Building a Social Progress-Adjusted Indicator of GDP Per Capita for the European Union’s Regions

Abstract

Boosting economic, social and territorial cohesion is a high priority for the European Union. The regional Cohesion Policy is its instrument for achieving this objective, with GDP per capita being the principal criterion for funding allocation. However, social cohesion also embraces a comprehensive range of non-economic features. This paper proposes a composite indicator of GDP per capita adjusted for social progress that incorporates both economic and non-economic issues. Notably, the indicator can account for policymakers’ preferences regarding the relative importance assigned to income and social progress. This novel indicator is used to simulate the eligibility status of European Union regions for funding in several policy scenarios. Regions’ simulated eligibility is then compared with their current eligibility for the period 2021–2027. Based on that comparison, winner and loser regions are identified, thus providing policymakers in charge of the European Cohesion Policy with an insight into the effect of incorporating social progress into funding allocation.

1 Introduction

The European Union (EU) has undergone profound changes in recent decades, following the enlargements that occurred from 2004 onwards. The integration of several ex-communist Eastern European countries gave rise a more complex and unequal EU, with wide disparities between the early members and the newcomers. The aim of the EU’s regional Cohesion Policy is to strengthen economic, social and territorial cohesion in the EU, by lessening imbalances across regions and Member States. Cohesion is thus a priority for the European Commission (EC); indeed, nearly a third of the EU’s budget is currently dedicated to addressing this issue (European Commission, 2020), with the European Structural and Investment Funds (ESIF) being the main financial instruments.

In the framework of the Cohesion Policy for the period 2021–2027, income per capita is still the principal criterion for funding allocation. Social and economic cohesion is, however, about much more than income. It also embraces non-economic issues such as personal freedom and choice, education, access to information and communication technologies (ICTs), tolerance and inclusion, and environmental quality, among others. In this respect, more than a decade has passed since the EC first acknowledged the need to produce composite indicators that include social facets of development along with income, in order to inform political decisions regarding the allocation of cohesion funds (European Commission, 2009). Likewise, the EC’s new policy orientations for the period 2019–2024 include—in addition to economic goals—essential objectives for social progress such as the European Green Deal, a Europe fit for the digital age, the protection of the rule of law, and a new impetus for European democracy (von der Leyen, 2019).

In this regard, in 2016 the EC launched the European Social Progress Index (EUSPI), which builds on the Social Progress Index delivered by the non-profit organization Social Progress Imperative (Porter et al., 2014). With two available releases, for the years 2016 and 2020,¹ the EUSPI represents a major attempt by the EC to understand EU citizens’ quality of life and to provide policymakers and stakeholders with a tool that can help them to design a more successful Cohesion Policy (Crescenzi et al., 2020). The index incorporates three broad non-economic dimensions of social progress, namely, (i) basic human needs, (ii) foundations of well-being, and (iii) opportunities. Each dimension is defined by more specific components, produced by aggregating a wide array of raw indicators. Interestingly, social progress is measured at the level of NUTS2 regions, the same territorial units the EU Cohesion Policy addresses. Nonetheless, despite its potential suitability for policy decisions, the EUSPI is currently used for information purposes only, and is not binding regarding the EC’s funding allocation.

To date, limited attention has been paid to the potential allocation of EU cohesion funds using composite indicators that account for social and environmental facets of development other than income. Some exceptions are Sánchez and Jiménez-Fernández (2022), who build a composite indicator of the socio-economic vulnerability of EU regions that summarizes their relative position regarding the five objectives of the Cohesion Policy for the period 2021–2027 (European Commission, 2021a); and Döpke et al. (2017), who find that allocating EU cohesion funds using alternative indicators based on the OECD’s regional Better Life Index (BLI) (Stiglitz et al., 2009) instead of income per capita would have a small effect on eligible regions. Compared to the BLI, the EUSPI is far more comprehensive in terms of both the coverage and regional availability of the quality of life dimensions; moreover, it avoids the inclusion of economic variables, which makes it particularly useful to complement GDP per capita (GDPpc). However, as far as the authors are aware, there have not yet been any initiatives that use the EUSPI for funding allocation.

The reason why the EC has not made more effective use of the EUSPI—beyond purely informative purposes—might be related to its novelty and, more importantly, to doubts about its robustness. In that regard, the recent work by Beltrán-Esteve et al. (2023) finds that the EUSPI is a robust indicator of social progress, given that the ranking of EU regions according to their social progress scores barely changes when the methodological choices in each step of its construction are altered. Accordingly, the authors suggest that using the EUSPI as a complement to GDPpc in establishing eligibility for Cohesion Policy funding could lead to a much more socially comprehensive and balanced allocation of funds. Nevertheless, a specific proposal is left for future research.

Against this background, the contribution of this paper is twofold. First, it presents a proposal for an indicator of the GDPpc of the EU regions adjusted by their level of social progress—measured by the EUSPI. This contributes to the Beyond GDP literature (Bleys, 2012) and adds to previous research focused on adjusting GDP for environmental issues (Lianos and Pseridis, 2021; Galiano Bastarrica et al., 2023) or other social matters (Malay, 2021). More specifically, the paper proposes a novel indicator that can be employed for funding allocation in the European regional context, thus complementing recent work analysing the robustness of the EUSPI as a useful instrument for policymaking. Notably, the indicator allows for a more comprehensive assessment and detailed understanding of regional disparities in the EU, which is a necessary step for improving Cohesion Policy decisions.

Secondly, the construction of the social progress-adjusted GDPpc indicator accounts for policymakers’ preferences regarding the relative importance assigned to income and social progress. In this regard, the eligibility of EU regions for funding from the regional Cohesion Policy is assessed in several simulated policy scenarios, and then compared with their actual eligibility status according to GDPpc for the period 2021–2027; by so doing, the potential winners and losers in these simulated scenarios can be identified. Winning regions would be those that improve their eligibility status when assessed with social progress-adjusted GDPpc instead of GDPpc, while losers would be the regions that see a decline in their status.² The results bring to light the fact that the eligibility status of several regions—mostly transition regions, according to the current EC criteria—would change when assessed with social progress-adjusted GDPpc; particularly when income is adjusted by social opportunities, which in the EUSPI framework represent the most advanced facets of social progress. This second contribution provides policymakers with valuable insights about the practical implementations of the proposed indicator, showing the sensitivity of the regional development status to the relative weight given to economic and non-economic aspects.

The remainder of the paper is organized as follows. Section 2 explains the EU regional Cohesion Policy for the period 2021–2027. Section 3 describes the main features of the EUSPI. Section 4 explains how the social progress-adjusted GDPpc indicator is computed. Section 5 comments on the results, paying particular attention to the comparison of regions’ eligibility for funding according to GDPpc and adjusted GDPpc. The final section concludes and highlights how this research may be of interest for policymaking.

2 The European Union’s Regional Cohesion Policy for the Period 2021–2027

The EU’s regional Cohesion Policy embraces a wide-ranging set of policies and funding programmes aimed at reducing disparities among regions and Member States by promoting economic, social and territorial cohesion. It particularly targets rural areas, zones affected by industrial transition, and areas suffering from severe and permanent natural or demographic handicaps.

The current Cohesion Policy for the period 2021–2027 has a long-term strategic vision aimed at moving towards a green, digital and fair Europe. To this end, policies and financial resources are targeted at five priority policy objectives, which are set out by the Common Provisions Regulation (European Commission, 2021a; Article 5):

• A more competitive and smarter Europe by promoting innovative and smart economic transformation and regional ICT connectivity.

• A greener, low-carbon transitioning towards a net zero carbon economy and resilient Europe by promoting clean and fair energy transition, green and blue investment, the circular economy, climate change mitigation and adaptation, risk prevention and management, and sustainable urban mobility.

• A more connected Europe by enhancing mobility.

• A more social and inclusive Europe implementing the European Pillar of Social Rights.

• A Europe closer to citizens by fostering the sustainable and integrated development of all types of territories and local initiatives.

The EU supports the achievement of these policy objectives through the ESIF. The funds directly linked to the implementation of the regional Cohesion Policy are the Cohesion Fund (CF), the European Regional Development Fund (ERDF), the European Social Fund Plus (ESF+) and, from 2021, the Just Transition Fund (JTF).

The JTF is a strategic tool for achieving the objective of moving towards a greener, low-carbon Europe from which all EU Member States can benefit. This fund co-finances investments in areas such as clean technologies and control of emissions, thereby mitigating the social, economic and environmental impacts of the transition towards the EU’s 2030 energy agenda and a climate-neutral economy by 2050. The CF provides support to Member States to deliver investments in the field of the environment and trans-European networks in the area of transport infrastructure.

The ERDF and the ESF+ are, however, the funds most directly linked to boosting economic and social cohesion across the EU regions. Resources from the ERDF are targeted at investing in the progress of regions and cities, with particular focus on improving the living conditions of citizens in less favoured areas—including regions with severe natural or demographic handicaps, and island regions. The ESF+ is the main financial instrument for supporting measures aimed at preventing and combating unemployment, developing human resources and promoting social integration in the labour market. To this end, the fund co-finances projects aimed at fostering employment, fostering equal opportunities for women and men and, in general, facilitating sustainable development and economic and social cohesion.

According to the European Court of Auditors (2019; p. 21), 81% of the budget of the EU regional Cohesion Policy for the period 2021–2027—including the CF, the ERDF and the ESF+ —will be allocated based on prosperity levels, measured by income per capita. The remaining 19% will be distributed according to a set of variables that account for the labour market, education and demography (15%); migration (3%); and climate change (1%).³

Regarding prosperity, the CF is targeted at the EU Member States with national income per capita below 90% of the EU27 average, whereas resources from the ERDF and the ESF+ are allocated across the EU regions according to their GDPpc, also relative to the average of the EU27. In all cases—regions and the EU27—GDPpc is calculated on the basis of figures for the period 2015–2017 and assessed in Purchasing Power Standards (PPS). The categories of regions, classification criteria and co-financing rates are as follows (European Commission, 2021a; Articles 108 and 112):

• Less developed regions, with a GDPpc of less than 75% of the EU27 average. The co-financing rate for investment projects may be up to 85% in these regions.

• Transition regions, whose GDPpc is between 75 and 100% of the average. The maximum co-financing rate is 60%; or 70% for regions that were classified as less developed regions for the 2014–2020 period.

• More developed regions, with a GDPpc above 100% of the average. In these regions, the co-financing rate may reach 40%; or 50% if they had been transition regions in the 2014–2020 period.

Applying these criteria, 78 out of the 240 EU regions at the NUTS2 level are categorized as less developed regions; 67 are transition regions; and the remaining 95 are more developed regions. Map 1 provides a picture of this classification. There is an evident geographical pattern in the distribution of regional development in the EU27, with the more developed regions mostly concentrated in the Centre of Europe and the Nordic countries, and the less developed ones mainly located in the Southern regions of Mediterranean countries such as Spain and Italy, in addition to regions throughout Greece, Portugal and Eastern countries that joined the EU from the 2000s onwards. In fact, among the four poorest regions in terms of their GDPpc, three are Bulgarian—Severozapaden (BG31), Severen tsentralen (BG32) and Yuzhen tsentralen (BG42)—and one is Romanian—Nord-Est (RO21).

Map 1

Source European Commission (2021b)

EU regions’ eligibility for funding from the ERDF and the ESF + for the period 2021–2027 according to GDPpc.

Prosperity levels and the abovementioned variables—labour market, education and demography; migration; and climate change—are combined to obtain a first distribution of the cohesion funds across the EU regions. The main factors shaping this initial allocation are: (i) the regional prosperity gap—i.e., the difference between the region’s GDPpc and that of the EU27 average—for less developed regions and transition regions; and (ii) the fixed aid intensity per capita—which is a pre-established amount of money per person per annum—for more developed regions and Member States. Population is also an influential factor in all cases. Additional adjustments are made on the basis of coefficients reflecting national prosperity levels, and several demographic and socio-economic criteria. Furthermore, premiums are allocated to less developed regions and transition regions depending on unemployment, youth unemployment, low levels of education, greenhouse gas emissions and migration, while more developed regions are awarded premiums based on greenhouse gas emissions and migration.

The ultimate allocation of the EU’s regional Cohesion Policy funds for the period 2021–2027 is decided on after applying some further minor adjustments to mitigate significant fluctuations in the resources received by each individual Member State. These include caps limiting the maximum amount receivable and safety nets guaranteeing a minimum level of funding. All these allocating mechanisms are described in detail in the report by the European Court of Auditors (2019; pp. 22–25; 35). In the final allocation, the CF accounts for 13% of the total funds, whereas the remaining 87% corresponds to the ERDF and the ESF+ —62% to less developed regions; 14% to transition regions; and 11% to more developed regions (European Court of Auditors, 2019; p. 26). Put another way, the less developed regions will receive 7 out of every 10 euros channelled through the ERDF and the ESF+, which are the funds with a specifically regional focus. These figures alone illustrate the crucial importance of GDPpc in determining the eligibility of EU regions for funding from the Cohesion Policy.

It is also worth highlighting that, despite the fact that the final amount of funds allocated to each EU region accounts for a few non-economic aspects of development, the initial categorization as less developed, transition and more developed regions entirely depends on income per capita relative to the EU27 average; and, according to current policy guidelines, that classification strongly conditions the maximum co-financing rate that a region can potentially receive.

3 The European Union Regional Social Progress Index

In 2016, the EC launched the EUSPI within the framework of a broader project aimed at contributing to the Beyond GDP agenda by assessing societal development and the quality of life in the EU regions. The EUSPI, the second edition of which was released in 2020, is a composite indicator based on a range of primary indicators that capture social and environmental issues relating to social progress, thus excluding economic facets. These indicators measure outcomes that matter for citizens’ lives, rather than inputs or efforts believed to lead to social progress (Porter et al., 2014). This means that the EUSPI is an output-oriented composite indicator, which is highly advisable for the purpose of assessing social progress (Fleurbaey & Blanchet, 2013). Many of the indicators making up the EUSPI refer to rights inspired by the Natural Rights theory (Locke, 1689) that are not culture-dependent and, as such, aspire to be universal. To an extent, they also represent the principle of equality of opportunities and the Rawlsian view of justice characteristic of Western democracies.

The 2020 release of the EUSPI assesses social progress in the 240 EU regions at the NUTS2 level, and is made up of 55 primary indicators—with data mainly coming from Eurostat—grouped into 12 components. The indicators included in each component are selected after having verified with Principal Component Analysis (PCA) techniques (Rencher & Christensen, 2012) that there is a strong multivariate correlation among them. Then, they are normalized and converted into a common scale using the min–max transformation with indicator-specific boundaries, which represent the best and worst performance on each indicator and are mostly set by utopian and dystopian values. Raw indicators are then aggregated into components using unweighted arithmetic means, which is a reasonable approach given that the internal consistency of the indicators within each component largely mitigates the effect of different weighting schemes (Decancq & Lugo, 2013), while also reducing compensability across indicators—i.e., the undesired partial or total offsetting of a deficit in one indicator with a surplus in another (Munda, 2008).

In a further step, components are aggregated into three broad dimensions using generalized means. These dimensions, which represent progressively more advanced features of social progress, are: (i) basic human needs—embodying issues that are necessary but not sufficient to achieve social development; (ii) foundations of well-being—which account for more advanced aspects of social and environmental progress; and (iii) opportunities—representing the most sophisticated facets of a cohesive and tolerant society. Finally, generalized means are also employed to aggregate the dimensions into the EUSPI. Given that the effect of compensability is expected to be larger across components and dimensions, the generalized mean is a sensible aggregation approach that allows this effect to be mitigated (Annoni & Weziak-Bialowolska, 2016). These and other technical issues of the EUSPI are detailed in Annoni and Bolsi (2020), while Table 1 summarizes the architecture of the index.⁴

Table 1 The architecture of the 2020 release of the European Social Progress Index: Indicators (I), Components (C) and Dimensions (D)

Composite indicators are useful tools that summarize the information collected on multidimensional realities into a single figure, greatly facilitating the communication of complex phenomena to policymakers and the public. Social progress is undoubtedly one of these phenomena, necessitating a multifaceted approach such as that proposed by the EUSPI. However, building composite indicators involves methodological choices and subjective decisions that could lead to widely different scores and rankings (Nardo, 2005; OECD, 2008); if this were the case with the EUSPI, it would be a poor instrument for policymaking. In this regard, Beltrán-Esteve et al. (2023) demonstrate that the 2020 release of the EUSPI is robust to multiple methodological designs and, therefore, suitable for policymaking. The authors also find that the EUSPI and GDPpc are strongly correlated, although they are in no way substitutes. This suggests that social progress could be used as a complement to income to determine the eligibility of EU regions for funding from the regional Cohesion Policy, thus achieving a fairer allocation based on economic as well as social and environmental facets of citizens’ lives.

The scatterplot in Fig. 1 depicts the relationship between EU regions’ social progress according to the 2020 release of the EUSPI, and their GDPpc calculated for the period 2015–2017 in Purchasing Power Standards (PPS) (European Commission, 2021a); in both cases, the EU-27 average is taken as a reference. The correlation between these two variables is 0.630 and is statistically significant at standard confidence levels. The relationship is, however, non-linear; while income and social progress are generally highly correlated in less developed and transition regions, once a sufficiently high level of income has been achieved, social progress stagnates. That said, there are a few cases of high (or very high) income and medium (or even low) social progress, and vice versa. For instance, Luxembourg (LU) is by far the European region with the highest GDPpc (276.4 relative to the EU27 average), but enjoys a level of social progress only slightly above the average (111.6); also, the Romanian region of Bucureşti-Ilfov (RO32) has a relative GDPpc of 140.9, but it scores just 81.1 on social progress. At the opposite extreme, citizens in the Finnish region of Pohjois- ja Itä-Suomi (FI1D) have a relative income per capita of 91.6, but enjoy high social progress (123.3). Some of these regions, along with other atypical ones, are analysed in more detail in Sect. 5.

Fig. 1

Source Own elaboration

EU regions GDPpc versus the EUSPI. (GDPpc in PPS; EU27 = 100). Note The data for GDPpc are the average of 2015–2017, which is the period of reference considered by the EC in the framework of the 2021–27 Cohesion Policy; and the data for social progress come from the 2020 edition of the EUSPI.

Map 2 plots the distribution of social progress across the EU regions according to their EUSPI. The different categories—from very low to very high—are set according to the quintiles of the distribution. At first glance, a marked geographical pattern can be discerned, with several Northern and Central regions recording the highest levels of social progress; in fact, the top five regions are Swedish—Övre Norrland (SE33), Mellersta Norrland (SE32) and Småland med öarna (SE21)—or Finnish—Helsinki-Uusimaa (FI1B) and Länsi-Suomi (FI19)—while the sixth is Danish—Midtjylland (DK04). On the other hand, the lowest levels of social progress are found in some Mediterranean regions of Southern Europe, and particularly in the regions of the Eastern European countries that joined the EU from the 2000s onwards. Specifically, the five regions with the poorest records are Bulgarian—Yugoiztochen (BG34) and Severozapaden (BG31)—and Romanian—Nord-Est (RO21), Sud-Muntenia (RO31) and Sud-Est (RO22).

Map 2

Source: Own elaboration

The distribution of social progress (EUSPI) across the EU regions. Note The different categories are set according to the quintiles of the distribution of the EUSPI. Figures in square brackets are the lower and upper cut-off points of each category.

Going beyond the patterns of social progress described above, Map 5 in the Appendix displays the distribution of the EUSPI dimensions of basic human needs, foundations of well-being and opportunities. Regional disparities are notable in foundations of well-being and even more so in opportunities, whereas they are much smaller in basic human needs. As a rule of thumb, regions with the highest levels of social progress according to the EUSPI also tend to score highly in its three dimensions. In this regard, 66.7%, 83.3% and 91.7% of the regions with a very high level of social progress are also in the group of top performers according to the distributions of basic human needs, foundations of well-being and opportunities, respectively. Figures are fairly similar for regions with very low social progress. In fact, only a few regions depart from this observed behaviour. An example is Eastern and Midland (IE06) in Ireland, where social progress, foundations of well-being and opportunities are very high, but the region scores poorly in basic human needs; also, the Hungarian region of Pest (HU12) scores very low in the EUSPI and the dimensions of foundations of well-being and opportunities, but ranks near the average in basic human needs.

4 Building an Indicator of Social Progress-Adjusted GDPpc

The social progress-adjusted GDPpc indicator for the EU regions proposed in this paper incorporates both GDPpc and the EUSPI, and is computed as:

$${\mathrm{Adjusted\,\, GDPpc}}_{\mathrm{r}}={\overline{\mathrm{GDPpc}} }_{\mathrm{r}} {\overline{\mathrm{EUSPI}} }_{\mathrm{r}}^{\upbeta }$$

(1)

where the sub-index r represents the 240 EU regions at the NUTS2 level.

The term ${\overline{\mathrm{GDPpc}} }_{\mathrm{r}}$ represents the GDPpc of region r relative to the average of the EU27 in the period 2015–2017, which is defined as:

$${\overline{\mathrm{GDPpc}} }_{\mathrm{r}}=\frac{{\mathrm{GDPpc}}_{\mathrm{r}}}{\mathrm{Average\,\, GDPpc \,\,of \,\,the\,\, EU}27}100$$

(2)

whereas ${\overline{\mathrm{EUSPI}} }_{\mathrm{r}}$ stands for the social progress of region r relative to the EU27 average, according to the 2020 edition of the European Social Progress Index. It is calculated as:

$${\overline{\mathrm{EUSPI}} }_{\mathrm{r}}=\frac{{\mathrm{EUSPI}}_{\mathrm{r}}}{\mathrm{Average\,\, EUSPI \,\,of\,\, the\,\, EU}27}$$

(3)

and is also used to classify the EU regions into those with a level of social progress above $\left({\overline{\mathrm{EUSPI}} }_{\mathrm{r}}>1\right)$ and below $\left({\overline{\mathrm{EUSPI}} }_{\mathrm{r}}<1\right)$ the average of the EU27 $\left({\overline{\mathrm{EUSPI}} }_{\mathrm{r}}=1\right)$.

Finally, $\upbeta$ is a policy parameter restricted to zero or greater than zero, which allows the incorporation of policymakers’ preferences about the relative importance of income per capita and social progress in the computation of the EU regions’ adjusted GDPpc. When this parameter is zero, social progress is assigned zero weight and adjusted GDPpc matches GDPpc. Conversely, when β is equal to one, social progress and GDPpc enter with the same weighting in the calculation of the adjusted GDPpc; a value of this policy parameter below one reduces the relative importance of social progress, while values above one increase it. Furthermore, regardless of regions’ relative income per capita, the larger the parameter β—meaning more importance is assigned to social progress—the lower the adjusted GDPpc for regions with social progress below the average. Equally, in regions where social progress is above the average, the larger this policy parameter, the higher the adjusted GDPpc.

Regarding the measurement of relative regional social progress, the 2020 release of the EUSPI provides information for 240 NUTS2 level EU regions, computed using regional-level indicators, as well as information for the EU27 Member States, using national-level indicators. Population-weighted averages of regional scores are, by construction, equal to national scores (Annoni & Bolsi, 2020; p.19). Accordingly, the EUSPI of the EU27 and its three dimensions have been calculated as population-weighted averages of all 240 regional scores. The calculation is based on the average population for the years 2015, 2016 and 2017 in order to be consistent with the way the EC calculates regional GDPpc to determine regions’ eligibility for regional Cohesion Policy funding for the period 2021–2027.

Leaving technical issues aside, the proposed indicator allows an analysis of how EU regions eligibility for funding from the European Cohesion Policy could change when income per capita is adjusted by the non-economic aspects of social progress considered in the EUSPI framework—which are tuned by a parameter representing the preferences of policymakers regarding income per capita and social progress. Furthermore, it is worth noting that in regions with a level of social progress below the EU27 average adjusted GDPpc will drop below GDPpc, which may improve their eligibility for funding. The exact opposite would happen in regions with social progress above the average, making them potential candidates for a decline in eligibility status. These regions are identified in Map 6 of the Appendix.

In order to further illustrate the performance of the social progress-adjusted indicator proposed, consider an EU transition region with ${\overline{\mathrm{GDPpc}} }_{\mathrm{r}}=90$ and ${\overline{\mathrm{EUSPI}} }_{\mathrm{r}}=1.1$ (above the EU27 average). In the scenario in which the policy parameter β is equal to 0.5, its adjusted GDPpc would be 94.4 (90 × 1.1^0.5), and the region would remain as a transition region in keeping with the EC’s current classification criteria; with β equal to 1 the adjusted GDPpc would be 99 (90 × 1.1¹) and the region would again maintain its status. Conversely, with a β of 2, thus giving much larger weight to social progress, the adjusted GDPpc would be 108.9 (90 × 1.1²) and the region would become more developed, thus changing its eligibility status; i.e., although according to its GDPpc it is a transition region, its relatively high social progress and the larger weight assigned to this factor when building the adjusted GDPpc would place the region above the EU27 average. As a point of comparison, a region with ${\overline{\mathrm{GDPpc}} }_{\mathrm{r}}=90$ and ${\overline{\mathrm{EUSPI}} }_{\mathrm{r}}=0.9$ (social progress below the average) would record an adjusted GDPpc of 85.4 for β equal to 0.5 (90 × 0.9^0.5); 81 if β is equal to 1 (90 × 0.9¹); and 72.9 for β equal to 2 (90 × 0.9²). Accordingly, the region would remain a transition region in the first two policy scenarios, but would become less developed in the last one. It should also be noted that whatever the value for the policy parameter β, the prosperity gap of the region with respect to the EU27 average would be altered, thus potentially affecting the final amount of funds received.

5 Results and Discussion

5.1 The Baseline Scenario

In a baseline scenario, regions’ GDPpc relative to the EU27 average—that is, the reference taken by the EC to determine their eligibility status for funding from the ERDF and the ESF+ in the period 2021–2027—has been adjusted by regional levels of relative social progress. To that end, four adjustment factors have been considered; namely, the EUSPI itself, and its dimensions of basic human needs, foundations of well-being, and opportunities. The policy parameter β is assumed to be equal to 1 in this scenario, thus giving income per capita and social progress the same importance in the calculation of the adjusted GDPpc indicator.⁵

Figure 2 shows the correlation between GDPpc and adjusted GDPpc, which is positive and statistically significant at 1% in all cases. The correlation is greater when GDPpc is adjusted with basic human needs (0.991), and decreases when the adjustment factors are either the foundations of well-being (0.973) or opportunities (0.959). In simpler terms, the relationship between GDPpc and adjusted GDPpc weakens as income is adjusted with more sophisticated facets of social progress.⁶ In addition, the average difference in the ranking of the 240 EU regions according to their GDPpc and adjusted GDPpc is 6.4 positions when income is adjusted with basic human needs, 11 with foundations of well-being, and 13.6 with opportunities. The average difference is 9.2 positions when GDPpc is adjusted with the EUSPI. These figures show that the dimensions of social progress that would most affect the EU regions’ eligibility for funding from the ERDF and the ESF+ are opportunities followed by foundations of well-being.

Fig. 2

Source Own elaboration

EU regions GDPpc versus adjusted GDPpc. Average 2015–2017 (PPS; EU27 = 100). Notes Luxembourg (LU) is not displayed in the scatterplots because the scale of both axes has been truncated at 275. This allows a better visualization of differences between relative GDPpc and adjusted GDPpc. ^*** means statistical significance at 1%.

Looking at Fig. 2 it can also be seen that the prosperity gap shown by less developed regions—i.e., their position relative to the EU27 average—assessed with adjusted GDPpc tends to be worse than when evaluated with GDPpc. This means that regions with an income per capita well below the average also often score lower in terms of social progress. Just the opposite happens, albeit with some exceptions, with more developed regions, the relative position of which normally improves when assessed with adjusted GDPpc. Therefore, the average prosperity gap increases when non-economic issues are accounted for, meaning the adjustment is useful for providing a more accurate view of actual EU regional disparities. As for transition regions, their relative position either improves or worsens with adjusted GDPpc, but there is no clear pattern; and this could affect their eligibility for funding from the ERDF and the ESF+. These findings underscore the relevance of this research as they suggest that income alone might not properly account for the structural preconditions for successful regional development.

The eligibility status of EU regions—i.e., less developed, transition or more developed regions—has been re-assessed with their GDPpc adjusted by the EUSPI, as well as with foundations of well-being and opportunities, which are the more influential dimensions of social progress in the construction of adjusted GDPpc. The assumption of neutral preferences regarding income and social progress is maintained. The transition probabilities matrix displayed in Table 2 shows the share of regions that would retain their current eligibility status—according to GDPpc—when assessed with adjusted GDPpc, and the percentage that would change status.

Table 2 GDPpc versus GDPpc adjusted by social progress (EUSPI), foundations of well-being and opportunities. The transition probabilities matrix

The largest probability always corresponds to the event that a region has the same eligibility status according to both GDPpc and adjusted GDPpc, although other interesting results arise. In this regard, 95.8% of the more developed regions according to GDPpc are classified in the same group with GDPpc adjusted for social progress, 4.2% would become transition regions, and no region would change category to less developed region. Also, 98.7% of the less developed regions according to GDPpc would maintain their status with adjusted GDPpc, and the remaining 1.3% would become transition regions.

The most interesting results, however, correspond to the group of transition regions. While 61.2% of them would maintain their status with social progress-adjusted GDPpc, 26.9% would become more developed regions and 11.9% less developed regions, meaning a change in their eligibility status to receive funds from the EU regional Cohesion Policy. The results are similar when GDPpc is adjusted for the foundations of well-being and opportunities. Notably, when income is adjusted by opportunities, the number of transition regions that would move to either less developed regions (20.9%) or more developed regions (29.9%) increases.

Panel (a) of Map 3 classifies the EU regions with their social progress-adjusted GDPpc, where less developed regions have an adjusted GDPpc of less than 75% of the EU27 average, transition regions between 75 and 100% of the average, and more developed regions have an adjusted GDPpc greater than 100% of the average. Panel (b) compares regions’ eligibility status calculated using adjusted GDPpc with their current status according to the EU regional Cohesion Policy for the period 2021–2027—depicted in Map 1 and based solely on GDPpc. According to this comparison, the losers are those regions that would move to a less favourable eligibility status for funding when assessed with adjusted GDPpc—either changing from a less developed region to a transition region, or from a transition region to a more developed region. Similarly, the winners are those regions that would change to a more favourable status—either changing from a more developed region to a transition region, or from a transition region to a less developed region. It is worth noting that even if a region does not change its eligibility status, its prosperity gap can widen or narrow when assessed with adjusted GDP; as noted in Sect. 4, this would also potentially affect funding from the EU’s Cohesion Policy.

Map 3

Source Own elaboration

Eligibility for funding from the ERDF and the ESF + for the period 2021–2027 according to GDPpc adjusted by social progress (EUSPI). Winner and loser regions. Note The losers are regions that change categories either from less developed regions to transition regions or from transition regions to more developed regions, whereas the winners are regions that change either from more developed regions to transition regions or from transition regions to less developed regions.

In this baseline scenario in which income per capita and social progress are given the same importance in building adjusted GDPpc, the losers (19 regions) and winners (12 regions) account for 12.9% of the 240 NUTS2 EU regions and are listed in Table 3. As already anticipated from the matrix of probabilities, they are mostly transition regions according to their GDPpc, and their adjusted GDPpc would either worsen their eligibility status by pushing them up into the category of more developed regions, or improve it by dropping them into the group of less developed regions. A geographical pattern is also apparent in the distribution of these regions. The winners are mostly Italian regions, although there are also Polish, Bulgarian and Spanish regions, in addition to the Belgian region of Hainaut (BE32). Conversely, the losers are Nordic regions, together with some French, German, Belgian and Dutch ones, as well as the Spanish region of La Rioja (ES23). Except for the Belgian region of Luxembourg (BE34), they are all transition regions according to their GDPpc with above-average levels of social progress, which leads them to be classified as more developed regions when categorized using adjusted GDPpc.

Table 3 Winner and loser regions according to eligibility with GDPpc adjusted by social progress (EUSPI) in the baseline scenario (policy parameter $\upbeta$ equal to 1)

The analysis of losing and winning regions has been repeated using GDPpc adjusted first with foundations of well-being and then with opportunities, to check the sensitivity of the simulated policy-reform to the index of social progress employed to adjust GDPpc. The results are quite similar to those noted above in terms of both the number of regions affected and their geographical distribution—see Maps 7 (foundations of well-being) and 8 (opportunities) in the Appendix. It is noteworthy that when income is adjusted by opportunities, the number of regions that would change their eligibility status rises to 40—half of them losers and the other half winners—representing 16.7% of the 240 NUTS2 EU regions. The new affected regions are mostly transition regions that, given their poor records in the most advanced facets of social progress, would become less developed regions.

5.2 Changing the Policy Preferences Regarding Income and Social Progress

In the previous baseline scenario, the preferences of policymakers were represented by assigning the same importance to income per capita and social progress in the construction of the adjusted GDPpc indicator; i.e., a parameter $\upbeta$ equal to 1 was assumed. In this Section, GDPpc is adjusted by social progress—through the EUSPI—in two further scenarios in which this policy parameter takes, respectively, the values of 0.5 and 2. Although these are rather ad hoc benchmarks, they make it possible to test how assigning different weightings to income per capita and social progress in the computation of adjusted GDPpc could affect EU regions’ status of eligibility for funding from the ERDF and the ESF+—information that might be useful for policymakers. The new losers and winners in these scenarios are plotted in Map 4.⁷

Map 4

Source Own elaboration

Winner and loser regions in alternative policy scenarios regarding the relative importance of income per capita and social progress in the building of social progress-adjusted GDPpc. Note The losers are regions that change categories either from less developed regions to transition regions or from transition regions to more developed regions, whereas the winners are regions that change either from more developed regions to transition regions or from transition regions to less developed regions.

Before commenting on these results, it is worth highlighting that, because of the nature of the adjusted indicator proposed in this paper, larger (smaller) values of the policy parameter $\upbeta$ increase (decrease) the number of regions affected by changes in their eligibility status. In the policy scenario in which social progress is half as important as income per capita, as indicated by a parameter $\upbeta$ equal to 0.5, the number of losers (9 regions) and winners (7 regions) noticeably decreases with respect to the baseline scenario, as might be expected (Map 4, panel a). The regions that would no longer be considered winners are located in Italy—Friuli-Venezia Giulia (ITH4), Liguria (ITC3), Toscana (ITI1) and Abruzzo (ITF1)—and Spain—Murcia (ES62). Generally speaking, they are all Southern European regions scoring poorly in terms of social progress. The first three of the aforementioned Italian regions are more developed regions according to GDPpc that had become transition regions in the baseline scenario; however, they now regain their initial status given the lesser importance given to social progress when calculating adjusted GDP. Abruzzo (ITF1) and Murcia (ES62) are transition regions according to their GDPpc, which became less developed regions in the baseline scenario, but now regain their initial eligibility status.

On the other hand, some regions stand to benefit from this new scenario as they would no longer be considered losers, as they were in the baseline scenario. Except for Luxembourg (BE34), they are transition regions in Central and Northern Europe that reached the status of more developed regions in the baseline scenario due to their higher social progress. They include Sjælland (DK02) in Denmark; Limburg (BE22) in Belgium; Trier (DEB2) and Dresden (DED2) in Germany; Friesland (NL12) and Drenthe (NL13) in The Netherlands; in addition to three French regions—Aquitaine (FRI1), Bretagne (FRH0) and Pays de la Loire (FRG0). The lesser weighting given to social progress in this new policy scenario would allow these regions to remain as transition regions.

Moving on to the scenario in which the policy parameter $\upbeta$ is set to 2—thus giving social progress double the weighting of income per capita—the number of regions affected noticeably increases, again as expected (Map 4, panel b). A total of 31 losers would see their status of eligibility become worse—12 more than in the baseline scenario—and 19 winners would improve—7 more. That is, one out of every five EU regions changes its eligibility status in this simulated scenario that rewards social progress in the computation of adjusted GDP. All regions would also register changes in their prosperity gap.

The new loser regions are mostly French—Auvergne (FRK1), Bourgogne (FRC1), Limousin (FRI2), Centre-Val de Loire (FRB0), Poitou–Charentes (FRI3) and Haute-Normandie (FRD2)—and German—Lüneburg (DE93) and Thüringen (DEG0). However, other losers are Burgenland (AT11), Northern and Western Ireland (IE04), and the Eastern regions of Malta (MT) and Severovýchod (CZ05). They are mostly transition regions with above-average social progress, which pushes them into the category of more developed regions. Conversely, the new winners are some Southern and Eastern more developed and transition regions with poor social progress, including Kýpros (CY), Attiki (EL30), Veneto (ITH3), Umbria (ITI2), Lazio (ITI4), Algarve (PT15) and Bucureşti-Ilfov (RO32). The case of the Romanian region of Bucureşti-Ilfov (RO32) stands out, as already noted in Sect. 3. It ranks 25th among the 240 EU regions with a GDPpc of 141 relative to the EU27 average in 2015–2017; nevertheless, a particularly poor record on social progress—0.811 with respect to the average—brings it down from a more developed region to a transition region in this scenario.

5.3 Explaining the Difference Between GDPpc and Adjusted GDPpc

One relevant issue for policymakers responsible for the EU’s Cohesion Policy concerns the factors that might help to explain the difference between regions’ relative scoring according to social progress-adjusted GDPpc and GDPpc. This is important because, even if the eligibility status of a region does not change, the amount of funds received may be altered since its prosperity gap—which, as explained in Sect. 2, is a main factor determining the funds assigned to less developed regions and transition regions—would actually change. In order to shed some light on this matter, two complementary analyses have been conducted with the results from the baseline scenario. The first relies on Bayesian Model Averaging (BMA) and the second on Relative Importance (RI) techniques. The methodological details of these approaches are described in the Appendix.

The BMA, on the one hand, provides a probabilistic ranking of the importance of the 12 components making up the EUSPI as explanatory variables of the difference in regions’ performance computed as adjusted GDPpc minus GDPpc, both expressed relative to the EU27 average. The findings when income per capita is adjusted by social progress (EUSPI) reveal that personal freedom and choice, access to advanced education and lifelong learning (LLL), environmental quality, and health and wellness are the leading predictors (Table 4). In simpler terms, in regions with better job opportunities, more tertiary-educated citizens, lower air pollution, less corruption and longer life expectancy, among other factors, adjusted GDPpc is expected to be much larger than GDPpc. This is the case for most regions in Nordic countries, including Stockholm (SE11), Hovedstaden (DK01) and Helsinki-Uusimaa (FI1B). At the opposite extreme, social progress-adjusted GDPpc is far below GDPpc in regions of some Southern European countries, and particularly Eastern European countries, such as Sud-Est (RO22), Bucureşti-Ilfov (RO32), Notio Aigaio (EL42) and Yugozapaden (BG41), to name just a few. Most of these regions are poorly endowed with the ingredients of social progress mentioned above.

Table 4 Explaining the difference between social progress-adjusted GDPpc and GDPpc. Bayesian Model Averaging (BMA)

It is worth mentioning that the most influential components of the EUSPI when it comes to explaining differences between regions’ adjusted GDPpc and GDPpc include most of the variables involved—besides income—in the allocation of funds from the regional Cohesion Policy in 2021–2027. As described in Sect. 2, these are: labour market—the component of personal freedom and choice includes variables closely related to employment and job opportunities; education—tertiary education is also included in the component of personal freedom and choice; migration—tolerance towards immigrants is an indicator of tolerance and inclusion; and climate change—the environmental quality component accounts for air pollution. This is good news for European policymakers, as it suggests that the non-income criteria employed to allocate the cohesion funds might also be contributing to boosting cohesion in social progress across the EU regions.

It should be noted, however, that a key feature of the indicator proposed in this research is that it directly affects the categorization of regions as less developed, transition or more developed, thus having much larger implications for funding allocation. In a nutshell, the analysis suggests that combining the indicator proposed in this research—which directly affects the initial categorization of regions and their prosperity gap—with the current EC criteria used for funding allocation following classification might result in a much-improved allocation of resources.

The RI analysis, on the other hand, assesses the relative contribution of each of the components of the EUSPI, or its dimensions, to explaining the difference in regions’ performance according to their adjusted GDPpc and GDPpc, as defined above. The results, which are shown in Table 5, reveal that the most influential components when GDPpc is adjusted with social progress—i.e., the EUSPI—are access to advanced education and lifelong learning (LLL) (20.8%), personal freedom and choice (18.9%), environmental quality (10.8%), and personal rights (9.5%). In the simulated scenario in which income per capita is adjusted by foundations of well-being, the contributions of environmental quality (45.4%) and access to ICTs (28.2%) stand out. Likewise, when the adjustment factor is opportunities, access to advanced education and lifelong learning (LLL) (39.8%) and personal freedom and choice (23.5%) are the most relevant components.

Table 5 Contribution of the social progress components to the difference between adjusted GDPpc and GDPpc. Relative Importance (RI) analysis

6 Summary, Conclusions and Policy Issues

European Union (EU) policymakers are unwavering in their commitment to fostering economic, social and territorial cohesion; in fact, the European Commission (EC) currently dedicates nearly one third of its budget to the regional Cohesion Policy. The allocation of cohesion funds across the EU regions and Member States for the period 2021–2027 is still principally based on income per capita. Cohesion, however, goes far beyond income to include other social and environmental issues that really matter for citizens’ lives. Whereas the limitations of income per capita as an indicator of the progress of societies have been widely recognized in the literature, recently renewed policy goals of the EC urgently call for new indicators to use in addition to income as criteria for the allocation of cohesion funds.

In this context, this paper contributes an indicator of the GDPpc of the EU regions adjusted by their level of social progress, which is measured by the 2020 release of the European Social Progress Index (EUSPI) produced by the EC. Notably, this indicator involves a policy parameter that makes it possible to account for policymakers’ preferences regarding the relative importance of income and social progress. Combining GDPpc and the EUSPI into a single composite indicator is in itself a novel contribution to the Beyond GDP literature.

The results bring to light the fact that less developed EU regions—according to the EC classification for the period 2021–2027—tend to score worse relative to the EU27 average when assessed with adjusted GDPpc instead of GDPpc; whereas more developed regions tend to do relatively better. In simpler terms, regions with higher income per capita tend to score better for their social progress; conversely, the poorest regions according to income per capita also seem to score lower regarding social progress. This result suggests that the actual development gap across EU regions is in fact larger than that driven by GDPpc. The findings are mixed for the group of transition regions, which either score better or worse with adjusted GDPpc. It is worth noting that assessing the level of development of EU regions with social progress-adjusted GDPpc changes their prosperity gap relative to the EU27 average; and it could also affect their eligibility status for funding from the EU’s Cohesion Policy. Both features could have an impact on the amount of funds received because regions’ prosperity gap and their eligibility status are key factors determining funding in the period 2021–2027.

Accordingly, a second contribution of the paper is the comparison of the current eligibility status of EU regions for funding—determined by their relative GDPpc—with their status based on several simulated policy scenarios in which eligibility is assessed using social progress-adjusted GDPpc. In this regard, the vast majority of less developed regions and more developed regions would maintain their current status. However, a number of transition regions would change either for the better or worse; particularly when income is adjusted with the most advanced facets of social progress represented by opportunities. The regions that would change to a less favourable status have been termed losers; they are mostly Central and Northern European transition regions highly endowed with the components of social progress, which push them to the status of more developed regions. Furthermore, the winners are transition regions that would become less developed regions due to their poor scoring regarding social progress; most of them are Italian, although there are also some Polish, Bulgarian and Spanish regions. The number of regions that would change their eligibility status increases as social progress is assigned greater importance in the construction of adjusted GDPpc.

In all, this research provides EU policymakers with sound information regarding the potential effects of allocating cohesion funds in simulated policy scenarios in which eligibility is determined with a composite indicator that accounts for both income per capita and social progress. In this regard, combining economic and non-economic facets of development into a single indicator could help to achieve a more comprehensive distribution of cohesion funds, in line with the renewed goals of the Cohesion Policy for the period 2021–2027 and the Agenda 2030; it could also bring funding allocation closer to the subjects that matter for citizens’ quality of life.

Appendix

1.1 The Bayesian Model Averaging Approach

The Bayesian Model Averaging (BMA) analysis is grounded in the estimation of a sequence of models in which the dependent variable is the adjusted GDPpc of region $\mathrm{r}= 1,\dots ,$ 240 minus its GDPpc, both relative to the EU27; accordingly, positive values mean better relative performance with the adjusted GDPpc. On the other hand, the explanatory predictors are the regional scores of the $\mathrm{C}= 1,\dots ,$ 12 components of the EUSPI:

$$\mathrm{y}=\mathrm{\alpha }{\upiota }_{\mathrm{n}}+\mathrm{X\delta }+\upvarepsilon$$

(A1)

with y being a $240\times 1$ dimensional vector containing the observations on each one of the 240 NUTS2 EU regions of the difference in performance in terms of adjusted GDPpc and GDPpc; the parameter $\mathrm{\alpha }$ represents a constant term; ${\upiota }_{\mathrm{n}}$ is a $240\times 1$ vector of ones; X is a $240\times 12$ matrix of component scores with associated response parameters $\updelta$ contained in a $12\times 1$ vector; lastly, $\varepsilon$ is a $240\times 1$ vector of disturbances.

There are many sub-models ${\mathrm{M}}_{\mathrm{k}}$ of the general model in (A1), which are given by the subsets of coefficients ${\upeta }^{\mathrm{k}}=\left(\mathrm{\alpha },{\updelta }^{\mathrm{k}}\right)$ and combinations of the regressors; in particular, the cardinality of the model space is of 2¹² = 4,096 models. The objective of the BMA approach is to estimate the probabilistic importance of the explanatory predictors by considering these 4,096 potential models. The results allow the assessment of the statistical significance of each predictor (Steel, 2020).

In this framework, computing Posterior Inclusion Probabilities—which are the standard Bayesian measure of variable importance—produces a probabilistic ranking of the importance of the 12 components of the EUSPI in explaining regional differences of performance according to adjusted GDPpc and GDPpc. Formally, the Posterior Inclusion Probability for each component $\mathrm{C}$ is computed as:

$$\mathrm{p}\left(\left.{\upeta }_{\mathrm{C}}\ne 0 \right|\mathrm{ y},\mathrm{X}\right)=\sum_{\mathrm{k}=1}^{\mathrm{4,096}}\mathrm{p}\left({\mathrm{M}}_{\mathrm{k}} |{\upeta }_{\mathrm{C}}\ne 0,\mathrm{y},\mathrm{X}\right)$$

(A2)

In addition, the Posterior Sign Certainty of component $\mathrm{C}$ is calculated as:

$$\mathrm{p}\left({\upeta }_{\mathrm{C}}\ge 0 |\mathrm{ y},\mathrm{X}\right)=\sum_{\mathrm{k}=1}^{\mathrm{4,096}}\mathrm{p}\left({\upeta }_{\mathrm{k},\mathrm{C}} |{\mathrm{ M}}_{\mathrm{k}},\mathrm{y},\mathrm{X}\right)\mathrm{ p}\left({\mathrm{M}}_{\mathrm{k}} |\mathrm{ y},\mathrm{X}\right)$$

(A3)

Values close to 1 for this Posterior Sign Certainty mean that the estimated parameter for component C is positive in the vast majority of models; conversely, values close to 0 suggest that the effect on the dependent variable is negative. The computation of these metrics requires calculating the Posterior Model Probability of each sub-model ${\mathrm{M}}_{\mathrm{k}}$ as:

$$\mathrm{p}\left({\mathrm{M}}_{\mathrm{k}} |\mathrm{ y},\mathrm{X}\right)=\mathrm{p}\left(\mathrm{y},\mathrm{ X }| {\mathrm{M}}_{\mathrm{k}}\right)\mathrm{ p}\left({\mathrm{M}}_{\mathrm{k}}\right) / \sum_{\mathrm{k}=1}^{\mathrm{4,096}}\mathrm{p}\left(\mathrm{y},\mathrm{ X }|{\mathrm{ M}}_{\mathrm{k}}\right)\mathrm{ p}\left({\mathrm{M}}_{\mathrm{k}}\right)$$

(A4)

where $\mathrm{p}\left(\mathrm{y},\mathrm{X }| {\mathrm{M}}_{\mathrm{k}}\right)$ is the marginal likelihood, and $\mathrm{p}\left({\mathrm{M}}_{\mathrm{k}}\right)$ the prior model probability.

The computation of all the abovementioned metrics requires the definition of priors on both the model and the parameter spaces. In this regard, a binomial prior is defined on the model space, adjusted such that every model has the same a priori probability, while a Zellgner g-prior based on the Bayesian Risk Inflation Criterion is employed in the parameter space.

1.2 Relative Importance Analysis

The Relative Importance (RI) analysis allows the assessment of the relative contribution of each component of social progress to the explained variability of the dependent variable—i.e., the adjusted GDPpc of region r minus its GDPpc—while considering its multivariate interactions with the other components. In formal terms, the RI decomposes the R-squared of the full model by considering all potential sub-models given the set of explanatory variables. In doing so, the Genizi metric, CaR scores (Zuber & Strimmer, 2011) and the LMG metric (Gromping, 2007) are employed. The output is summarized with the average over the R-squared decompositions in all potential models.

1.3 Maps

See Maps

Map 5

Source Own elaboration

The distribution of the three dimensions of the EUSPI across the EU regions. Note The different categories are set according to the quintiles of the distribution of each dimension of the EUSPI. Figures in square brackets are the lower and upper cut-off points of each category.

5,

Map 6

Source Own elaboration

EU regions with social progress (EUSPI) below and above the EU27 average.

6,

Map 7

Source Own elaboration

Eligibility for funding from the ERDF and the ESF + for the period 2021–2027 according to GDPpc adjusted by foundations of well-being. Winner and loser regions. Note The losers are regions that change category either from less developed regions to transition regions or from transition regions to more developed regions, whereas the winners are regions that change either from more developed regions to transition regions or from transition regions to less developed regions.

7 and

Map 8

Source Own elaboration

Eligibility for funding from the ERDF and ESF + for the period 2021–2027 according to GDPpc adjusted by opportunities. Winner and loser regions. Note The losers are regions that change category either from less developed regions to transition regions or from transition regions to more developed regions, whereas the winners are regions that change either from more developed regions to transition regions or from transition regions to less developed regions.

8