Systemic Interactions Among Digital Transformation, Sustainable Orientation, and Economic Outcomes in EU Countries

Abstract

Digital transformation and sustainable orientation have become key drivers of economic development within the European Union. This study investigates how progress in digitalization and sustainable orientation influences economic outcomes. To address this objective, we apply a combination of techniques, including factor analysis to reduce dimensionality and identify underlying structures, generalized linear models to estimate causal connections and cluster analysis to group countries with similar profiles. The findings highlight strong complementarities between digital transformation and sustainable development in nurturing higher levels of economic outcome, with digital readiness amplifying the effects of sustainable development practices. Moreover, cluster analysis methods reveal significant asymmetries among EU countries, underlining persistent regional disparities in the pace of digital and sustainable transitions. The study concludes that a systems-based approach to managing the twin transition is essential for promoting convergence, competitiveness, and resilience in the EU economic system.

1. Introduction

Today, digital transformation (DT) and sustainable orientation (SO) are two central pillars of Europe’s economic and political agenda. Increasingly, policymakers and scholars see them as closely connected processes that influence both economic outcomes (EO) and regional convergence in complex ways. Treating them as separate issues would hide their tightly linked dynamics: digitalization simultaneously promotes innovation while increasing environmental pressures, whereas SO requires a reevaluation of resource use amid a time when new technologies are changing production and consumption patterns [1]. Early studies introduced the idea of “digital sustainability,” focusing on integrating social and ecological concerns into the design and use of digital infrastructures [2,3]. More recent research shows not only the potential of digital technologies to support the green transition but also the risks of worsening inequalities or increasing dependence on non-renewable resources [1,4,5].

This ambivalence becomes especially evident in the European Union, where digital and sustainability agendas merge under the framework of the so-called “twin transition.” However, the literature highlights a series of structural paradoxes. On one hand, DT creates opportunities for new sustainable business models and enhances corporate value when paired with firm environmental commitments [6,7]. On the other hand, efficiency improvements often lead to rebound effects that reduce or even negate ecological benefits [8,9]. Comparative analyses also show that EU member countries display highly uneven levels of digital readiness and SO, which in turn create significant disparities in how these processes produce tangible results [10].

The current study examines the systemic relationship among digital transformation (DT), sustainable orientation (SO), and economic outcomes (EO) across European Union member states. While previous research has looked at the links between the Digital Economy and Society Index (DESI) and the Sustainable Development Goals Index (SDGi), this study pushes the discussion further by analyzing how these two aspects interact to influence economic performance, measured by GDP per capita.

The analysis employs a multidimensional mixed-methods framework that combines factor analysis to identify underlying structures, generalized linear models to estimate causal effects, and cluster analysis to uncover patterns of convergence and divergence among member states. This approach goes beyond traditional correlation-based studies by capturing both the structural and dynamic relationships that define the European twin transition. The study’s originality lies in conceptualizing this transition not just as the intersection of two policy areas but as a deeply interconnected process that affects competitiveness, resilience, and long-term economic sustainability within the EU. Therefore, the paper addresses two main research questions: What is the nature of the relationship between DT, SO, and EO in the European Union? Also, how can member states be grouped based on their digital, sustainable, and economic profiles to identify patterns of convergence and divergence?

The structure of the article follows this order: the introduction sets the context and highlights the importance of the topic, then it moves to the literature review and hypothesis development, a detailed presentation of materials and methods, the results, a discussion, and finally the conclusions.

2. Theoretical Background and Hypotheses

2.1. DT, SO and EO in the European Union

The interaction between digital transformation (DT), sustainable orientation (SO), and economic outcomes (EO) is a rapidly changing area in the European Union, reflecting the concept of the “twin transition” that combines digitalization and sustainability as overlapping strategic goals. Research increasingly shows that DT and SO are not separate phenomena but interconnected processes whose interaction impacts growth, innovation, and social cohesion in complex ways [2,3]. The idea of “digital sustainability” has emerged from this dialog, emphasizing the conditions where digitalization can promote environmental and social goals at the same time.

Empirical studies demonstrate that firms combining digital and ecological strategies can turn social and regulatory pressures into opportunities for innovation [11,12,13]. This dual focus has led to notable impacts in industries such as agriculture [14,15] and healthcare [16,17], and supports the development of digital tools—like digital twins and smart cities—as drivers of sustainable urban growth [18,19,20]. Still, the literature also highlights risks tied to these technologies, especially the chance of increasing social and technological divides [21].

Despite progress, the academic debate remains divided. Alcott [8] and York [9] revisit classical paradoxes—such as the rebound effects of efficiency improvements and the ongoing level of material consumption—showing that technological progress does not automatically result in decreased resource use. Related arguments by Berkhout and Hertin [22] and Andersen et al. [23] support that digitalization can both reduce and increase materialization in production, emphasizing the need for integrated analyses of DT and SO rather than treating them separately, which could lead to misleading conclusions [24].

In economic terms, studies examining the connection between digitalization and productivity show mixed results. While Cardona et al. [25] and Brynjolfsson [26] describe the “productivity paradox,” later work by Brynjolfsson et al. [27] expands on it to include artificial intelligence, highlighting the gap between technological investments and measurable benefits. Similar uncertainty exists in the literature on circular economy models [28,29], where DT both promotes innovation and creates new vulnerabilities.

From a societal standpoint, digital progress raises issues of equity and inclusion. DiMaggio and Hargittai [30] introduced the concept of “digital inequality,” and Klein and Kleinman [31] showed that institutional and cultural settings influence technological outcomes. Recent studies warn that, without redistribution mechanisms, DT could worsen inequalities and threaten social justice [32,33,34].

In supply chain management and industrial organization, the literature emphasizes both opportunities and risks. Studies by Bag et al. [35] and Yadav et al. [36] show how big data and Industry 4.0 solutions improve efficiency but also create challenges for equity and labor conditions. Research on blockchain-based circular economy systems [37,38] further demonstrates that technological results heavily depend on governance frameworks and policy decisions.

The overall consensus is that DT is not neutral. Its contribution to sustainability and economic performance depends on strategic coordination, institutional quality, and governance capacity [39,40,41,42,43,44,45]. Comparative research confirms that DT’s effects vary widely across EU member states, influenced by differences in digital maturity, industrial structure, and institutional resilience [46,47,48]. Methodological advances—such as bibliometric and scientometric analyses [49,50]—help map these dynamics and emphasize the contextual nature of the DT–SO–EO nexus [51,52].

Recent studies on EO show partial convergence between digital and sustainable dimensions. Magoutas et al. [53] and Hristoski and Kostoska [54] find that ICT investments boost GDP growth, while Harangozó and Fakó [10] demonstrate correlations between the Digital Economy and Society Index (DESI) and the Sustainable Development Goals Index (SDGi), revealing uneven but progressing alignment. Nevertheless, regional disparities persist, indicating that structural and institutional barriers still hinder full convergence [55].

Building on this complex relationship between DT, SO, and EO, there is a need for empirical testing of their complementarity within the EU.

Hypothesis H1.

DT and SO together improve the EO of EU countries, as measured by GDP per capita.

2.2. Cluster Analysis and Digital–Sustainable–Economic Profiles in the European Union

Beyond their direct connection to EO, DT, and SO reveal deep structural asymmetries among EU member countries. These differences, which go beyond GDP or access to technology, are rooted in variations in human capital, governance quality, and institutional capacity [55,56]. Comparative studies consistently show that Nordic nations tend to incorporate digitalization and sustainability into cohesive growth models, while Central and Eastern European countries often display scattered progress and ongoing gaps [55]. Such differences highlight the need for analytical tools—like cluster analysis—that can capture the multidimensional nature of these disparities and identify patterns of convergence and divergence across the Union.

Research highlights that the paths of DT and SO are influenced not only by infrastructure but also by social inclusion and labor dynamics. While digital transformation can improve civic participation and transparency [57,58], it may also increase inequalities and lead to job displacement as automation and robotics reshape employment structures [59,60]. These tensions also affect energy transitions, where DT supports efficiency, monitoring, and renewable integration but benefits are unevenly distributed across regions [61].

Urban and regional settings further illustrate this uneven progress. Smart city initiatives show how digital technologies can boost sustainable development [62,63,64], yet their spread remains geographically uneven, reflecting differences in institutional readiness and policy coherence. Likewise, responses to external shocks—such as the COVID-19 pandemic—demonstrate that digital tools can improve governance resilience [65,66], although their effectiveness mainly depends on leadership and organizational culture [67,68,69].

Environmental and resource-related challenges add another layer to this analysis. Studies indicate that digital infrastructures lead to increasing energy costs [70,71] and expanding carbon footprints, especially from blockchain and data center growth [72,73,74,75]. Additionally, the spread of IoT systems and digital agriculture enhances environmental pressures [76,77], and the increased use of rare resources and electronic waste worsens global inequalities [5,78,79]. These trends divide Europe between countries with strong governance and adaptive capacity and those facing systemic vulnerabilities.

At the policy level, integrating circular economy principles into digital agendas becomes a key driver of convergence. Countries that align digital transformation with circularity show greater resilience and stronger economic and social performance [80,81]. Similarly, governance quality and business models influence how well digital and sustainable strategies promote competitiveness and innovation [82,83,84,85]. Sectoral studies, especially in healthcare and manufacturing, indicate that while DT improves efficiency and adaptability [86,87,88], its success depends on ethical, institutional, and cultural factors.

The literature on organizational culture and SMEs highlights the importance of intangible factors in explaining national differences. Organizational values that promote innovation and environmental awareness improve the effectiveness of DT and SO [89,90,91,92,93], while SMEs—despite resource limitations—act as vital catalysts for experimentation and adaptation within local ecosystems [94,95,96,97]. These human and cultural aspects reinforce the case for using cluster analysis, which can detect subtle variations that traditional economic indicators miss.

Finally, research on innovation ecosystems reinforces the idea that DT and SO develop as interconnected systems influenced by governance, technology, and social goals [98,99,100,101,102,103]. Clusters, therefore, are not just statistical models but representations of the various configurations through which European countries pursue the twin transition.

Building on these insights, the second research hypothesis (H2) suggests that EU member states can be grouped into distinct clusters based on their levels of DT, SO, and EO. These clusters show both convergence and ongoing asymmetries in how digital and sustainable transformations develop across Europe.

Considering these debates on structural disparities among European countries and the need to employ clustering methods to identify convergence and divergence patterns, the study introduces a second hypothesis:

Hypothesis H2.

EU member states can be grouped into distinct clusters based on their levels of DT, SO, and EO. These clusters highlight not only regional convergences but also persistent asymmetries in the development of the twin transition.

3. Materials and Methods

3.1. Research Design

The conceptual and empirical framework of this study is based on the idea that digital transformation (DT) and sustainable orientation (SO), although often studied separately, have combined systemic effects on economic outcomes (EO) when considered together. In the European context—marked by structural diversity and ongoing regional disparities—the research aims not only to identify direct relationships among variables but also to explore the broader patterns that develop as DT and SO interact dynamically.

The research design adopts a transnational comparative perspective focusing on EU member states from 2017 to 2024. This timeframe was selected to ensure the availability of harmonized data and to capture a period of intensified European policy efforts related to the digital and green transitions. These years also include major external shocks, such as the COVID-19 pandemic and the energy crisis, which tested the resilience and adaptability of national systems.

The selected indicators—primarily drawn from the Digital Economy and Society Index (DESI), the Sustainable Development Goals Index (SDGi), and Eurostat’s macroeconomic datasets—were chosen for their conceptual relevance, data reliability, and cross-country comparability. Together, they offer a consistent empirical foundation for assessing how digital readiness, sustainable development, and economic performance interact on a systemic level. While this set of indicators highlights key aspects of the twin transition, it naturally simplifies a complex reality, as factors like governance quality, innovation ecosystems, or institutional culture remain outside the model’s direct scope.

It is also important to recognize the limitations inherent in the cross-sectional nature of the dataset. While this design allows for a comparative snapshot of EU countries and reveals structural patterns of convergence or divergence, it does not fully capture the temporal dynamics or causal feedback mechanisms that develop over longer periods. As a result, the findings should be viewed as indicative of current structural relationships rather than conclusive evidence of long-term causality. Despite these limitations, the combination of methods and indicators offers a coherent and empirically grounded framework for analyzing how digital and sustainable transitions together influence economic outcomes across Europe.

To address the research objectives, the study employs a three-stage model. The first stage uses factor analysis to simplify data and identify underlying structures that influence the relationships between DT and SO variables. The second stage applies generalized linear models (GLM) to estimate causal links between digital and sustainable progress, on the one hand, and EO on the other. The third stage involves cluster analysis, grouping EU countries into similar categories with comparable DT-SO-EO profiles. This methodological approach improves the robustness of the research and allows for a detailed interpretation of the studied phenomenon.

Figure 1 presents the conceptual model developed by the authors.

3.2. Selected Variables

The selection of variables depends on both their theoretical relevance and the availability of harmonized European data. Primary sources include indicators from the European Commission [104] via the Digital Economy and Society Index (DESI), SO measures published by Sachs et al. [105] in the Sustainable Development Goals Index (SDGi), real GDP per capita (GDPc), and the number of enterprises adopting at least one artificial intelligence technology, as reported by Eurostat [106,107].

Digital variables include Connectivity (CON), Digital Public Services (DPS), Human Capital (HC), and Integration of Digital Technology (IDT), representing the main dimensions of DESI and measured as standardized composite scores across the EU. Additionally, we introduce an emerging indicator, AIT, which measures the percentage of enterprises adopting at least one AI technology. Including this variable emphasizes the growing importance of AI as both a driver of digital transformation and a differentiator among European economies. It plays a key role in the cluster analysis that incorporates AI, GDP, and SDGi data for 2024.

On the SO dimension, the main variable is SDGi, a composite score that merges national progress across the SDGs. Economic success is evaluated by GDP per capita, shown as real spending per person and adjusted for purchasing power standards (PPS_EU27_2020=100). This approach guarantees comparability between countries while avoiding distortions from price differences or economic structural disparities.

Table 1. Variables used, measures, and sources.

Variable	Data	Measures	Sources
CON	Connectivity	Score	[104]
DPS	Digital public services	Score	[104]
HC	Human capital	Score	[104]
IDT	Integration of digital technology	Score	[104]
SDGi	SDG Index	Score	[105]
GDPc	Gross domestic product	Gross domestic product—Volume indices of real expenditure per capita (in PPS_EU27_2020=100)	[106]
AIT	Enterprises use at least one of the AI technologies	Percentage of enterprises	[107]

Source: author’s design based on European Commission [104], Sachs et al. [105], and Eurostat [106,107].

presents the variables, measures, and data sources.

The combination of these indicators builds a strong foundation for analyzing the interaction between DT, SO, and EO, while also providing a solid basis for testing the proposed hypotheses.

3.3. Methods

The methodological approach combines two complementary tools for testing hypothesis H1: factor analysis and generalized linear models (GLM). This combination addresses the limitations of each method individually and offers a more comprehensive understanding of the phenomenon.

Factor analysis serves as the initial step, reducing the dataset’s dimensionality and identifying latent structures [108] that influence the relationship between digital and SO variables. The general formula for factor analysis can be expressed as (1):

$$X=LF+\in$$

(1)

$X$—observed variables (GDPc, CON, DPS, HC, IDT, SDGi).

$L$—matrix of factor loadings.

$F$—latent factors.

$\in$—errors.

This technique allows the identification of common constructs, such as a latent factor of “digital readiness” or “institutional sustainability,” while minimizing risks of redundancy and multicollinearity.

The second step uses GLM to estimate causal relationships between variables. GLM builds on traditional linear regression by incorporating link functions that capture more complex interactions between the independent variables and the dependent variable [109]. The general formula is (2):

$$Y={\beta }_0+{\beta }_1{X}_1+{\beta }_2{X}_2+\cdots +{\beta }_n{X}_n+\in$$

(2)

$Y$—dependent variable (GDPc);

$X_1,X_2,\dots ,X_n$—independent variables (CON, DPS, HC, IDT, SDGi);

${\beta }_0$—intercept;

${\beta }_1,{\beta }_2,\dots ,{\beta }_n$—regression coefficients;

$\in$—error.

GLM thus identifies causal relationships and assesses how strongly digital and sustainable progress impact EO.

Along with these methods, the study utilizes cluster analysis to investigate hypothesis H2, grouping member states based on their DT, SO, and EO profiles. The clustering process follows Ward’s method [110,111], as described by the function (3):

$$\Delta \left(A,B\right)=\sum _{i\in A\cup B}{\left|\left|{\overrightarrow{x}}_i-{\overrightarrow{m}}_{A\cup B}\right|\right|}^2-\sum _{i\in A}{\left|\left|{\overrightarrow{x}}_i-{\overrightarrow{m}}_A\right|\right|}^2-\sum _{i\in B}{\left|\left|{\overrightarrow{x}}_i-{\overrightarrow{m}}_B\right|\right|}^2={\displaystyle \frac{n_A{n}_B}{n_A{+n}_B}}{\left|\right|{\overrightarrow{m}}_A-{\overrightarrow{m}}_B\left|\right|}^2$$

(3)

${\overrightarrow{m}}_{\mathrm{j}}$—the center of cluster j.

$n_{\mathrm{j}}$—number of points in cluster j.

Δ—merging cost of combining the clusters A and B.

i—cases.

The results of this analysis reveal patterns of convergence and divergence within the European Union, showing groups of states with either similar or different paths.

This methodological approach offers a comprehensive understanding of the systemic interactions linking DT, SO, and EO. By combining traditional statistical methods with machine learning, the analysis goes beyond just quantifying effects, providing a systemic view of the mechanisms that drive convergence and divergence in the European space.

4. Results

The factor analysis conducted in this study demonstrates the robustness of the dataset and the relevance of the variables chosen to examine the relationship between DT, SO, and EO. The correlation matrix highlights significant interdependencies among digital variables (CON, DPS, HC, and the IDT). These indicators show strong correlations both with each other and with GDP per capita (

Table 2. Correlation Matrix.

		CON	DPS	HC	IDT	SDGi	GDPc
Correlation	CON	1.000	0.588	0.379	0.559	0.315	0.243
	DPS	0.588	1.000	0.759	0.749	0.378	0.438
	HC	0.379	0.759	1.000	0.797	0.570	0.577
	IDT	0.559	0.749	0.797	1.000	0.460	0.418
	SDGi	0.315	0.378	0.570	0.460	1.000	0.129
	GDPc	0.243	0.438	0.577	0.418	0.129	1.000
Sig. (1-tailed)	CON		0.000	0.000	0.000	0.000	0.001
	DPS	0.000		0.000	0.000	0.000	0.000
	HC	0.000	0.000		0.000	0.000	0.000
	IDT	0.000	0.000	0.000		0.000	0.000
	SDGi	0.000	0.000	0.000	0.000		0.051
	GDPc	0.001	0.000	0.000	0.000	0.051

Source: author’s design with SPSS v.27.0.

).

The strongest link is between HC and the IDT (r = 0.797, p < 0.001). This shows that building digital skills helps to use emerging technologies. At the same time, positive links between GDP and digital indicators—especially HC and DPS—confirm that digital growth directly affects EO, although the strength of these links varies. The sustainability index is the only area with a weaker link to GDP (r = 0.129, p = 0.051), suggesting a more complex relationship likely influenced by other factors.

The application of factor analysis is justified by the KMO value of 0.708, which exceeds the minimum threshold, and Bartlett’s test, which indicates high statistical significance (χ² = 562.192, p < 0.001). Together, these factors confirm that the dataset is suitable for uncovering underlying structures and reducing dimensionality.

The results on communalities (

Table 3. Communalities and factor matrix.

	Initial	Extraction	Factor 1
CON	0.467	0.322	0.568
DPS	0.703	0.733	0.856
HC	0.829	0.836	0.914
IDT	0.727	0.786	0.887
SDGi	0.427	0.261	0.511
GDPc	0.410	0.253	0.503

Source: author’s design with SPSS v.27.0.

) show that digital variables, especially HC and digital technology integration, are highly explained by the extracted factor, with values above 0.8. In contrast, SDGi and GDP have lower communalities (0.261 and 0.253, respectively).

This pattern indicates that the latent factor primarily captures the digital dimension of progress, while SO only partially fits into this model. The factor loadings support this interpretation (Table 3): digital variables load very strongly (from 0.856 for DPS to 0.914 for HC), whereas GDP and the SDGi show more moderate loadings (0.503 and 0.511). This distribution suggests that EO and SO are influenced indirectly, with digital progress being the main driver within the European Union.

The table on total variance explained shows that one factor accounts for 53.19% of the total variance (

Table 4. Total variance explained.

Factor	Initial Eigenvalues			Extraction Sums of Squared Loadings
	Total	% of Variance	Cumulative %	Total	% of Variance	Cumulative %
1	3.553	59.214	59.214	3.191	53.190	53.190
2	0.888	14.802	74.015
3	0.758	12.638	86.653
4	0.448	7.466	94.119
5	0.243	4.052	98.171
6	0.110	1.829	100.000

Source: author’s design using SPSS v.27.0.

).

Although this percentage exceeds the minimum threshold for interpretability, it also shows that nearly half of the variance remains unexplained. This indicates that the phenomenon is complex and involves additional factors not captured by the model. The single factor can be viewed as a latent dimension of “digital and institutional readiness,” where SO and EO are present, though in a more subtle form.

Regarding Hypothesis H1, the results support the idea of complementarity between DT and SO in explaining EO, while also highlighting differences in their levels of influence. Digitalization emerges as the most important and influential factor, while SO plays a supporting role, somewhat embedded within the core structure. However, the strong correlations between GDP per capita and both digital indicators, and to a lesser extent, the sustainability index, confirm that digital and sustainable development are interconnected. Digitalization, though, has a more immediate impact. Therefore, H1 is supported, with the finding that DT is currently the main driver, while SO mainly contributes through indirect and mediated mechanisms. This aligns with the European context, where the twin transition does not happen in a completely symmetrical manner but shows complementarity as a clear way to enhance EO.

The generalized linear model further illustrates how DT and SO dimensions influence progress toward the SDGs, as well as EO, which is measured through GDP per capita. The tests of between-subject effects show that the models explain a significant amount of the variance in the dependent variables, with R² values of 0.362 for SDGi and 0.342 for GDP per capita. These findings demonstrate strong explanatory power, especially given the complexity of cross-national data (

Table 5. Tests of between-subjects effects.

Source	Dependent Variable	Type III Sum of Squares	df	Mean Square	F	Sig.	Partial Eta Squared	Noncent. Parameter	Observed Power
Corrected Model	SDGi	612.463	4	153.116	22.285	0.000	0.362	89.140	1.000
	GDPc	102,416.879	4	25,604.220	20.416	0.000	0.342	81.663	1.000
Intercept	SDGi	23,769.319	1	23,769.319	3459.491	0.000	0.957	3459.491	1.000
	GDPc	4263.068	1	4263.068	3.399	0.067	0.021	3.399	0.449
CON	SDGi	47.532	1	47.532	6.918	0.009	0.042	6.918	0.743
	GDPc	1080.131	1	1080.131	0.861	0.355	0.005	0.861	0.152
DPS	SDGi	39.211	1	39.211	5.707	0.018	0.035	5.707	0.661
	GDPc	0.328	1	0.328	0.000	0.987	0.000	0.000	0.050
HC	SDGi	245.293	1	245.293	35.701	0.000	0.185	35.701	1.000
	GDPc	38,788.480	1	38,788.480	30.928	0.000	0.165	30.928	1.000
IDT	SDGi	0.591	1	0.591	0.086	0.770	0.001	0.086	0.060
	GDPc	2465.993	1	2465.993	1.966	0.163	0.012	1.966	0.286
Error	SDGi	1078.709	157	6.871
	GDPc	196,899.566	157	1254.137
Total	SDGi	1,030,331.290	162
	GDPc	2,001,944.000	162
Corrected Total	SDGi	1691.171	161
	GDPc	299,316.444	161

Source: author’s design with SPSS v.27.0.

).

For SDGi, HC is the most influential contributor, with a substantial and statistically significant effect (F = 35.701, p < 0.001). This indicates that digitally skilled and technologically adaptable human resources are key to advancing sustainable development goals. Connectivity (F = 6.918, p = 0.009) and DPS (F = 5.707, p = 0.018) also have significant, though less pronounced, effects. Meanwhile, digital technology integration shows no notable impact on SDGi.

The prominence of human capital (HC) as the only statistically significant variable in explaining both GDP per capita and SDGi outcomes highlights the mediating role of skills and digital literacy in turning technological potential into real socio-economic progress. This finding suggests that investments in education and workforce adaptability are essential for achieving the full benefits of the twin transition.

Interestingly, DPS has a negative coefficient for SDGi (B = −0.212, p = 0.018), suggesting potential tensions between administrative DT and SO growth, possibly caused by a focus on efficiency rather than social or environmental concerns (

Table 6. Parameter Estimates.

Dependent Variable	Parameter	B	Std. Error	t	Sig.	95% Confidence Interval		Partial Eta Squared	Noncent. Parameter	Observed Power
						Lower Bound	Upper Bound
SDGi	Intercept	69.589	1.183	58.817	0.000	67.252	71.926	0.957	58.817	1.000
	CON	0.225	0.085	2.630	0.009	0.056	0.394	0.042	2.630	0.743
	DPS	−0.212	0.089	−2.389	0.018	−0.388	−0.037	0.035	2.389	0.661
	HC	0.994	0.166	5.975	0.000	0.666	1.323	0.185	5.975	1.000
	IDT	−0.044	0.151	−0.293	0.770	−0.342	0.253	0.001	0.293	0.060
GDPc	Intercept	−29.471	15.985	−1.844	0.067	−61.044	2.102	0.021	1.844	0.449
	CON	1.072	1.155	0.928	0.355	−1.210	3.353	0.005	0.928	0.152
	DPS	0.019	1.200	0.016	0.987	−2.350	2.389	0.000	0.016	0.050
	HC	12.502	2.248	5.561	0.000	8.062	16.943	0.165	5.561	1.000
	IDT	−2.853	2.035	−1.402	0.163	−6.872	1.166	0.012	1.402	0.286

Source: author’s design with SPSS v.27.0.

).

Regarding the negative coefficient of DPS in relation to the sustainability index, the result should be interpreted with caution. One possible explanation is the efficiency-focused nature of administrative digitalization, which may initially emphasize cost savings and process improvements over inclusive or environmentally sustainable outcomes. However, this relationship might also reflect data artifacts or transitional effects common in cross-country comparisons, where rapid digitalization can temporarily coincide with lower sustainability scores due to structural or institutional mismatches. To address these limitations, we recommend that future research verify this finding through longitudinal studies or by including governance quality and institutional variables that could influence this relationship.

Turning to GDPc, the model confirms the decisive role of HC (F = 30.928, p < 0.001; B = 12.502). This result emphasizes that investments in digital skills and education directly contribute to increases in GDPc. In contrast, CON, DPS, and IDT show no significant direct effects on GDPc. This outcome indicates that merely having digital infrastructure or electronic services does not guarantee immediate economic benefits unless supported by a capable HC. The non-significant intercept for GDPc reinforces the idea that EO relies on the explanatory variables in the model, especially HC.

Taken together, these findings clarify that H1 is validated but in a nuanced way. DT and SO progress both contribute to EO, although the strength of this support varies across different dimensions. Human capital serves as the connecting element, driving both SO and EO. Connectivity and DPS play a larger role in promoting SO, although their economic effects remain indirect. Digital technology integration, however, has not yet made a significant impact, possibly due to uneven levels of digital maturity and implementation capacity across EU member states.

The GLM thus confirms H1, while emphasizing that DT mainly supports both SO and EO through HC development. This finding highlights the critical role of education and training as key areas for achieving a genuine twin transition in Europe. The factor analysis results showed a stronger clustering of digital indicators compared to sustainability and GDP variables, suggesting that the analytical model more robustly captures the digital dimension of transformation. This pattern indicates that, while digitalization serves as a primary structural driver, progress related to sustainability may follow a more indirect and context-dependent path. Accordingly, the complementarity between DT and SO observed in the model should be seen as partial and evolving rather than fully symmetrical.

Cluster analysis based on variables related to artificial intelligence use in enterprises (AIT), the sustainability index (SDGi), and GDP per capita (GDPc) reveals significant structural differences among EU member states. The results confirm the presence of patterns of convergence and divergence in the twin transition process, forming three main groups, each with its own internal logic and implications for European dynamics (Figure 2 and

Table A1. Cluster data.

Country	AIT	SDGi	GDPc
France	9.91	73.9	99
Italy	8.20	72.3	98
Czechia	11.26	73.7	91
Spain	11.31	71.2	92
Lithuania	8.76	67.8	88
Cyprus	7.90	62.7	95
Finland	24.37	81.1	103
Slovenia	20.89	73.8	91
Cluster 1 mean	12.83	72.06	94.63
Latvia	8.83	70.6	71
Slovakia	10.78	70.6	75
Greece	9.81	66.5	70
Bulgaria	6.47	62.9	66
Croatia	11.76	72.4	77
Estonia	13.89	71.7	79
Poland	5.13	73.3	79
Portugal	8.63	70.6	82
Hungary	7.41	68.8	77
Romania	3.07	64.2	78
Cluster 2 mean	8.58	69.15	75.40
Denmark	27.58	79.7	128
Netherlands	23.06	71.9	136
Austria	20.27	77.3	116
Sweden	25.09	79.4	113
Germany	19.75	75.0	115
Belgium	24.71	72.2	116
Malta	17.30	69.3	109
Cluster 3 mean	22.54	74.97	119.00
Ireland	14.90	71.5	211
Luxembourg	23.73	67.6	242
Eu mean	14.28	71.55	103.59

Source: author’s design with SPSS v.27.0.

in Appendix A).

The initial group includes countries with moderate levels of digital development, as measured by the percentage of enterprises adopting AI technologies, and shows values close to the EU average for both the SDGi and GDP per capita. France, Italy, Spain, the Czech Republic, Lithuania, Cyprus, and Slovenia are part of this group, with averages of 12.83% AIT, 72.06 SDGi, and a GDP per capita of 94.63. These countries hold a balanced position where digital and sustainable sectors are fairly developed but have yet to drive significant economic growth. Their situation indicates a phase of institutional consolidation, where digital and sustainability strategies are somewhat mature but often fragmented across different administrative levels. Policy implementation is often hindered by bureaucratic inertia or inconsistent governance, which slows down turning technological capacity into broad economic progress.

The second cluster, mainly made up of Central and Eastern European countries such as Romania, Bulgaria, Hungary, Poland, and Greece, exhibits a distinct structural profile. With an average AI adoption rate of 8.58%, an SDGi of 69.15, and GDP per capita of 75.40—substantially below the EU average—these nations face persistent institutional challenges that slow down the twin transition. Although they benefit from substantial EU structural funds and ongoing digitalization initiatives, weak administrative capacity, limited absorptive ability, and inconsistent regulatory environments often weaken the impact of policy measures. However, countries like Estonia and Croatia, which are closer to the EU average, show that institutional learning and targeted reforms can gradually reduce these gaps, indicating that governance improvements are crucial in shaping national progress.

The third group includes high-performing countries such as Denmark, Sweden, Germany, Austria, Belgium, the Netherlands, and Malta. These economies have an average AI adoption rate of 22.54%, an SDGi of 74.97, and a GDP per capita of 119, which are all well above the EU average. Their success comes from strong institutional frameworks characterized by transparent governance, policy continuity, and effective collaboration between the public and private sectors. These countries have built integrated systems where digitalization policies are closely aligned with sustainability goals, creating synergies that boost innovation, competitiveness, and resilience. Ireland and Luxembourg, with outstanding GDP levels (211 and 242, respectively), are outliers that further show how institutional coherence and long-term strategic planning can enhance the effects of the twin transition.

This distribution confirms Hypothesis H2, demonstrating that EU member states can be grouped into distinct clusters based on their levels of digital transformation, sustainability orientation, and economic results (Figure 3).

The analysis shows that structural differences are rooted not only in economic capacity but also in institutional quality, policy coordination, and governance effectiveness. Northern and Western European countries, backed by stable institutions and strategic policy alignment, lead the process, while Central and Eastern regions continue to face institutional barriers that hinder convergence. Therefore, the twin transition progresses unevenly, influenced by both institutional maturity and technological progress. The complementarity between DT and SO exists across all clusters, but its advantages are greater where governance systems are coherent and long-term planning frameworks are integrated into national strategies.

5. Discussion

The analysis in this study offers a detailed view of the relationship between DT, SO, and EO in the European Union. Overall, the findings support both Hypothesis H1—that progress in DT and SO together helps boost GDP per capita—and Hypothesis H2, which groups European countries based on their digital, sustainable, and economic profiles. However, the observed complementarities are not consistent across all regions, as structural differences create both points of convergence and ongoing disparities.

The factor analysis indicates that the digital dimension—represented by HC, DPS, CON, and technology integration—constitutes the core of the latent factor. SO and EO are only moderately connected to this factor, suggesting that DT is the main driver of the twin transition, while SO plays a more indirect role. Practically, this implies that member states investing in digital skills, infrastructure, and technologies not only realize direct economic benefits but also build a stronger foundation for achieving sustainable development goals. Similar findings were reported by Magoutas et al. [53], who found that digital investments directly enhance GDP growth, while their link to SO is more indirect and relies on supportive public policies. The fact that both GDP per capita and the sustainability index load onto the same latent factor, even if only moderately, echoes Harangozó and Fakó’s [10] argument that there is a partial convergence between digital maturity and sustainable progress.

The generalized linear models improve this perspective by clarifying how complementarity develops. Human capital emerges as the most important factor for both SO and EO. Its high coefficient and strong statistical significance confirm that technological progress has little impact without people who can adopt and utilize it. This finding aligns with Brynjolfsson’s [26] concept of the “productivity paradox,” which states that significant investments in technology do not automatically lead to economic gains unless combined with organizational changes and skilled HC. Similarly, Cardona et al. [25] and Brynjolfsson et al. [27] showed that the economic potential of digital technologies only becomes real when paired with investments in skills and institutions.

A particularly notable result from the GLM is the negative impact of DPS on the sustainability index. Although it may seem counterintuitive at first, this result could stem from administrative DT that emphasizes efficiency and cost-cutting while unintentionally causing social or environmental externalities that hinder sustainable development. Researchers like Alcott [8] and York [9] have already warned about this conflict, showing how technological advancements can lead to rebound effects that cancel out ecological gains. Similarly, Beier et al. [24] argued that DT should be analyzed alongside SO, as examining them separately risks oversimplified conclusions. This finding supports those concerns and suggests that e-government policies need careful adjustment to avoid replacing social and ecological aspects with a purely technocratic view of progress. As discussed earlier, this unexpected negative correlation between DPS and SO may reflect transitional institutional dynamics, where rapid digitalization of administration initially prioritizes efficiency and procedural modernization over broader social and environmental goals. Over time, as governance frameworks mature and digital reforms become more inclusive, this relationship may shift toward better alignment between technological progress and sustainability outcomes.

Hypothesis H1 is therefore supported, but in a nuanced way. Digitalization and SO do contribute complementarily to GDP per capita growth, but this mainly occurs through HC rather than uniformly across all digital sectors. For example, technology integration did not show a significant effect in the model, which might indicate that these technologies are still in early stages or that their impacts only become evident after some delay.

The cluster analysis offers further insights, confirming H2 and revealing the structural polarizations within the European Union. The three identified groups show different paths of the twin transition. The cluster of Nordic and Western states, characterized by high AI adoption, strong SO scores, and superior EO, represents the ideal model of complementarity. In these countries, DT and SO not only coexist but also reinforce each other, creating significant competitive advantages. This supports the observations of Skvarciany et al. [55], who noted that Western European countries lead the twin transition, while Central and Eastern regions continue to lag behind.

The intermediate cluster, which includes countries like France, Italy, and Spain, shows partial convergence: their levels of DT and SO are close to the EU average, while EO stays moderate. We believe these countries are in a stage of consolidation, consistent with Ricci et al. [6] and Ukko et al. [7], who emphasized that integrating DT and SO can boost economic value but only if supported by clear policies and strong environmental commitments.

The group of Central and Eastern European countries, characterized by low AI adoption and GDP per capita below the EU average, highlights the challenges these nations face in transforming digital development (DT) into a driver for sustainable and economic growth. This situation reflects Grybauskas et al. [32]’s concerns about the social risks of DT, as well as those of Nosratabadi et al. [57], who argued that DT could worsen inequalities, especially in countries with fragile infrastructure and governance.

Based on these results, H2 is confirmed: member states form distinct groups, and these groups show not only regional convergence but also persistent asymmetries. The gap between Nordic and Western states and Central and Eastern states highlights that the twin transition remains fragmented and its benefits are unevenly distributed.

Compared to the existing literature, this study offers a systemic view that emphasizes how DT and SO complement each other with varying degrees of intensity and effects shaped by HC and the institutional context. The critique of green growth by Hickel and Kallis [34] is relevant here: DT does not guarantee sustainable progress, and SO does not always lead to significant economic growth. However, their complementarity becomes clear in high-performing clusters, showing that synergies do exist but need favorable conditions to fully develop.

5.1. Theoretical Implications

The findings from the factor analysis, generalized linear models, and cluster analysis offer valuable contributions to the expanding literature on the European twin transition. From a theoretical perspective, the study supports the idea of complementarity between digital transformation (DT) and sustainable orientation (SO), while also showing that this relationship is neither uniform nor unconditional. The analysis confirms that human capital (HC) functions as a key mediating factor—its presence enhances the economic and sustainability outcomes of digitalization, while its absence limits them.

In this sense, the study advances the debate on Brynjolfsson’s “productivity paradox” [26] by showing that technology alone does not ensure progress in economic or sustainable terms. Instead, it is the interaction between technological development, institutional capacity, and human adaptability that determines whether innovation benefits society. By identifying distinct clusters across Europe, the analysis enhances theoretical discussions on regional convergence and divergence, supporting Harangozó and Fakó’s [10] argument that the twin transition remains fragmented. The research, therefore, emphasizes the need for a systemic, context-aware approach that connects digital and sustainable shifts to wider institutional dynamics.

5.2. Empirical Interpretations

Empirically, the results show partial support for the hypothesized complementarity between DT and SO. Hypothesis H1 is confirmed, but with qualifications: while digital and sustainable dimensions jointly influence economic outcomes (EO), their effects differ significantly across indicators and regions. Human capital stands out as the most consistent and statistically significant factor affecting both GDP per capita and progress toward the Sustainable Development Goals Index (SDGi), indicating that digitalization provides tangible benefits mainly where workforce skills and adaptability are high.

On the other hand, the negative link between Digital Public Services (DPS) and sustainability, as shown in Section 4, indicates a complex and transitional pattern. This unexpected outcome might be due to efficiency-focused reforms that initially emphasize administrative modernization rather than inclusive or environmental goals. It could also result from temporary measurement distortions or institutional imbalances during the early phases of digital reform. These findings emphasize that complementarities are not automatic; they rely on governance quality, policy sequencing, and institutional readiness—factors that need further long-term investigation.

Hypothesis H2 is also supported, as the cluster analysis identifies three structural groups within the EU: a high-performance cluster in the North and West, a moderate group in the center, and a lagging cluster in the East and South. However, the ongoing regional disparities show that convergence is still incomplete and that institutional legacies and administrative capacity influence the twin transition as much as economic or technological factors.

5.3. Policy Implications

The policy implications of this research must be approached with caution, given the limited timeframe and diverse dataset. Nevertheless, the findings highlight several important directions. First, the importance of human capital emphasizes that investments in digital education, reskilling, and training are essential for achieving both economic growth and sustainable development. These efforts should be regarded not as supplementary but as the foundation of competitiveness.

Second, the tensions observed between administrative digitalization and sustainability highlight the need for responsible digital governance. Efficiency and automation goals must be balanced with social inclusion and environmental responsibility, ensuring that digital reforms do not unintentionally reinforce inequality or ecological harm.

Finally, the persistence of regional disparities requires tailored policy approaches. European and national strategies must stay adaptable, recognizing diverse institutional capacities instead of enforcing one-size-fits-all solutions. Customized interventions, particularly those that strengthen local governance, innovation systems, and absorptive capacities, are crucial to ensure that the twin transition promotes cohesion rather than division.

5.4. Limitations and Further Research

While the research design provides valuable insights, it is important to acknowledge its limitations. The study depends on a limited set of harmonized indicators (DESI, SDGi, AI adoption, and macroeconomic measures) that cannot fully represent the multidimensional nature of the twin transition. Additionally, the cross-sectional approach restricts causal inference and prevents observing long-term effects. Future research should incorporate longitudinal and sectoral perspectives to track the development of complementarities over time and across different economic sectors. Including variables related to governance quality, institutional innovation, and cultural adaptability would also enhance understanding of how digital and sustainable transitions interact.

6. Conclusions

This study explored how digital transformation and sustainable orientation interact in shaping economic outcomes across EU member states. The analysis confirms that the relationship between these processes is complex, asymmetric, and mediated by human capital. Hypothesis H1 is thus supported in a qualified way: digital and sustainable progress reinforce each other, but the strength of their complementarity depends on institutional maturity and workforce adaptability. Hypothesis H2 is also supported, as the identified clusters show persistent regional polarization despite shared policy goals.

Overall, the results indicate that the twin transition in Europe should be seen as an evolving and context-dependent process rather than a linear path. Digitalization can boost sustainability only when supported by coherent governance, inclusive education, and stable institutional structures. The study’s novelty lies in combining both dimensions into a single analytical framework and emphasizing how structural asymmetries and institutional conditions influence their interaction.

Treating the twin transition as both an economic and institutional phenomenon, this research offers a framework for understanding current disparities and shaping future policies. Although limited in scope and duration, the findings promote a more detailed understanding of Europe’s transformation—one that emphasizes the interconnectedness of technological, human, and governance factors in creating a sustainable digital future.