Digital Transformation of Enterprises and Employment in Technologically Advanced and Knowledge-Intensive Sectors in the European Union Countries

Abstract

The digital transformation of enterprises constitutes a crucial element of modernization processes in the countries of the European Union, supporting the achievement of sustainable development goals through increased innovation, economic efficiency and the development of high-tech sectors and knowledge-intensive services. In empirical analyses, multidimensional statistical analysis methods were applied, with particular emphasis on linear ordering, classification methods and regression analysis. Ten monitoring indicators adopted by the European Union constitute the basis for assessing the digital transformation advancement of enterprises. The study classified EU countries into three groups according to the level of digital transformation: low, medium and high. Based on the employment structure in high-tech manufacturing, medium–high-tech manufacturing and knowledge-intensive services, four distinct types of employment structures were identified. A positive relationship was found between the level of digital transformation and the share of employment in knowledge-intensive services, while no significant relationship was observed for the other sectors. The study shows that the EU countries are clearly differentiated both in terms of the level of digital transformation of enterprises and the structure of employment in high-tech manufacturing, medium–high-tech manufacturing and knowledge-intensive services. A higher level of digital transformation supports the growth of employment in knowledge-intensive services, while its impact on high-tech and medium–high-tech manufacturing remains limited.

1. Introduction

Digital transformation constitutes one of the key processes shaping the modern economy, industry and labor market. It is defined as a comprehensive organizational and technological change aimed at improving the functioning of businesses by means of using modern digital technologies [1]. Its importance is growing along with advancements in the areas of automation, artificial intelligence (AI), Internet of Things (IoT) and data analytics, which transform business models and competency requirements in the labor market [2,3]. Digital transformation includes not only the implementation of new technologies, but also changes in management, operating models and organizational culture [4].

One of the most important effects of digital transformation is its impact on employment. On the one hand, digitalization is creating new jobs and increasing the demand for highly skilled professionals in such areas as data analytics, cybersecurity and software engineering [5]. On the other hand, automation and the use of AI may result in the elimination of certain professions, raising concerns about higher technological unemployment and labor market polarization [6,7].

Digital transformation can stimulate changes in the structure of employment, affecting the share of employment in high- and medium–high-tech manufacturing and knowledge-intensive services, e.g., through increased demand for specialized skills, implementing technological innovation, changes in business models or process optimization. Countries characterized by a high level of technological advancement are more likely to record employment growth in high-tech manufacturing and knowledge-intensive services, which promotes the innovative development of the economy [8,9].

The digital transformation of enterprises and the development of employment in high- and medium–high-tech manufacturing and knowledge-intensive services are essential to achieving many of the Sustainable Development Goals (SDGs) [10,11,12]. The main ones are as follows:

–

SDG 8: Decent work and economic growth. Digital transformation leads to the creation of new jobs; however, at the same time, it requires the adaptation of workers’ competencies to the digital economy. Jobs in high- and medium–high-tech manufacturing and knowledge-intensive services typically provide higher wages, better working conditions and greater stability of employment. These sectors support innovation-driven economic growth.

–

SDG 9: Industry, innovation and infrastructure. Strengthening digital transformation promotes the development of modern industry and knowledge-intensive services, which is the foundation of an innovative economy.

–

SDG 10: Reduced inequalities. Differences in the level of digitalization and employment structure can exacerbate inequalities between the EU countries. The HTM, MHTM and TKIS sectors can foster social and professional inclusion of different social groups through flexible forms of employment, e.g., remote work.

–

SDG 12: Responsible consumption and production. Digital transformation allows the optimization of supply chains, reduction in waste and more efficient use of raw materials. Innovative technologies developed in high-tech manufacturing support creating greener solutions (e.g., sustainable manufacturing, energy-efficient materials, recycling).

–

SDG 13: Climate action. Enterprises implementing digital tools can reduce their carbon footprint through smart energy management, digital documentation and logistics optimization. Knowledge-intensive services and high-tech manufacturing stimulate the development of green technologies and sustainable solutions, such as renewable energy, smart city management systems and efficient waste management.

–

SDG 4: Quality education. Digital transformation and the development of technology sectors force changes in the education system—it is necessary to develop technical and scientific education and strengthen digital competence and lifelong learning.

In summary, both the digital transformation of enterprises and increased employment in high- and medium–high-tech manufacturing and knowledge-intensive services are contributing to the construction of a modern, innovative and sustainable economy of the European Union countries. In the economic dimension, they increase competitiveness, improve efficiency and stimulate innovation, contributing to economic growth. In the social dimension, they improve the quality of jobs, promote digital and vocational education and reduce inequality in the labor market. In the environmental dimension, they enable the optimization of resource consumption, reduction in emissions and implementation of sustainable technologies.

2. Literature Review

2.1. Digital Transformation

Digital transformation is a concept that can be approached in both broad and narrow terms. In the narrow sense, digital transformation is focused on the level of organizations, especially enterprises. It is a fundamental change in the operations, processes, competencies and business models of an enterprise aimed at taking full advantage of the opportunities offered by digital technologies [1,13,14,15,16].

It is not just the implementation of technology, but a strategic and operational shift affecting external relations, internal processes and resource allocation. It includes the modification of business models and value chains, as well as the creation of new products and services [1,16].

In the case of small and medium-sized enterprises, digital transformation is crucial as it affects many aspects of their operations. It contributes to improving competitiveness and productivity through the adaptation of business processes and practices to the digital world and the use of technologies such as artificial intelligence, big data, cloud computing, e-commerce and the Internet of Things enabling them, i.a., to optimize the use of resources and increase the quality of products and services [13,17,18].

Digital transformation involving the application of digital platforms and e-commerce opens up new channels for SMEs to sell and communicate with customers, facilitating the acquisition of new customers and markets, including overseas ones [1,19].

Previous research shows that digital transformation increases the innovation capacity of an enterprise. In turn, the implementation of innovations involving, e.g., the launch of new products and services (product innovation) or new processes (process innovation) improves company productivity [20]. Digital technologies enable SMEs to implement innovations in their business models and create new business patterns and methods for generating value [1,19].

Digital transformation exerts a significant impact on corporate human resource management. It includes, i.a., an innovative approach to recruitment, the digital management of personnel data and the development of digital competencies among employees. The implementation of digital solutions in the HRM area promotes the creation of more flexible and effective models for personnel management [21].

The application of data and digital analyses allows SMEs to make better-informed and strategic business decisions [1]. Digital transformation can help reduce regulatory burdens by using e-government and online public services. SMEs that have implemented digital transformation are frequently more flexible and resilient to external shocks such as pandemics or economic downturns [19].

The digital transformation of enterprises creates new opportunities for the realization of sustainability goals. Digital technologies support companies in managing the environmental impact of their operations more effectively [22]. At the same time, enterprises implementing advanced digital technologies demonstrate greater readiness to improve their sustainability and social responsibility practices [23].

SMEs also face challenges and barriers in the process of digital transformation, such as limited resources, including financial ones, competency gaps, a lack of awareness of the related benefits, and the rigidity of traditional hierarchical structures rooted in the command-and-control management styles [24].

A lack of technological infrastructure or the need to upgrade it, resulting in limited access of an enterprise to modern IT systems or digital tools, are often identified as the main external factors that can hinder or slow down the implementation of digital technologies. Regulatory complexity—unclear or unstable laws regarding digital security, privacy and artificial intelligence, or the absence of such regulations—can constitute a significant external barrier. Absent or insufficient government and institutional support and low levels of digital maturity of the corporate environment, including business partners, are considered further barriers [25,26,27]. Digital transformation is not fostered by a turbulent market environment, intense market competition or the technological intensity of the industry either [28].

The involvement of entrepreneurs, the development of their competencies (including digital ones), organizational capacity building, support from digital platform providers and government policies are critical to the success of digital transformation in SMEs [16,17,18,19]. Factors such as a well-planned and organized digital transformation implementation process, properly selected technology tools and access to high-quality data are also important [25,29]. Above all, entrepreneurs should remain absolutely confident about the need for and benefits of implementing digital transformation. It can turn out to be a significant challenge for them to overcome the so-called resistance to change on the part of employees despite the occurrence of significant changes in the work environment [30,31]. Managers should understand the importance of strategic planning for digital transformation, investing in the development of employees’ digital skills, and taking advantage of all the available forms of support, external programs and grants, as well as cooperation with experts outside the enterprise [30].

In broad terms, digital transformation covers deep and accelerating changes in society and various sectors of the economy through the application of digital technologies [1,4,14]. It refers to the transformation of entire economies, constituting the key factor in economic development. Digital transformation is a driving force stimulating innovation, productivity growth and competitiveness of economies in the European Union [32,33]. It enables the development of new, innovative industries and services. It affects many aspects of society, such as education, the labor market, salary levels, social inequality, health, resource efficiency and security [14,34]. It enhances the sustainability of the economy by using resources more efficiently, reducing waste, improving the monitoring of the environment, supporting the concept of a circular economy and contributing to the reduction in carbon emissions [18]. The digitalization of public administration and health services (e-government and e-health) improves the accessibility and quality of services for citizens [2,19,35].

In order for digital transformation to benefit all citizens and not exacerbate the existing social inequalities, it is crucial to consider its social sustainability (SOSDIT). This means ensuring equal access to digital technologies and digital competencies, addressing digital exclusion and taking into account the impact of new technologies on employment [14,35]. Thus, digital transformation goes beyond an enterprise level, creating both enormous opportunities and complex challenges. Addressing them effectively requires taking coordinated actions at the political, economic and social levels across the European Union [33,36,37].

The actions supporting the EU’s digital transformation began in 2010 with the Digital Agenda for Europe initiative [38] carried out within the framework of Europe 2020 strategy. In March 2021, as part of the implementation of the “Path to the Digital Decade” program, the European Commission adopted the 10-year action plan 2030 Digital Compass: the European way for the Digital Decade, including goals and methods for monitoring them. The digital transformation of businesses remains one of the pillars in the program. The EU has also introduced a number of regulations, e.g., on data management, digital services, digital markets or cybersecurity [39]. The EU financial support for digital transformation is provided through the Digital Europe Program and a range of budgetary instruments, including cohesion programs, the Technical Support Instrument and funds from the Recovery and Resilience Facility. European Digital Innovation Hubs (EDIHs) have been organized in the EU countries, offering specialized consulting services and technology support to small and medium-sized enterprises [40,41].

Based on the above considerations, the level of digital transformation of enterprises can vary significantly among the countries of the European Union due to the occurrence of different institutional, economic and technological conditions. Such diversity creates the need to identify disproportions in this area and to study its impact on socio-economic structures. It constitutes the basis for formulating the research hypotheses presented in Section 2.3 of this paper.

2.2. Digital Transformation and the Labor Market

Driven by technologies such as artificial intelligence, big data, cloud computing and the Internet of Things, digital transformation is fundamentally changing the way businesses and economies operate [1,14,42]. While causing changes in various sectors of the economy, it also affects the labor market and employment, including high- and medium–high-tech manufacturing and knowledge-intensive services.

Digital transformation exerts a complex, multifaceted impact on the labor market, both at the macro- and micro-economic levels. At the macroeconomic level, it leads to changes in the structure of employment, wage levels and labor productivity, as well as the emergence of new areas of activity [43,44]. At the enterprise level, digital transformation is forcing companies to make profound changes in their organizational structure, management models and operational processes. It results in the disappearance of some jobs and the emergence of others, changing the scope of responsibilities and requiring employees to continuously develop competencies, especially the digital and analytical ones [22,27].

When considering the impact of digital transformation on the labor market, it is important to take into account both the positive and negative effects of this process. Positive outcomes include the emergence of new employment opportunities. Digital transformation generates new occupations, such as data analyst or AI engineer [14,42,43,44,45]. It contributes to new industries being established, e.g., e-commerce, digital logistics, cybersecurity, big data and data analytics, and Fin-Tech digital platform management. Thus, it stimulates the creation of new jobs [44,46]. The particularly beneficial effect is that these tend to be high-quality, talent-attracting positions [26]. Another positive consequence of digital transformation is perceived in improved efficiency and productivity, directly related to better use of resources due to the application of modern technologies [37,42,43]. Taking advantage of digital technologies offers opportunities to expand remote work and develop flexible forms of employment. Through the development of e-learning and digital platforms, it facilitates access to knowledge and training, which is extremely important in view of the growing demand for employees with university education and highly qualified specialists [44].

An important negative effect of the implementation of new technologies is the destabilization of traditional employment structures, leading to a mismatch between the market demand for certain competencies and the actual qualifications presented by workers, which may eventually result in the gradual disappearance of jobs in sectors requiring low skills [22].

A natural consequence of the digital transformation development is the automation of jobs, which may bring about downsizing in some areas or sectors of the economy [44,47]. Workers performing repetitive and routine tasks are more likely to be replaced by digital technologies, while a growing demand for specialists in areas requiring knowledge and analytical and interpersonal skills will increase [12,35]. Digital transformation is changing the nature of work in these sectors, requiring workers not only to have technical expertise, but also to be adaptable, continuously improve their competencies and learn new digital skills [14,19].

Digital transformation can exacerbate digital and competence-based inequalities, also resulting in unequal access to the labor market and a widening wage gap [46]. The need to invest in digital skills and upskill or retrain workers comes at an additional cost, which may influence companies’ decisions to cut jobs or wages [48,49,50].

Digital transformation is a key driver of growth and innovation in high- and medium–high-technology manufacturing. The development of these sectors translates directly into opening new high-tech jobs. Knowledge-intensive services also experience significant changes and higher employment resulting from digital transformation. Digital technologies enable better data collection, their analysis and use, which is the foundation of many knowledge-intensive services. In the EU countries, the pursuit to build a uniform digital market as well as investments in the development of digital infrastructure and digital skills offer favorable conditions for higher employment in high-tech manufacturing and knowledge-intensive services [14,35].

In summary, digital transformation, in a broad sense, exerts a complex impact on the labor market. While it may lead to downsizing in traditional sectors, it creates significant opportunities for increased employment in high- and medium–high-tech manufacturing and knowledge-intensive services. However, it is crucial to invest in education and develop the digital skills of society so that workers can adapt to the changing demands of the labor market and take full advantage of the potential offered by digital transformation. Governments and businesses in the EU countries must actively support these processes to ensure that digital transformation has a beneficial and sustainable impact on the labor market and the economy [14,35]. The European Union, as part of the aforementioned European Digital Agenda, is undertaking initiatives and activities aimed at reducing digital inequalities between the Member States and supporting digital transformation in all sectors of the economy. In the long term, they contribute to a better adaptation of the labor market to the challenges posed by the growing digital economy [38,51].

The listed source literature highlights the multifaceted impact of digital transformation on the labor market. It justifies the need to study potential employment patterns related to different stages of digital advancement. This rationale is the basis for formulating the research hypotheses presented in Section 2.3, which assume the existing interdependence between the level of digital transformation and diverse employment patterns in high-tech industries and knowledge-intensive services in the EU countries.

2.3. Development of Research Hypotheses

The source literature raises questions about the diversity of countries regarding the level of digital transformation determined by technological differences, which can also affect changes in employment structure and the labor market. Strategies implemented by the EU support the development of a digital single market, investment in infrastructure and the development of digital competencies. The Digital Economy and Society Index (DESI) constitutes the key tool for monitoring progress made by the Member States in this regard [52]. However, studies of the digital transformation impact on the labor market focus primarily on the microeconomic aspects, especially on the effects of implementing digital solutions in enterprises, such as the growing demand for ICT specialists, highly skilled workers presenting digital skills, changes in employment forms and workforce educational structure, etc. [27].

At the macroeconomic level, the source literature provides analyses addressing digital transformation, its determinants and links to the socio-economic phenomena in the European Union Member States. The conducted research was focused on international comparisons covering the development level of the economy and digital society against the level of education and the life satisfaction degree of the population [53]. The convergence of the level of digital integration in the enterprises of the EU countries was also analyzed [49]. Moreover, attention was paid to the relationship between the economy’s digitalization advancement and sustainable development [51] as well as the links between the level of digital transformation, measured by the DESI, and economic growth [47]. Comparative studies were also focused on analyzing the competitiveness of European economies, considering their innovation potential and the sophistication of digital processes [46] and comparing the level of digital transformation in different-sized enterprises in Central and Eastern European countries [50]. The literature also took into account the issue of digital convergence of the EU Member States’ markets approached as a process of bridging the differences in the development of the economy and digital society [33]. The relationship between the regional differentiation of the sectoral employment structure and the level of economic development was also analyzed [54]. Subsequent research studies attempted to capture the convergence processes occurring in the employment structure by sector and by enterprise size [55].

While the above studies provide valuable information on the degree of digitalization and its macroeconomic implications, as well as changes in the structure of employment in the EU countries, they do not offer analyses addressing the relationship between the level of digital transformation and the structure of employment in technologically advanced sectors, particularly in the context of small and medium-sized enterprises. Although some individual studies point to possible relationships between digitalization and employment in these sectors, there is still a lack of systematic and comparative analyses examining the relationship between the level of digital transformation and employment structures in high-tech manufacturing (HTM), medium–high-tech manufacturing (MHTM) and knowledge-intensive services (TKISs) across the EU Member States. In particular, there is a shortage of empirical research focused on how digital transformation in small and medium-sized enterprises affects the employment distribution in these advanced and innovation-driven sectors. Therefore, it can be concluded that there is a research gap in identifying differences in employment structure in high- and medium–high-tech manufacturing and knowledge-intensive services among the European Union countries characterized by different levels of advancement of digital transformation in small and medium-sized enterprises. Filling this gap can contribute to a better understanding of the digital transformation implications for the European labor market and more effective design of policies to support digital sustainability.

The scope of the conducted research refers to small and medium-sized enterprises operating in the European Union Member States; therefore, a narrow presentation of digital transformation was applied and approached as a process for implementing modern technologies in various company areas, focusing on the intensity of the implemented changes. The purpose of the study is to assess the degree of differentiation and to classify the European Union countries in terms of the level of digital transformation advancement of enterprises and the share of employment in high-tech manufacturing, medium–high-tech manufacturing and knowledge-intensive services, and to identify the existing relationships between these phenomena. For empirical analyses, multidimensional statistical analysis methods were applied, with a particular emphasis on linear ordering and classification methods.

The following research questions were formulated in the conducted studies:

• What is the degree of differentiation between the EU countries in terms of digital business transformation, and how can they be grouped based on their level of advancement?
• What types of employment structure in high- and medium–high-tech manufacturing and knowledge-intensive services occur in the EU countries?
• What differences in the structure of employment in high- and medium–high-tech manufacturing and knowledge-intensive services occur between the groups of EU countries presenting different levels of digital business transformation?
• Is there a relationship between the level of digital transformation advancement and the employment structure of high- and medium–high-tech manufacturing and knowledge-intensive services in the EU countries?

The research addressing the above questions seems important from both an academic and a practical perspective. The identification of countries presenting different levels of digital transformation allows assessing the progress of digitalization in the EU and indicating both leaders and laggards. It can help develop effective strategies to support digitalization at the national and EU levels.

Digital transformation, understood as the process of implementing modern technologies, is currently one of the key pillars of the modern economy. The economic literature, including Solow’s neoclassical growth theory [56], Romer’s endogenous growth theory [57] and the theory of evolution [58], emphasizes the fundamental importance of technological progress for long-term economic development and productivity enhancement. These theories indicate that technology—which today can be referred to as digital transformation—is the driving force behind growth and innovation.

In Rostow’s [59] modernization approach, the implementation of modern technologies is perceived as a prerequisite for moving to higher stages of socio-economic development. However, the pace and scope of transformation are not homogeneous—different countries reach various levels of advancement, which may result from their historical technological background, the economic structure and the availability of resources (financial, human, institutional). This diversity is also evident within the European Union, where its member countries differ regarding their level of digital technology implementation in the business sector. Therefore, one of the objectives of the presented study was to determine the degree of such diversity and to identify possible groups of countries—leaders, intermediates and laggards—based on a synthetic indicator of digital transformation. On this basis, the following hypothesis was formulated:

H₁. 

EU countries differ significantly in terms of the level of digital transformation of enterprises, which allows for the identification of distinct groups characterized by high, medium and low levels of advancement.

The structure of employment in high-tech manufacturing and knowledge-intensive services reflects the level of innovation and competitiveness of the economy. An analysis of this aspect enables assessing which EU countries are developing toward a knowledge-intensive and high-tech economy and which are still relying on less innovative sectors. It is also important in terms of forecasting future labor market trends and planning educational activities in the field of digital skills.

From Porter’s [60] theory of competitive advantage and Balassa’s [61] concept of comparative advantage, one can conclude that countries develop based on different comparative advantages, which leads to different employment structures between countries. Some economies may maintain a strong position in medium-tech industries, such as machinery manufacturing or the automotive sector, whereas others focus on developing knowledge-intensive and high-tech services, such as Fin-Tech or big data services. In addition, the literature on economic integration and structural convergence in the EU indicates that despite common digital or industrial policies, persistent differences occur between the economic structures of Member States. In practice, there are manifestations of both economic convergence and divergence [54,62]. Accordingly, it can be considered that the EU countries may be characterized by different types of employment structures in high-technology manufacturing (HTM), medium–high-technology manufacturing (MHTM) and knowledge-intensive services (KISs)—from industrial technology, through sustainable technology, to highly service-oriented and digitally advanced structures. Such reasoning justifies the formulation of the following research hypothesis:

H₂. 

European Union countries are characterized by different types of employment structures in the high-tech, medium–high-tech and knowledge-intensive service sectors, reflecting different models of economic transformation—from industrial specialization, through a balanced structure, to the dominance of knowledge-intensive services.

By comparing the employment structure of countries characterized by different levels of digital transformation, it is possible to determine whether digitalization promotes employment growth in technologically advanced sectors and identify differences in the level of these sectors’ development between groups of countries. It can help in understanding the mechanisms based on which digitalization affects the economy and labor market and in developing strategies to enable more diverse technological development across the European Union.

Examining the relationships between the level of digital transformation advancement and the structure of employment in high- and medium–high-tech manufacturing and knowledge-intensive services in the EU countries is essential to determine how digitalization affects the labor market and employment structure. If it turns out that a higher level of digital transformation is associated with a higher share of employment in high-tech manufacturing, it will prove that digitalization can act as a tool supporting the development of the knowledge-intensive economy.

The diversity of employment structures in high- and medium–high-technology manufacturing (HTM, MHTM) as well as knowledge-intensive services (KISs) is justified by the theory of the knowledge-based economy (KBE), according to which knowledge, innovation and their diffusion in the manufacturing and service sectors constitute the main drivers of growth [63]. Since countries integrate these elements in different ways, they develop diverse employment structures. As a result, differences in the level of digital business transformation between countries can lead to significant diversification in the employment structure of high-tech and knowledge-intensive sectors. Moreover, unlike manufacturing sectors (HTM, MHTM), where digitalization can also result in automation and downsizing in some areas, in the KIS sector, digital transformation generates an increased demand for highly skilled personnel and promotes the opening of new jobs [6,44,47]. This may constitute the basis for the formulation of the following research hypotheses:

H₃. 

Employment structures in the high- and medium–high-tech and KIS sectors differ significantly between the groups of EU countries presenting high and low levels of digital transformation—the countries with higher levels of digitalization are characterized by a higher share of employment in the TKIS sector.

H₄. 

There is a positive relationship between the level of digital business transformation in the EU countries and the employment structure of the HTM, MHTM and TKIS sectors, with the strongest impact observed in the case of knowledge-intensive services.

Analyzing these issues is important both from the EU economic policy perspective and in the context of pursuing sustainable development. Digital transformation has the potential to increase business innovation and efficiency, but its impact on the labor market may vary depending on the national context. The analysis of differences between the EU countries as well as the relationships between digitalization and the employment structure allows a better understanding of the mechanisms underlying economic development and the adoption of policies that support both innovation and labor market stability in Europe.

The findings can provide crucial guidelines for the EU economic policy in supporting digitalization and shaping employment strategies tuned to the new labor market reality. A better understanding of the mechanisms related to the digitalization of the economy and its impact on employment can contribute to the development of more effective adaptation strategies, enabling balanced technological development across the European Union.

3. Materials and Methods

The conducted research analyzed the advancement of digital transformation of enterprises, as well as employment in high- and medium–high-tech manufacturing and knowledge-intensive services, which allowed carrying out the research purpose and verifying the hypotheses regarding the diversity of the EU countries in the area of digital business transformation (H₁), the occurrence of different types of employment structures in high-tech and knowledge-intensive sectors (H₂), the relationship between the level of digitalization and employment structure (H₃) and the impact of digital business transformation on employment in individual sectors (H₄).

The basis of empirical analyses aimed at diagnosing the level of digital transformation was the data on the following set of 10 indicators adopted by the European Union to monitor the advancement of digital transformation in enterprises (10 or more employees and self-employed persons):

X₁—enterprises with at least a basic level of digital intensity (% of enterprises);

X₂—enterprises using an ERP (enterprise resource planning) software package to share information between different functional areas (e.g., accounting, planning, production, marketing) (% of enterprises);

X₃—enterprises using two or more of the following social media: social networks, enterprise blogs or microblogs, multimedia content sharing websites, wiki-based knowledge sharing tools (% of enterprises);

X₄—enterprises performing data analytics (internally or externally) (% of enterprises);

X₅—enterprises purchasing at least one of the following cloud computing services: hosting of the enterprise database, accounting software applications, CRM software, computing power (% of enterprises);

X₆—enterprises using at least one AI technology (% of enterprises);

X₇—enterprises sending e-invoices, suitable for automated processing (% of enterprises);

X₈—cumulative number of unicorns;

X₉—enterprises’ total turnover from e-commerce sales (% of turnover);

X₁₀—enterprises selling at least 1% of their turnover online (% of enterprises).

The analysis of employment in high- and medium–high-technology manufacturing and knowledge-intensive services was conducted based on statistical information regarding the following elements of the sectoral structure of employment identified based on the technological intensity level:

HTM—employment in high-tech manufacturing (%);

MHTM—employment in medium–high-tech manufacturing (%);

TKIS—employment in total knowledge-intensive services (%).

The spatial scope of the study included the 27 Member States of the European Union. The temporal scope of the research covered 2023, which was determined by the availability of statistical data on the broadest and most up-to-date range of variables, ensuring a comprehensive diagnosis of the digital transformation level and the timeliness of the presented research results. Statistical information on the values of all indicators of digital transformation and elements of the sectoral employment structure was retrieved from the Eurostat database [64].

The selection of a sample covering all 27 EU Member States in 2023 is fully adequate to carry out the research purpose and verify the formulated research hypotheses. The spatial scope of the study corresponds to the content scope of the hypotheses, which refer to the differences and dependencies occurring in the EU countries. The sample is internally diversified, since the European Union is an internally differentiated political and economic environment. It creates the necessary conditions for analyzing the statements included in hypotheses H₁–H₄ regarding the level of digital transformation and employment structures in high-tech and knowledge-intensive sectors.

When selecting the indicators for the digital transformation of enterprises and the period of the study, the most current and comprehensive approach to measuring digital transformation was taken into account. Thus, the completeness, comparability and timeliness of statistical data were considered. The new available set of indicators monitoring the level of digital transformation adopted by the EU in 2023 was used. This assumption also determined the temporal scope of the study. It was impossible to maintain the availability and comparability of data over a longer period. The availability of information in terms of the time range in the cross-section of indicators was very limited and diverse. For example, in the case of the share of companies performing data analytics (internally or externally) (X₄), it was limited to 2023 only. The research findings reflect the current status of digitalization, which is a definite advantage of this approach as it increases their usefulness for economic and social policy.

In addition, the necessary statistical data were almost fully available. The exceptions were missing information on the X₉ indicator for Italy and the shares of those working in the HTM and MHTM sectors in Luxembourg. In such cases, the indicator values from the year prior to the year under study were adopted. In addition, the information originates from a single source, the Eurostat database, which ensures its consistency, completeness and comparability across the EU countries. As a result, the adopted research sample allows for a reliable and credible verification of the research hypotheses.

The applied research procedure was designed to effectively achieve the research goal, i.e., to assess the degree of differentiation and classify the European Union countries by the advancement of digital transformation of small and medium-sized enterprises and the structure of employment in the HTM, MHTM and TKIS sectors and identify their mutual relationships.

The research was carried out according to the following research procedure consisting of five stages:

Stage 1. Description of the research sample.

This stage is the assessment of the diversity of the EU countries based on indicators monitoring the advancement of digital transformation of enterprises and on the structure of employment in high-tech manufacturing, medium–high-tech manufacturing and knowledge-intensive services using statistical descriptive parameters.

Stage 2. Linear ordering and division of the EU countries into relatively homogeneous classes according to the level of digital transformation intensity.

Stages 1 and 2 lead to the empirical verification of hypothesis H₁: EU countries differ significantly in terms of the level of digital transformation of enterprises, which allows for the identification of distinct groups characterized by high, medium and low levels of advancement.

More information on linear ordering methods, with particular emphasis on the ways to normalize variables and aggregation formulas that can be used, is described, i.a., in the studies [64,65,66,67,68,69,70,71,72,73,74,75,76].

• Matrix construction:

$X={\left[x_{ij}\right]}_{\left(nxm\right)}$—values of indicators X₁–X₁₀ monitoring the advancement of digital transformation in enterprises;

$Z={\left[z_{ij}\right]}_{\left(nxm\right)}$—normalized values of indicators X₁–X₁₀.

The zero unitarization formula was used to normalize digital transformation intensity indicators, expressed by Formula (1)—all indicators play the role of stimulants:

$$z_{ij}={\displaystyle \frac{x_{ij}-\underset{i}{\mathit{min}}{x}_{ij}}{R_j}}$$

(1)

where $z_{ij}—$normalized value of the j-th indicator for the i-th object; $x_{ij}$—value of the j-th indicator for the i-th object; $R_j—$range of j-th indicator values; i = 1, …, n (n = 27); j = 1, …, m (m = 10).

The construction and properties of the null normalization formula were characterized by K. Kukuła in his work [72].

• The average standardized sums method was applied as the function aggregating the normalized values of digital transformation intensity indicators:

$${SM\left(DT\right)}_i={\displaystyle \frac{1}{m}}\sum _{j=1}^m{z}_{ij}$$

(2)

where ${SM\left(DT\right)}_i$—synthetic measure of the level of digital transformation intensity of enterprises for the i-th object country; j = 1,…, m—the number of digital transformation indicators; i = 1, 2,…, n—object-country number.

In the study, the individual indicators of digital transformation intensity were assigned a system of equal weights. Currently, there is no consensus in the source literature on objectively differentiating their value. The equal weights system recognizes the complementarity of components of digital business transformation and reduces the subjectivity of the applied approach. Equal weights were treated as a neutral approach to individual areas of business transformation, which turns out to be particularly important when applying the newly adopted set of digitalization indicators. More on the weighting of variables is included in the studies [67,70,75].

The reduction in dimensions referring to the digital transformation of enterprises was abandoned in the analysis, as the study aimed not only at using an aggregate measure to assess the level of development of a complex phenomenon such as digital transformation, but also at maintaining transparency and comprehensibility for the final recipients (e.g., policymakers or government representatives). Equal weighting and consideration of all indicators allows for an intuitive understanding of the contribution of various components of digital transformation. In addition, the set of monitoring indicators for 2023 was modified by the EU due to the indicators’ informative value and relevance for measuring the current state of digital transformation. Abandoning the reduction in dimensions enabled them to be fully incorporated into the construction of a synthetic measure and thus enabled the level of digitalization of enterprises to be assessed.

The indicators monitoring the digital transformation intensity of enterprises, normalized using the zero unitarization method, take values in the [0, 1] interval; therefore, synthetic measures of digital transformation intensity are characterized by the following value:

$${SM\left(DT\right)}_i\in [0,1]$$

(3)

• Sensitivity analysis of the synthetic measure of digitalization development.

Sensitivity analysis was conducted to assess the robustness of the synthetic measure to changes in the set of sub-indicators. For this purpose, a procedure was used that involved sequentially eliminating each of the ten indicators of digital transformation and recalculating the value of the synthetic measure for all countries. The resulting rankings were then compared with the baseline ranking (based on the full set of digitalization indicators) using Spearman’s rank correlation coefficient. The values of this coefficient for each of the analyzed versions of the synthetic measure allow verifying its stability and the absence of dominance of individual indicators in the structure of the synthetic measure.

• The division of the European Union countries into classes presenting different levels of digital transformation intensity was carried out by specifying the following ranges of aggregate measures values:

Class 1—countries characterized by the lowest level of digital transformation intensity:

$${SM\left(DT\right)}_i\le \underset{i}{\mathit{min}} \left\{{SM\left(DT\right)}_i\right\}+{\displaystyle \frac{1}{k}}R$$

(4)

Class 2—countries presenting a higher level of digital transformation intensity than the ones grouped in Class 1:

$$\underset{i}{\mathit{min}} \left\{{SM\left(DT\right)}_i\right\}+{\displaystyle \frac{1}{k}}R<{SM\left(DT\right)}_i\le \underset{i}{\mathit{min}} \left\{{SM\left(DT\right)}_i\right\}+{\displaystyle \frac{2}{k}}R$$

(5)

Analogically, the subsequent classes and the last identified k-th class include countries with the highest level of digital transformation intensity:

$$\underset{n}{\mathit{min}} \left\{{SM\left(DT\right)}_i\right\}+{\displaystyle \frac{k-1}{k}}R<{SM\left(DT\right)}_i\le 1$$

(6)

where R—range of synthetic measure value of the complex digital transformation intensity; k—the number of classes adopted a priori.

The research adopted an arbitrary division into 3 classes of countries characterized by low, medium and high levels of digital transformation of enterprises, referring to research hypothesis H₁. This approach also struck a balance between detail and clarity of interpretation, allowing the formulation of recommendations arising from the research.

Stage 3. Classification of the EU countries into groups with similar employment structures in high-tech, medium–high-tech and knowledge-intensive service sectors.

The third research stage focused on assessing hypothesis H₂: European Union countries are characterized by different types of employment structures in the high-tech, medium–high-tech and knowledge-intensive service sectors, reflecting different models of economic transformation—from industrial specialization, through a balanced structure, to the dominance of knowledge-intensive services.

Ward’s hierarchical classification method and k-means clustering were used to carry out this stage of the research. More information on the classification methods can be found in the studies [77,78,79,80,81,82,83]. This stage included the following steps:

• Using Ward’s hierarchical classification method, based on the dendrogram structure and the course of the agglomeration process, to determine the optimal number of classes best reflecting the diversity of the analyzed countries in terms of employment structure. The squared Euclidean distance was used to determine the distance between countries.
• Division of the EU countries into relatively homogeneous classes using k-means clustering.
• Determining the typology classes of the EU countries regarding the structure of employment in high-tech manufacturing, medium–high-tech manufacturing and knowledge-intensive services.

Stage 4. The identification of differences in the structure of employment in the HTM, MHTM and TKIS sectors between the groups of EU countries characterized by different levels of digital transformation.

The fourth research stage allowed verifying hypothesis H₃: Employment structures in the high- and medium–high-tech and TKIS sectors differ significantly between the groups of EU countries presenting high and low levels of digital transformation—the countries with higher levels of digitalization are characterized by a higher share of employment in the TKIS sector.

Stage 5. The assessment of relationships between the level of digital transformation advancement and the structure of employment in the HTM, MHTM and TKIS sectors between the groups of EU countries with different levels of digital transformation in the EU countries using regression analysis.

The robustness of the models was assessed in two ways: by identifying and, if necessary, excluding influential outliers based on scatterplot diagnostics, and by estimating extended regression models that included control variables such as GDP per capita and GERD per capita. These steps were undertaken to ensure the statistical reliability and substantive validity of the obtained results.

Stage 5 of the research allowed the verification of research hypothesis H₄: There is a positive relationship between the level of digital business transformation in the EU countries and the employment structure of the HTM, MHTM and TKIS sectors, with the strongest impact observed in the case of knowledge-intensive services.

The Eurostat database was the information basis of the empirical research. The research procedure was conducted using STATISTICA 13.3.

4. Results

4.1. Description of the Research Sample

The assessment of the diversity of EU countries was conducted based on the indicators monitoring the advancement of digital transformation in enterprises.

Table 1. Descriptive parameters of the digital transformation of business indicators in 2023 in the EU countries.

Parameter	X1	X2	X3	X4	X5	X6	X7	X8	X9	X10
min	27.90 RO	21.70 BG	38.00 BG	19.10 SI	17.50 BG	1.50 RO	14.80 PL	0.00 BG, LV, HU, RO, SI, SK	6.30 BG	9.40 LU
max	86.00 FI	67.30 DK	87.10 MT	53.20 HU	78.30 FI	15.20 DK	97.50 IT	65.00 DE	30.50 CZ	36.50 DK
R	58.10	45.60	49.10	34.10	60.80	13.70	82.70	65.00	24.20	27.10
$\overline{x}$	59.14	41.29	62.96	34.06	46.70	8.22	36.70	9.74	18.12	22.01
Md	59.00	41.90	60.80	35.00	46.50	7.90	29.20	3.00	17.70	20.90
CV (%)	24.14	28.73	22.00	29.73	35.00	50.36	58.74	158.21	36.73	34.90

Codes for the EU countries: Bulgaria (BG), Czech Republic (CZ), Germany (DE), Denmark (DK), Finland (FI), Hungary (HU), Italy (IT), Luxembourg (LU), Latvia (LV), Malta (MT), Poland (PL), Romania (RO), Slovenia (SI), Slovakia (SK); R—range; $\overline{x}$—arithmetic mean; Md—median; CV—coefficient of variation (%). Source: author’s compilation based on the Eurostat database.

shows the selected descriptive parameters of the digital transformation of business indicators (for the output data before normalization). Table 1 also presents information on the countries characterized by minimum or maximum values of individual indicators showing the intensity of digital transformation of enterprises.

It is worth noting that the smallest, and therefore the lowest rated, values of indicators forming the basis for assessing the advancement of digital transformation were characterized by the countries of the so-called new EU enlargement. Luxembourg is the exception in the share of enterprises selling at least 1% of their turnover online (X₁₀). In turn, the most favorable values for the majority of indicators referred to the so-called old EU countries. Only Hungary and the Czech Republic stood out in this respect regarding the % of enterprises performing data analytics (X₄) and % of enterprises’ total turnover from e-commerce sales (X₉), respectively.

Figure 1 shows the values of coefficients of variation for all analyzed digital transformation intensity indicators. Its analysis shows that the diversification of the EU countries can be considered significant. In the case of each of the analyzed transformation indicators, the coefficient of variation exceeds 20%. The EU countries are characterized by extremely high variability regarding the cumulative number of unicorns (X₈).

Next, the assessment of the EU countries’ diversity based on the structure of employment in high-tech manufacturing, medium–high-tech manufacturing and knowledge-intensive services was performed.

Table 2. Descriptive parameters of employment in high- and medium–high-technology manufacturing and knowledge-intensive services in 2023 in the EU countries (%).

Parameter	HTM	MHTM	TKISs
min	0.30	0.30	24.4
	LU	LU	RO
max	3.40	9.30	58.10
	IE	CZ	LU
R	3.10	9.00	33.7
$\overline{x}$	1.15	3.80	41.86
Md	0.90	3.30	38.90
CV (%)	57.51	67.45	18.73

HTM—high-technology manufacturing; MHTM—medium–high-technology manufacturing; TKISs—total knowledge-intensive services; LU—Luxembourg; RO—Romania; IE—Ireland; CZ—Czech Republic; R—range; $\overline{x}$—arithmetic mean; Md—median; CV—coefficient of variation (%). Source: author’s compilation based on the Eurostat database.

shows the values of descriptive parameters regarding the share of employment in high- and medium–high-technology manufacturing and knowledge-intensive services. Table 2 also presents information on the EU countries characterized by the minimum or maximum values of the share of employment in the analyzed sectors. The structure of employment in Luxembourg is characteristic and features the lowest share of employment in the high- and medium–high-tech sectors and the highest share of employment in knowledge-intensive services.

Figure 2 illustrates the clear difference in the differentiation of the European Union countries regarding the share of employment in each sector of technological advancement of the economy.

It can be observed that the EU countries are characterized by a very large and by far dominant diversification in the share of those working in the medium–high tech sector, followed by the high-tech sector. The countries also differ in the share of employment in knowledge-intensive services, but the dispersion in this area is much smaller.

The research sample covering 27 EU countries is clearly diversified both in terms of indicators showing digital intensity and the share of employment in the high-tech sectors and knowledge-intensive services.

4.2. Linear Ordering and Division of the EU Countries into the Relatively Homogeneous Classes According to the Level of Digital Transformation Intensity

This section presents findings obtained after the application of stage 2 of the research procedure, which provide the empirical basis for the verification of hypothesis H₁: EU countries differ significantly in terms of the level of digital transformation of enterprises, which allows for the identification of distinct groups characterized by high, medium and low levels of advancement.

The results of the linear ordering and division of the EU countries into relatively homogeneous classes according to the level of digital transformation intensity are presented in Figure 3,

Table 3. Classification results of the European Union countries using the linear ordering method according to the digital transformation intensity of enterprises in 2023.

Classes of Countries	Class Name	Class Composition	Class Size	Average Value of SM(DT)
Class 1	Low digital transformation intensity of enterprises	Bulgaria (0.0417), Romania (0.0741), Greece (0.2176), Poland (0.2338), Latvia (0.2526), Slovakia (0.2529)	6	0.1788
Class 2	Medium digital transformation intensity of enterprises	Czech Republic (0.3326), Estonia (0.3444), Slovenia (0.3549), France (0.3645), Portugal (0.3793), Austria (0.3812), Hungary (0.3870), Lithuania (0.4010), Cyprus (0.4028), Luxembourg (0.4234), Croatia (0.4264), Italy (0.4349), Spain (0.4943), Germany (0.5252), Ireland (0.5373)	15	0.4126
Class 3	High digital transformation intensity of enterprises	Malta (0.5860), the Netherlands (0.6591), Belgium (0.6640), Sweden (0.7442), Finland (0.7652), Denmark (0.8038)	6	0.7037

SM(DT)—synthetic measure of the level of digital transformation intensity of enterprises. Source: author’s compilation based on the Eurostat database.

and Scheme 1. The class of countries characterized by a low digitalization level of enterprises included the countries for which the synthetic measure of digital intensity level ${SM\left(DT\right)}_i$ did not exceed the value of 0.295 (cf. Formula (4)); the class presenting a medium level of digitalization covered the countries where the measure value was in the range (0.295; 0.550] (cf. Formula (5)). The countries featuring higher values of the measure of digitalization development formed a class characterized by a high digitalization level of enterprises (cf. Formula (6)).

The above calculations and presentations show that more than half of the EU Member States were qualified in the class characterized by the average intensity of digital trans-formation of small and medium-sized enterprises. This class primarily includes Western and Southwestern European countries. The most favorable position is occupied by Northern European countries. The average values of the synthetic measure of digital transformation development presented in Table 3 indicate clear differences between countries classified in the low, medium and high digital transformation development classes. The average values of SM(DT) are 0.1788, 0.4126 and 0.7037, respectively.

The next research step was to analyze the sensitivity of the synthetic measure of the digital transformation level to changes in the set of sub-indicators for digitalization. Figure 4 shows the values of Spearman’s rank correlation coefficients between the rankings of EU countries obtained after sequential elimination of individual indicators with the baseline ranking, arranged in descending order. The analysis includes normalized values of indicators X₁ to X₁₀ (Z₁–Z₁₀, cf. Formula (1)).

All Spearman’s rank correlation coefficients take values greater than 0.95, which means that the synthetic measure of digital transformation development SM(DT) is relatively stable with no dominance of any indicator of digitalization. Removing indicator X₄ (enterprises performing data analytics) had the greatest impact on the ranking, whereas removing indicator X₁ (enterprises with at least a basic level of digital intensity) exerted the least impact.

In summary, the findings confirm research hypothesis H₁, according to which the EU countries differ significantly in terms of the digital transformation level of enterprises, thus allowing the identification of distinct groups characterized by high, medium and low levels of advancement.

4.3. Classification of EU Countries into Groups with Similar Employment Structures in High-Tech, Medium–High-Tech and Knowledge-Intensive Service Sectors

This section presents the results of analyses obtained following the application of the third stage of the research procedure, which enable the empirical verification of hypothesis H₂: European Union countries are characterized by different types of employment structures in the high-tech, medium–high-tech and knowledge-intensive service sectors, reflecting different models of economic transformation—from industrial specialization, through a balanced structure, to the dominance of knowledge-intensive services.

The third research stage consisted in the classification of the EU countries based on the share of employment in high-tech sectors of the economy. The optimum number of country classes was selected based on the dendrogram of connections, integration distances and classification stages obtained as a result of applying Ward’s hierarchical classification method, using squared Euclidean distance, as presented in Figure 5a,b.

Based on the analysis of Figure 5a,b, the EU countries were classified, according to the values of high- and medium–high-technology manufacturing and knowledge-intensive services in 2023, into four relatively homogeneous classes using k-means clustering. The classification results are shown in

Table 4. Classification results of the European Union countries using the k-means method according to employment in high- and medium–high-technology manufacturing and knowledge-intensive services in 2023 in the EU countries.

Classes of Countries	Class Description (Share of Employment)	Class Composition (Distance from the Middle of the Class)	Class Size		Average Share of Employment
				SUM	HTM	MHTM	TKISs
Class 1	Average SUM Max MHTM	Czech Republic (0.0939), Germany (0.0940), Hungary (0.1636), Slovenia (0.0749), Slovakia (0.1354)	5	47.46	1.58	8.00	37.88
Class 2	Min SUM Min TKISs	Bulgaria (0.0648), Estonia (0.1134), Spain (0.0759), Croatia (0.1004), Italy (0.0735), Austria (0.1378), Poland (0.0537), Portugal (0.0853), Romania (0.2392)	9	40.80	0.92	4.08	35.80
Class 3	Average SUM Min HTM Min MHTM	Greece (0.1368), France (0.1143), Cyprus (0.0799), Latvia (0.1209), Lithuania (0.1230), Luxembourg (0.2394), the Netherlands (0.1438)	7	47.29	0.61	1.44	45.24
Class 4	Max SUM Max HTM Max TKISs	Belgium (0.1084), Denmark (0.0623), Ireland (0.3271), Malta (0.1073), Finland (0.1463), Sweden (0.2130)	6	54.69	1.77	2.62	50.30

HTM—high-technology manufacturing; MHTM—medium–high-technology manufacturing; TKISs—total knowledge-intensive services. Source: author’s compilation based on the Eurostat database using STATISTICA 13.3 package.

, Scheme 2 and Figure 6. Table 4 also provides information on the characteristics of the obtained classes of countries.

Based on the obtained classification results, the types of identified employment structures in the HTM, MHTM and TKIS sectors occurring in the European Union countries in 2023 were determined. Then, for each type of employment structure (for each of the identified classes of countries), the typological characteristics were developed.

Table 5. Typology of employment structure in high-tech, medium–high tech and knowledge-intensive services of the European Union countries.

Classes of Countries	Class Composition	Type of Employment Structure	Typological Characteristics
Class 1	Czech Republic, Germany, Hungary, Slovenia, Slovakia	Industrial and technological	High share of employment in the industrial sector (especially medium–high technology) with a significant share of employment in knowledge-intensive services
Class 2	Bulgaria, Estonia, Spain, Croatia, Italy, Austria, Poland, Portugal, Romania	Sustainable low-tech	Lack of dominance of a specific sector; relatively low share of employment in modern sectors
Class 3	Greece, France, Cyprus, Latvia, Lithuania, Luxembourg, the Netherlands	Service and innovation	High share of employment in knowledge-intensive services with a very low share in high- and medium–high-technology industries—post-industrial structure
Class 4	Belgium, Denmark, Ireland, Malta, Finland, Sweden	Service and technologically advanced	The highest share of employment in knowledge-intensive services and simultaneously in high-tech industry—a highly innovative and digitally advanced structure

Source: author’s compilation.

shows the results of the above analysis.

The findings summarized in Table 5 confirm research hypothesis H₂. In the European Union countries, different types of employment structures can be identified in the high-technology, medium–high-technology and knowledge-intensive service sectors, reflecting different models of economic transformation—from industrial specialization, through a balanced structure, to the dominance of knowledge-intensive services. Indeed, the conducted research made it possible to identify, in the EU countries, both industrial and technological structures (class 1), balanced-low-technological structures (class 2), and structures characterized by advanced development of knowledge-intensive services. The latter type of employment structure occurred in two classes of countries. Indeed, the EU countries characterized by the presence of service and innovation (class 3) and service and technologically advanced structures were identified, differing in their shares of employment in industrial sectors.

4.4. Employment Structures in High-Tech, Medium–High-Tech and Knowledge-Intensive Services Among EU Countries with Differing Intensities of Digital Transformation

The research results presented in this section allow assessing hypothesis H₃: Employment structures in the high- and medium–high-tech and TKIS sectors differ significantly between the groups of EU countries characterized by high and low levels of digital transformation—the countries featuring higher levels of digitalization have a higher share of employment in the TKIS sector.

To identify differences in the structure of employment in high-tech manufacturing, medium–high-tech manufacturing and knowledge-intensive services between groups of the EU countries presenting different levels of digital business transformation,

Table 6. Comparison of the classification of the EU countries according to the intensity of digital transformation and the structure of employment in the HTM, MHTM and TKIS sectors.

Classes of Countries	Type of Employment Structure	The Level of Intensity of Digital Transformation
		Class 1, Low	Class 2, Medium	Class 3, High
Class 1	Industrial and technological	Slovakia	Czech Republic, Slovenia, Hungary, Germany	no countries
Class 2	Sustainable low-tech	Bulgaria, Romania, Poland	Italy, Estonia, Spain, Portugal, Austria, Croatia	no countries
Class 3	Service and innovation	Greece, Latvia	France, Lithuania, Cyprus, Luxembourg	The Netherlands
Class 4	Service and technologically advanced	no countries	Ireland	Malta, Belgium, Sweden, Finland, Denmark

Source: author’s compilation.

summarizes the results of the two classifications.

The analysis of Table 6 shows that all EU countries characterized by a high level of digital transformation were classified only in the classes presenting a high and very high share of employment in knowledge-intensive services, i.e., Class 4, service and technologically advanced (five countries), and Class 3, service and innovation (one country). It can also be noted that none of the EU countries with a low level of digital transformation intensity were included in the class of countries with service and innovation structures (Class 4) characterized by the highest share of employment in both the TKIS and HTM sectors. However, this relationship is disrupted by Greece and Latvia, the countries featuring low levels of digital transformation of enterprises with service and innovation employment structures (Class 3). These countries show a high share of employment in knowledge-intensive services and very low shares in the HTM and MHTM sectors.

On this basis, it should be concluded that the results of the study only partially confirm hypothesis H₂. The structures of employment in the high- and medium–high-tech and KIS sectors differ between the groups of EU countries presenting high and low levels of cyber transformation—generally, the countries featuring higher levels of digitalization are characterized by a higher share of employment in the KIS sector. However, in this study, such a relationship was disrupted by Greece and Latvia.

4.5. Relationship Between Digital Transformation Level and Employment Share in High-Tech and Medium–High-Tech Manufacturing and Knowledge-Intensive Services in EU Countries

This section presents findings allowing the verification of hypothesis H₄: There is a positive relationship between the level of digital business transformation in the EU countries and the employment structure of the HTM, MHTM and TKIS sectors, with the strongest impact observed in the case of knowledge-intensive services.

Figure 7 presents the relationships between the level of digital transformation advancement and the share of employment total in high- and medium–high-tech manufacturing and knowledge-intensive services in the EU countries using regression analysis.

Table 7. OLS regression analysis of the relationship between digital transformation intensity (SM(DT)) and employment structures in the HTM, MHTM and TKIS sectors across the EU Member States.

Regression Parameters and Statistics	Explained Variable
	HTM	MHTM	TKISs
Intercept (β0)	0.7243	4.9316	28.4677
Standard Error (Intercept)	0.3060	1.2118	2.3980
Slope (β1)	1.0051	−2.6692	31.4750 *
Standard Error (Slope)	0.6574	2.6033	5.1513
p-Value (Slope)	0.1388	0.3150	2.19 × 10−6
R-Squared	0.0855	0.0404	0.5989
Standard Error of Regression (RMSE)	0.6460	2.5582	5.0620
Shapiro–Wilk W	0.8764	0.9478 **	0.9062 ***
Shapiro–Wilk p-Value	0.0040	0.1895	0.0186

* significant for α = 0.05, ** consistency of the distribution of the residuals with the normal distribution for α = 0.05, *** consistency of the distribution of the residuals with the normal distribution for α = 0.01. Source: author’s compilation based on the Eurostat database.

presents the results of the ordinary least squares (OLS) regression analysis, in which the level of digital transformation is the explanatory variable SM(DT) and employment shares in the HTM (high-tech manufacturing), MHTM (medium–high-tech manufacturing) and TKIS (total knowledge-intensive service) sectors are the explanatory variables.

The analysis addressing the relationship between the level of digital transformation intensity and the share of employment in the HTM, MHTM and TKIS sectors indicated significant differences between sectors. In the HTM and MHTM sectors, the regression coefficients turned out to be statistically insignificant (p = 0.1388 and p = 0.3150, respectively), and the R-squared coefficients of determination were very low (0.0855 and 0.0404). This means that there is no statistically significant linear relationship in the HTM and MHTM sectors.

The situation was different in the case of the model explaining the relationship between the level of digital transformation intensity in SMEs and the share of employment in the TKIS sector. The regression analysis results indicate a positive and statistically significant relationship (β₁ = 31.475, p = 2.19 × 10⁻⁶). The coefficient of determination R-squared = 0.5989, meaning that approx. 60% of the variation in the share of employment in knowledge-intensive services can be explained by the level of digital transformation (SM(DT)). The standard error of regression RMSE indicates that the average deviation of empirical TKIS values from those predicted by the model amounts to approx. 5 percentage points, which is 12.1% of the average value for TKISs. The regression coefficient value (β₁ = 31.475) shows that increasing the level of digital transformation intensity SM(DT) by 0.1 results in an average increase of approx. 3.15 percentage points in the share of employment in the TKIS sector of the EU countries. The normality of the distribution of residuals was assessed using the Shapiro–Wilk test. The test confirmed the normality of residuals in the MHTM model (p = 0.1895), which supports the reliability of inference for this model. In the case of the TKIS sector, the null hypothesis of normality was not rejected at the stricter α = 0.01 level (p = 0.0186), which still indicates an acceptable fit for the assumption of normality. For the HTM model, the low p-value (p = 0.0040) suggests that the normality assumption is not satisfied, which may limit the validity of parametric inference in this case.

A dot plot showing the relationship between the level of digital transformation development in the EU countries (SM(DT)) and the share of employment in the TKIS sector (Figure 7) reveals the presence of a potential outlier observation, clearly distant from the regression line due to the relatively high value of the TKIS explanatory variable, despite the average value of the SM(DT) explanatory variable. This observation was identified as Luxembourg. Such observations may affect the results of the regression analysis; therefore, a robustness analysis was performed by estimating the model excluding this observation. For comparison, the calculation results for both regressions are presented in

Table 8. Comparison of regression results of the relationship between digital transformation intensity (SM(DT)) and employment structures in TKISs with and without the outlier (Luxembourg).

Regression Parameters and Statistics	Model Including Luxembourg (n = 27)	Model Excluding Luxembourg (n = 26)
Intercept (β0)	28.4677	27.8260
Standard Error (Intercept)	2.3980	1.8522
Slope (β1)	31.4750 *	31.5091 *
Standard Error (Slope)	5.1513	3.965884
p-Value (Slope) (α = 0.05)	2.19 × 10−6	3.56 × 10−8
R-Squared	0.5989	0.7245
Adjusted R-Squared	0.5828	0.7131
Standard Error of Regression (RMSE)	5.0620	3.8971
Shapiro–Wilk W	0.9062 ***	0.9677 **
Shapiro–Wilk p-Value	0.0186	0.5636

* significant for α = 0.05, ** consistency of the distribution of the residuals with the normal distribution for α = 0.05, *** consistency of the distribution of the residuals with the normal distribution for α = 0.01. Source: author’s compilation based on the Eurostat database.

. In contrast to this case, influential outlier observations were not detected in the models for the HTM and MHTM sectors. The distribution of observations in these models did not indicate any outliers justifying their exclusion.

A comparison of the regression results for the full sample of 27 EU countries with the results obtained after eliminating Luxembourg as an outlier shows the relative stability of the regression coefficient (β₁), which was 31.4750 and 31.5091, respectively, and was statistically significant in both cases. A certain change was also recorded in the intercept (28.4677 and 27.8260). At the same time, the fit of the model improved. The R-squared value increased from 0.5989 to 0.7245, and the adjusted R-squared increased from 0.5828 to 0.7131. In addition, the consistency of the distribution of residuals with the normal distribution improved in the model without Luxembourg. According to the Shapiro–Wilk test, the residuals from the full model deviated significantly from normality (p = 0.0186), indicating potential issues with the validity of parametric inference at the standard α = 0.05 level. After the outlier was excluded, the test yielded p = 0.5636, suggesting that the residuals in the refined model do not significantly differ from a normal distribution and that the assumption of normality is met. In the model without Luxembourg, over 72% of the variation in the employment share in knowledge-intensive services can be explained by the level of digital transformation. The standard error of regression RMSE indicates that the average deviation of the empirical TKIS values from those predicted by the model amounts to approx. 3.9 percentage points, i.e., 9.45% of the average value for TKISs, and improves the predictive characteristics of the model. The findings indicate that the observed positive relationship between the level of digital transformation and employment in knowledge-intensive services (TKISs) can be considered statistically robust and resistant to the influence of outlier observations.

To increase the reliability of the regression analysis results, the original models were extended to include additional control variables:

–

GDP per capita—Gross Domestic Product per capita in PPS at current prices;

–

GERD per capita—Gross Domestic Expenditure on Research and Development per capita in PPS at constant 2005 prices.

The selection of control variables is justified by the theoretical basis indicating a possible relationship between the level of economic development and the intensity of R&D expenditure vs. the structure of employment in high-tech and knowledge-intensive sectors.

For each of the analyzed HTM (high-tech manufacturing), MHTM (medium–high-tech manufacturing) and TKIS (total knowledge-intensive service) sectors, regression models were estimated with one control variable (GDP per capita) and two control variables (GDP per capita and GERD per capita). Regarding the analyses for the TKIS sector, due to Luxembourg having been identified as an outlier in the simple regression model, the models were estimated based on the data for 26 EU Member States (excluding Luxembourg). For the HTM and MHTM sectors, the conducted analyses covered the full sample including 27 EU countries. The results of the regression analysis are presented in

Table 9. OLS regression analysis of the relationship between digital transformation intensity (SM(DT)) and employment structures in the HTM, MHTM and TKIS sectors across the EU Member States with one and two control variables.

Regression Parameters and Statistics	Explained Variable
	HTM	MHTM	TKISs a
Models with one control variable (GDP per capita)
Intercept (β0)	0.5022	6.2058	25.0549
Standard Error (Intercept)	0.3796	1.4691	2.5917
Slope SM(DT) (β1)	0.6918	−0.8720	27.2262 *
Standard Error SM(DT)	0.7298	2.8248	4.8179
p-Value SM(DT)	0.3527	0.7602	9.42 × 10−6
Slope GDP per capita (β2)	9.02 × 10−6	−5.17 × 10−5	0.0001
Standard Error GDP per capita	9.11 × 10−6	3.53 × 10−5	8.24 × 10−5
p-Value GDP per capita	0.3320	0.1553	0.1494
R-Squared	0.1214	0.1193	0.7488
Adjusted R-Squared	0.0482	0.0459	0.7270
Standard Error of Regression (RMSE)	0.6462	2.5012	3.8014
Shapiro–Wilk W	0.9347 **	0.9508 **	0.9780 **
Shapiro–Wilk p-Value	0.0901	0.2242	0.8220
Models with two control variables (GDP per capita, GERD per capita)
Intercept (β0)	0.5794	7.0894	26.1133
Standard Error (Intercept)	0.3995	1.4154	2.8631
Slope SM(DT) (β1)	0.2465	−5.9679	24.3992 *
Standard Error SM(DT)	0.9778	3.4641	5.7920
p-Value SM(DT)	0.8032	0.0984	0.0004
Slope GDP per capita (β2)	6.50 × 10−6	−8.06 × 10−5 *	8.17 × 10−5
Standard Error GDP per capita	9.90 × 10−6	3.51 × 10−5	9.48 × 10−6
p-Value GDP per capita	0.5183	0.0310	0.3981
Slope GERD per capita (β3)	0.0004	0.0047 *	0.0033
Standard Error GERD per capita	0.0006	0.0021	0.0037
p-Value GERD per capita	0.4947	0.0349	0.3838
R-Squared	0.1394	0.2773	0.7575
Adjusted R-Squared	0.0272	0.1830	0.7245
Standard Error of Regression (RMSE)	0.6533	2.3146	3.8189
Shapiro–Wilk W	0.9165 ***	0.9390 **	0.9706 **
Shapiro–Wilk p-Value	0.0325	0.1154	0.6401

a—excluding Luxembourg, * significant for α = 0.05, ** consistency of the distribution of the residuals with the normal distribution for α = 0.05, *** consistency of the distribution of the residuals with the normal distribution for α = 0.01. Source: author’s compilation based on the Eurostat database.

.

The analysis of regression models for the HTM sector showed that both in the original model and in the models with one (GDP per capita) and two control variables (GDP per capita, GERD per capita), the regression coefficient for the SM(DT) variable was positive but not significant (p > 0.1). In addition, the R-squared coefficient of determination values ranged from 0.0855 to 0.1394, indicating a very poor fit of the model to the data. The original model did not meet the assumption of normality of the distribution of residuals (p = 0.0040), but when a control variable (GDP per capita) was added, the distribution became consistent with a normal distribution at the 0.05 significance level (p = 0.0901). The model extended with two control variables did not show such consistency (p = 0.0325). Despite the improved diagnostics of the model, in both cases, no statistically significant relationship was identified between the intensity of digital transformation and the share of employment in the HTM sector.

In the case of the medium–high-tech manufacturing (MHTM) sector, no statistically significant relationship was identified between the digital transformation intensity (SM(DT)) and employment share in any of the analyzed models. In the original model, the regression coefficient for SM(DT) was negative and insignificant (p = 0.3150), and the degree of model fit was very low (R-squared = 0.0404). The distribution of residuals followed a normal distribution. Regarding the model extended with one control variable (GDP per capita), the regression coefficient for SM(DT) remained insignificant (p = 0.7602), and the R-squared increased only slightly to 0.1193. The control variable was also found to be insignificant (p = 0.1553), indicating that the level of economic development alone does not explain the variation in employment in MHTM. Nonetheless, the model remained consistent with the assumption of normality of residuals. When the model was expanded to include two control variables (GDP per capita and GERD per capita), an improvement in model fit was observed (R-squared = 0.2773; adjusted R-squared = 0.1830), while the main explanatory variable SM(DT) remained insignificant (p = 0.0984). At the same time, the coefficients at both control variables, GDP per capita (p = 0.0310) and GERD per capita (p = 0.0349), proved to be statistically significant, which may indicate their stronger impact on employment in this sector. At the same time, a negative effect of GDP per capita and a positive effect of GERD per capita on the share of employment in the MHTM sector were identified, which may indicate the complex nature of the relationship. The distribution of residuals was consistent with a normal distribution (p = 0.1154).

The analysis did not confirm a statistically significant direct impact of digital transformation on employment in the MHTM sector, but the improvement in model fit after the inclusion of control variables indicates the possibility of indirect dependencies, requiring further in-depth research.

In the case of the total knowledge-intensive service (TKIS) sector, the results of all regression analyses indicate a statistically significant and strongly positive relationship between the level of digital transformation of SM(DT) enterprises and the share of employment in this sector. Even in the original model, without control variables, the regression coefficient for the SM(DT) variable was high (β₁ = 31.4750) and statistically significant (p = 2.19 × 10⁻⁶), and the fit of the model was good (R-squared = 0.5989). However, the residuals showed consistency with the normal distribution only at the 0.01 significance level (W = 0.9062; p = 0.0186). In the model extended with one control variable (GDP per capita), the regression coefficient for the SM(DT) variable remained statistically significant (β₁ = 27.2262; p = 9.42 × 10⁻⁶), and the model fit increased to R-squared = 0.7488 (adjusted R-squared = 0.7270), indicating a very good fit quality. The control variable had no significant effect (p = 0.1494), and the distribution of residuals followed a normal distribution (p = 0.8220). In the model with two control variables (GDP per capita and GERD per capita), the significance of the digital transformation impact was confirmed (β₁ = 24.3992; p = 0.0004). Although the control variables did not reach the significance level (for GDP per capita, p = 0.3981; for GERD per capita, p = 0.3838), the model fit remained high (R-squared = 0.7575; adjusted R-squared = 0.7245). The residuals were consistent with a normal distribution (p = 0.6401).

All models confirmed the occurrence of a strong and statistically significant positive relationship between the level of digitalization and the share of employment in knowledge-intensive services. These results fully support hypothesis H₄ for the TKIS sector and indicate that digital transformation can be an important factor enhancing higher employment in this area.

Based on the conducted research, it can be concluded that hypothesis H₄ has been partially confirmed—there is a positive and significant relationship between the level of digital transformation and employment in the TKIS sector in the EU countries, whereas in the HTM and MHTM sectors, such a relationship has not been confirmed.

5. Discussion

The purpose of the study is to assess the differentiation and to classify the European Union countries in terms of the level of digital transformation advancement of enterprises and the share of employment in high-tech manufacturing (HTM), medium–high-tech manufacturing (MHTM) and total knowledge-intensive services (TKISs) and to identify the existing relationships between these phenomena.

The starting point of the conducted research was characterizing the research sample in terms of the degree of its diversity. The results clearly show a wide diversity among the EU countries in terms of both the intensity of digital transformation and the share of employment in high-tech manufacturing.

The dispersion of the EU countries based on individual criteria for assessing the intensity of digital business transformation in 2023 was very diverse (see Table 1 and Figure 1). A very strong variation and, at the same time, the largest variation occurred in the case of X₈ indicator, determining the cumulative number of unicorn companies, i.e., startups that reach a valuation of USD 1 billion and are not listed on the stock market. The coefficient of variation of this indicator exceeded the value of 158% (CV = 158.21%). In six EU countries, including Bulgaria, Latvia, Hungary, Romania, Slovenia and Slovakia, the value of this indicator was 0. The highest number of unicorn companies, i.e., as many as 65, operated in Germany. The median number of unicorns was 3, meaning that 50% of the EU countries had no more than 3 operating unicorns and 50% of countries had between 3 and 65 such companies.

There was also a strong variation in the share of enterprises using at least one AI technology (X₆) and sending e-invoices (X₇) in the analyzed countries; the values of CV were 50.36% and 58.74%, respectively. The largest number of SMEs using at least one AI technology was in Denmark (15.2%), and the smallest was in Romania (1.5%). On average, across the EU countries, approx. 8.22% of SMEs used AI technology. As for the share of enterprises sending e-invoices, the largest was in Italy (approx. 97.5%), and the smallest was in Poland (14.8%). On average, in the EU countries, more than 36.7% of SMEs use e-invoices.

A very small variation occurred for SMEs with at least a basic level of digital intensity (X₁) and using two or more social media (X₃) (CV = 24.14% and CV = 22.00%, respectively). On average, about 60% of SMEs in the EU countries present at least a basic level of transformation and use at least two social media in their operations. The minimum values of these indicators were recorded in Romania (27.9%) and Bulgaria (38%), respectively, whereas the highest were recorded in Finland (86%) and Malta (87.10%).

The variation among EU countries in terms of other indicators can be considered average (CV took values from approx. 28% to 35%). It can be observed that, taking into account the specific indicators of the digital transformation intensity, their minimum values were most often recorded in Bulgaria (five times), followed by Romania (three times). In turn, the highest value of indicators was most often observed in Denmark (three times), followed by Finland (twice).

The differentiation assessment of the EU countries regarding the share of employment in the analyzed sectors of technological intensity was carried out based on the results presented in Table 2 and Figure 2. The variation among EU countries in terms of the share of employment in the HTM, MHTM and TKIS sectors was low for the knowledge-intensive service sector (CV = 18.73%) and high for the HTM (CV = 57.51%) and MHTM (CV = 67.45%) sectors. In the case of high- and medium–high-tech manufacturing, by far the smallest share of employment was recorded in Luxembourg, amounting to 0.30% in each of the sectors. The maximum share of employment in high-tech manufacturing was in Ireland (3.4%), and that in medium–high-tech manufacturing was in the Czech Republic (9.3%). On average, the EU countries had an employment share of approx. 1.15% in high-tech manufacturing and 3.8% in medium–high-tech manufacturing. The EU countries were characterized by a significantly higher average share of employment in knowledge-intensive services, amounting to approx. 41.86%. In this sector, the dominant share of employment occurred in Luxembourg (58.10%), whereas the smallest was in Romania (24.4%). In half of the EU countries, the share of employment in knowledge-intensive services exceeded 38.9%.

Based on the above considerations, it can be concluded that the research sample, consisting of 27 European Union countries, used in the empirical study shows a clear diversity in terms of the level of digital transformation and the structure of employment in high-tech sectors. This diversity makes the sample suitable for carrying out the research objective and allows for a reliable and credible verification of the adopted research hypotheses.

Hypothesis H₁ assumes that the EU countries differ significantly in terms of the digital transformation level of enterprises, which enables the identification of distinct groups characterized by high, medium and low levels of advancement. The analysis conducted based on the synthetic measure value of the digital transformation level (SM(DT)) in 2023 allowed the identification of three classes of EU Member States characterized by clearly diversified levels of digital transformation: the classes of countries with low, medium and high levels of advancement (cf. Figure 3, Table 3 and Scheme 1). The Nordic countries—Denmark, Finland and Sweden—were the leaders in this respect. The class of countries characterized by the highest level of digital transformation also included Belgium, the Netherlands and Malta. By far the lowest levels of digital business transformation were recorded in Bulgaria and Romania. These countries can be considered the most lagging in terms of the degree of digital transformation process implementation in small and medium-sized enterprises in the European Union. Greece, Poland, Latvia and Slovakia were also included in this class. The remaining EU countries were placed in the most numerous class of countries (15 countries), presenting an average level of digital transformation. The classes of countries characterized by high and low digital transformation intensity were equal-sized and included six EU Member States each. The findings confirm hypothesis H₁ and indicate a significant diversification among the EU countries in terms of the digital transformation level of enterprises. The identification of separate groups of countries with low, medium and high levels of advancement confirms the occurrence of such a diversification, which is partly in line with earlier studies carried out by the European Commission indicating persistent regional differences in the level of digitalization of the economy and a clear digital division in the EU between the countries of the “old” and “new” EU. The obtained results are consistent with the findings of the European Commission’s DESI reports, which confirm persistent differences in the level of digitalization presented by enterprises in the EU countries. Denmark, Finland and Sweden regularly score highest in terms of digital integration, while Bulgaria and Romania remain at the bottom of the list, reflecting the persistent digital division [84,85]. The above findings are also consistent with broader research addressing the digital maturity of EU countries, which points to the Nordic countries not only leading the way in terms of digital infrastructure, governance and digital competence, but also offering particularly favorable conditions for digital transformation across all sectors [86]. The aforementioned studies do not directly refer to small and medium-sized enterprises, but allow the conclusion that the generally developed digital environment and high digital skills of the population create favorable conditions for achieving high digital intensity of all enterprises. In turn, countries such as Bulgaria, Romania, Slovakia and Greece have been identified as requiring comprehensive support, especially in the development of digital infrastructure and digital education, which also confirms the consistency of results with this study [86]. These results also provide evidence of the need for a differentiated approach in the EU digital policy that takes into account the existing differences in the digital development of small and medium-sized enterprises across the EU countries.

Hypothesis H₂ assumed the occurrence of different types of employment structures in the high-technology, medium–high-technology and knowledge-intensive service sectors in the European Union countries, reflecting different models of economic transformation—from industrial specialization, through balanced structure, to the dominance of knowledge-intensive services. The conducted research showed that the structures of employment in the HTM, MHTM and TKIS sectors showed significant differentiation in small and medium-sized enterprises operating in the EU countries. In 2023, the division of the EU countries into four classes was found to be optimal in terms of the share of employment in the HTM, MHTM and TKIS sectors (cf. Table 4, Scheme 2 and Table 5). On average, all the identified classes of countries are characterized by the highest share of employment in knowledge-intensive services and the lowest share in high-tech manufacturing. In turn, the particular classes differ both in the total share of employment in the three analyzed sectors of the economy and in the relationships between the shares of employment in each of the sectors. Class 1 of the EU countries, representing the type of industrial and technological employment structure, is characterized by an above-average total share of employment in all analyzed sectors (47.46%) and the largest share of employment in medium–high-tech manufacturing among all identified classes (8.00%). It is the smallest class, including five countries: the Czech Republic, Germany, Hungary, Slovenia and Slovakia. Hungary and Slovakia are the most distant from the center of the class. On the map, these countries are concentrated in Central Europe.

Class 2, defined as a type of sustainable low-tech employment structure, covers the countries presenting the smallest average total share of employment in all surveyed sectors (40.80%) and in knowledge-intensive services (35.80%). This structure lacks dominance of any modern sector. It is the most numerous class, including nine EU countries. The class comprises both the countries of the so-called old EU 15, such as Spain, Italy, Portugal and Austria, and the countries of the new EU enlargement, including Bulgaria, Romania, Estonia, Croatia and Poland. Romania turned out to be the furthest from the center of the class. Geographically, the countries included in this class are located in Southern and Central–Eastern Europe.

Class 3, characterized by a service and innovation employment structure, contains countries with the highest share of employment in the surveyed sectors (similar to Class 1), amounting to 47.29%, and the lowest share of employment in both medium–high manufacturing (0.61%) and high-tech manufacturing (1.44%). This class consists of seven countries, including Greece, France, Cyprus, Latvia, Lithuania, Luxembourg and the Netherlands. Luxembourg is placed by far the furthest from the center of the class. On the map, these countries are located in Western and Southern Europe.

Class 4 covers the EU countries characterized by the service and technological employment structure, where the average share of employment is the highest (among all the identified classes) both in all the analyzed sectors combined (54.69%) and in high-tech manufacturing (1.77%) and knowledge-intensive services (50.30%). The employment structures in this class of countries are the most innovative and digitally advanced. The following six countries were included in this class: Belgium, Denmark, Ireland, Malta, Finland and Sweden, located mainly in Northern and Western Europe. Sweden is the country most distant from the center of the class.

The findings confirm hypothesis H₂ and indicate the occurrence of four types of employment structures in the EU countries, reflecting various economic transformation strategies and different stages of their implementation. There are typologically diverse employment structures, ranging from the least modern low-tech sustainable structures, through industrial and technological as well as service and innovation structures, to the most digitally advanced service and technological structures, characteristic of knowledge economies. In addition, the geographic distribution of the classes indicates clear regional differences: the countries located in Central Europe dominate the industrial and technological class, the countries in the south and east of the EU dominate the sustainable low-technological class, while the countries in the north and west of the EU are the leaders of the most advanced service and technological structures (see Scheme 2).

Hypothesis H₃ states that employment structures in the high-tech, medium–high-tech and TKIS sectors differ significantly between the groups of EU countries characterized by high and low levels of digital transformation—the countries featuring higher levels of digitalization have a higher share of employment in the KIS sector. The conducted research allowed a comparison of the classification of the European Union countries by the level of digital transformation with the types of employment structures in the HTM, MHTM and TKIS sectors (see Table 6). The analysis shows that all the countries classified as those featuring a high level of digital transformation were assigned only to the classes characterized by a high or very high share of employment in knowledge-intensive services (TKISs), i.e., the service and technologically advanced class (Malta, Belgium, Denmark, Finland and Sweden) or the service and innovation class (the Netherlands). In turn, none of the countries with a low level of digital transformation intensity were included in the service and technologically advanced class, typical of the economies characterized by the highest share of employment in knowledge-intensive services and high-technology industries (see Table 4). Diversification was observed in the countries covered by the class featuring a medium level of digital transformation, with all the distinguished types of employment structures present there.

Significantly, the analyzed data showed deviations from the adopted relationship. Greece and Latvia, despite belonging to the group of countries with a low level of digitalization, were classified in the class of service and innovation structures, characterized by a high share of employment in the TKIS sector.

The findings partially confirm hypothesis H₃. Employment structures in the HTM, MHTM and TKIS sectors differ between the classes of countries presenting high and low levels of digitalization, generally; however, the countries with higher levels of digital transformation are characterized by a higher share of employment in knowledge-intensive services. The identified deviations indicate that the level of digital transformation is not the only factor shaping the structure of employment in the high-tech and knowledge-intensive service sectors. The results of conducted studies confirm a moderate correlation between the intensity of digital transformation and the structure of employment in high-tech manufacturing and knowledge-intensive services. Particularly clear relationships were observed in the countries characterized by a high level of digitalization, where politically and economically supported digital transformation processes contribute to employment growth in knowledge- and technology-intensive sectors, primarily in knowledge-intensive services. In the countries featuring medium levels of digital advancement, these relationships were less consistent, which may result from different national strategies, diverse structures of the economy and an asymmetric pace of digital implementation.

The countries featuring the most advanced digital development often invest in cutting-edge technology, R&D and education, which can have a direct impact on a higher share of employment in high-tech manufacturing (HTM, TKISs). Digital transformation is strategically supported there both politically and economically. Finland is an example of a country with extremely high efficiency in digital transformation owing to the implementation of the Digivisio 2030 digital development program and the Artificial Intelligence 4.0 program [32], whereas Poland, e.g., belonging to the group of countries that present a low level of digital transformation, is currently at the stage of intensive work focused on developing a strategy for the country’s digitalization by 2035 [27].

Countries with medium levels of digitalization may be more diverse in terms of digitalization policies, economic structure and development priorities; therefore, their employment structure does not directly reflect the level of digital transformation. This can also result from an asymmetric pace of technology implementation or different starting points. Moderate correlation may also stem from the fact that digital transformation does not always immediately translate into changes in the employment structure; it often takes time for the effects of investments in digitalization to be reflected in the labor market.

Hypothesis H₄ assumed the occurrence of a positive relationship between the level of digital business transformation in the EU countries and the structure of employment in the HTM, MHTM and KIS sectors, with the strongest impact expected in knowledge-intensive services. The results of the conducted OLS regression analysis indicate a diversified relationship between the level of digital transformation and the share of employment in each sector (see Figure 7 and Table 7). In the HTM and MHTM sectors, the regression coefficients proved statistically insignificant (p = 0.1388 and p = 0.3150). The calculated coefficients of determination took on very low values (R-squared = 0.0855 and 0.0404). This means that in the high- and medium–high-technology sectors, a significant linear relationship between the digital transformation of enterprises and the structure of employment in the European Union countries was not confirmed.

In the case of knowledge-intensive services, the results of regression analysis confirmed the statistically significant positive relationship between the level of digital transformation and the share of employment in knowledge-intensive services (β₁ = 31.4750; p < 0.001). The R-squared coefficient of determination value, amounting to 0.5989, means that approx. 60% of the variation in the share of employment in TKISs can be explained by the level of digital transformation intensity. Robustness analysis, carried out by excluding an outlier observation (Luxembourg), showed stability in the regression coefficient and improved model characteristics (see Table 8). In the model without Luxembourg, the value of the R-squared coefficient of determination increased to 0.7245, and the distribution of residuals was found to follow a normal distribution at the significance level of α = 0.05. The standard error of the regression was reduced, which translated into better predictive characteristics of the model. The findings indicate that the positive relationship between the level of digital transformation and employment in the TKIS sector is statistically robust and resistant to the influence of outlier observations, hence strengthening its reliability.

To confirm the reliability of findings, additional analyses were carried out by including control variables: GDP per capita and GERD per capita. The results of the extended regression models with control variables (Table 9) provided additional information allowing a better assessment of the strength and nature of the relationship between the level of digital transformation and the structure of employment in the analyzed sectors.

For the HTM sector, none of the models showed statistically significant relationships, and the level of model fit remained low. The variable SM(DT), as well as GDP per capita and GERD per capita, proved to be statistically insignificant. This means that the relationship between the level of digitalization and employment in this sector was not confirmed.

In the MHTM sector, in turn, despite an increase in model fit after the inclusion of control variables, the regression coefficient for the SM(DT) variable remained statistically insignificant. Instead, a significant negative effect of GDP per capita and a positive effect of GERD per capita on employment were recorded in this sector. This may suggest that digital transformation affects employment indirectly or in interaction with other factors. However, such an interpretation requires further in-depth analysis.

In the TKIS sector, the relationship between the level of digital transformation and the share of employment maintained statistical significance in all models, and model fit improved after the inclusion of GDP per capita and GERD per capita. In both extended models, the control variables (GDP per capita and GERD per capita) were not statistically significant, and their inclusion did not affect the significance of the main variable. The increase in model fit and positive results of normality tests of the residuals (Shapiro–Wilk) confirm the stability and robustness of the original model.

In summary, a positive and statistically significant relationship was confirmed only for the knowledge-intensive service sector. In the case of the medium–high-tech industry sector, the results are inconclusive, and for the high-tech industry sector, no empirical evidence of such a relationship was found.

Thus, it can be concluded that hypothesis H₄ has been partially confirmed. There is a significant and positive relationship between the level of digital transformation and the structure of employment in knowledge-intensive services, while in high- and medium–high-tech industries, such a relationship has not been confirmed. This means that digital transformation processes are more conducive to the development of knowledge-intensive services than high- and medium–high-tech industries.

The lack of significant correlations for the other sectors suggests that the digital transformation of small and medium-sized enterprises in a short-term perspective does not translate directly into employment changes in technology sectors. It should be highlighted that knowledge-intensive services may be more directly linked to digital transformation than high- and medium–high-tech manufacturing, whose growth depends on other factors too, such as capital investment or the development of R&D infrastructure. Digital transformation has a greater impact on increasing the share of employment in knowledge-intensive services than in high- and medium-tech manufacturing, because knowledge-intensive services have a more extensive potential for transformation in terms of the provision and scope of offered services owing to digital technologies. In knowledge-intensive services, such as digital education, financial services, consulting or IT, digitalization enables the formation of new industries and markets [32,37]. In the “data-driven” sectors, including media and entertainment services and financial services, relying most heavily on the Internet of All Things, i.e., networks connecting people, objects, data and processes for the digital exchange of value, digital transformation enables offering new services [37]. The digitalization of services allows new markets and customers to be reached, which increases the demand for workers in this sector. In knowledge-intensive services, employees need to be competent in using digital technologies to provide services, analyze data, communicate with customers online, etc. As digital transformation keeps advancing, the demand for professionals presenting these skills is also increasing, contributing to employment growth in the sector [27,32].

In contrast, high- and medium–high-tech manufacturing, as the already technologically advanced sectors, may experience transformation focused on optimization and automation of production, which does not necessarily lead to a proportionally higher employment growth compared to the service sector intensely adapting these technologies to its operations [32]. In addition, digital transformation in knowledge-intensive services affects the way work is organized and the employment profile more quickly than in industry, where changes are often more costly and complicated.

Previous studies describe digital transformation as a gradual process, involving automation and requiring digital maturity [2,27,28,37]. These aspects suggest that changes in employment are likely to occur with some delay from the implementation of digital technologies, as the full effects of transformation, including its impact on employment structure, manifest themselves as the process progresses and are spread out over time. Thus, a time gap may occur between the implementation of new technologies and their impact on the employment structure. Hence, it would be necessary to monitor the analyzed relationship over a longer period of time. Currently, the lack of statistical information is a limitation for conducting such research.

Despite drawing important conclusions, several key limitations of the conducted research, which affect the scope and interpretation of the findings, should be mentioned. One of them is the short time frame of available statistical information on the digital transformation rates of small and medium-sized enterprises. This limited the ability to capture the dynamic changes and the long-term impact of digitalization on the employment structure. Another limitation of the presented research is the narrow focus on the small and medium-sized enterprise sector. This choice was dictated both by practical considerations, including the availability of statistical data at the level of European Union countries, and by the substantive intention to focus the analysis on entities that, due to their scale of operation, may react differently to digital transformation processes than large enterprises. In addition, the empirical study was based on quantitative data, whereas a better understanding of the mechanisms influencing the impact of the level of digital transformation on the structure of employment in high- and medium–high-tech manufacturing and knowledge-intensive services could be achieved through qualitative research (e.g., interviews with entrepreneurs, case studies).

Identifying the reasons underlying the occurrence of relationships between the level of digital transformation and the structure of employment in high-tech and medium–high-tech manufacturing and knowledge-intensive services requires further in-depth research. The current results allow identifying general correlations, but due to their limitations, they do not fully explain the mechanisms behind the observed differences, especially in the group of countries characterized by a medium level of digital advancement. It is therefore reasonable to continue and deepen studies covering this area—in particular, long-term studies capturing the lagged impact of digitalization processes on the structure of employment, as well as analyses integrated with other aspects of the transformation—such as digital skills, infrastructure and the quality of digital public services. Then it will be possible to fully understand the dynamics and consequences of digital transformation in the social, economic and political context.

In the future, it is worth undertaking spatial–temporal analyses to capture the regional and dynamic dimensions of digital transformation. Expanding the methodological scope to include multivariate regression, panel models and more advanced estimation techniques would allow better identification and interpretation of the complex relationships between digital transformation and employment in high-tech sectors and knowledge-intensive services.

It may also be a valuable research direction to combine the results of empirical analyses with the effects of implementing specific European Union initiatives and strategies on digital transformation and employment development in high-tech and knowledge-intensive sectors. Such an approach will enable a deeper interpretation of the research results and make them more useful in the context of public policies.

Future research should also consider expanding the scope of analysis to include the large enterprise sector, thus allowing a comparison of the dynamics of digital transformation and its effects according to the scale of companies’ operations. Such an approach could also enable the identification of good practices in the implementation of digitalization to be adapted by smaller entities.

6. Conclusions

The following conclusions can be formulated based on the conducted research:

• The level of digital transformation of enterprises in the European Union countries is clearly diversified, which allows three classes of countries characterized by low, medium and high levels of digitalization advancement to be distinguished. The identification of these differences provides a valuable rationale for the formulation of public policies aimed at supporting less digitally advanced countries in the processes of economic transformation and digital integration within the EU single market.
• The results of the conducted analyses confirmed the significant diversification of employment structures in the high-technology (HTM), medium–high-technology (MHTM) and knowledge-intensive service (TKIS) sectors in the European Union Member States. Four types of employment structures were distinguished, reflecting different economic transformation strategies and stages of their implementation—from the low-technology sustainable structures, through industrial and technological and service and innovation structures, to the most modern service and technology structures. The classification also indicated the occurrence of regional differences—the countries of Central and Eastern as well as Southern Europe dominate the lower technologically advanced classes, whereas Northern European countries have structures characterized by the highest share of employment in both knowledge-intensive services and high-tech industries. The results indicate the need for a differentiated approach in policies supporting the development of the digital economy, taking into account the specificity of employment structures in individual EU countries.
• The research has partially confirmed that employment structures in the high- and medium–high-tech and knowledge-intensive service (TKIS) sectors differ between the groups of EU countries presenting different levels of digital transformation. In general, the countries featuring higher levels of digitalization were characterized by a higher share of employment in the TKIS sector. However, the observed deviations highlight the need to take into account additional factors, such as the specificity of the national economy structure or historical determinants.
• The European Union countries are characterized by a significant positive relationship between the digital transformation level of small and medium-sized enterprises and the share of employment in knowledge-intensive services (TKISs), while no such relationship was confirmed in the high- and medium–high-technology sectors (HTM and MHTM). The results emphasize the growing role of knowledge-intensive sectors in the digital transformation processes of the EU economies.
• The research results confirmed a clear and stable positive relationship between the level of digital transformation of small and medium-sized enterprises and higher employment in knowledge-intensive services (TKISs) in the European Union countries. This highlights the growing importance of these services in the digital economy. However, regarding the medium–high technology manufacturing (MHTM) industry, the relationship was inconclusive. In turn, for high-technology manufacturing (HTM), no relationship was found between the level of digital transformation and employment structure. This means that in the case of this sector, other factors may play a greater role than digital transformation.

The results of the conducted research provide important conclusions for both economic practice and public policy. They constitute an important source of knowledge for policymakers, institutions supporting entrepreneurial development and individual businesses. They can serve as a reference for designing public policies aimed at enhancing digital transformation, especially in countries lagging behind in its implementation. In addition, the data can help educational and R&D institutions determine investment and educational priorities that respond to the changing needs of the labor market.

In light of the research, it is clear that further support for the digital transformation of small and medium-sized enterprises is indispensable, both at the national and the EU level. At the national level, these activities should respond to the local needs for the development of digital infrastructure, competence and innovation. At the European Union level, there is a need for a differentiated approach in the EU digital policy, taking into account the existing differences in the digital development of small and medium-sized enterprises in the individual EU countries. This need is further substantiated by empirical research indicating that digital transformation has a positive and multidimensional impact on business functioning by improving performance, stimulating innovation, increasing operational efficiency and enhancing long-term competitiveness [87]. These benefits can turn out particularly important for SMEs, which frequently operate with limited resources and require targeted support to fully exploit the potential of digital technologies.

It is necessary to continue active political, economic and financial support for the digital transformation of Member States currently reflected in both strategic documents and development programs, such as the Digital Agenda for Europe [88], the Europe’s Digital Decade: Digital Targets for 2030 [89] and Digital Europe [90]. These initiatives not only promote innovation and the development of digital infrastructure but also support the development of digital competencies and the transformation of public and private sectors, including small and medium-sized enterprises.

The importance of digital transformation and its impact on the structure of employment has a global dimension as well—it plays an important role in achieving the Sustainable Development Goals. Therefore, the intensification of digitalization efforts should be perceived as an essential component in the implementation of the sustainable development agenda, in which digital transformation becomes not only a determinant of competitiveness, but also a factor of social integration and sustainable growth. The unfolding digital transformation is not only changing the way we do business, but also shaping new career paths, models of work and social relationships; therefore, its continued monitoring and support remain crucial for sustainable economic and social development across the European Union.