technologies influence energy efficiency and environmental performance in emerging economies?
• Q4: How does advanced digital infrastructure contribute to reducing carbon emissions and enhancing environmental sustainability?

These questions are crucial for understanding how the adoption of digital technologies can positively influence environmental performance in these regions and for identifying the mechanisms through which digitalization can support the transition towards more sustainable economic practices.

In this framework, the structure of the paper is organized as follows: Section 2 provides a comprehensive review of the existing literature, exploring in detail the impact of digitalization on environmental performance. Section 3 outlines the methodology adopted and describes the datasets used in our analysis. The results obtained and their corresponding interpretations are presented in Section 4, which includes an in-depth discussion of the implications of these findings. Finally, Section 5 synthesizes the main conclusions of the study, offers policy recommendations, and highlights the limitations of the current research while also suggesting directions for future investigations.

2. Literature Review

The environmental implications of digital transformation have become a focal point of research, particularly in the context of global efforts to combat climate change. Digital technologies can have both positive and negative effects on the environment, depending on how they are deployed and managed. On the positive side, digital technologies can contribute to environmental sustainability by improving energy efficiency, enabling the transition to renewable energy, and supporting the development of smart cities [21].

Digital transformation has a significantly positive impact on the environment, playing a crucial role in enhancing energy consumption efficiency and reducing carbon emissions [22]. Numerous studies have highlighted that the use of smart technologies contributes to increased efficiency in resource and energy management, optimizing production processes and energy distribution, leading to a substantial reduction in resource and energy waste [23]. Additionally, researchers such as Chong et al. [24] have demonstrated that digital transformation is essential for reducing carbon emissions by facilitating teleworking and optimizing supply chains, thereby decreasing the carbon footprint associated with transportation. Moreover, it fosters ecological innovations, such as electric vehicles and renewable energy technologies, reducing dependence on fossil fuels. Digital technologies also enhance environmental monitoring capabilities, enabling rapid response actions. Recently, Hao et al. [14] demonstrated that digitalization and the growth of the green economy represent a steadily increasing trend, with digitalization significantly promoting the latter, exhibiting a marginal effect of 1.648.

However, digitalization also presents challenges, including the increase in electronic waste and energy consumption in data centers [25]. Due to its diverse effects—sometimes beneficial, other times detrimental—on the environment, digital transformation has been extensively studied in the literature, with its long-term benefits being consistently confirmed.

According to the literature, environmental performance can be measured through a variety of indicators and methods, though it is most often evaluated using greenhouse gas emission indicators and energy efficiency indicators, frequently expressed through the use of renewable energy. In this context, we will analyze studies investigating the role of digitalization in reducing greenhouse gas emissions and studies examining its impact on increasing the use of renewable energy.

Several studies have demonstrated that digital transformation influences greenhouse gas emissions both directly, through the production, use, and disposal of hardware and infrastructure required for information and communication technologies (ICT), and indirectly, through ICT’s impact on emissions in other sectors such as transportation and energy, by applying these technologies to optimize and streamline processes in those sectors [26]. Preliminary studies have supported the idea that net effects (calculated as the difference between indirect and direct effects) can lead to a significant reduction in these emissions [27,28]. For instance, Pamlin and Szomolányi [27] estimated that the potential for reducing greenhouse gas (GHG) emissions in the EU through flexible work, audio and video conferencing, online telephone bills, virtual voicemail, and online tax returns is 48.37 Mt CO₂ compared to 4.73 Mt CO₂ directly caused by ICT.

Furthermore, in the Global e-Sustainability Initiative [28], the ICT industry’s association for sustainability asserts that, on a global scale, ICT applications could avoid up to 20% of annual GHG emissions by 2030 (indirect effect), while the ICT sector causes approximately 2% of global GHG emissions (direct effect). Similarly, Malmodin and Bergmark [29] explored the potential for reducing global greenhouse gas emissions by 2030, with results indicating that ICT solutions could contribute to a reduction of about 8 gigatons of CO₂e, equivalent to 12% of global GHG emissions under a high reduction potential scenario. In a medium reduction scenario, the potential was estimated at approximately 4 gigatons of CO₂e or 6% of global GHG emissions in 2030. Thus, both studies assessing the effects of ICT solutions on greenhouse gas emissions indicate positive outcomes, highlighting the potential of digital transformation to contribute positively to environmental sustainability by reducing GHG emissions.

A more focused study applied to Switzerland, conducted by Bieser and Hilty [26], confirmed that ICT has the potential to reduce GHG emissions in the country, particularly in the building, transportation, and energy sectors. The hypothesis that the digital transformation of enterprises promotes significant synergistic reductions in pollution and carbon emissions was recently validated by Chong et al. [24] through a series of robustness tests, confirming the positive impact of digitalization on environmental performance. Additionally, a favorable impact of digital transformation on the reduction of carbon intensity in urban infrastructure was observed in China, according to a study by Wang et al. [22]. The study demonstrated that digitalization, besides reducing local emissions, also contributed to the reduction of carbon emissions in neighboring cities, exhibiting a spatial spillover effect, with the limit of this effect being determined to be 600 km.

Consequently, the impact of digitalization on reducing greenhouse gas emissions is considerable and positive, as evidenced by numerous studies and research. Digital transformation, through the use of smart technologies and ICT solutions, significantly contributes to the efficiency of industrial processes, the optimization of energy consumption, and the reduction of the carbon footprint. Studies show that, in addition to direct effects, digitalization also has an important indirect impact on reducing greenhouse gas emissions, by improving production processes and logistics, facilitating telework, and stimulating green innovation. Overall, digitalization emerges as an essential factor in global efforts to reduce emissions and promote environmental sustainability.

Regarding the impact of digitalization on renewable energy, it is argued that digital technology triggers technological innovation in the energy industry and promotes innovative production processes throughout its entire life cycle, particularly in the renewable energy sector [30].

The results of studies conducted by Zhao et al. [31] and Ren et al. [32] both indicated that digital transformation has increased the efficiency of renewable energy use in China. Furthermore, the latter study emphasizes that when a renewable energy enterprise adopts digital transformation, it achieves higher operational efficiency, lower costs, and better innovation success, leading to improved performance. In another study conducted in China, Cao et al. [33] investigated the impact of digital finance on energy and environmental performance using panel data for the period 2011–2017. The results demonstrated that digital finance significantly contributes to improving energy and environmental performance in China, with the conclusions remaining robust even after rigorous testing. The positive effects of digital transformation on the performance of the renewable energy sector have also been identified by other researchers [34,35,36,37].

Ning and Xiong [38] analyzed the dynamic evolution process of the digital transformation of renewable energy enterprises, and the results showed that the “digital transformation” of renewable energy power plants and renewable energy sales companies brings benefits through cooperation and concertedly promotes the adoption of digital transformation.

Xu et al. [39] identified an interesting impact: digitalization has a greater impact on energy in low-income countries and underdeveloped regions and a lesser impact on energy in high-income countries and developed regions.

Recently, Pandey et al. [40] conducted a study with the primary aim of evaluating the potential of Industry 4.0 technologies in optimizing renewable energy production. Following analyses and research, the authors concluded that Industry 4.0 could make intermittent renewable energy more accessible in several industries, including solar, wind, hydropower, and biomass energy generation. Specifically, they argue that virtual power plants and microgrids can facilitate energy distribution, ultimately increasing the number of renewable energy users.

Based on existing studies, it can be inferred that digitalization has a significant and positive impact on the performance of the renewable energy sector, contributing to increased efficiency, cost reduction, and the stimulation of technological innovation in this field. The studies demonstrate that the adoption of digital technologies by renewable energy enterprises not only improves operational performance but also facilitates cooperation among various entities in the sector, leading to a broader and concerted adoption of these technologies.

As observed in the literature, most studies focus solely on the impact of digital transformation on greenhouse gas emissions or the use of renewable energy, without considering the complex role that digitalization can play in both dimensions of environmental performance. This limited approach neglects the way in which digitalization can influence and mediate the interaction between these two aspects, highlighting the need for a more integrated analysis that fully explores the potential of digital transformation in promoting ecological sustainability. Additionally, there is a significant gap in research on the impact of digital transformation on environmental performance in developing countries in Europe. Most existing studies focus on China or more heterogeneous groups, such as the entire European Union, leaving the specific situation of emerging European economies largely unexplored.

The contribution of this study lies in addressing this gap by providing a detailed analysis of how digital transformation influences environmental performance in developing European countries, such as Romania, Bulgaria, Poland, and Hungary. This research not only extends the theoretical understanding of the interactions between digitalization and environmental performance but also offers empirical results that support practical recommendations for policymakers in these regions, thereby contributing to a more balanced and sustainable development. This underscores the importance of implementing digital policies in emerging European economies, highlighting the ways in which digital technologies can support the transition towards more sustainable economic practices.

3. Data and Methods

3.1. Data and Variables

To achieve the objective of this study, which examines the impact of digital transformation on the environmental performance of emerging economies, a comprehensive panel dataset was constructed. This dataset encompasses four European Union member states—Bulgaria, Hungary, Poland, and Romania—that are classified as emerging economies by the International Monetary Fund (IMF) [2,3,4,5,41]. The analysis of digital transformation is operationalized through three key dimensions of digitalization: infrastructure, affordability, and user adoption. These dimensions were meticulously selected and calculated based on the rigorous methodology put forth by Cámara and Tuesta [42], incorporating a set of 10 carefully chosen indicators, which are detailed in

Table 1. Data information.

Dimension		Acronym	Indicator	Unit of Measure	Source
Digital transformation	User adoption	AMBS	Active mobile-broadband	Subscriptions per 100 inhabitants	ITU [43]
		HOUSE	Households with internet access at home	%	ITU [43]
		INDIV	Individuals using the Internet, total	%	ITU [43]
		MOBS	Mobile-cellular	Subscriptions per 100 inhabitants	ITU [43]
		FIXB	Fixed broadband	Subscriptions per 100 inhabitants	ITU [43]
	Infrastructure	POP3G	Population covered by at least a 3G mobile network	%	ITU [43]
		INTBAND	International bandwidth per internet user	kbit/s	ITU [43]
		SECURE	Secure internet servers	Per 1 million people	WBG [44]
	Affordability	FIXTAR	Fixed-broadband basket	PPP$	ITU [45]
		MOBTAR	Mobile-cellular low-usage basket	PPP$	ITU [45]
Environmental performance	Air emissions	GHG	Annual greenhouse gas emissions	CO2 equivalents	IEA [46]
	Renewable energy	REN	Share of energy from renewable sources	%	Eurostat [47]

Source: Own processing, 2024.

.

The environmental performance of these economies is evaluated using two critical metrics: greenhouse gas emissions and the proportion of energy generated from renewable sources. Greenhouse gas emissions are a direct measure of the environmental impact, reflecting the levels of carbon dioxide and other harmful gases released into the atmosphere. In contrast, the percentage of energy derived from renewable sources serves as an indicator of the sustainable energy practices adopted by these economies.

The data employed in this study were meticulously sourced from reputable international databases to ensure accuracy and reliability for the period from 2010 to 2022. Specifically, the digitalization indicators were obtained from the International Telecommunication Union (ITU) and World Bank Group (WBG), which provides comprehensive data on global telecommunication and information technology trends. The environmental performance data, including GHG emissions and renewable energy statistics, were sourced from the International Energy Agency (IEA) and Eurostat, ensuring that the analysis is grounded in the most up-to-date and reliable data available.

The digital transformation of the four emerging economies within the European Union—Bulgaria, Hungary, Poland, and Romania—was rigorously assessed in the present study through the calculation of a comprehensive digitalization index (DIGIX) for the period 2010–2022. This index encapsulates the three critical dimensions of digital transformation: infrastructure (INFR), affordability (AFFORD), and user adoption (USAD), as illustrated in Figure 1.

The relative weighting of these dimensions within the overall index was meticulously determined based on an extensive review of existing literature, with specific reference to the values assigned to the underlying indicators across multiple empirical studies [48,49,50]. These studies provided a robust framework for understanding the structural composition of digitalization and guided the methodological approach in assigning appropriate weights to each dimension. By incorporating these well-established methodologies, the study ensures a nuanced and precise measurement of digital transformation, enabling a more accurate evaluation of its impact on the environmental performance of these emerging economies.

In light of the research objectives, the proposed econometric model will incorporate the three dimensions of the digitalization index—infrastructure, affordability, and user adoption—as the primary explanatory variables. These dimensions are critical to understanding the multifaceted nature of digital transformation and its potential influence on environmental outcomes.

The model will further utilize two key indicators of environmental performance—greenhouse gas emissions and the percentage of energy derived from renewable sources—as dependent variables. These indicators have been selected based on their relevance and capacity to capture the broader environmental impact of digitalization within emerging economies.

To provide a comprehensive foundation for the subsequent analysis, descriptive statistics for all variables involved in the model have been meticulously calculated and are presented in

Table 2. Descriptive statistics of the variables.

	GHG	REN	AFFORD	INFR	USAD
Mean	1.44 × 108	17.3528	28.1678	96.4807	72.6803
Median	8.02 × 107	15.9695	27.4550	99.8500	71.6600
Maximum	3.82 × 108	25.0320	43.6800	100.0000	107.6800
Minimum	4.50 × 107	9.2810	12.8600	61.2000	45.3400
Std. Dev.	1.29 × 108	4.7838	9.5904	8.5946	14.5435
Skewness	1.0793	0.2550	0.1901	−2.9462	0.4775
Kurtosis	2.2909	1.7393	1.8656	10.7930	3.0700
Jarque–Bera	11.1866	4.0071	3.1014	206.8140	1.9873
Probability	0.0037	0.1348	0.2120	0.0000	0.3702

Source: Own processing, 2024.

. These statistics offer an essential overview of the data, highlighting the central tendencies, variability, and distributional properties of both the explanatory and dependent variables.

This preliminary analysis is crucial for understanding the underlying data structure and ensuring the robustness of the econometric model to be applied. Moreover, it facilitates the identification of potential patterns, outliers, and correlations that may influence the interpretation of the results and the validity of the conclusions drawn from the analysis. Taking into account that the data do not follow a normal distribution, the appropriate model for the analysis is a generalized linear model using the Newton–Raphson algorithm and the Marquardt estimation method.

3.2. Methodology

The generalized linear model (GLM) is an extension of traditional linear regression that allows response variables to follow distributions other than the normal distribution. GLMs are used for a wide range of data types, including binary outcomes (logistic regression), count data (Poisson regression), and more. The model is characterized by specifying a linear predictor and a link function that connects the linear predictor to the expected value of the response variable.

The GLM framework introduced by Nelder and Wedderburn [51] extends linear regression by allowing the mean component ${\mu }_i$ to be linked to a linear predictor via a nonlinear function, while the distribution of the stochastic component ε_i can belong to any member of the linear exponential family. Specifically, a GLM specification includes:

• a linear predictor or index ${\eta }_i=X_i^{\prime }\beta +o_i$, unde oi is an optional offset term;
• a distribution for Y_i belonging to the linear exponential family;
• a smooth, invertible link function, $g\left({\mu }_i\right)={\eta }_i$, relating the mean ${\mu }_i$ and the linear predictor ${\eta }_i$

The GLM assumptions imply that the first two moments of Y_i may be written as functions of the linear predictor:

$${\mu }_i=g^{-1}\left({\eta }_i\right)$$

(1)

$$V_i=\left({\displaystyle \frac{\varphi }{{\omega }_i}}\right)V_{\mu }\left(g^{-1}\left({\eta }_i\right)\right)$$

(2)

where $V_{\mu }\left(\mu \right)$ is a distribution-specific variance function describing the mean–variance relationship, the dispersion constant ϕ > 0 is a possibly known scale factor, and ω_i > 0 is a known prior weight that corrects for unequal scaling between observations.

In GLMs, the Newton–Raphson and Marquardt methods are used to iteratively estimate the model parameters by maximizing the log-likelihood function [52]. The Newton–Raphson algorithm is a root-finding algorithm used to solve equations by iteratively improving guesses. In the context of GLMs, it is employed to find the parameter estimates that maximize the likelihood function. The method converges quickly near the optimum if the initial guess is close enough to the true parameter values and requires computation of the Hessian matrix. The Marquardt method combines gradient descent and the Newton–Raphson algorithm to improve convergence, especially when the initial parameter estimates are far from the optimum. The method is less sensitive to poor initial guesses and can handle nonlinearities better [53].

The selection of GLM in this study is based on several key considerations related to the characteristics of the data and the research objectives. Traditional linear regression models operate under the assumption of normally distributed residuals, constant variance, and linear relationships between variables. However, environmental and economic data often exhibit deviations from these assumptions due to inherent complexities, nonlinearity, and heteroskedasticity. GLM provides a more flexible modeling framework that accommodates a wider range of data distributions and relationships, making it particularly well-suited for analyzing the impact of digital transformation on environmental sustainability indicators.

One of the primary advantages of GLM is its ability to handle nonnormal data distributions. In this study, greenhouse gas (GHG) emissions and renewable energy adoption, the key environmental performance indicators, exhibit skewed distributions and varying dispersion levels across observations. GLM allows for different error structures and link functions, ensuring that the relationships between variables are appropriately captured without imposing restrictive assumptions about normality. This flexibility enhances the robustness of the estimated coefficients and improves the interpretability of the results.

Another critical factor in choosing GLM is its resilience to nonstationary data. As demonstrated by the stationarity tests conducted in this study, some of the independent variables exhibit unit roots, indicating potential nonstationary behavior. In contrast to conventional time-series models, which require strict stationarity conditions to avoid spurious regression results, GLM remains robust in the presence of nonstationary panel data. By using appropriate link functions and variance structures, GLM mitigates biases associated with time-dependent trends, allowing for more reliable inference.

Moreover, the flexibility of GLM in modeling panel data structures makes it particularly relevant for this research. The dataset encompasses multiple countries over an extended period, requiring an analytical framework capable of capturing both cross-sectional and temporal variations. GLM accommodates varying data structures while maintaining statistical efficiency, making it an effective tool for evaluating the role of digital transformation in shaping environmental performance outcomes.

Additionally, the precedent set by previous empirical research further supports the application of GLM in this context. Several studies examining the relationship between technological advancements and environmental sustainability have utilized GLM due to its ability to model complex interactions and account for nonlinearity in economic and environmental phenomena. The methodological rigor and empirical validity of GLM in similar domains reinforce its appropriateness for the current study.

In alignment with the overarching objective of this research, we have developed two distinct econometric models, each specifically designed to address the key research questions identified in the earlier stages of our study:

$$\mathrm{M}1:{lnGHG}_i=\exp \left({\beta }_1+{\beta }_2{AFFORD}_i+{\beta }_3{INFR}_i+{\beta }_4{USAD}_i\right)+{\epsilon }_i$$

(3)

$$\mathrm{M}2:{REN}_i=\exp \left({\beta }_1+{\beta }_2{AFFORD}_i+{\beta }_3{INFR}_i+{\beta }_4{USAD}_i\right)+{\epsilon }_i$$

(4)

4. Results and Discussion

Conducting stationarity tests before applying an econometric model is crucial for avoiding spurious regressions and ensuring the validity of the results. Stationarity of a time series, which implies a constant mean, variance, and autocovariance over time, is essential because using nonstationary series can lead to apparently significant relationships between variables that are, in reality, coincidental and do not reflect a true connection. Therefore, stationarity tests help confirm that the estimated econometric relationships are robust and reliable.

In our study, we employed two distinct unit root tests to assess the stationarity of the panel data: the Levin, Lin, and Chu (LLC) test [54] and the Im, Pesaran, and Shin (IPS) test [55]. The LLC test is designed to evaluate the presence of a unit root under the assumption of a common unit root process across all cross-sectional units within the panel. This means that the LLC test assumes a homogeneous autoregressive parameter across the entities, making it suitable for panels where the cross-sectional units share similar dynamics over time. On the other hand, the IPS test is more flexible, as it allows for heterogeneity in the autoregressive coefficients across different cross-sectional units. This test accommodates individual unit root processes, recognizing that each entity in the panel may have its own unique time-series properties. By applying both the LLC and IPS tests (

Table 3. Unit root test for the data in the model.

	LLC		IPS
	Statistic	Probability	Statistic	Probability
lnGHG	−3.3149	0.0005	−1.9863	0.0235
REN	−3.6247	0.0001	−1.6173	0.0529
AFFORD	0.5406	0.7056	1.0675	0.8571
INFR	−122.0910	0.0000	−77.5371	0.0000
USAD	−3.4510	0.0003	−2.5603	0.0052

Source: Own processing, 2024.

), we aim to capture a comprehensive picture of the stationarity properties of our data, accounting for both common trends and individual variations within the panel. This dual approach enhances the robustness of our analysis and ensures that our econometric modeling is based on reliable and properly diagnosed data.

Considering that not all variables employed in the analysis exhibit stationarity, the use of the GLM is not only justified but also advantageous for obtaining consistent and robust results. Unlike traditional time series models, which often require strict stationarity assumptions to avoid spurious relationships, the GLM framework is designed to handle a broader range of data structures.

GLM does not rely on the time-dependent properties of the data in the same way that models such as ARIMA or VAR do, making it particularly suitable for analyzing datasets where variables may be integrated at different orders or exhibit nonstationary behavior. By utilizing appropriate link functions and error distributions, GLM allows for flexibility in modeling complex relationships between variables, thereby mitigating potential biases or inefficiencies that might arise from nonstationary data. This capability ensures that the estimations remain reliable and that the inferences drawn from the model are valid, even in the presence of nonstationarity among the variables.

The correlation matrix between variables presented in

Table 4. Correlation matrix for the two models.

M1					M2
	LNGHG	AFFORD	INFR	USAD		REN	AFFORD	INFR	USAD
lnGHG	1.0000				REN	1.0000
	-----					-----
AFFORD	−0.4929	1.0000			AFFORD	−0.3093	1.0000
	0.0002	-----				0.0257	-----
INFR	−0.1710	−0.1123	1.0000		INFR	0.4030	−0.1123	1.0000
	0.2255	0.4278	-----			0.0031	0.4278	-----
USAD	0.4341	−0.4161	0.3545	1.0000	USAD	−0.0733	−0.4161	0.3545	1.0000
	0.0013	0.0022	0.0099	-----		0.6055	0.0022	0.0099	-----

Source: Own processing, 2024.

is a statistical tool that displays the correlation coefficients between all pairs of variables within a dataset, thereby providing a comprehensive view of the linear relationships among them. These coefficients range from −1 to 1, indicating, respectively, a perfect negative correlation, a perfect positive correlation, or the absence of a linear relationship. The matrix is symmetric with respect to the main diagonal, and the elements along this diagonal are always 1, as each variable is perfectly correlated with itself. Analyzing the correlation matrix allows for the identification of the strength and direction of relationships between variables, which is crucial for understanding interdependencies and for subsequent statistical modeling.

In Model M1, the analysis of the correlation between greenhouse gas emissions and other variables reveals a nuanced set of relationships characterized by a combination of moderate positive and negative associations. Specifically, the correlation between greenhouse gas emissions and affordability is moderately negative, suggesting that increased emissions are associated with reduced affordability. Conversely, the correlation between greenhouse gas emissions and user adoption is moderately positive, indicating that higher emissions correspond with higher user adoption rates. The relationship between greenhouse gas emissions and infrastructure is weak and statistically insignificant, implying that infrastructure does not have a significant linear effect on emissions within this model.

In Model M2, the emphasis is placed on renewable energy, which exhibits moderate correlations with both infrastructure and affordability, yet has a minimal impact on user adoption. Specifically, renewable energy is positively correlated with infrastructure, suggesting that better infrastructure supports increased use of renewable energy. Similarly, there is a negative correlation between renewable energy and affordability, implying that higher renewable energy usage may be associated with lower affordability levels. However, the correlation between renewable energy and user adoption is very weak and not statistically significant, indicating that renewable energy does not substantially influence user adoption rates.

The estimation results from

Table 5. GLM estimations for M1.

Dependent Variable: lnGHG
Method: Generalized linear model (Newton–Raphson/Marquardt steps)
Sample: 2010 2022
Family: Normal
Link: Identity
Dispersion computed using Pearson Chi-Square
Convergence achieved after 0 iterations
Coefficient covariance computed using observed Hessian
Variable	Coefficient	Std. Error	z-Statistic	Prob.
AFFORD	−0.029424	0.009841	−2.990033	0.0028
INFR	−0.032348	0.010679	−3.029205	0.0025
USAD	0.021947	0.006896	3.182481	0.0015
C	20.80956	1.064793	19.54329	0.0000
Mean dependent var	18.45490	S.D. dependent var		0.778523
Sum squared resid	17.99738	Root MSE		0.588306
Log-likelihood	−46.27946	Akaike info criterion		1.933825
Schwarz criterion	2.083921	Hannan–Quinn criter.		1.991368
Deviance	17.99738	Deviance statistic		0.374945
Restr. deviance	30.91103	LR statistic		34.44142
Prob (LR statistic)	0.000000	Pearson SSR		17.99738
Pearson statistic	0.374945	Dispersion		0.374945

Source: Own processing, 2024.

reveal the influence of digital transformation indicators on greenhouse gas emissions. The negative and statistically significant coefficient for digital infrastructure (INFR) (−0.0323, p-value = 0.0025) confirms that improvements in digital infrastructure contribute to a reduction in GHG emissions. This finding supports the notion that enhanced digital connectivity facilitates energy-efficient practices, smart grid optimization, and reduced reliance on carbon-intensive activities. The results align with previous studies suggesting that technological advancements can drive emissions reductions through increased efficiency in industrial processes and urban management [22,23].

Conversely, the positive and statistically significant coefficient for user adoption (0.0219, p-value = 0.0015) indicates that higher rates of digital technology adoption are associated with increased GHG emissions. This suggests that while digitalization enables efficiency gains, it may also lead to higher electricity demand, especially in regions where the energy mix is dominated by fossil fuels. This result highlights a potential trade-off between digital expansion and environmental sustainability, reinforcing the need for parallel investments in renewable energy infrastructure to offset increased energy consumption.

The negative coefficient for affordability (−0.0294, p-value = 0.0028) further supports this interpretation. As digital technologies become more affordable, their widespread adoption appears to drive energy demand, which, in turn, affects emissions levels. However, affordability may also enable the adoption of energy-efficient digital solutions, thus contributing to lower emissions in the long run. These findings suggest that policies promoting energy efficiency within the digital sector are crucial for ensuring that digital transformation leads to net-positive environmental outcomes.

In

Table 6. GLM estimations for M2.

Dependent Variable: REN
Method: Generalized linear model (Newton–Raphson/Marquardt steps)
Sample: 2010 2022
Family: Normal
Link: Identity
Dispersion computed using Pearson Chi-Square
Convergence achieved after 0 iterations
Coefficient covariance computed using observed Hessian
Variable	Coefficient	Std. Error	z-Statistic	Prob.
AFFORD	−0.215804	0.062879	−3.432035	0.0006
INFR	0.282854	0.068234	4.145334	0.0000
USAD	−0.142603	0.044065	−3.236194	0.0012
C	6.506090	6.803710	0.956256	0.3389
Mean dependent var	17.35287	S.D. dependent var		4.783833
Sum squared resid	734.8027	Root MSE		3.759098
Log-likelihood	−142.7232	Akaike info criterion		5.643201
Schwarz criterion	5.793297	Hannan–Quinn criter.		5.700744
Deviance	734.8027	Deviance statistic		15.30839
Restr. deviance	1167.138	LR statistic		28.24172
Prob (LR statistic)	0.000003	Pearson SSR		734.8027
Pearson statistic	15.30839	Dispersion		15.30839

Source: Own processing, 2024.

, the results highlight the relationship between digital transformation and the share of renewable energy in total energy consumption. The positive and statistically significant coefficient for digital infrastructure (0.2828, p-value = 0.0000) suggests that improvements in digital infrastructure directly support the integration of renewable energy sources. This aligns with theoretical arguments that digitalization enhances energy grid efficiency, improves demand forecasting, and facilitates the adoption of smart energy solutions, thereby increasing the share of renewable energy in the energy mix.

However, the negative and statistically significant coefficient for affordability (−0.2158, p-value = 0.0006) suggests that as digital services become more affordable, the immediate impact on renewable energy adoption may be limited. This could be due to the fact that lower digital costs drive broader adoption of technology, which initially increases overall energy demand before renewable sources can sufficiently meet that demand. This finding underscores the importance of aligning digital expansion with investments in clean energy infrastructure to ensure that affordability does not inadvertently lead to greater reliance on fossil fuels. Similarly, the negative coefficient for user adoption (−0.1426, p-value = 0.0012) indicates that increased digital usage does not necessarily correlate with a higher share of renewable energy. This suggests that, in many emerging economies, digital expansion is still largely dependent on conventional energy sources, reinforcing the need for targeted policies that incentivize the integration of renewables into digital ecosystems.

Taken together, the results from Table 5 and Table 6 provide empirical evidence that digital transformation has a measurable and significant impact on both GHG emissions and renewable energy adoption. While digital infrastructure improvements contribute to emissions reduction and facilitate renewable energy integration, the effects of affordability and user adoption remain complex. These findings emphasize that the environmental benefits of digitalization are not automatic and depend on the energy composition of the economy, regulatory frameworks, and parallel investments in clean energy technologies.

To ensure that digital transformation supports climate objectives, policymakers should prioritize sustainable digital infrastructure, promote energy-efficient technologies, and accelerate the transition toward renewable energy sources. The results also suggest that future research should further explore the interplay between digitalization, energy policies, and emissions reduction strategies in different economic contexts.

The results of the two models, M1 and M2, provide valuable insights into how digital infrastructure, affordability, and user adoption of digital technologies impact greenhouse gas emissions and the use of renewable energy in emerging economies. Figure 2 presents how these models respond to the research questions provided in this study.

In the case of the first research question, Q1, the results from Model M1 indicate a statistically significant negative relationship between improved digital infrastructure and greenhouse gas emissions. This suggests that enhancements in digital infrastructure contribute to a reduction in emissions, likely through more efficient energy use, improved resource management, and the integration of low-carbon technologies.

For the second research question, Q2, Model M2 results indicate a negative and statistically significant relationship between the affordability of digital technologies and the use of renewable energy. This implies that as digital technologies become more affordable, the share of renewable energy may decrease. This could suggest that affordability drives broader digital adoption that, paradoxically, increases reliance on nonrenewable energy due to the higher overall energy demand.

For Q3, in Model M1, user adoption of digital technologies is positively correlated with greenhouse gas emissions. This indicates that higher user adoption might be associated with activities that increase energy consumption and emissions, potentially due to the higher energy demands of widespread digital usage. Conversely, in Model M2, user adoption shows a negative correlation with renewable energy adoption, suggesting that increased digital usage may not align with renewable energy use, possibly due to the current energy mix in emerging economies.

Based on the results from Models M1 and M2, advanced digital infrastructure contributes to enhancing environmental sustainability in nuanced ways. Model M1 demonstrates a statistically significant negative relationship between digital infrastructure and greenhouse gas emissions. This suggests that improvements in digital infrastructure lead to a reduction in emissions, likely due to more efficient energy use, better resource management, and the integration of cleaner technologies. However, the results from Model M2 present a more complex picture. While digital infrastructure positively correlates with the adoption of renewable energy, the relationship between user adoption and renewable energy is negative. This implies that while advanced infrastructure enables the deployment of renewable energy, other factors, such as how digital technologies are used, might counterbalance these gains by increasing overall energy demand and reliance on non-renewable sources.

The results of this study align with and expand upon several theoretical perspectives on digital transformation and environmental sustainability. The findings support the technology-push and market-pull theories [56], which argue that technological advancements drive efficiency gains and structural transformations within economies. The significant negative correlation between digital infrastructure and greenhouse gas emissions (GHG) in Model M1 corroborates prior studies suggesting that improved digital infrastructure can enable energy-efficient solutions, smart grid integration, and optimized resource utilization [22,23].

At the same time, the positive relationship between user adoption and GHG emissions highlights an important deviation from the green technology paradigm, which generally posits that digitalization leads to sustainability improvements [14]. Our results suggest that digital adoption, particularly in emerging economies, may increase energy demand faster than the energy efficiency gains derived from digital technologies. This finding is in line with Jevons Paradox [57], which argues that efficiency improvements can sometimes lead to greater overall resource consumption rather than conservation. This phenomenon underscores the need for policy interventions to ensure that digital expansion is accompanied by investments in clean energy sources to offset rising energy demands.

Additionally, our findings contribute to the diffusion of innovation theory [58] by demonstrating how the different dimensions of digitalization—infrastructure, affordability, and user adoption—exert distinct influences on environmental performance. While infrastructure improvements positively impact renewable energy adoption (Model M2), affordability and user adoption show a negative correlation with the share of renewable energy. This suggests that digital expansion in emerging economies may not inherently align with sustainability goals unless guided by policies that prioritize green energy integration.

Furthermore, the study builds upon the institutional theory of digital transformation [8], which emphasizes that digitalization is shaped by socioeconomic and regulatory contexts. Our results indicate that while digital infrastructure advancements can contribute to lower emissions, the overall environmental impact of digitalization is contingent upon the regulatory and energy policy frameworks within each economy. This supports the notion that digital transformation alone is not a sufficient driver of sustainability; instead, its effectiveness depends on complementary policies that promote sustainable energy transitions.

5. Conclusions

This study employs two distinct econometric models, M1 and M2, to examine the relationship between digital transformation and environmental sustainability. The fundamental distinction between these models lies in their dependent variables and the specific environmental dimensions they assess. Model M1 investigates the impact of digitalization—measured through infrastructure (INFR), affordability (AFFORD), and user adoption (USAD)—on greenhouse gas emissions (GHG). The findings indicate that improvements in digital infrastructure and affordability contribute to emission reductions, whereas increased user adoption is positively correlated with higher emissions, likely due to heightened energy consumption.

In contrast, Model M2 evaluates the influence of digitalization on the share of renewable energy (REN) in total energy consumption. The results suggest that enhanced digital infrastructure facilitates the integration of renewable energy sources, while greater affordability and user adoption exhibit a negative association with renewable energy uptake, indicating potential challenges in aligning broader digital adoption with sustainability goals.

These findings underscore the nuanced role of digital transformation in environmental sustainability. While advancements in digital infrastructure appear to support emission reductions and the integration of renewable energy, their overall effectiveness is contingent on how affordability and user adoption influence energy demand. This distinction is critical for policymakers seeking to leverage digital transformation as a strategic tool for achieving environmental objectives in emerging economies.

The study contributes to the broader theoretical discourse on digital transformation and environmental sustainability by offering empirical insights into the nuanced effects of digitalization in emerging economies. The results align with existing literature suggesting that digital infrastructure development can play a pivotal role in enhancing sustainability outcomes by improving resource efficiency, facilitating renewable energy integration, and reducing greenhouse gas emissions. However, the study also highlights the dual impact of digitalization, demonstrating that while infrastructure and affordability can drive positive environmental change, increased digital adoption may lead to higher energy consumption, thereby partially offsetting these benefits.

From a theoretical perspective, these findings suggest a need to re-evaluate existing digital transformation models in the context of sustainability. While prior studies have often emphasized the role of digitalization in fostering economic growth and efficiency, this study underscores its complex relationship with environmental performance. The results indicate that digital transformation cannot be viewed solely as a linear enabler of sustainability; rather, its impact depends on how different dimensions—infrastructure, affordability, and user adoption—interact with energy systems and economic behaviors.

Furthermore, the study contributes to technology adoption and environmental transition theories by illustrating how digitalization dynamics vary across different economic contexts. In emerging economies, where energy infrastructures are still evolving, the expansion of digital adoption without parallel investments in green energy sources may inadvertently exacerbate environmental pressures. This finding suggests that digitalization should be analyzed within an integrated framework that considers both its direct efficiency-enhancing effects and its indirect demand-driven consequences on energy use.

These insights provide a theoretical foundation for future research exploring the interplay between digitalization, economic structures, and sustainability policies. By identifying the mechanisms through which digital transformation influences environmental outcomes, this study offers a framework that can be adapted and tested in other regional contexts, contributing to the broader discourse on digital sustainability and green economic transitions.

Policymakers should ensure that the expansion of digital infrastructure is accompanied by strong environmental regulations. This includes promoting energy-efficient technologies and ensuring that digital expansion does not lead to increased overall energy consumption. Governments should focus on making green digital technologies more affordable. Subsidies or tax incentives could be provided for energy-efficient devices and renewable energy technologies to ensure that affordability contributes positively to both digital adoption and environmental outcomes. Educational campaigns and incentives could encourage users to adopt sustainable digital practices, such as using energy-efficient devices and minimizing unnecessary digital consumption. Policymakers could also introduce regulations to guide companies in developing more sustainable digital products and services. Continuous monitoring of the environmental impact of digital growth is necessary. Governments should be prepared to adjust policies based on observed outcomes, ensuring that the positive impacts of digital infrastructure improvements are not negated by increased energy use due to higher user adoption. To align the expansion of digital infrastructure with sustainability goals, investments in renewable energy integration with digital technologies should be prioritized. This could involve upgrading power grids to handle more renewable energy and encouraging the development of digital solutions that facilitate renewable energy use.