Digitalization of the EU Economies and People at Risk of Poverty or Social Exclusion

: Despite the fact that a comprehensive analysis of digitalization processes in the EU member states has been carried out, the impact of a country’s digitalization level on the risks of poverty and social exclusion requires further investigation. The purpose of the paper is to verify a hypothesis that a higher level of national digitalization provides positive trends in reducing the risks of poverty and social exclusion for the population. The Digital Economy and Society Index (DESI) was used to evaluate the digitalization levels of the EU countries. The indicator “People at risk of poverty or social exclusion” (AROPE) was applied to estimate the poverty level. As the main research methods, the authors used a comparative and correlation analysis with respect to the above-mentioned indicators, as well as the Monte Carlo method in order to evaluate the probability of a change in the indicator “population at risk of poverty or social exclusion” in 2021. The EU countries with higher digitalization levels have a lower percentage of the population at risk of poverty and social exclusion. However, a higher digitalization level of the EU member states does not provide an accelerated risk reduction of poverty and social exclusion. Statistical calculations with respect to the entire population of these countries mainly indicate reverse processes. At the same time, a further reduction of poverty and social exclusion level is less probable in the countries with a higher level of digitalization. For relatively poor segments of the population (the 1st and 2nd quintiles by income) in the EU member states, the level of digitalization does not play a signiﬁcant role. For relatively wealthy segments of the population (the 3rd and 4th quintiles by income) the authors noticed a pattern: the higher the level of digitalization is, the lower the risk of poverty and social exclusion becomes. A pairwise comparison of countries with initially similar AROPE values showed that in most cases (3 out of 5), the countries with higher levels of digitalization showed a more signiﬁcant reduction in poverty and social exclusion. However, the probability of further positive changes in this area is higher for the countries with a lower level of digitalization.


Introduction
The development of Industry 4.0 and the formation of a single digital market are among the key items on the EU agenda. The EU is constantly expanding their legal support to stimulate economic Digitalization of the economy is a complex process. Therefore, there are two alternatives for its description. The first alternative involves exploring a set of different indicators. The second alternative involves the use of existing indexes, which, in turn, are a convolution of a set of indicators. Given that various indexes already exist, one of them was chosen.
To evaluate the digitalization level of the EU countries, the Digital Economy and Society Index (DESI) was used, which is regularly calculated by the European Commission for all the EU member states (European Commission 2019). Unlike other indexes that characterize digitalization processes (for example the ICT Development Index; Digital Adoption Index; UN Global E-Government Development Index; Networked Readiness Index) and are calculated by various international organizations, companies, and research groups for all countries of the world, DESI is compiled exclusively for the EU member states taken as the target group. This ensures maximum consideration of the socio-economic specificity of these EU countries and the relevance of comparing the data obtained.
A connectivity dimension includes 8 indicators ( Table 1). All of them describe access to broadband coverage (fixed and mobile). A higher level of connectivity creates more opportunities for using digital skills, internet services, digital technology, and digital public services. A human capital dimension includes 6 indicators (Table 2), which are connected with Internet user skills and employment of the ICT specialists. Possession of these skills creates income-generating opportunities that allow people to be outside the poverty zone. The use of an internet services dimension includes 11 indicators (Table 3), which are connected with frequency of using the internet for online activities and transactions. The use of the internet services does not only create extra ways for earning, but it multiplies social capital and ensures wider opportunities for finding a job.  A dimension of digital technology integration comprises 7 indicators (Table 4), which characterize business and commerce digitalization. This dimension creates opportunities for running your own business more efficiently and increasing the quantity of work places. A dimension of digital public services includes 5 indicators (Table 5), which characterize the e-Government domain. E-Government allows for businesses to cut costs when they communicate with authorities. This frees up resources to expand business activities.

Indicator
Description 5a e-Government 5a1 e-Government users Individuals who sent filled forms to public authorities over the internet in the previous 12 months 5a2 Pre-filled forms Amount of data that is pre-filled in public service online forms 5a3 Online service completion The share of administrative steps that can be done online for major life events (birth of a child, new residence, etc.) 5a4 Digital public services for businesses The indicator broadly reflects the share of public services needed to start a business and to conduct regular business operations that are available online for domestic as well as foreign users. Services provided through a portal receive a higher score, services which provide only information (but have to be completed offline) receive a more limited score. Most of the indicators examined characterize processes and states that directly or indirectly create conditions for reducing poverty. This is primarily due to the expansion of opportunities for taking online income, business development, and employment in the ICT sector. An indicator of the people at risk of poverty or social exclusion (AROPE) was used to measure poverty. This indicator is regularly calculated by the statistical service of the European Union (Eurostat 2020a(Eurostat , 2020b. It "corresponds to the sum of persons who are either at risk of poverty or severely materially deprived or living in a household with a very low work intensity. Persons are only counted once even if they are present in several sub-indicators. The AROPE rate, the share of the total population who are at risk of poverty or social exclusion, is the headline indicator to monitor the EU 2020 Strategy poverty target" (Eurostat 2020c). AROPE, as well as DESI, is calculated specifically for the EU member states.
In order to test the research hypothesis, DESI and AROPE were compared comprehensively and their dependence was evaluated, which covered the following steps: Step 1. Countries from the top to the bottom of the sample are compared in terms of digitalization relative to AROPE.
Step 2. SThe general relationships between DESI and AROPE are analyzed.
Step 3. AROPE is compared by income and the level of digitalization in the EU countries.
Step 4. The dynamics of DESI and AROPE changes are compared.
Step 5. The probability of positive changes in AROPE against the background of DESI values is estimated.
Step 6. The dynamics of country pairs with the AROPE close values is compared, taking the average DESI level into account.
The authors used a comparative and correlation analysis with respect to the above-mentioned indicators as the main research method, as well as the Monte Carlo method, in order to estimate the probability of a change in the AROPE indicator in 2021.

Results and Discussion
An increase of the digitalization level is a characteristic found in all EU member states. Moreover, the digital dominance (and, accordingly, the backlog) of individual countries is a very stable process.
During the observation period (2014)(2015)(2016)(2017)(2018), only a few countries moved up or down by several positions. None of the countries moved from the first position to the last or vice versa. The level of autocorrelation with a lag of 1 year is 0.99 and higher. The initial comparison of the level of digitalization and poverty in the EU (Table 6) suggests that the countries from the TOP-5 (Denmark, Finland, Sweden, the Netherlands, and Luxembourg) have a lower poverty level with digitalization than countries occupying the last five positions (Italy, Poland, Bulgaria, Greece, and Romania). For the former, the percentage of their population at risk of poverty or social exclusion is in the range 16.5-21.8% (on average, 18.1%), and for the latter, the range is 18.4-32.8% (on average, 28.6%).  (2019) and Eurostat (2020a). Note: * Rank 1 shows the best national digitalization situation. Rank 28 shows the worst national digitalization situation; ** The lower AROPE, the better the changes of low poverty or social exclusion.
Consequently, the Spearman's rank correlation coefficient between the DESI rank and the AROPE indicator has a plus sign throughout the observation period, although the coefficient value decreases over the years ( Analyzing the share of people at risk of poverty or social exclusion by income quintile against digitalization (Table 7), we can see the ambiguity of the digitalization role, which depends on the quintile of the population income. So, in the first quintile by income (20% of the population with the lowest incomes) the group of countries with a low digitalization level (an average rank during the monitoring period from 21 to 28) shows a lower risk of poverty or social exclusion than a group of countries with a medium-high digitalization level (rank 8-14), 86.81% and 86.86%, respectively.
The lowest risk of the population's poverty or social exclusion occurs in the group of countries with a medium-low level of digitalization-78.16%.
At the same time, the group of countries with a high level of digitalization (ranks 1-7) does not always show a better result than countries with a relatively lower level of digitalization. In this group, 83.03% of the population are at risk of poverty or social exclusion. In the meantime, this value is 78.16% on average in countries with a medium-low level of digitalization (ranks 15-21).
An ambiguous situation also takes place in relation to the 2nd quintile by income. The population of low-digitalized countries is more at risk of poverty and social exclusion (25.44%), and the population of countries with a high level of digitalization is less at risk of poverty and social exclusion (9.74%). However, the population of countries with a medium-high digitalization level has greater risks of poverty and social exclusion than the population of a group of countries with a medium-low digitalization level (14.06% and 11.6%, respectively).
According to the research hypothesis, the logical dynamics occur only for the 3rd and 4th quintiles, where the higher the digitalization level of a group of countries is, the lower the risk is of poverty or social exclusion of their populations.
However, a direct relationship between the risk of poverty and the level of digitalization is not entirely clear, because more economically developed countries have a higher level of digitalization (Micic 2017). Therefore, we can say that the level of economic development, being a key factor, affects both the country's level of digitalization and the level of poverty. The higher the economic development is, the higher the level of digitalization and the lower the risks of poverty or social exclusion. Therefore, it is advisable to use other research approaches: (1) to compare progress in digitalization with progress in reducing the risks of poverty or social exclusion; (2) to estimate the probability of poverty reduction; (3) to divide the EU member states into groups according to their level of risks of poverty or social exclusion and to compare the dynamics of these level changes depending on the level of digitalization.
For the period from 2014 to 2018, the change in the countries' DESI ranking varies from +4 to -3, at the same time ranking varies from −12.7 to +2.8 pp by percentage of the population at risk of poverty and social exclusion ( Table 8). The greatest progress in reducing the share of the population according to the AROPE indicator was observed for the countries that moved down one place due to increased digitalization in 5 years and amounted to −9.3 pp. The second most successful country in terms of reducing poverty risks for the population is Ireland (−6.6 percentage points), which moved its digitalization rate up by three places from 11 to 8. At the same time, Lithuania, which showed the greatest positive change in the DESI index (+4 positions), is among the countries where the situation outlined by the AROPE indicator worsened (+1 pp in 2014-2018). In this case, progress in digitalization has failed to ensure there is progress in the fight against poverty and social discrimination. Thus, a relatively high level of digitalization does not provide the accelerated poverty alleviation. This is confirmed graphically (Figure 1).
3, at the same time ranking varies from −12.7 to +2.8 pp by percentage of the population at risk of poverty and social exclusion ( Table 8). The greatest progress in reducing the share of the population according to the AROPE indicator was observed for the countries that moved down one place due to increased digitalization in 5 years and amounted to −9.3 pp. The second most successful country in terms of reducing poverty risks for the population is Ireland (−6.6 percentage points), which moved its digitalization rate up by three places from 11 to 8. At the same time, Lithuania, which showed the greatest positive change in the DESI index (+4 positions), is among the countries where the situation outlined by the AROPE indicator worsened (+1 pp in 2014-2018). In this case, progress in digitalization has failed to ensure there is progress in the fight against poverty and social discrimination. Thus, a relatively high level of digitalization does not provide the accelerated poverty alleviation. This is confirmed graphically (Figure 1).   [−0.6; 0]. To estimate the changes within one forecast year, 10,000 independent scenarios were generated for each country. Accordingly, the probability is the ratio of the number of positive outcomes (a reduction in the percentage of the population at risk of poverty or social exclusion) to the total number of outcomes generated. Aggregated results for 2019-2021 are presented in Table 9. Table 9. The Probability that the Value of the AROPE Indicator in 2021 will be lower than in 2018, %. For a group of countries with a high digitalization level, the average probability of the AROPE index to decrease in 2021 compared to 2018 is 75.61%, which is lower than for a group of countries with a medium-high (81.99%), medium-low (81.05%), and low (76.45%) digitalization level.

Rank of DESI
The Spearman's rank correlation coefficient (0.32) shows that the lower the country's digitalization level is, the higher the probability of poverty reduction and social exclusion. This contradicts the original hypothesis.
The approaches discussed above have their limitations because a different percentage of the population at risk of poverty or social exclusion forms different basic conditions, and a direct comparison of these populations is not always correct. For example, in Bulgaria in 2013 the AROPE level was 48%, and in 2018 it fell to 32.8% (minus 15.2 pp). In Czechia, such a significant decline is impossible even theoretically, because in 2013, this percentage amounted for 14.6% and theoretically could fall only to 0% or only by 14.6 pp. Consequently, the probability of a decrease in AROPE is to some extent determined by initial values.
Therefore, to further estimate the impact of the digitalization level on the risks of poverty or social exclusion, the authors selected pairs of the countries where initial values in 2013 were the same or differed no more than 0.1 pp. Thus, we can say that these countries were at the same starting position. Five pairs were selected for comparison (Table 10), and the authors analyzed the change of the percentage of the people at risk of poverty or social exclusion for 2014-2018 against the rank of the average digitalization level (DESI indicator) for the same period, as well as the probability of further positive changes. In pair No. 1 (the Netherlands and Finland), a country with a higher digitalization level (Finland) has less rapid growth of the percentage of the population at risk of poverty or social exclusion (+0.5 pp) than a country with a lower level of digitalization (+0.8 pp). Thus, we observe a positive effect, which manifested itself in a lesser degree of deterioration of the situation.
In pair No. 2 (France and Sweden), a country with a higher level of digitalization (Sweden) showed lower rates of poverty risk reduction (−0.3 percentage points) than France (−0.9 percentage points). Therefore, in this pair, the positive effect of a higher level of digitalization is not observed.
In pair No. 3 (Sweden and Denmark), a country with a higher level of digitalization (Denmark) showed a higher rate of poverty reduction (−0.9 percentage points) than Sweden (−0.3 percentage points). Therefore, in this pair, the positive effect of a higher level of digitalization is observed.
In pair No. 4 (Slovenia and Germany), a country with a higher level of digitalization (Germany) showed lower rates of poverty risk reduction (−1.6 percentage points) than Slovenia (−4.2 percentage points). Therefore, in this pair, the positive effect of a higher level of digitalization is not observed.
In the last pair No. 5 (Ireland and Croatia), a country with a higher level of digitalization (Ireland) showed higher rates of poverty risk reduction (−8.8 percentage points) than Croatia (−5.1 percentage points). Therefore, in this pair, the positive effect of a higher level of digitalization is observed.
Thus, three pairs out of five demonstrated a positive effect of the digitalization level. Moreover, the probability of further positive changes within each of the five pairs is higher for a country with a lower level of DESI.

Conclusions
Comprehensive studies and statistics show that two processes are simultaneously taking place in the EU: deepening of economic digitalization and poverty reductions. Given its potential, it was logical to assume that digitalization has a positive effect on reducing the risks of poverty and social exclusion. However, on average, a higher level of digitalization of an EU member state does not provide an accelerated reduction of the risks of poverty and social exclusion.
Statistical calculations for the EU-28 countries mainly indicate the opposite; a higher level of digitalization provides a less rapid rate of poverty risk reduction. At the same time, countries with a higher level of digitalization are less likely to reduce poverty and social exclusion further.
For relatively poor segments of the population (1st, 2nd quintile by income) in the EU member states, the level of digitalization does not play a significant role. For relatively wealthy segments of the population (the 3rd and 4th quintiles by income) there is a pattern: the higher the level of digitalization is, the lower the risk of poverty and social exclusion.
A pairwise comparison of countries with similar initial AROPE levels showed that in most cases (three out of five), countries with a higher digitalization level showed a more significant reduction in poverty and social exclusion. However, the probability of further positive changes in this area was higher for countries with a lower level of digitalization.
Given the ambiguity of the results and the EU's focus on further deepening of the digitalization level, there still remains the need for a more detailed study of the digitalization impact on social processes, although there is no doubt that further research, aimed at improving the social efficiency of introducing digital technology, is necessary.