Abstract
OBJECTIVE:
This study aims to determine the effects of the DESI (connectivity, human capital/digital skills, use of internet service by citizens, integration of digital technology and digital public services) on labor market indicators (labor market insecurity, long-term unemployment rate, employment rate, and personal earnings).
METHODS:
In the study, the 2018 DESI data, the 2018 Better Life Index (job) data and 23 EU countries were analyzed. In the analysis, the SmartPLS 3.0 program was executed with 23 cases and 5000 bootstraps.
RESULTS:
It was concluded that an increase in the DESI has increased employment rate and personal earnings, which are positive labor market indicators (t: 15,929; 0,849) and has decreased long-term unemployment rate and labor market insecurity, which are negative labor market indicators (t: 3,685; –0,535).
CONCLUSIONS:
As a result, digitalization in EU countries for the year 2018 has proved that the labor market indicators were improving.
Introduction
One of the leading concerns in society after the Industrial Revolution was that use of machines in factories and businesses could lead to the loss of jobs for the mass populous. However, this was not the case; while some professions were lost over time, new occupations replaced them. A similar criticism is also being discussed in literature today. According to this thought, technological transformation could lead to the loss of some professions and leave millions or even billions of people unemployed. However, it is likely that a transformation similar to what happened during the Industrial Revolution will occur, the creation of more jobs will happen after the loss of some.
The concept of technological transformation, which has perturbed billions of people, has now gone beyond the concept of the differentiation of machinery and equipment and has started to be discussed as the digitization of everything [1, 2]. As a matter of fact, digitalization has now manifested itself in all areas of industries. A very wide range of matters, from the use of social media by individuals to digital literacy, from companies carrying all their assets and brands to digital media and public services being transferred into electronic environments, has been associated with the concept of information technologies or the concept of digital transformation. In other words, transformation is no longer due to technology or computers, but to a more integral perspective of digitalization itself. Digitalization manifests itself in all areas of life and continues to expand in all areas, from the proliferation of mobile devices to the delivery of the internet to wider audiences, from e-commerce applications to e-government services. In this context, digitalization is regarded as a dynamic that has a significant and widespread effect transforming the individual, society, state, and economy.
In this context, the aim of this study was to determine the effects of the Digital Economy and the Society Index (DESI) on labor market indicators for 23 EU countries in the year of 2018. In the first part, the definition of the Digital Economy and Society Index is given; in the second part, the possible effects of digital transformation on labor markets are discussed; and then the method and findings of the study are given.
Digital economy and society index (DESI)
The digital economy, defined as the use of information technologies by the public, business, and society, is massive, encompassing states, societies, businesses, and individuals [1]. A digital society is the transformation of society’s lifestyle into a digital one through the increased use of digital technologies; or in other words, the transformation of a traditional lifestyle into a digital lifestyle [2]. The increase in the internet usage, the expansion of information technologies, the companies’ increased use of e-commerce, and the digital transformation in public services raises the question of how much today’s societies will accept and keep up with digitalization. To answer this question, the European Union has established the DESI.
The DESI is a composite index of 5 dimensions that summarizes indicators related to the digital performance of European Union member states and monitors the level of digital competitiveness of countries [3]. These 5 sub-dimensions are connectivity, human capital, use of internet services, integration of digital technology and digital public service.
The first sub-dimension of the DESI is connectivity. Connectivity itself consists of 5 sub-dimensions. These are fixed broadband, mobile broadband, fast broadband, ultrafast broadband and broadband price index. Connectivity dimension looks at both the demand and the supply side of fixed and mobile broadband and also the price [3, 4]. Connectivity represents the assessment of how connected a country is to the internet. While the increase in this index value indicates that a larger segment of society is able to connect to the internet, the decrease in this value indicates that access to the internet in society remains low due to factors such as technical infrastructure or cost. The results reveal that the country with the highest value for 2018 is Finland and the country with the lowest value is Bulgaria.
The second sub-dimension of the DESI is human capital. This dimension consists of 2 sub-indicators; internet user skills and advanced skills and development. The first indicator, internet user skills, is calculated based on the number and complexity of activities involving the use of digital devices and the internet. The second, advanced skills and development, includes indicators related to the employment of information technology specialists and information technology graduates [4, 5]. An increase in this index value indicates an increase in digital skills in society and the quality of information technology experts and the education provided to them, whereas a decrease in this value indicates low digital skills and expert employment due to reasons such as a lack of need for the internet, a lack of talent and high internet costs. The results report that the country with the highest value for 2018 is Finland and the country with the lowest value is Bulgaria.
The third sub-dimension of the DESI is the use of internet services. This dimension is measured with 3 sub-indicators. The first is internet use. This ratio shows what portion of society are internet users. The second is activities online, and this indicator shows the ratio of the society that has used online activities, such as news, social networking and shopping. The third indicator is transactions. This ratio shows the portion of society engaged in online banking transactions [4, 6]. The increase of this index indicates that society’s multi-purpose internet usage practices are increasing, and the decrease of this value indicates that society’s internet usage practices are not sufficiently developed, perhaps because there is a problem with trust in online systems. The results report that the country with the highest value for 2018 is Denmark and the country with the lowest value is Romania.
The forth sub-dimension of the DESI is integration of digital technology. This dimension consists of 2 sub-indicators. The first indicator is the business digitization. This ratio shows the ratio of initiatives using electronic information sharing, social media use, big data analysis, and cloud solutions. The second indicator is e-commerce. This ratio shows the proportion of the small and medium-sized businesses which sell online, how much of the small and medium-sized businesses sales consist of online sales, and finally how much of the small and medium-sized businesses sales are derived from international online trade [4, 7]. A rise in this index value indicates that businesses are successful in their digital transformation and that digital transformation has an economic equivalent, while a decrease in this value indicates that businesses are less successful in achieving a digital transformation and transferring their assets to a digital economy. The results report that the country with the highest value for 2018 is Ireland and the country with the lowest value is Bulgaria.
The final sub-dimension of the DESI is digital public services. This dimension consists of 2 sub-indicators. The first of these indicators is e-government. This rate indicates the proportion of the documents submitted to the government online, the proportion of the forms used in government transactions arriving completed (digitally integrated with the information of the individuals), how much of the public services can be carried out online, how much of the public services are available online for companies and the level of digital open data sharing of the public sector. The second indicator is e-health. This rate shows the rate of health services that can be obtained without going to a doctor or health institution, the sharing of health records with other health service providers and the prevalence of the e-prescription application [4, 8]. An increase in this index value indicates that a significant portion of public services are transferred to digital channels, and a decrease in this value indicates that the public sector still provides public services through traditional platforms. The results report that the country with the highest value for 2018 is Finland and the country with the lowest value is Bulgaria.
The DESI, which includes all indicators, measures a wide range of digital skills, from being connected to the internet to digital employment and literacy on the one hand, to the active use of the internet by individuals, companies and the public on the other hand. In this context, the increase of the DESI value indicates a rising digitalization level of a country both economically and as a society; while the decrease indicates a decline in the level of digital competence and competition of individuals, companies and the public. The results show that Finland has the highest level of digital competitiveness in 2018 and Bulgaria has the lowest level of digital competitiveness [6, 8].
Effect of digitalization on labor market
As can be seen, digitalization has a close impact on labor markets. With this ever changing condition, the skills available in the labor markets have become differentiated. As a natural result of this, the competencies of the future have differentiated from those of the present, and thus digitization of everything emerges. Accordingly, the concept of the new industry has led to the emergence of many new competencies, and almost all of the new competencies needed are considered in relation to digital skills [9–15].
This change, of course, re-opens the debate on the impact of a technological transformation on labor markets, a classical topic that has been discussed for more than a hundred years. This seemingly new topic has actually been debated since the 1800s [16]. The first opinion in this sense was put forward by Jean-Baptiste Say. Say stated that some occupations may be lost with technological change, but in their place, new occupations will appear, with increasing technology, the unit cost of products falls, which is reflected as a decrease in product and service prices, and, accordingly, the increase in demand and employment increase as a result for the labor markets [17]. A similar opinion has been put forward by David Ricardo. Ricardo stated that in the event that there is no sudden drop in product prices, costs can be reduced through technological progress, which can in turn increase profits and investments, and as a result employment growth is possible for labor markets [18]. As a matter of fact, Joseph Schumpeter believed that new jobs will emerge with technological change [19]. But then economists, such as John Maynard Keynes and Wassily Leontief, have argued that machines can replace humans in parallel with technological changes, which may create technological unemployment [20]. Indeed, a report by Acemoğlu & Restrepo in 2017 supported a similar finding [21]. Accordingly, the use of technology can leave people, who do basic jobs, unemployed and lower their income. As can be seen, there are different views on the effects of technology on labor markets.
The areas discussed above do not explain digital transformation but shed light on the technological transformation of two centuries ago. The technological transformation, which occurred two centuries ago, was considered to be innovative in both machinery and equipment, and today presents itself as digitalization. Therefore, today’s technological transformation refers to digital transformation, and the subject of this study is the reflection of the digitalization of society on labor markets.
When examined in chronological order, van Reenen [22], Regev [23], Greenan & Guellec [24], Piva & Vivarelli [25], Yang & Lin [26], Coad & Rao [27] have revealed that different technological developments for different countries and dates have shown a positive effect on employment. The effects of digital transformation, which demonstrates today’s technological change on labor markets, are similar. For example, estimates by the World Economic Forum [28] suggest that the number of jobs will increase in the future with digital transformation. In fact, studies in the literature suggest that the digitalization of the economy and society will have a positive impact on employment and earnings [29–34]. Therefore, the following hypothesis has been proposed.
H1: The higher the Digital Economy and Society Index, the higher positive employment indicators will be.
In addition, studies in literature indicate that digital competencies are very important in the future and that the employment of individuals providing digital competencies will be easily obtained and the employment of individuals, who do not provide them, will be difficult [9, 35–40]. The following hypothesis has therefore been put forward.
H2: The higher the Digital Economy and Society Index, the lower negative employment indicators will be.
Method
SmartPLS 3.0 structural equation modeling program was used to analyze the data [41]. This is a preferred structural equation modeling analysis program, where the sample size is limited, the number of variables (items) constituting the factors is less than 3, and the normal distribution of data is not expected. In the models predicted by SmartPLS, goodness of fit criteria are not sought for the suitability of the model; instead, item reliability, structure reliability, common variance, square root of common variance and “t” values are used [42, 43]. In addition, there are also studies where the structural equality modeling technique, which is thought to be used more in the analysis of micro-data, is used in the analysis of macro-data [44].
The DESI data required to carry out the analyses were obtained from the policy unit of the European Commission (https://ec.europa.eu/digital-single-market/en/desi). Labor market indicators were taken from the job indicators in the OECD’s Better Life Index database, both positive and negative (https://stats.oecd.org/). While the DESI contains data on 28 countries, as the OECD database only contains data on labor market indicators for 23 EU member states, this model is designed for 23 EU member states. It is suggested that the suitable sample size for SmartPLS is between 30–100 [45, 46]. However, in some studies, it is recommended that between 20–30 sample size is also accepted [47]. The aforementioned countries in the model are; Austria, Belgium, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Netherlands, Poland, Portugal, Slovakia, Slovenia, Spain, Sweden and the UK. The variables used in the model can be examined in Table 1.
Variables Subjected to Analysis
Variables Subjected to Analysis
Table 1 shows the factors and indicators subject to analysis. Accordingly, the Digital Economy and Society Index (DESI) was measured with 5 sub-dimensions. These are; connectivity, digital public services, human capital, integration of digital technology skills and use of internet services by citizens. The definitions for these variables are given above. The indicators of the labor market are evaluated in two dimensions as negative employment indicators and positive employment indicators. The negative employment indicator factors consist of two sub-dimensions. The first factor is the labor market insecurity. Labor market insecurity is defined in terms of the expected earnings loss associated with unemployment and is measured as a percentage. An increase in labor market insecurity shows that the confidence in labor markets is low. According to the results, Luxembourg has the lowest rate with 1.7% and Greece has the highest rate with 29.8%. The second factor is long-term unemployment. This indicator refers to the number of persons who have been unemployed for one year or more as a percentage of the labor force and is measured as a percentage. An increase in this rate shows that the rate of long-term unemployment in the country has increased. According to the results, the lowest long-term unemployment rate is in the Czech Republic with 1.04%, while the highest long-term unemployment rate is in Greece with 15.65%. This factor is designed to measure negative indicators in labor markets, and an increase in this value indicates that the confidence of citizens in labor markets in the country has decreased and long-term unemployment has increased. The final factor is positive employment indicators and this factor consists of 2 sub-dimensions. The first of these sub-dimensions is personal earnings. This indicator refers to the average annual wages per full-time equivalent dependent employee, which are obtained by dividing the national-accounts-based total wage bill by the average number of employees in the total economy, which is then multiplied by the ratio of average usual weekly hours per full-time employee to average usually weekly hours for all employees. This rate is measured in US Dollars. According to the results, Luxembourg has the highest work-related earnings average of 63,062 dollars while Hungary has the lowest work-related average of 22,576 dollars. The second sub-dimension is the employment rate. This indicator is the ratio of employed persons aged 15 to 64 over the population of the same age and this indicator is measured as a percentage. According to the results, Sweden has the highest rate with 77% and Greece has the lowest rate score with 53%. This factor is designed to measure positively-rated indicators in labor markets, and the increase in value here is indicative of increased employment and job-related earnings.
For SmartPLS 3.0, items in a model need to have an outer load of more than 0.70 in order to be reliable [48]. In order for the factors in the model to be reliable, composite reliability values and internal consistency values should be higher than 0.70 [49]. The variables, items, item reliability, composite reliability and internal consistency values in the predicted model can be examined in Table 2.
Reliability results of the model
Reliability results of the model
Table 2 shows, factors, indicators, outer loadings and composite reliability of the model. According to the results of the analysis, the outer loads of the items in the model were higher than 0.70 and the composite reliability values of the variables were higher than 0.70. Accordingly, the items and variables in said model are reliable.
In order for the predicted model to be valid in SmartPLS, the average variance extracted (AVE) value should be higher than 0.50 [49] and the square root of the average variance extracted should be higher than the correlation values in the corresponding column [50]. The validity values in the predicted model can be examined in Table 3.
Validity Results of the Model
*Square root of AVE value.
Table 3 shows the validity values in the predicted model. Accordingly, it was determined that the average variance extracted (AVE) was greater than 0.50 and that the average variance extracted was higher than the correlation values in the corresponding column. Therefore, the predicted model is a valid model.
A reliable and valid model in SmartPLS should also be checked for multicollinearity. Hence, the correlations between the independent variables should be below 0.70 and variance inflation factor (VIF) values should be below 10 [51]. Accordingly, it was determined that the correlations between independent variables were below 0.70 and the variance inflation factor (VIF) was lower than 10. As a result, the issue of multicollinearity is not present among the variables as can be seen in Table 4.
Multicollinearity results
*p < 0.05; **p < 0.01.
After the model was found to be reliable and valid in SmartPLS, it was evaluated by considering the “t” values to determine whether the paths in the model were significant or not. If the value of “t” was greater than 1.96, the path in the model is significant; if the value of “t” was less than 1.96, the path in the model is interpreted as insignificant. The values of the paths and “t” values in the estimated model can be examined in Fig. 1.

t values for the predicted model.
Figure 1 shows the “t” values for the predicted model. According to this; all paths in the predicted model are significant (t > 1,96). According to this; the DESI has a statistically significant effect on positive employment indicators and negative employment indicators.
Figure 2 shows the coefficient values for the predicted model. Accordingly, in SmartPLS, the path coefficients in the model vary between –1 and +1. When the path coefficient approaches 1, the effect level increases; when it approaches 0, the effect level decreases. It is concluded that when the path coefficient values are negative (–), they have a negative effect on the variable; when the path coefficient values are positive (+), they have a positive effect on the variable. The coefficient values for the predicted model can be examined in Fig. 2.

Coefficient values for the predicted model.
Figure 2 shows the coefficient values for the predicted model. Accordingly, the DESI positively affects positive employment indicators (0,849) and negatively affects negative employment indicators (–0.535) for 23 European Union member countries in 2018. Furthermore, 72.1% of the change in positive employment indicators and 28.7% of the change in negative employment indicators are explained by the DESI.
The current study was conducted to determine the effects of the DESI on negative employment indicators and positive employment indicators for 23 EU member countries with 2018 data. According to the predictions made with the SmartPLS 3.0 program, it has been determined that the DESI positively affects positive employment indicators (0,849; t: 15,929) and negatively affects negative employment indicators (–0,535; t: 3,685). The result is that the digital transformation of the economy and society increases the employment rate and job-related earnings, which are positive indicators of the labor markets, and reduces long-term unemployment and insecurity in the labor markets. In other words, the widespread use of the internet, the increase in information technology-based employment, the active use of the internet by individuals, companies and the government, increases employment rates and job-related earnings; and reduces long-term unemployment and insecurity towards labor markets.
This result is consistent with the findings of the studies in literature, and the literature gives four important conclusions about the results of the current study. The first is the increase in employment and earnings in societies that successfully manage digital transformation. For example, research by Balsmeier & Woerter [34] covering the years 2014-2015 for Switzerland concluded that digitalization investments increased the employment of highly skilled workers. Research carried out by Titan et al. [29] concluded that the increase in the digital economy in Romania between 2006 and 2011 increased information technology employment. The study conducted by Lee & Clark [32] that covered 2009–2015 in the UK and the result has shown that the development of advanced technology has increased the employment rate and wages of middle-skilled workers while lowering the average salary of low-skilled workers. A study by Bauer [31] concluded that digitalization is effective in the emergence of new industries, employment opportunities, and occupations with high-income potential. Simic [33] demonstrated that in Serbia, digital competences support employment and digital inadequacies create unemployment. The report by Piva & Vivarelli [30] demonstrated a similar situation. Accordingly, technology-based R & D expenditures realized for 11 EU countries between 1998 and 2011 increased employment in the medium and high technology sectors and did not affect the low-tech sectors.
Other findings in the literature include the evaluation and reflection of digitalization in terms of new century competencies. Research carried out by van Laar et al. [9, 11] focuses on 21st-century skills and states that in the future digital skills may have an important place in labor markets and facilitate the employment of individuals. Indeed, a similar conclusion was reached in a study carried out by Misra & Khurana [38] and Fleaca & Stanciu [39]. Research by Campos et al. [35] covering the years 2007–2011 in Spain reached the conclusion that internet users had a higher advantage in finding work. Also, research by Frey & Osborne [52] has shown that with technological progress, low-skilled workers will turn to tasks that are not computerized, i.e. tasks that require creative and social intelligence. Another study by Nemeskéri et al. [37] focuses on Hungarian labor markets and highlights that a lack of digital skills creates problems with entering labor markets. From here, it is possible to say that digitalization will be an important skill in the future and that it will be easier for societies that enable digital transformation to enter labor markets.
Some of the research in literature have focused on the reflections of digital transformation on performance, efficiency, and competition. For example, Martin-Pena et al. [53] examined the relationship between digitalization and business performance in Spain between 2014–2017 and showed a positive relationship between digitalization and business performance. Research carried out by Jalava & Pohjola [54] also determined that information and communication technology increased production and labor productivity in Finland in 1995–2005. Saculescu [55] examined the impact of development in the information and communication technology sector on competitiveness in European Union member states and concluded that the increase in the share of information technology in the economy in almost every country increased the competitiveness of the country. Research conducted by Dobrolyubova et al. [56] focused on digital transformation in Russia, and the results show that economic and social digital transformation made the country’s economy and labor market more competitive. From this point, it would be appropriate to infer that the societies that successfully carry out the digital transformation have improved performance and productivity and that the labor markets of these countries may also be competitive [57, 58].
Finally, digitalization increases the self-confidence of individuals from a professional standpoint. For example, research by Lissitsa & Chachashvili-Bolotin [36] and Nam [40] showed that internet use increases the self-confidence of the individual from a professional standpoint. From here, it is possible to say that the widespread use of the internet is a factor in reducing job insecurity.
When all the results are evaluated, it is possible to say that the digitalization of the economy and society has a corrective effect on the dynamics of labor markets. However, it is not true to consider digitalization a magic wand. Digitalization can only yield accurate results if equal opportunities are created. Therefore, it is important that the economic and social digital transformation created is inclusive, achievable and widespread.
Regarding the limitations of the study, current research has reached this conclusion (the effects of the DESI on negative employment indicators and positive employment indicators) based on data from 23 EU member states for 2018. It should be noted that results for different years and different countries may vary. This study has used the structural equation modeling technique with macro data. It should also be taken into consideration that the results may be different in studies where different techniques are used.
Determination of the effect of digitalization on labor markets, which is the subject of this study, is one of the important pursuits in current human history. Concerning the further research, it can be suggested that the researchers interested in the matter, may aim to answer the question “Does the impact of digitalization on the labor market differ according to welfare regimes?”. In addition, a similar analysis for the pre-and post-crisis periods may be important in terms of revealing “the role of the crisis in the impact of digitalization on labor markets”.
