Regional employment rates in Europe, 2024–2025: intra-national divergence, measurement limits, and why convergence did not occur automatically

Scope — This assessment examines EU-27 NUTS-2 employment outcomes principally for 2025, using the EU Labour Force Survey and Eurostat regional series, with 2024 used for temporal comparison and Q2 2026 used only as an aggregate EU contextual observation; EFTA states, candidate countries, the United Kingdom and other non-EU territories are excluded from the EU ranking and treated separately wherever analytically necessary.

---

Executive Summary / BLUF

The 2025 EU Labour Force Survey does not describe one integrated European labour market but a distribution of regional labour markets operating under common European rules while producing radically different employment outcomes, because among people aged 20–64 the recorded employment rate ranges from 86.9% in Warszawski stołeczny (PL91) to 36.9% in Mayotte (FRY5), a 50.0 percentage-point observed spread, while continental southern regions such as Calabria (ITF6, 50.3%), Campania (ITF3, 50.8%) and Sicilia (ITG1, 51.4%) remain close to the bottom of the EU distribution.

The evidence nevertheless requires a qualification to the proposition that “convergence did not occur,” because Eurostat’s population-weighted dispersion measure indicates that some sigma-convergence did occur at EU level, with the coefficient of variation of regional employment rates declining from 12.5% in 2015 to 9.2% in 2025; the defensible conclusion is therefore that convergence has been partial, geographically asymmetric and emphatically non-automatic, rather than absent everywhere.

The strongest counterexample to national-average analysis is Italy, where the 2025 population-weighted coefficient of variation in regional employment rates reached 15.1%, the highest among the EU countries for which Eurostat publishes that dispersion measure, while the national 20–64 rate was 67.6%, Provincia Autonoma di Bolzano/Bozen stood at 80.0%, and Calabria at 50.3%, meaning that the national statistic compresses radically different territorial equilibria into a number that corresponds poorly to the labour-market conditions encountered in either region.

At EU level, 115 of 244 NUTS-2 regions, or 47.1%, reached or exceeded the 78% 2030 employment target in 2025, whereas 64 of 244 remained below 74%, demonstrating simultaneously that the Union has moved closer to its aggregate employment objective and that the gains remain strongly spatially concentrated.

The policy implication is not that cohesion instruments produced no convergence, because the official record contradicts such an absolute proposition, but that market integration and fiscal redistribution cannot be assumed to eliminate differences rooted in productivity, skills, firm concentration, demographic structure, accessibility and agglomeration dynamics; the Commission’s own cohesion evidence records strong convergence in much of Central and Eastern Europe alongside exceptionally weak long-run growth in parts of southern Europe and acknowledges that national averages conceal regional disparities in employment and other social indicators.

The measurement qualification is also material, because the EU-LFS is a sample survey of residents rather than a census of jobs, regional results refer to respondents’ place of residence rather than necessarily their place of work, and unregistered economic activity cannot be assumed to be observed exhaustively; consequently, the southern employment deficit should be interpreted as a combination of genuine non-employment, inactivity and some degree of measurement incompleteness rather than treated either as a perfect census or dismissed because informal activity exists.

Finally, the Q2 2026 EU employment rate of 76.4% for people aged 20–64, accompanied by labour-market slack of 11.0% of the extended labour force, demonstrates continuing improvement in the aggregate without changing the analytical conclusion about regional heterogeneity, because a quarterly EU-wide mean cannot identify whether the additional employment is occurring in Calabria, Utrecht, Prague or another already high-employment region.

Israel Under Institutional Pressure in Europe: Italy’s Anti-Israel Turn and the Strategic Cost of Technological Disengagement

Europe’s employment divide is no longer mainly between countries

Europe’s labour-market problem is no longer captured by the familiar contrast between stronger northern economies and weaker southern ones, because the 2025 EU-LFS regional map shows a 50.0 percentage-point spread between Warszawski stołeczny at 86.9% and Mayotte at 36.9% for people aged 20–64, while Calabria, Campania and Sicilia remain near 50–51% despite operating inside the same Single Market, the same EU cohesion architecture and, for Italy, the same national fiscal and social-security system. The central policy failure is therefore not the absence of convergence, because Eurostat’s population-weighted coefficient of variation fell from 12.5% in 2015 to 9.2% in 2025, but the persistence of regional productive structures that transfers and integration have narrowed without equalising. The immediate stakes are fiscal and industrial: Europe can sustain weaker territories through redistribution, but it cannot manufacture convergence unless spending changes the productive capacity that determines where firms invest, where skilled workers remain and where employment is generated. Eurostat — Labour market statistics at regional level, 2025

The EU average has improved while its regional floor remains structurally low

The European Union reached an employment rate of 76.1% for people aged 20–64 in 2025, with 197.7 million employed, 12.1 million unemployed and 49.8 million outside the labour force, yet that aggregate sits above a distribution in which 115 of 244 NUTS-2 regions, or 47.1%, had already reached the EU’s 78% employment target while 64 regions remained below 74%. The numbers therefore describe two realities at once: the Union is close to its aggregate policy objective, but a large territorial bloc remains substantially below it, which means that a rising European mean can coexist with entrenched regional underperformance rather than dissolve it. Eurostat — EU employment rate grew above 76% in 2025 Eurostat — Labour market statistics at regional level, 2025

The upper tail is geographically diverse rather than confined to one national model, because Warszawski stołeczny recorded 86.9%, Praha 86.0%, Åland 85.5%, Utrecht 85.4%, Oberbayern 85.1% and Mellersta Norrland 85.0%, while the lower tail included Mayotte at 36.9%, Guyane at 50.0%, Calabria at 50.3%, Campania at 50.8% and Sicilia at 51.4%. The structural message is that the Single Market has not produced one labour market, but a network of regional labour markets whose employment intensity depends on local productive density, participation, human capital and accessibility far more than national averages imply. Eurostat — Labour market statistics at regional level, 2025

Italy shows why the national average is an administrative statistic, not an economic equilibrium

Italy is the clearest demonstration of the analytical limits of national averages because its 20–64 employment rate was 67.6% in 2025, the lowest among EU Member States, while Provincia Autonoma di Bolzano/Bozen stood at 80.0% and Calabria at 50.3%, producing a 29.7 percentage-point internal gap inside one state. Eurostat’s population-weighted coefficient of variation for Italian regional employment rates reached 15.1%, the highest level among the EU countries covered by that comparison, against 8.0% in Romania, 7.9% in Belgium and 6.1% in Spain. Eurostat — Labour market statistics at regional level, 2025

The Italian divide is also embedded in labour-force composition rather than explained only by unemployment, because the 2025 gender employment gap reached 28.7 percentage points in Puglia, 28.2 in Campania, 28.1 in Sicilia, 25.4 in Calabria and 25.3 in Basilicata, compared with an EU-wide gap of 9.7 points. That distribution means that a policy designed around the national 67.6% rate obscures not only the north-south divide but also the mechanism sustaining it, since southern underperformance combines weak labour demand with much lower female participation and employment. Eurostat — Labour market statistics at regional level, 2025

Convergence happened, but it did not create comparable labour-market structures

The evidence does not support the claim that European convergence failed outright, because the EU-wide population-weighted coefficient of variation in regional employment rates fell from 12.5% in 2015 to 9.2% in 2025, while disparities narrowed in 15 of the 17 Member States for which Eurostat reports a comparable series. Spain and Portugal are particularly important because both recorded major reductions in internal dispersion, while Portugal finished 2025 with seven of nine NUTS-2 regions at or above 78% and a regional coefficient of variation of 2.0% or less, demonstrating that southern or peripheral geography does not mechanically condemn a country to persistent employment divergence. Eurostat — Labour market statistics at regional level, 2025

Yet sigma-convergence is not the same as structural equalisation, because a narrower distribution can coexist with persistent low-employment clusters and fundamentally different productive systems. Romania illustrates that contradiction particularly well: București-Ilfov reached 81.1% in 2025, compared with the national 69.0%, while Sud-Vest Oltenia stood at 61.7% and Sud-Est at 62.1%; at the same time, the European Commission records the broader Central and Eastern European convergence story in which GDP per head rose from roughly 45% of the EU average in 1995 to nearly 80% by 2021. National convergence toward Europe can therefore occur while capital regions pull away from domestic peripheries. European Commission — Ninth Report on Economic, Social and Territorial Cohesion Eurostat — Labour market statistics at regional level, 2025

Transfers protect income, but productive geography decides whether convergence becomes self-sustaining

The EU has not ignored regional divergence fiscally, because cohesion policy amounted to approximately €347 billion in current prices for 2007–2013, approximately €351.8 billion for 2014–2020, and €392 billion in current-price global resources for 2021–2027, yet those totals cannot be interpreted as a single “anti-divergence” budget because they finance infrastructure, competitiveness, human capital, environmental transition and social objectives across different programming periods. What they do establish is that persistent regional underperformance has survived repeated, large-scale intervention, making the composition and transmission of spending more important than its gross volume. European Commission — 2021–2027 cohesion policy initial allocations European Commission — Cohesion Policy 2007–2013 guide

The Commission’s own RHOMOLO framework makes the distinction explicit by separating short-run demand effects from long-run supply effects transmitted through physical-capital accumulation, human capital and technological progress, which means that a transfer capable of raising household consumption does not automatically alter the region’s production function. The decisive question is whether public resources create infrastructure that firms use, skills that local employers demand, supplier networks that retain value added, innovation capacity that raises productivity and conditions attractive enough to prevent trained workers from leaving. European Commission — Macroeconomic impact of 2021–2027 cohesion-policy programmes

Agglomeration explains why integration can reinforce the winners as well as lift the laggards

Regional divergence persists because productive assets are complementary, which means that firms do not locate where wages are lowest or fiscal need is greatest, but where labour, suppliers, infrastructure, finance, research and markets combine to maximise expected returns. OECD regional evidence records productivity gaps in which the most productive region within a country produces, on average, almost twice the labour productivity of the least productive region, while dense metropolitan economies systematically benefit from better matching, deeper supplier networks and stronger knowledge spillovers. OECD — Productivity and innovation in regions

This mechanism turns the Single Market into both a convergence machine and a sorting mechanism, because lower barriers allow weaker regions to attract capital and integrate into continental value chains, but they also allow firms and skilled labour to move more easily toward already-dense productive centres. The success of Praha at 86.0% and Warszawski stołeczny at 86.9% proves that integration can generate powerful catching-up outside the old western core, while the persistence of southern Italian and Greek weakness proves that integration does not supply missing complementarities automatically. Eurostat — Labour market statistics at regional level, 2025

The policy divide is now between regions that can compound investment and regions that cannot

Germany and the Netherlands show what distributed productive density looks like in employment terms, because 34 of Germany’s 38 NUTS-2 regions and all 12 Dutch regions were at or above the 78% target in 2025, while Utrecht reached 85.4% and Oberbayern 85.1%. The difference with Italy is therefore not only one of national averages but of geographic diffusion: high employment is spread across multiple regional economies rather than concentrated in one national growth pole. Eurostat — Labour market statistics at regional level, 2025

The opposite risk is visible in regions caught between low employment and human-capital loss, because the Commission’s Ninth Cohesion Report identifies development traps where weak productivity, employment and income growth reinforce one another, while its demographic analysis identifies talent-development traps in territories struggling to develop, attract or retain skilled workers. Once skilled workers leave because firms are absent and firms stay away because skilled workers are absent, transfer dependence becomes a symptom of failed productive coordination rather than its cause. European Commission — Ninth Cohesion Report, Economic Cohesion European Commission — Demographic transition and talent-development traps

Over the next 12–24 months, the cost of inaction will be paid in lost labour, not only lost output

The immediate policy risk is that the EU employment rate can continue rising from the 76.1% annual level recorded in 2025 toward the 78% 2030 target while the territorial structure beneath it changes far more slowly, because aggregate improvement does not establish that Calabria, Campania, Sicilia, low-employment Greek regions or French outermost territories are closing their gap with Utrecht, Praha or Oberbayern. The September 2026 quarterly EU rate of 76.4% for people aged 20–64 shows that aggregate labour-market conditions continued to improve, but it cannot identify whether the lower tail is converging or merely being outweighed statistically by stronger regions. Eurostat — Employment rate up in Q2 2026

Over the next 12–24 months, the decisive test will therefore be whether cohesion expenditure and national policy convert rising employment into stronger productive capacity in the lagging regions, because failure will impose costs first on younger and more mobile workers, who face incentives to leave, then on local employers, which inherit a thinner labour pool, and finally on national budgets, which must finance larger compensatory transfers from a weaker regional tax base. The political cost will fall on governments that continue to legislate around national averages, but the economic cost will be borne inside the regions where employment remains around 50–60% while the European frontier remains above 85%, because every year in which human capital, firms and investment continue to compound elsewhere makes later convergence more expensive rather than more automatic.

Amazon cocaine corridors are shifting Europe’s port risk

---

Navigational Index

I. The map beneath the national averages

The decisive empirical question is not merely which Member State has the highest employment rate, but how a distribution extending from approximately 87% to approximately 37% can coexist inside the same regulatory and economic union, and why large parts of southern Italy, southern Spain, Greece and the French outermost territories continue to occupy the lower tail.

II. Partial convergence without equalisation

The relevant analytical distinction is between measurable reduction in overall dispersion and convergence toward comparable labour-market structures, because Eurostat records declining EU-wide regional dispersion while simultaneously recording persistent extreme gaps, especially inside Italy and in several southern and peripheral regional systems.

III. From transfers to productive geography

The principal policy question is whether fiscal transfers and integration can themselves overcome regional differences in human capital, productive density, investment attraction and agglomeration, or whether spatially self-reinforcing productive structures require policies capable of changing the regional production function rather than merely household disposable income.

---

Master Abstract

The central finding is divergence within integration

The 2025 regional evidence supports a stronger and more precise proposition than the familiar claim that Europe is divided between high-employment northern states and low-employment southern states, because some of the largest economically meaningful differences occur within Member States rather than between them, and Italy provides the clearest example through an observed 20–64 employment-rate difference of 29.7 percentage points between Provincia Autonoma di Bolzano/Bozen, at 80.0%, and Calabria, at 50.3%, while Eurostat places Italy’s population-weighted regional coefficient of variation at 15.1%, compared with 8.0% in Romania, 7.9% in Belgium and 6.1% in Spain. The inference is not that borders have ceased to matter, because taxation, social protection, labour regulation and national institutions remain decisive, but that the Member State is often too aggregated a statistical unit for understanding the opportunity set actually facing a company choosing where to invest, a household considering migration or a graduate deciding where to begin a career; indeed, the European Commission’s own cohesion analysis states that within-country disparities can be as large as cross-country disparities and explicitly notes that national averages conceal differences in employment rates and social indicators.

The European distribution contains two very different realities

For people aged 20–64 in 2025, Eurostat’s current official regional material identifies Warszawski stołeczny (PL91) at 86.9%, Praha (CZ01) at 86.0%, Åland (FI20) at 85.5%, Utrecht (NL31) at 85.4%, Oberbayern (DE21) at 85.1% and Mellersta Norrland (SE32) at 85.0% as the six NUTS-2 regions recording rates of at least 85%; at the opposite end, Mayotte (FRY5) records 36.9%, Guyane (FRY3) 50.0%, Calabria (ITF6) 50.3%, Campania (ITF3) 50.8% and Sicilia (ITG1) 51.4%, meaning that even after separating the exceptional circumstances of French outermost territories, continental EU regions remain almost 35 percentage points below the most employment-intensive European regions. These observations matter because 115 of 244 regions had reached the 78% EU target by 2025, with dense concentrations in Czechia, Denmark, Germany, Ireland, Hungary, the Netherlands, Portugal, Slovakia and Sweden, while the lower tail remained concentrated in French outermost regions, southern Italy, much of Greece and numerous Spanish regions.

Convergence occurred, but it did not abolish structural geography

The data do not sustain the categorical formulation that convergence simply “did not occur,” because Eurostat’s population-weighted coefficient of variation across EU regional employment rates declined from 12.5% in 2015 to 9.2% in 2025, while 15 of the 17 Member States for which comparable national dispersion results are available recorded declining internal disparities during that interval; Finland, Portugal and Spain registered some of the largest proportional declines, while Denmark and Romania moved in the opposite direction. The stronger conclusion, which survives the data, is that integration did not generate automatic equalisation, because reduced dispersion at the level of the EU distribution can coexist with persistent territorial traps, large internal spreads and enduring bottom-tail clusters; the Commission’s Ninth Cohesion Report evidence is consistent with that interpretation, because it records striking convergence in Central and Eastern Europe while finding average annual real GDP-per-head growth of approximately 0.1% in southern regions over the long-run comparison considered, compared with much stronger catching-up elsewhere, and identifies Italy, Spain, Greece and France among the countries containing substantial numbers of regions whose GDP per head had still not returned to their 2008 level by 2021.

The mechanism is cumulative productive advantage rather than a simple transfer deficit

The available evidence is consistent with an endogenous-growth and agglomeration interpretation in which productive density, skills, firm networks, infrastructure and access to specialised labour operate as complementary assets, so that successful metropolitan and industrial regions acquire additional advantages from already possessing them while weaker regions can lose skilled labour and investment when their local returns remain inferior; Eurostat itself notes that capital regions frequently attract highly qualified workers through a combination of employment, educational and social opportunities, while its 2025 regional material identifies low-employment peripheral regions as places often characterised by restricted opportunities, including for intermediate- and high-skilled labour. This mechanism does not demonstrate that transfers are ineffective, because cohesion spending can improve infrastructure, education, household welfare and investment conditions, and the Commission attributes part of Central and Eastern Europe’s strong convergence to integration and cohesion policy; it does demonstrate, however, why transfers cannot be equated analytically with transformation of the regional production function, because a transfer that sustains household income does not by itself create an industrial cluster, a dense employer base, an innovation ecosystem or a labour market capable of retaining highly skilled workers.

ChatGPT Under the DSA: Europe Regulates AI Search

---

Key Evidence Table

Indicator	Value/status	Reference date	Definition/scope	Issuer	Exact source
EU employment rate	76.1%	2025	Persons aged 20–64, employed / population of same age	Eurostat	Employment and labour force by sex and age — lfsi_emp_a
NUTS-2 regions at/above EU 78% target	115 / 244, 47.1%	2025	Employment rate, persons 20–64	Eurostat	Labour market statistics at regional level — 2025 regional material; lfst_r_lfe2emprtn
Warszawski stołeczny PL91	86.9%	2025	Employment rate, 20–64	Eurostat	Regional LFS / lfst_r_lfe2emprtn
Praha CZ01	86.0%	2025	Employment rate, 20–64	Eurostat	Regional LFS / lfst_r_lfe2emprtn
Åland FI20	85.5%	2025	Employment rate, 20–64	Eurostat	Regional LFS / lfst_r_lfe2emprtn
Utrecht NL31	85.4%	2025	Employment rate, 20–64	Eurostat	Regional LFS / lfst_r_lfe2emprtn
Oberbayern DE21	85.1%	2025	Employment rate, 20–64	Eurostat	Regional LFS / lfst_r_lfe2emprtn
Mellersta Norrland SE32	85.0%	2025	Employment rate, 20–64	Eurostat	Regional LFS / lfst_r_lfe2emprtn
Mayotte FRY5	36.9%	2025	Employment rate, 20–64; EU outermost region	Eurostat	Regional LFS / lfst_r_lfe2emprtn
Guyane FRY3	50.0%	2025	Employment rate, 20–64; EU outermost region	Eurostat	Regional LFS / lfst_r_lfe2emprtn
Calabria ITF6	50.3%	2025	Employment rate, 20–64	Eurostat	Regional LFS / lfst_r_lfe2emprtn
Campania ITF3	50.8%	2025	Employment rate, 20–64	Eurostat	Regional LFS / lfst_r_lfe2emprtn
Sicilia ITG1	51.4%	2025	Employment rate, 20–64	Eurostat	Regional LFS / lfst_r_lfe2emprtn
Italy national employment rate	67.6%	2025	Persons 20–64	Eurostat	Employment and labour force by sex and age — lfsi_emp_a
Italy national employment rate	62.5%	2025 annual average	Persons 15–64, therefore not directly comparable with the 20–64 rows above	ISTAT EU-LFS implementation	Labour market – Q4 2025
Italy regional dispersion	15.1% CV	2025	Population-weighted coefficient of variation of regional employment rates	Eurostat	Regional labour-market disparities / lfst_r_lmder
EU regional dispersion	12.5% → 9.2%	2015→2025	Population-weighted coefficient of variation	Eurostat	Regional labour-market disparities / lfst_r_lmder
EU employment rate	76.4%	Q2 2026	Seasonally adjusted, persons 20–64, EU aggregate rather than regional	Eurostat	Employment rate up in Q2 2026 — lfsi_emp_q
EU labour-market slack	11.0%	Q2 2026	Unmet employment need / extended labour force, persons 20–64	Eurostat	Employment rate up in Q2 2026 — lfsi_sla_q

Source-control finding

A material dataset distinction must be preserved throughout the dossier, because tgs00007 and the corresponding lfst_r_lfe2emprt presentation describe the 15–64 regional employment rate, whereas Eurostat’s regional 2024 and emerging 2025 map material for the European Pillar of Social Rights’ 20–64 employment target cites lfst_r_lfe2emprtn; consequently, the report will not relabel a 20–64 observation as though it originated from a 15–64 series merely to conform mechanically to the initial dataset list.

A second provenance issue is equally important for audit purposes, because Eurostat’s currently accessible live Statistics Explained page still displays the 2024 regional release, whereas an official Eurostat PDF artifact generated in 2026 contains the 2025 observations, including the 86.9% Warszawski stołeczny, 36.9% Mayotte, 15.1% Italian dispersion and 115-of-244 target-achievement results; the dossier will therefore identify the latter figures as current Eurostat 2025 regional material while preserving the publication-status distinction rather than presenting the live webpage and draft-generation artifact as though they were identical vintages.

---

The 2025 distribution at a glance

Competing explanations

A formal ACH table is not justified at this stage because the evidence does not support three mutually exclusive explanations for the observed pattern, since agglomeration, demographic composition, skills, industrial structure, institutional differences and informality can operate simultaneously rather than representing falsifiable alternatives; the dossier will therefore test them as complementary mechanisms and countervailing forces rather than manufacture false analytical exclusivity.

The principal working judgment is that the persistence of low-employment regional clusters is most consistent with a cumulative-production mechanism in which low productive density, weaker labour demand, outward migration of human capital, demographic structure, sectoral composition and imperfect institutional capacity interact, while cohesion investment, labour mobility, infrastructure improvements and national redistribution operate as countervailing forces; the empirical fact that EU-wide regional dispersion has fallen materially since 2015 prevents any claim that the countervailing forces have failed universally, but Italy’s 2025 dispersion and the continuing southern tail demonstrate equally clearly that they have not generated uniform convergence.

Measurement limits that materially affect interpretation

The EU-LFS definition of employment is conceptually distinct from both unemployment and labour-market slack, because the employment rate measures employed persons relative to the comparable population, the unemployment rate measures persons without work who are actively seeking and available for work relative to the labour force, and labour-market slack extends the analytical perimeter to additional persons with unmet employment needs; a region can therefore have a low employment rate without displaying an equivalently extreme unemployment rate if a large share of its working-age population is economically inactive.

Regional LFS data are annual averages produced from a sample survey whose respondents are allocated according to permanent residence, meaning that commuting can separate the region in which an employed person is counted from the region in which production actually takes place, while sampling and reliability issues become particularly important for small regions; Eurostat additionally notes that annual regional observations are derived from quarterly information rather than from a territorial census.

Undeclared activity requires a more careful qualification than either assuming that it disappears from the survey or assuming that the LFS captures it completely, because the Commission defines undeclared work as otherwise legal paid activity not reported to public authorities and recognises partially and fully undeclared forms across sectors including agriculture, construction and household services; because labour-force surveys obtain individual responses rather than relying exclusively on tax registrations, some informal activity can be reported as employment, but hidden work that respondents do not disclose or sampling frames do not adequately represent remains intrinsically difficult to measure, which means that recorded employment rates in high-informality economies should be read as survey estimates rather than exhaustive counts of all market production.

The evidence assembled so far is sufficient to establish this measurement limitation, but not sufficient to assign a defensible numerical “informality correction” to Calabria, Campania, Spanish southern regions or Greek regions; no such adjustment will therefore be fabricated in subsequent chapters, and any regional comparison will retain the official LFS figure while discussing informal employment separately.

EU-27 versus candidate, EFTA and neighbourhood Europe

The EU ranking in this dossier contains EU-27 NUTS-2 regions only, because Iceland, Norway and Switzerland belong to EFTA rather than the EU, while Bosnia and Herzegovina, North Macedonia, Serbia, Montenegro and Türkiye are outside the EU and use Eurostat statistical-region classifications rather than EU NUTS membership; the United Kingdom is likewise outside the EU and will appear only in the dedicated UK analytical lens where comparison adds decision value.

This distinction is methodologically important because inserting a low Turkish or candidate-country observation into an “EU employment floor” would silently change the geographic universe and exaggerate the spread being attributed to the Single Market; the EU-27 evidence already produces a sufficiently large and policy-relevant dispersion without doing so.

Europe’s Cyber Risk: Ransomware and Critical Systems

Italy, France, Germany and the United Kingdom: preliminary institutional lenses

Italy is the decisive intra-national case because its 20–64 national employment rate was 67.6% in 2025, the lowest among EU Member States, yet that national mean combines regions above or near the Union’s 78% benchmark with Calabria, Campania and Sicilia around 50–51%, while the independently reported 15–64 annual employment rate of 62.5% must remain explicitly separated from the 20–64 series rather than blended into the same comparison.

France cannot be characterised adequately through a mainland national average because its EU territorial universe includes outermost regions, and the 2025 regional record places Mayotte at 36.9% and Guyane at 50.0% for people aged 20–64, while Eurostat reports 11 French regions below 74%, including all five outermost regions considered in the 2025 regional assessment; French territorial divergence therefore combines peripheral metropolitan weaknesses with a qualitatively different overseas component that should not be collapsed into one causal category.

Germany presents almost the inverse configuration because 34 of its 38 NUTS-2 regions had reached or exceeded 78% in 2025, with Oberbayern at 85.1%, although Bremen, Berlin, Düsseldorf and Arnsberg remained below the target; the German case therefore matters not because territorial differences have disappeared but because high employment is much more widely distributed across regional systems than in Italy, providing an institutional and productive comparison for the later chapters.

The United Kingdom must not be included in the EU-27 2025 ranking after withdrawal from the Union, and any UK comparison must therefore be treated as an external European benchmark with compatible definitions established before numbers are compared; this methodological separation is particularly important because Eurostat’s regional-dispersion metadata historically include the UK in certain coefficient-of-variation calculations, which does not make it part of the EU-27 analytical universe.

Principal Gaps and Watch Indicators

The first material gap concerns publication vintage, because the currently live Eurostat Statistics Explained regional page remains on the 2024 regional picture while official 2026 Eurostat PDF material contains the 2025 map and tables; subsequent chapters will therefore anchor quantitative statements to the dataset vintage and exact official artifact rather than implying that all Eurostat presentation layers were updated synchronously.

The second gap concerns informality at NUTS-2 level, because the existence of undeclared work is officially established but a harmonised 2025 regional correction capable of transforming LFS employment rates into a comprehensive estimate of formal-plus-hidden employment has not been established in the verified record examined here; the appropriate analytical treatment is therefore a directional measurement qualification rather than an invented correction coefficient.

The principal watch indicator is the EU-wide population-weighted coefficient of variation, because continuation of the decline from 12.5% in 2015 to 9.2% in 2025 would constitute evidence of further sigma-convergence even if individual bottom-tail regions remained structurally weak, whereas renewed increases combined with persistent low-employment clusters would indicate that aggregate labour-market gains were again becoming spatially concentrated.

A second decisive indicator is whether southern Italian regional rates rise faster than the EU frontier rather than merely rise in absolute terms, because Calabria moving from approximately 50% to 55% while high-employment regions move from approximately 85% to 90% would represent improvement without convergence, whereas sustained closure of the percentage-point and proportional gaps would provide evidence of genuine catching-up.

A third indicator is the relationship between employment convergence and productivity, investment, skills and firm creation, because employment-rate equalisation achieved predominantly through low-productivity or precarious activities would have different implications from convergence associated with stronger productive capacity, while a transfer-induced increase in disposable income without corresponding labour-demand formation would not constitute the productive transformation tested by the core hypothesis.

Controlling assessment

The 2025 map does not validate the proposition that European labour markets have failed to converge in every statistical sense, because measurable EU-wide regional dispersion has declined substantially; it does validate the more consequential proposition that European integration has not produced automatic territorial equalisation, because nearly three decades of common-market deepening and twenty-five years of monetary union coexist with approximately 50 percentage points between the highest and lowest EU NUTS-2 20–64 employment rates, with a nearly 30-point gap inside Italy alone, and with geographically persistent bottom-tail clusters that national averages systematically attenuate.

The analytical burden of the full dossier will therefore be to explain why aggregate convergence and persistent regional divergence can coexist, rather than forcing the evidence into an absolute “convergence versus no convergence” dichotomy, because that distinction is necessary if the report is to establish what the data actually show rather than what either a pro-integration or anti-integration political narrative requires.

Principal official sources: Eurostat — Employment rates by NUTS 2 region, lfst_r_lfe2emprt · Eurostat — Employment rate, persons aged 15–64, tgs00007 · Eurostat — Labour market statistics at regional level · Eurostat — EU’s employment rate grew above 76% in 2025 · European Commission — EU regional convergence trends, Ninth Cohesion Report evidence

The map beneath the national averages

Principal judgment

The 2025 regional employment map demonstrates that the European Union possesses an increasingly integrated legal and economic space without possessing a correspondingly homogeneous labour-market geography, because the same Single Market that guarantees the movement of people, goods, services and capital contains NUTS-2 regions whose recorded employment rates for persons aged 20–64 range from 86.9% in Warszawski stołeczny to 36.9% in Mayotte, while several large continental regions in southern Italy remain close to 50%; the resulting 50.0 percentage-point EU-wide observed range, or roughly 35–37 percentage points even when the exceptional French outermost territories are removed from the continental comparison, is too large to be treated as statistical noise around a common European equilibrium and instead identifies territorially distinct systems of labour demand, participation, human capital, demographic structure and productive capacity operating inside common European institutional frameworks. The official 2025 Eurostat regional release records Warszawski stołeczny at 86.9%, Praha at 86.0%, Åland at 85.5%, Utrecht at 85.4%, Oberbayern at 85.1% and Mellersta Norrland at 85.0%, while identifying Mayotte at 36.9%, Guyane at 50.0%, Calabria at 50.3%, Campania at 50.8% and Sicilia at 51.4% at the opposite end of the distribution. Eurostat — 2025 regional employment-rate release

The evidence nevertheless does not support the stronger assertion that regional convergence has been wholly absent, because Eurostat's population-weighted measure of employment-rate dispersion declined from 12.5% in 2015 to 9.2% in 2025, while disparities narrowed in 15 of the 17 Member States for which the published comparison is available; what the data support is the more analytically important proposition that convergence has been incomplete, spatially uneven and non-automatic, because a reduction in the variance of the European distribution can coexist with persistent low-employment regional clusters, severe within-country differences and development paths that remain structurally separated after decades of market integration, monetary integration and cohesion expenditure. Eurostat — regional labour-market statistics and disparities

A common economic space does not imply a common labour market

The institutional experiment against which the 2025 employment map must be read is unusually deep by international standards, because the Single Market has existed since 1 January 1993 around the four freedoms of movement of goods, services, persons and capital, while the third stage of Economic and Monetary Union began on 1 January 1999, when the currencies of the initial participating states were irrevocably linked and responsibility for a common monetary policy passed to the Eurosystem; euro notes and coins subsequently entered circulation on 1 January 2002, meaning that Italy, France, Germany, Spain and several other economies appearing today at radically different points of the regional employment distribution have operated for more than two decades without bilateral exchange-rate adjustment among themselves. The institutional significance of that history is not that a common currency should mechanically equalise employment rates, which neither the treaties nor economic theory promise, but that several adjustment mechanisms available between sovereign national economies have been substantially altered: currency depreciation between euro-area members has disappeared, capital can move across borders with comparatively low institutional friction, workers possess extensive mobility rights, monetary policy responds to euro-area conditions rather than individual regional labour markets, and fiscal as well as structural adjustment remains predominantly national and regional. European Union — history of the Single Market and the four freedoms ECB — Economic and Monetary Union history

This matters for the interpretation of regional employment because economic integration removes important barriers to movement without removing the geography of production, and the distinction is visible directly in the 2025 data: all eight Czech NUTS-2 regions, four of Denmark's five, 34 of Germany's 38, all three Irish regions, six of Hungary's eight, all twelve Dutch regions, seven of Portugal's nine, three of Slovakia's four and all eight Swedish regions had reached or exceeded the EU's 78% employment-rate benchmark for people aged 20–64, while 11 Italian regions, 11 French regions, 11 of Greece's 13 regions, 11 Spanish regions and seven of Romania's eight regions were below 74%. The contrast therefore cannot be reduced to a simple north-versus-south national hierarchy, because Portugal performs strongly across most of its regions despite its southwestern periphery, Romania combines an 81.1% capital-region employment rate in București-Ilfov with rates of 61.7% in Sud-Vest Oltenia and 62.1% in Sud-Est, and Germany itself contains four regions below the 78% threshold despite very high national and regional employment elsewhere. Eurostat — Labour market statistics at regional level, 2025 data

The correct unit of analysis is consequently not simply “the country,” because firms hire inside accessible labour markets, workers commute within particular transport geographies, graduates choose among concrete metropolitan and regional opportunity structures, and households make participation decisions under locally specific combinations of wages, childcare, housing, transport and employment probability; national institutions matter enormously, but the labour market actually encountered by an individual or employer is generally narrower than the national territory and, in metropolitan economies, can cross regional administrative boundaries altogether. Eurostat's own regional methodology underscores this limitation because NUTS is an administrative-statistical classification rather than a map of self-contained functional labour markets, while metropolitan regions are separately defined through cities and commuting zones precisely because administrative territorial units and economic interaction areas do not always coincide. Eurostat — NUTS principles

NUTS-2 is the policy map, not a literal map of local labour markets

The choice of NUTS-2 as the central territorial scale is nevertheless institutionally unavoidable, because NUTS 2024 contains 244 NUTS-2 regions across the EU, and Eurostat explicitly defines this level as the set of “basic regions for the application of regional policies,” while NUTS-1 identifies major socio-economic regions and NUTS-3 smaller units for more specific diagnoses; NUTS-2 regions are also the level through which much of cohesion-policy eligibility and regional statistical comparison is organised, which makes them the correct unit for assessing whether the policy geography on which European regional intervention operates corresponds to convergence in employment outcomes. Under the NUTS principles, the normal population range for a NUTS-2 unit is 800,000 to 3 million inhabitants, although exceptions are permitted for geographical, socio-economic, historical, cultural or environmental reasons, so the system achieves considerable harmonisation without creating territorially identical economic units. Eurostat — NUTS overview and NUTS 2024 classification

That distinction places an important boundary around every inference made from the regional map, because an employment rate attached to Lombardia, Calabria, Oberbayern, Utrecht or Warszawski stołeczny describes the employment status of the resident population within a statistical territory rather than the total jobs physically located inside that territory, while commuting can connect adjoining regional labour markets and make the place of residence different from the place where production occurs. Eurostat's commuting statistics illustrate the scale of this issue, because the agency recorded more than 12.5 million intra-country commuters crossing regional boundaries in 2022, alongside 2.09 million employed people aged 15–64 commuting from their region of residence to another country, and defines the relevant commuter for that exercise as a person travelling at least once a week from the region of usual residence to a different region for work. Eurostat — intra-country and cross-border regional commuting

The implication is consequential for high-employment metropolitan regions and surrounding commuter belts, because a NUTS-2 rate should not automatically be interpreted as the employment-generating capacity of firms physically headquartered within the same boundary, yet the commuting problem does not invalidate the observed regional disparities because the employment rate remains a valid measure of the proportion of residents of the specified age group who are employed; rather, it changes the causal question from “how many jobs exist inside the boundary?” to “what proportion of the resident working-age population succeeds in participating in employment, whether locally or through accessible commuting networks?”, which is the more relevant question for territorial welfare and labour-market inclusion.

The 2025 distribution is not a marginal variation around the EU average

The EU aggregate employment rate for people aged 20–64 reached 76.1% in 2025, representing 197.7 million employed people out of 259.7 million residents in that age group, while the labour force comprised 209.9 million people, including 12.1 million unemployed, and another 49.8 million people were outside the labour force; these figures are essential because they show why the regional employment rate cannot be substituted with the unemployment rate, since a low-employment region can combine unemployment with a very large inactive population rather than displaying an unemployment rate numerically equivalent to the employment deficit. Eurostat also reports that the 76.1% EU rate represented the highest value in the series beginning in 2009 and left the Union 1.9 percentage points below the 78% European Pillar of Social Rights target for 2030, which means the EU aggregate itself was comparatively close to the policy objective even while substantial parts of the regional distribution remained far below it. Eurostat — EU employment rate in 2025

2025 EU labour-market quantity	Official value	Definition
Population aged 20–64	259.7 million	Core working-age population used in the 2030 employment target
Employed aged 20–64	197.7 million	EU-LFS employed persons
Unemployed aged 20–64	12.1 million	Not employed, actively seeking and available for work
Labour force aged 20–64	209.9 million	Employed plus unemployed
Outside the labour force, aged 20–64	49.8 million	Neither employed nor unemployed under the ILO/EU-LFS definition
Employment rate, aged 20–64	76.1%	Employed divided by population aged 20–64
EU 2030 employment target	78.0%	European Pillar of Social Rights Action Plan

Source and definition note: the values above are taken from Eurostat's 2025 regional labour-market release and national employment publication, while the 78% benchmark is the official European Pillar of Social Rights Action Plan target applying to the population aged 20–64 rather than the older 15–64 convention. European Commission — European Pillar of Social Rights Action Plan

Against that EU mean, the regional distribution reveals a scale of heterogeneity that an aggregate of 76.1% necessarily suppresses, because 115 of 244 NUTS-2 regions, equivalent to 47.1%, were already at or above 78% in 2025, whereas 64 regions, equivalent to 26.2%, were below 74%, so the Union was simultaneously close to its aggregate target and still contained more than one quarter of its NUTS-2 territories in a substantially weaker employment band. The corresponding 2024 release had reported 113 of 243 regions with available data, or 46.5%, at or above the target, which indicates continued aggregate improvement between the two releases but does not demonstrate that the lower tail was being eliminated, particularly because the composition and data availability underlying the denominator must be respected when comparing vintages. Eurostat — 2024 regional employment-rate release

The upper tail is geographically diverse, but it is not random

The six regions with employment rates of at least 85% in 2025 are analytically significant because they do not belong to one country or one institutional model, yet they share features associated with either dense high-productivity economic systems, strong labour-force participation, favourable demographic structures, or combinations of these mechanisms: Warszawski stołeczny, PL91, 86.9%; Praha, CZ01, 86.0%; Åland, FI20, 85.5%; Utrecht, NL31, 85.4%; Oberbayern, DE21, 85.1%; and Mellersta Norrland, SE32, 85.0%, while a further 23 regions recorded at least 83.5%, including nine in Germany and six in the Netherlands. The existence of Polish and Czech capital regions at the top of the ranking also contradicts any residual interpretation of the EU labour market as a simple old-West/high-employment versus new-East/low-employment divide, because some of the strongest labour-market outcomes are now located in the Member States that entered the Union during the 2004 enlargement cycle. Eurostat — Employment rate by NUTS-2 region, 2025

Capital-region effects are particularly important because Eurostat reports that every multi-regional northern, eastern and southern EU Member State had a capital-region employment rate above its national average in 2025, while Belgium, Germany and Austria were the notable western exceptions where the capital region lay below the national mean; the most striking positive capital premium cited by Eurostat is Romania, where București-Ilfov reached 81.1%, 12.1 percentage points above Romania's 69.0% national rate, whereas Brussels-Capital at 63.9% and Vienna at 71.3% were the lowest-employment regions of Belgium and Austria respectively. This mixed pattern matters because it demonstrates that agglomeration is powerful but not mechanically sufficient, since capital status can concentrate high-productivity employment and human capital while simultaneously coexisting with socioeconomic exclusion, high housing costs, commuter effects, demographic composition or institutional features that depress the resident employment rate.

The lower tail is geographically persistent and institutionally heterogeneous

At the bottom of the 2025 distribution, Eurostat identifies a combination of outermost territories, southern regions, rural or sparsely populated territories, peripheral regions and former industrial areas, which means that one causal story cannot adequately explain all low employment; Mayotte's 36.9% and Guyane's 50.0% are embedded in demographic, geographic and institutional conditions unlike those of Calabria, while southern Italian regions share neither the same territorial configuration nor the same production history as north-eastern French former industrial regions or Spanish autonomous cities. Eurostat nevertheless observes a common descriptive feature across many low-employment regions, namely restricted employment opportunities, including for intermediate- and high-skilled workers, while separately identifying former industrial heartlands extending from north-eastern France into Wallonia as another category affected by structural change in sectors such as coal, steel and textiles.

The lower-tail concentration is sufficiently large to be treated as a structural European pattern rather than a collection of isolated outliers, because in 2025 the sub-74% group included 11 French regions, all five French outermost regions among them; 11 Italian regions concentrated overwhelmingly in central and southern Italy; 11 of Greece's 13 regions; 11 Spanish regions; seven of Romania's eight regions; Brussels-Capital at 63.9%; Lazio at 69.3%; Vienna at 71.3%; Attiki at 73.1%; and Luxembourg at 73.9%. The geographical implication is therefore more complex than “southern Europe underperforms,” because the map contains a southern and peripheral concentration but also capital-region failures, deindustrialisation belts and sharp within-country differences that require separate mechanisms rather than a single cultural or macroeconomic explanation.

Selected 2025 NUTS-2 employment-rate extremes

NUTS-2 region	Code	Member State	Employment rate	Age group	Position in distribution
Warszawski stołeczny	PL91	Poland	86.9%	20–64	Highest EU rate
Praha	CZ01	Czechia	86.0%	20–64	Second highest
Åland	FI20	Finland	85.5%	20–64	Upper extreme
Utrecht	NL31	Netherlands	85.4%	20–64	Upper extreme
Oberbayern	DE21	Germany	85.1%	20–64	Upper extreme
Mellersta Norrland	SE32	Sweden	85.0%	20–64	Upper extreme
Mayotte	FRY5	France	36.9%	20–64	Lowest EU rate; outermost
Guyane	FRY3	France	50.0%	20–64	Second-lowest group; outermost
Calabria	ITF6	Italy	50.3%	20–64	Lowest continental Italian rate
Campania	ITF3	Italy	50.8%	20–64	Bottom EU tail
Sicilia	ITG1	Italy	51.4%	20–64	Bottom EU tail

The figures in the table are taken from Eurostat's 2025 NUTS-2 release and therefore use the 20–64 denominator required by the European Pillar employment target, while they must not be combined silently with the 15–64 rate published in tgs00007.

Italy is the strongest demonstration that the national average is analytically insufficient

Italy constitutes the most powerful empirical case for treating the regional cleavage as independently important because the country simultaneously records the lowest national 20–64 employment rate in the EU, at 67.6% in 2025, and the largest population-weighted regional coefficient of variation, at 15.1%, while its internal regional range extends from 80.0% in Provincia Autonoma di Bolzano/Bozen to 50.3% in Calabria, producing a 29.7 percentage-point internal gap inside one legal, fiscal, monetary and social-security state. Eurostat — national employment rates in 2025

That 67.6% national rate therefore does not identify the employment environment encountered either in Bolzano or in Calabria, and its policy usefulness depends entirely on the question being asked, because it is appropriate for an EU-level comparison of Member States and for assessing Italy's aggregate progress toward the 78% target, but it is inadequate as a representation of territorial labour-market conditions when deciding where labour shortages exist, where inactive labour reserves are concentrated, where childcare and participation constraints are binding, where skilled workers are leaving, or where firms face a scarcity rather than an excess of labour. A national mean remains statistically valid while becoming economically misleading when it is interpreted as though it described the central tendency of a spatial distribution with extreme internal dispersion, and Eurostat's coefficient-of-variation result quantifies precisely why Italy requires that qualification.

The Italian case becomes still more consequential when gender is introduced, because Eurostat's 2025 regional release records gender employment gaps of 28.7 percentage points in Puglia, 28.2 in Campania, 28.1 in Sicilia, 25.4 in Calabria and 25.3 in Basilicata, compared with an EU-wide male 20–64 employment rate of 80.9%, a female rate of 71.2%, and an EU gender employment gap of 9.7 percentage points; this indicates that the southern Italian employment deficit cannot be interpreted exclusively as insufficient male labour demand, because low female participation and employment form a major component of the territorial divergence.

The policy significance is substantial because a region whose headline employment rate is depressed by a 25–28 point gender gap presents a different adjustment problem from a region in which male and female employment are both moderately low, since childcare availability, care infrastructure, tax-and-benefit incentives, transport accessibility, labour-demand composition, occupational segregation and social norms become part of the mechanism rather than peripheral social variables; Eurostat explicitly identifies unpaid care, childcare availability, tax-benefit disincentives, occupational segregation and labour-market informality among the mechanisms associated with gender employment differences, while noting that the widest regional gaps are concentrated in Greece, southern Italy and parts of Romania.

Education reveals that the employment deficit is not distributed evenly across human capital groups

The 2025 regional evidence also demonstrates that educational attainment changes employment probabilities dramatically, because among people aged 25–64 the EU employment rate was 58.9% for those with no more than lower-secondary education, 78.5% for those with upper-secondary or post-secondary non-tertiary education, and 88.1% for tertiary-educated people, while 83 NUTS-2 regions recorded employment rates of at least 90% among tertiary-educated residents.

This pattern is analytically important because it shows that some low-employment regions do not suffer a uniform inability to employ all categories of labour, but instead display much larger deficits among low- and medium-skilled populations, while highly educated workers who remain in those territories can have much stronger employment outcomes; among low-education 25–64-year-olds, Eurostat records 53 regions below 50% employment, with Campania at 38.6% and Sicilia at 38.8%, whereas for the medium-education group Calabria stood at 56.0% and Campania at 59.2%, yet employment among tertiary-educated people exceeded 90% in large numbers of eastern, Portuguese and other EU regions.

The interaction between education and migration therefore becomes central to the later causal analysis, because a weak region can simultaneously display poor overall employment, reasonably strong employment probabilities for those with tertiary education, and continued outward migration of skilled young people if the number and quality of local high-skill opportunities remain insufficient relative to opportunities elsewhere; under those conditions, the regional labour market can lose precisely the workers whose human capital would otherwise support innovation, firm formation and productivity growth, thereby reinforcing the divergence between low-density and high-density productive systems. The European Commission's Ninth Cohesion Report reaches a closely related structural conclusion when it finds that regions caught in long-lasting development traps systematically display weaker human-capital indicators and argues that the ability to invest in, attract and retain tertiary-educated people is a recurrent feature distinguishing regions that avoid persistent stagnation. European Commission — Ninth Report on Economic, Social and Territorial Cohesion, Economic Cohesion chapter

The convergence question must distinguish sigma-convergence from structural equalisation

The statement that “convergence did not occur” becomes analytically defensible only after the meaning of convergence is specified, because the European Commission distinguishes sigma convergence, in which the dispersion of regional outcomes narrows; beta convergence, in which poorer or weaker regions grow faster than richer ones toward a common or conditional steady state; and club convergence, in which groups of regions with similar structural characteristics converge among themselves without necessarily converging toward a single European equilibrium.

On the narrow sigma-convergence measure relevant to employment rates, the evidence shows measurable progress, because the population-weighted coefficient of variation across EU regions declined from 12.5% in 2015 to 9.2% in 2025, while Finland, Portugal and Spain each recorded reductions of at least 40% in their internal coefficient and only Denmark and Romania experienced increases among the 17 Member States covered by the comparison; describing that record as “no convergence” would therefore be factually incorrect.

The stronger conclusion is that sigma-convergence has not produced structural equalisation, because the reduced dispersion coexists with Italy's 15.1% internal coefficient, a 29.7-point Bolzano–Calabria spread, large southern and outermost clusters below 74%, and an overall EU NUTS-2 distribution that still spans 50 percentage points; under a club-convergence interpretation, the possibility arises that high-employment metropolitan, industrial and highly connected regions increasingly resemble one another across national borders while low-employment peripheral or development-trapped regions remain clustered around a different trajectory, meaning that EU-wide dispersion can decline without the disappearance of structurally separate regional equilibria. The Ninth Cohesion Report expressly warns that geography, sectoral specialisation, productivity, investment and human capital affect convergence paths, while its analysis finds evidence of differentiated regional convergence patterns rather than one universal trajectory.

Cohesion policy has been large enough that “insufficient integration” cannot be the whole explanation

The persistence of territorial employment gaps has occurred alongside very substantial European regional-policy intervention, because the EU allocated approximately €347 billion in current prices to cohesion policy for 2007–2013, approximately €351.8 billion for 2014–2020, and €392 billion in current-price global cohesion-policy resources for 2021–2027, of which approximately €378 billion remained available for programming under the main Investment for Jobs and Growth, Just Transition and Interreg headings after specified transfers and Commission-managed amounts. European Commission — 2021–2027 cohesion policy initial allocations

These figures cannot legitimately be summed into a simple “amount spent to eliminate regional employment gaps,” because they cover different programming periods, objectives, prices, funds and implementation horizons, while cohesion policy pursues infrastructure, competitiveness, environmental, social, human-capital, territorial and transition objectives rather than a single employment-rate target; nevertheless, the order of magnitude establishes that persistent divergence has survived not an absence of regional policy but repeated, large-scale intervention intended to strengthen growth and cohesion. The Commission's own RHOMOLO assessment for the 2021–2027 programming period describes cohesion policy as operating through physical capital accumulation, human capital and technological progress, while also modelling interactions among regions and distinguishing short-run demand effects from longer-run supply-side effects, which is precisely why simple transfer arithmetic is analytically inadequate. European Commission — expected macroeconomic impact of 2021–2027 cohesion policy

The report's central mechanism therefore requires a narrower claim than “transfers do not work,” because transfers and public investment can raise disposable income, maintain services, improve transport, finance education, reduce financing constraints and raise productivity, while successful convergence in substantial parts of Central and Eastern Europe demonstrates that integration and cohesion policy can coexist with rapid catching-up; the relevant proposition is instead that income support or expenditure alone does not mechanically alter the regional production function, because durable convergence requires the accumulation and retention of productive capital, human capital, organisational capability, firm density, infrastructure and innovation capacity in combinations capable of generating self-sustaining private labour demand. This distinction is also embedded in the Commission's modelling framework, which treats cohesion spending as acting through long-run supply mechanisms rather than assuming that every euro of expenditure permanently raises regional productive capacity by construction.

National averages are administratively necessary but economically incomplete

National employment averages remain necessary for Council-level surveillance, country-specific recommendations, national target setting, fiscal planning and comparisons among Member States, yet their analytical limitation becomes severe whenever within-country dispersion approaches or exceeds typical differences between Member States; Italy's 67.6% national rate for people aged 20–64, for example, lies 12.4 points below Bolzano's 80.0% and 17.3 points above Calabria's 50.3%, so using “Italy” as the unit of analysis creates an artificial labour market that no worker or employer actually inhabits in that form.

The same principle applies elsewhere even when internal dispersion is smaller, because Spain's regional coefficient of variation was 6.1% in 2025, Belgium's 7.9%, Romania's 8.0%, while Finland, Sweden, the Netherlands and Portugal were at or below 2.0%, indicating that the informational loss generated by national aggregation varies materially by country; an EU comparison based only on national averages therefore treats the Netherlands, Italy and Romania as though the relationship between national mean and regional reality were statistically equivalent when Eurostat's dispersion measure demonstrates that they are not.

For government analysis, the appropriate practice is therefore hierarchical rather than substitutive, because national data should be retained for legal authority, fiscal incidence, social-security rules and country targets, while regional data should govern analysis of territorial labour availability, participation, development traps and productive geography; functional metropolitan or commuting-zone evidence should then be introduced wherever NUTS boundaries materially distort the economic area under study. This hierarchy preserves both institutional reality and spatial economic reality instead of treating either the nation or the region as universally sufficient.

The age-group problem is not a technical footnote

The report will maintain an explicit separation between 20–64 and 15–64 employment rates because they answer related but non-identical questions, and mixing them can create several percentage points of artificial difference that is then mistakenly interpreted as economic change; the European Pillar of Social Rights uses 20–64, whereas tgs00007 is the traditional regional 15–64 employment-rate table, and the younger denominator captures people aged 15–19 whose participation is heavily affected by secondary education and whose inclusion systematically lowers comparability with the 20–64 policy series.

A provenance complication in Eurostat's 2026 regional publication must also remain visible because the text and graphical elements exposed through the Statistics Explained PDF are not fully uniform in their displayed online-data-code labelling: the 2025 20–64 map itself identifies lfst_r_lfe2emprtn, while another presentation element associated with the same chapter is surfaced as lfst_r_lfe2emprt; rather than silently harmonising those labels, this dossier will cite the exact official 2025 publication and state the age group beside every material rate, while tgs00007 will be used only when a 15–64 observation is explicitly required.

This distinction is particularly important for Italy, because a 62.5% 15–64 employment rate and a 67.6% 20–64 employment rate can both be correct for 2025 without describing conflicting realities, while presenting them without the denominator would create a false five-point contradiction; every later table will consequently carry year, age range, territorial unit and source series as part of the observation rather than relegating those definitions to a footnote.

Informality qualifies the southern gap but does not erase it

The EU-LFS is a household sample survey rather than an administrative count of registered employment, which means that the presence of undeclared work does not imply that every informal worker is automatically absent from the employment numerator, because respondents can report work that is not fully reflected in tax, social-security or business registers; at the same time, hidden activity that is not disclosed to interviewers, populations imperfectly covered by the sampling frame and irregular work arrangements can generate undermeasurement, while regional estimates also carry sampling uncertainty that becomes more consequential in smaller territories. Eurostat describes the regional labour-market database as deriving from the EU-LFS household sample, publishes regional annual values as averages of quarterly observations and applies reliability flags where estimates do not satisfy the relevant precision thresholds. Eurostat — EU Labour Force Survey methodology

The appropriate analytical treatment is therefore neither to accept recorded southern employment rates as a census-level measure of every remunerated activity nor to “correct” them upward using an unsupported informal-economy multiplier, because no harmonised 2025 NUTS-2 adjustment exists in the verified Eurostat record that would permit Calabria's 50.3%, Campania's 50.8% or Sicilia's 51.4% to be transformed into alternative comparable employment rates. The existence of informality should consequently reduce confidence in literal point estimates at the margin while leaving the direction and scale of the territorial gap intact unless an official harmonised adjustment demonstrates otherwise, particularly because Eurostat's own regional analysis associates the widest southern gender employment gaps with weaker and more informal labour markets rather than treating informality as evidence that recorded differences are meaningless.

Why the map matters more than the aggregate Q2 2026 improvement

The latest aggregate observation does not overturn the structural regional finding because the EU employment rate for persons aged 20–64 reached 76.4% in Q2 2026, but that quarterly number answers a different question from the annual NUTS-2 map: it establishes that employment continued to increase at Union level after the 76.1% annual 2025 result, while providing no evidence that Calabria, Campania, Sicilia, Mayotte or other bottom-tail regions were closing their distance from Warszawski stołeczny, Praha, Utrecht or Oberbayern. An increase in the European mean can result from improvements concentrated in already strong regions, improvements concentrated in lagging regions, or simultaneous movements in both groups, and only regional data can distinguish those trajectories. Eurostat — Q2 2026 employment and labour-market slack release

This distinction between level improvement and convergence will govern the remainder of the report, because an increase from 50% to 55% in a low-employment region is economically significant even if the frontier rises from 85% to 90%, but it does not reduce the absolute territorial gap; conversely, a region can converge toward the EU mean during a period in which its employment rate remains low in absolute terms, so policy evaluation must examine both the direction of local change and the change in the distance from the relevant benchmark.

Decision-useful baseline

The evidence establishes five facts that later chapters must treat as constraints rather than hypotheses, beginning with the finding that the EU aggregate labour market improved materially through 2025 and into 2026, so the report is not analysing a Union-wide employment collapse; continuing with the finding that regional dispersion nevertheless remains economically large, including a 50-point EU-wide NUTS-2 observed range and nearly 30 points within Italy; recognising that some sigma-convergence has occurred, because the EU coefficient of variation fell from 12.5% to 9.2% between 2015 and 2025; acknowledging that national averages conceal very different degrees of internal dispersion, with Italy at 15.1% while several northern and western Member States record coefficients around or below 2%; and preserving the distinction that low employment has multiple territorial mechanisms, including southern peripherality, outermost-region geography, former industrial decline, capital-region exclusion, gender participation differences and human-capital structure, which prevents any single explanation from being applied uniformly across the map.

The principal analytical inference at this stage is therefore that the European labour market should be understood as an integrated system of connected regional labour markets rather than a single spatially equilibrated market, because common rules, capital mobility, labour mobility and, for 21 EU Member States as of 2026, a common currency coexist with regional employment outcomes whose persistence demonstrates that mobility and market access do not automatically equalise productive structures, labour-force participation or local demand for labour.

Key judgments

The strongest finding is that the national boundary is no longer sufficient as the primary descriptive boundary for European employment analysis, because several Member States contain internal regional gaps comparable with or larger than cross-country differences, while Italy's 29.7-point Bolzano–Calabria divide provides the clearest case of a single national institutional system containing fundamentally different employment equilibria.

The second judgment is that the evidence supports partial convergence rather than either full convergence or complete failure, because declining EU-wide dispersion demonstrates measurable sigma-convergence while the continuing bottom tail, Italy's extreme internal coefficient and persistent southern-peripheral clusters demonstrate that integration has not generated spontaneous territorial equalisation.

The third judgment is that cohesion policy must be evaluated through its capacity to modify productive conditions rather than through transfer volumes alone, because the EU has deployed hundreds of billions of euros across successive programming periods while its own modelling framework distinguishes temporary demand support from longer-lived supply-side effects operating through physical capital, human capital and technological progress.

The fourth judgment is that southern low-employment rates cannot be interpreted correctly without incorporating female employment, skills, demographic and informal-economy considerations, yet none of these qualifications justifies discarding the official employment-rate gap, because the verified EU-LFS record continues to show exceptionally low participation in employment even after the methodological limitations are acknowledged.

What would change the assessment

The assessment that convergence remains incomplete would weaken materially if successive regional releases showed a sustained decline not only in the EU coefficient of variation but also in the distance between the persistent bottom-tail regions and the European frontier, particularly if Calabria, Campania, Sicilia, southern Spanish regions, Greek regions and French outermost territories recorded several years of employment growth systematically faster than already high-employment regions while their human-capital retention, female participation and productive investment indicators improved in parallel.

The assessment would strengthen if the EU aggregate employment rate approached or exceeded the 78% 2030 target while the number of sub-74% regions remained broadly unchanged, because such a configuration would demonstrate that improvement in the European mean was being generated disproportionately outside the persistent low-employment cluster rather than through broad territorial convergence.

The endogenous-growth interpretation would weaken if official regional evidence demonstrated durable employment convergence in lagging territories without corresponding improvements in productive investment, human capital, firm density, accessibility or other supply-side variables, while it would strengthen if outmigration of skilled workers, weaker investment and lower innovation capacity continued to co-move with persistent low employment in the same territorial clusters documented by the Commission's development-trap analysis.

Open official record

The first unresolved issue concerns the final publication status and labelling of Eurostat's 2025 regional Statistics Explained material, because the 2026 official PDF contains the full 2025 employment map and associated findings while some presentation layers and dataset-code labels remain inconsistent; this does not invalidate the reported values, but a final stable Statistics Explained release and fully synchronised Data Browser vintage would improve auditability for a certified derivative report.

The second unresolved issue concerns the magnitude of informal-employment undercoverage at NUTS-2 level, because the existence and relevance of undeclared work are established while a harmonised official regional correction capable of translating EU-LFS employment rates into formal-plus-unreported employment estimates has not been identified; without such an instrument, the report will preserve the official survey values and treat informality as an uncertainty qualifier rather than as a numerical adjustment.

The third unresolved issue concerns functional labour-market geography, because NUTS-2 remains the correct policy and statistical unit for this dossier but does not perfectly map commuting zones, meaning that a later robustness test should distinguish regional resident employment from metropolitan and cross-regional accessibility wherever commuting materially alters interpretation, particularly around capital regions and dense cross-border labour markets.

Partial convergence without equalisation

Principal judgment

The European regional labour market has converged in a measurable statistical sense without converging toward a common territorial employment structure, because Eurostat's population-weighted coefficient of variation for regional employment rates fell from 12.5% in 2015 to 9.2% in 2025, while the same 2025 distribution still contained Italy at 15.1% regional dispersion, Romania at 8.0%, Belgium at 7.9% and Spain at 6.1%, alongside much tighter regional systems in Finland, Sweden, the Netherlands and Portugal, all at or below 2.0%; the result is therefore not “failure of convergence” in the strict statistical sense, but partial sigma-convergence combined with persistent structural non-equalisation, which means that the distribution has narrowed while remaining composed of regional labour markets with materially different employment intensities, participation structures, productive capacities and demographic constraints. Eurostat — Labour market statistics at regional level, 2025

That distinction is central because a declining coefficient of variation does not establish that lagging regions have become economically similar to the high-employment regions toward which the distribution is nominally moving, and it certainly does not establish that the underlying production systems have converged; a region can reduce its distance from the EU average while retaining a much lower employment rate, lower productivity, weaker private investment, a smaller employer base, higher inactivity and stronger outward migration of skilled workers, just as the EU as a whole can experience falling regional dispersion while Italy continues to display an internal 29.7 percentage-point difference between Provincia Autonoma di Bolzano/Bozen at 80.0% and Calabria at 50.3%. Eurostat — regional employment disparities, 2025

The most defensible conclusion is consequently that Europe has achieved convergence in some dimensions, regions and periods, while failing to eliminate deeply embedded territorial differences, which is consistent with the European Commission's broader cohesion assessment: Central and Eastern Europe has experienced substantial long-run income convergence, with GDP per head rising from roughly 45% of the EU average in 1995 to nearly 80% by 2021, yet regional disparities stopped narrowing after the global financial crisis and some less-developed and transition regions, particularly in southern Europe, began to diverge from the EU average. European Commission — Ninth Report on Economic, Social and Territorial Cohesion, Economic Cohesion

Sigma-convergence is real, but it is a distributional statistic rather than a structural verdict

The clearest evidence of partial convergence is Eurostat's population-weighted coefficient of variation, which measures how dispersed regional employment rates are around the relevant mean after allowing larger-population regions to carry greater weight; on that measure, the EU-wide coefficient fell from 12.5% in 2015 to 9.2% in 2025, and disparities narrowed in 15 of the 17 Member States for which Eurostat reports a comparable regional series, with the largest proportional declines occurring in Finland, Portugal and Spain, each of which reduced its regional employment-rate coefficient by at least 40%, while Denmark and Romania were the only countries in the comparison whose regional disparities increased, by 22.2% and 2.6% respectively. Eurostat — Labour market statistics at regional level, Figure 1, dataset lfst_r_lmder

This is substantive evidence rather than statistical decoration, because a reduction from 12.5% to 9.2% means that regional employment outcomes became less dispersed relative to the EU mean over the decade considered; however, the coefficient of variation contains no information about which regions converged, through what mechanism, whether convergence occurred through improvement at the bottom or compression at the top, whether employment quality changed, or whether productive capacity and labour demand became structurally more similar, and for those reasons it cannot serve as a proxy for economic equalisation. A coefficient can decline while the worst-performing regions remain deeply separated from the frontier, just as a distribution can become statistically narrower because middle-ranking regions move toward the mean while an entrenched lower tail persists.

The limitation is especially visible in the 2025 map, because 47.1% of EU NUTS-2 regions, equivalent to 115 of 244 regions, had already reached or surpassed the 78% employment target for people aged 20–64, while 64 regions remained below 74%, so the Union simultaneously contained a large group already above the 2030 headline threshold and a substantial lower tail that remained at least four percentage points beneath it. Eurostat — Employment rate by NUTS-2 region, 2025

The relevant analytical inference is therefore that sigma-convergence captures movement in the shape of the distribution, not convergence in the structure generating the distribution, and those two questions must remain separate throughout the report because the latter requires evidence on productivity, industrial composition, firm density, skills, demographic participation, investment and accessibility rather than one statistical dispersion measure.

Convergence in levels does not imply convergence in mechanisms

A region with a 72% employment rate and another with an 82% rate can move numerically closer without becoming structurally alike, because the first might converge through increased low-productivity service employment, temporary work, commuting or higher female participation while the second remains anchored in dense industrial, technological or business-service ecosystems; equally, two regions can record similar headline employment rates while differing sharply in productivity, wages, labour-force composition, hours worked, sectoral concentration and dependence on public employment. The employment rate therefore measures an outcome that is policy-relevant and comparable, but it does not by itself identify the production function producing that outcome.

Eurostat's 2025 evidence reinforces this point through educational attainment, because among people aged 25–64, the EU employment rate was 58.9% for those with low educational attainment, 78.5% for those with medium attainment and 88.1% for those with tertiary education, while 83 NUTS-2 regions recorded tertiary-educated employment rates of at least 90%, including all Romanian regions and almost every Polish region, even though several of those countries still contained low overall regional employment rates. Eurostat — regional employment rates by educational attainment, 2025

That coexistence matters because a region can simultaneously possess a very high probability of employment for tertiary-educated residents and a weak aggregate employment rate if the highly educated are a relatively small share of the population, if employment among low- and medium-skilled workers is weak, if female participation is low, or if a substantial share of working-age residents remains outside the labour force; southern Italy offers precisely this type of structural imbalance, since Campania recorded only 38.6% employment among 25–64-year-olds with low education and Calabria only 56.0% among those with medium education, while tertiary-educated employment rates across Europe were generally much higher. Eurostat — Labour market statistics at regional level, education section

A policy strategy aimed solely at increasing the aggregate employment rate can therefore improve the headline statistic without necessarily eliminating the underlying structural asymmetry, because structural equalisation would require convergence in the distribution of skills, access to productive employment, sectoral composition, participation, private capital formation and the capacity to retain human capital rather than only convergence in the final ratio of employed residents to working-age population.

Italy is the clearest case of convergence without equalisation

Italy's 2025 position is analytically decisive because Eurostat records the highest regional employment-rate disparity in the EU at 15.1%, almost twice Romania's 8.0% and substantially above Spain's 6.1%, while the country's regional pattern remains organised around a pronounced north-south divide, with Provincia Autonoma di Bolzano/Bozen at 80.0% and Calabria at 50.3% for people aged 20–64. Eurostat — Labour market statistics at regional level, 2025

This is not merely an extreme-case comparison between two statistical endpoints, because 11 Italian regions were below 74% in 2025, and the lowest rates were concentrated in the south, with Calabria at 50.3%, Campania at 50.8% and Sicilia at 51.4%, whereas northern and Alpine regions were substantially closer to or above the EU benchmark; the Italian national employment rate of 67.6% for people aged 20–64 therefore sits between regional systems that differ by almost 30 percentage points and should be interpreted as a weighted national accounting statistic rather than as a representative labour-market equilibrium. Eurostat — EU employment rate grew above 76% in 2025

The persistence of this internal divergence is especially consequential because Italy operates under one national labour-law framework, one currency, one central fiscal authority, one social-security system and decades of national and European territorial redistribution, which means that institutional uniformity at the state level has not been sufficient to produce similar employment outcomes across space; the most plausible interpretation is not that institutions are irrelevant, but that common institutions interact with regionally heterogeneous productive structures, leaving local labour demand, participation, employer density, infrastructure, demographic structure and human-capital retention as independent determinants of employment performance.

The Italian case also demonstrates why female participation must be included in any serious convergence assessment, because Eurostat records 2025 gender employment gaps of 28.7 percentage points in Puglia, 28.2 in Campania, 28.1 in Sicilia, 25.4 in Calabria and 25.3 in Basilicata, compared with an EU-wide gap of 9.7 points, which indicates that the lower southern employment equilibrium is partly produced by a much lower rate of female participation and employment rather than only by male unemployment or sectoral job scarcity. Eurostat — regional gender employment gaps, 2025

The implication is that Italian convergence cannot be evaluated only by asking whether southern GDP or aggregate employment rises, because a genuine structural narrowing would require simultaneous progress in female employment, low- and medium-skill employment, private labour demand, human-capital retention and sectoral productivity; without those complementary changes, an improving headline rate could still coexist with a persistent development gap.

Spain demonstrates genuine narrowing without complete territorial equalisation

Spain provides a useful counter-case to Italy because Eurostat reports a 2025 regional employment-rate coefficient of variation of 6.1%, while identifying Spain among the Member States where internal disparities fell by at least 40% between 2015 and 2025, which constitutes credible evidence that territorial convergence occurred over the decade rather than merely being asserted in policy language. Eurostat — regional employment-rate disparities, 2015–2025

Yet the same 2025 regional release identifies 11 Spanish regions below 74%, concentrated among peripheral, southern, island or predominantly rural territories, while northern, eastern and capital regions generally recorded higher employment rates; Spain therefore illustrates precisely why convergence and equalisation must be separated, because a substantial fall in dispersion can coexist with a persistent territorial hierarchy in which southern and peripheral regions remain systematically weaker than metropolitan and northern ones. Eurostat — Labour market statistics at regional level, Spain regional pattern

The Spanish result also shows why the report should not adopt a deterministic thesis that cohesion policy or integration “failed,” because the fall in regional dispersion is evidence of meaningful progress; the more defensible inference is that convergence mechanisms can operate successfully without fully overcoming inherited geographic, demographic and productive differences, leaving a residual lower tail even after the overall national distribution becomes materially less unequal.

This distinction is important for policy evaluation because if Spain's internal dispersion falls while southern regions remain below the EU benchmark, the next policy problem is no longer simply how to reduce national inequality, but how to identify the remaining regional mechanisms that prevent lagging territories from completing the transition toward higher employment equilibria.

Germany and the Netherlands demonstrate that high aggregate employment can coexist with different degrees of regional concentration

Germany's 2025 profile differs sharply from Italy's because 34 of its 38 NUTS-2 regions reached or exceeded the 78% EU employment target, while Oberbayern stood at 85.1%, placing it among the six highest-employment EU regions; this broad territorial distribution suggests that Germany's high employment performance is not confined to one capital or metropolitan enclave but is spread across a large number of regional economies. Eurostat — 2025 NUTS-2 employment rates

The Netherlands displays an even more compressed high-employment configuration because all 12 Dutch NUTS-2 regions were above the 78% benchmark in 2025, while Eurostat places the country's regional coefficient of variation at 2.0% or less, and Utrecht ranked among the EU leaders at 85.4%; the Netherlands therefore represents a case in which high national employment and low regional dispersion reinforce rather than offset one another, which is much closer to territorial equalisation than the Italian configuration. Eurostat — regional employment rates and dispersion, 2025

Germany nevertheless reminds us that high average performance does not imply complete symmetry, because Bremen, Berlin, Düsseldorf and Arnsberg remained below the 78% benchmark, while the capital region itself had a lower employment rate than the national average; this matters because it demonstrates that metropolitan concentration does not always generate the strongest resident employment outcome, particularly where commuting, population composition, housing, inactivity or deindustrialisation complicate the relationship between economic density and the employment rate of residents.

The comparative lesson from Germany and the Netherlands is therefore not that one national model automatically guarantees convergence, but that broad geographic diffusion of productive activity and high participation can produce both high mean employment and low dispersion, whereas the Italian pattern combines a low mean with unusually wide regional variance; that difference becomes more informative for policy than a simple national league table.

Romania shows that catching-up can coexist with pronounced capital-region divergence

Romania offers a different structural configuration because its regional employment-rate coefficient stood at 8.0% in 2025, the second-highest level after Italy among the values highlighted by Eurostat, while the capital region of București-Ilfov recorded 81.1% employment, 12.1 percentage points above Romania's national 69.0% rate, compared with 61.7% in Sud-Vest Oltenia and 62.1% in Sud-Est. Eurostat — Labour market statistics at regional level, Romania

This is a particularly important case because the Commission's broader cohesion evidence simultaneously identifies Central and Eastern Europe as the Union's most successful long-term convergence story, with GDP per head rising from around 45% of the EU average in 1995 to nearly 80% in 2021, meaning that cross-country convergence toward the EU average can occur at the same time as widening or persistent within-country regional divergence. European Commission — Ninth Cohesion Report, Economic Cohesion

The Romanian pattern therefore warns against treating convergence as a single scalar process, because national catching-up can be driven disproportionately by capital and metropolitan regions that attract investment, skilled labour and internationally integrated activities, while secondary and peripheral regions move more slowly; in such a configuration the Member State converges toward Europe while its internal spatial economy becomes more polarised.

That mechanism is highly relevant to the wider European convergence debate because it suggests that integration can reward already-connected regional nodes strongly enough to raise national averages while simultaneously increasing the distance between those nodes and their domestic periphery, which is compatible with agglomeration economics and with the Commission's finding that capital regions often exert strong pull effects on highly qualified workers. Eurostat — regional employment by educational attainment and capital-region pull effects

Portugal complicates any simple core-periphery narrative

Portugal is especially valuable analytically because Eurostat places it among the Member States with a regional employment-rate coefficient of variation of 2.0% or less in 2025, while also identifying Portugal among the countries where internal regional disparities fell by at least 40% over 2015–2025, and seven of its nine NUTS-2 regions had already reached or surpassed the 78% employment-rate target. Eurostat — Labour market statistics at regional level, 2025

This result matters because Portugal is peripheral in geographic terms and historically lagged richer northern European economies, yet its current regional employment distribution is both relatively compressed and comparatively strong, which directly contradicts any explanation that attributes low regional employment mechanically to southern latitude, peripheral location or membership in the euro area; geography remains relevant, but geography operates through institutions, demographics, connectivity, sectoral structure, education and investment rather than determining outcomes independently.

Portugal's employment profile also shows why the report's core mechanism must remain probabilistic rather than deterministic, because agglomeration pressures can generate divergence, but public investment, labour-market reform, outward integration, sectoral change and human-capital improvement can offset those pressures sufficiently to produce a more balanced regional distribution; the policy question is therefore not whether agglomeration exists, but under what conditions its cumulative effects can be counteracted.

France contains both metropolitan convergence and outermost-region structural discontinuity

France requires a particularly careful interpretation because its NUTS-2 distribution includes outermost regions whose geography, demography, accessibility and economic structure are not comparable to metropolitan France, with Mayotte at 36.9% and Guyane at 50.0% in 2025 while all five French outermost regions were included among the 11 French regions below 74%. Eurostat — regional employment rates, France and outermost regions

Treating these territories as simple observations on the same continuous distribution as metropolitan regions would underestimate the institutional and geographic discontinuity involved, because outermost territories combine remoteness, small or atypical labour markets, different demographic pressures and higher transport as well as connectivity costs; however, excluding them from the EU distribution altogether would also be analytically inappropriate because they are constitutionally part of the Member State concerned and form part of the Union's territorial cohesion obligations.

The correct interpretation is therefore two-layered: France exhibits a genuine internal regional distribution in metropolitan Europe, while its outermost regions form an additional extreme-periphery category that materially increases the national and EU territorial range; this is one reason why the raw 50-point EU spread should be supplemented by a continental comparison rather than used as the sole indicator of European labour-market divergence.

The existence of former industrial low-employment regions in north-eastern France adds a further structural mechanism, because Eurostat identifies a belt extending into Wallonia where traditional sectors such as coal, steel and textiles were adversely affected by economic restructuring and globalisation; French regional divergence is therefore produced by at least two distinct territorial processes, one linked to remoteness and outermost status and another linked to industrial transition. Eurostat — low-employment former industrial regions

Greece remains a bottom-tail concentration despite aggregate European improvement

Greece remains one of the clearest examples of persistent territorial underperformance because 11 of its 13 NUTS-2 regions recorded employment rates below 74% in 2025, with only Kriti and Peloponnisos outside that group, while the national 20–64 employment rate stood at 71.0%, the third-lowest among EU Member States after Italy and Romania. Eurostat — 2025 regional employment rates Eurostat — EU employment rate grew above 76% in 2025

The Greek pattern is significant because it differs from Romania's capital-led divergence and Italy's extreme north-south split: the weakness is geographically broad, suggesting that the principal problem is not merely one underperforming periphery within an otherwise high-employment national system but a national labour market with extensive regional underperformance; this distinction matters for policy because a territorially concentrated problem can be attacked through spatially targeted intervention, whereas a broadly distributed weakness requires national reforms in addition to place-based instruments.

The gender dimension again deepens the diagnosis, because Eurostat reports that 10 of Greece's 13 regions had gender employment gaps of at least 18 percentage points, with Sterea Elláda reaching 29.0 points, the widest gap in the EU in 2025; the structural convergence problem is therefore inseparable from labour-force participation and the economic integration of women rather than being reducible to conventional unemployment alone. Eurostat — regional gender employment gap, 2025

The 2009 break matters because convergence has not been linear

The European Commission's Ninth Cohesion Report provides the historical context necessary to avoid reading the 2015–2025 decline in employment-rate dispersion as evidence of one uninterrupted convergence process, because its long-run regional analysis finds that EU regional disparities narrowed before the global financial crisis but subsequently stagnated, largely because less-developed regions in Central and Eastern Europe slowed and some less-developed and transition regions, particularly in southern Europe, began to diverge. European Commission — Ninth Report on Economic, Social and Territorial Cohesion

The Commission further reports that by 2021 around one third of EU regions had still not returned to their 2008 level of GDP per head, with many of these regions located in Italy, Spain, Greece and France, and associates the post-2009 slowdown in convergence with comparatively weak productivity, investment and employment performance in regions that had previously been catching up. European Commission — Economic Cohesion chapter

This historical break is crucial because it indicates that convergence is path-dependent and shock-sensitive, rather than an automatic consequence of time spent inside the Single Market or monetary union; regions can converge during one phase of integration, stall after a financial or structural shock, and then resume or fail to resume convergence depending on their productive resilience, fiscal capacity, investment response and demographic dynamics.

The post-pandemic employment rebound should consequently be interpreted as evidence of resilience rather than proof that pre-existing structural divides have disappeared, because the EU employment rate rose to 76.1% in 2025, its highest level in the series beginning in 2009, even while Italy, Greece, southern Spain, parts of Romania and French outermost territories remained concentrated in the lower part of the regional distribution. Eurostat — EU employment rate in 2025

Eastern European convergence is genuine and analytically indispensable

Any argument that European integration failed to produce convergence would be contradicted by the Commission's official long-run evidence for Central and Eastern Europe, where GDP per head increased from approximately 45% of the EU average in 1995 to 52% at the time of the 2004 enlargement and nearly 80% by 2021, a transformation the Commission explicitly associates with European integration and cohesion policy. European Commission — Ninth Cohesion Report, Economic Cohesion

The employment map is consistent with that structural catching-up, because all eight Czech regions reached or exceeded 78% employment in 2025, while Warszawski stołeczny at 86.9% and Praha at 86.0% occupied the first two positions in the EU NUTS-2 employment ranking, and six of Hungary's eight regions as well as three of Slovakia's four were also above the 78% threshold. Eurostat — 2025 employment rates by NUTS-2

That evidence makes the geographic contrast with southern Europe more rather than less important, because integration has demonstrated that substantial catching-up is possible under the common European institutional framework; the persistence of low-employment clusters in southern Italy, Greece and parts of Spain must therefore be explained through region-specific growth constraints, demographic patterns, investment histories, institutional capacity and productive structure rather than by assuming that the European framework mechanically suppresses convergence everywhere.

This distinction also strengthens the endogenous-growth interpretation, because when the same broad European rules coexist with rapid catching-up in Prague, Warsaw and other Central European centres but stagnation in parts of the Mezzogiorno, the explanatory burden shifts toward the interaction between integration and local productive capacity rather than toward integration alone.

Social convergence and economic convergence are related but not identical

The Commission's social-cohesion assessment states that regional disparities in employment have narrowed and that increased labour-market participation by under-represented groups has contributed both to convergence and to the easing of labour shortages, while simultaneously noting that disparities persist and that graduates remain strongly concentrated in large cities. European Commission — Ninth Report on Economic, Social and Territorial Cohesion, Social Cohesion

This distinction matters because employment convergence can occur through rising participation even when productivity convergence is weaker, and productivity convergence can occur without an equivalent increase in the employment rate if demographic decline, migration or low participation offset labour demand; the two processes should therefore be analysed jointly but not treated as interchangeable.

The concentration of graduates in large cities is particularly relevant to the persistence of regional disparities because highly educated workers generally have much higher employment rates, while capital and metropolitan regions can attract them through stronger labour demand, universities, professional networks and social amenities; Eurostat's 2025 evidence explicitly observes that capital regions frequently have higher employment rates among tertiary-educated residents than their national averages and attributes this pattern in part to their ability to attract highly qualified people. Eurostat — regional employment by educational attainment, 2025

The resulting feedback loop is central to the later agglomeration analysis because skilled workers move toward regions where high-skill jobs are already concentrated, firms move toward regions where skilled workers and supplier networks are available, and the combined process can improve national productivity while widening the productive distance between leading and lagging regions.

Employment convergence should be decomposed into extensive and structural margins

For decision-grade analysis, changes in regional employment rates should be decomposed conceptually into at least four margins, because a region can raise its employment rate through higher labour-force participation, lower unemployment, stronger net job creation, or demographic and migration changes that alter the composition of the denominator; without that decomposition, identical headline improvements can reflect fundamentally different processes.

The first margin is participation, which is particularly important in southern Italy and Greece because large gender employment gaps indicate substantial unused labour potential; the second margin is unemployment absorption, which captures whether people already seeking work are being employed; the third is job-generation capacity, which depends on the expansion, creation and survival of firms and public-sector employment; and the fourth is demographic selection, under which outward migration of unemployed or low-employment groups can mechanically improve the employment rate without equivalent local job creation, while inward migration of working-age residents can have the opposite short-run effect.

Eurostat's labour-market framework makes these distinctions analytically necessary because in 2025 the EU contained 197.7 million employed people aged 20–64, 12.1 million unemployed and 49.8 million people outside the labour force, meaning that the employment rate is determined by the joint behaviour of employment, unemployment, inactivity and population rather than by job creation alone. Eurostat — Labour market statistics at regional level, 2025

A serious convergence assessment should therefore reject any claim that a narrowing employment-rate gap automatically proves convergence in productive capacity, because the same numerical result can emerge from structurally different combinations of participation, migration, inactivity and labour demand.

A lower coefficient of variation does not eliminate the bottom tail

The most important mathematical property of the coefficient of variation for this report is that it summarises dispersion without identifying the distribution's tails, meaning that the EU coefficient can fall from 12.5% to 9.2% even if some very low-employment regions remain almost unchanged, provided enough other regions move closer to the mean; that is why the 2025 tail must be analysed independently of the aggregate coefficient.

Eurostat reports that 64 of 244 NUTS-2 regions remained below 74% in 2025, including 11 regions in France, 11 in Italy, 11 of 13 in Greece, 11 in Spain and seven of eight in Romania, which represents a geographically coherent and economically meaningful lower segment rather than isolated statistical exceptions. Eurostat — Labour market statistics at regional level, 2025

The persistence of this tail is the clearest reason that “convergence without equalisation” is the correct formulation, because the centre of the distribution has moved and overall dispersion has narrowed, yet a large block of regions remains structurally separated from the high-employment frontier; in policy terms, the relevant question is therefore no longer whether convergence exists in the abstract, but why some regions participate in it much more effectively than others.

The distance from the frontier remains economically large

A regional convergence process should be tested not only against the EU average but also against the employment frontier, because a region can approach the mean while remaining far from the best-performing areas, and the frontier represents the observable employment intensity already achieved under European institutional conditions.

In 2025, Warszawski stołeczny at 86.9% exceeded Calabria at 50.3% by 36.6 percentage points, Campania at 50.8% by 36.1 points, and Sicilia at 51.4% by 35.5 points, while even Italy's strongest region, Bolzano/Bozen at 80.0%, remained 6.9 points below the EU leader; these gaps are sufficiently large that they cannot plausibly be described as residual convergence noise. Eurostat — Employment rate by NUTS-2 region, 2025

The analytical relevance of frontier distance is that productive agglomerations are not static benchmarks, because high-employment regions can continue to improve while lagging regions improve as well, meaning that positive local growth does not guarantee catch-up; true convergence requires the lagging region's employment rate to rise persistently faster than the frontier or, at minimum, faster than the relevant high-performing comparator over a sufficiently long period.

This is precisely why future monitoring should track gap closure rather than level change alone, because Calabria increasing from 50.3% to 53% would constitute progress, but not convergence if leading regions increased by the same or a greater amount.

Club convergence is a plausible interpretation of the European map

The 2025 regional distribution is compatible with a form of club convergence, under which regions increasingly converge toward different clusters rather than toward one common equilibrium, because high-employment regions are concentrated across Germany, the Netherlands, Czechia, Sweden, Hungary, Slovakia and Ireland, while a separate low-employment cluster persists across southern Italy, much of Greece, parts of Spain, Romania and the French outermost territories. Eurostat — regional employment map, 2025

This interpretation should be treated as an analytical hypothesis rather than as a demonstrated econometric result, because confirming club convergence would require a formal longitudinal test of regional transition paths rather than visual clustering alone; nevertheless, the official map is strongly consistent with the proposition that regions sharing similar combinations of productive density, human capital, connectivity and participation increasingly resemble one another across national boundaries more than they resemble lower-performing regions inside their own state.

The implications are significant because policy designed exclusively around national averages risks misidentifying both the problem and the peer group, whereas Calabria's most relevant comparison may not be with Lombardia alone but with other low-employment peripheral regions facing similar structural constraints, while Utrecht or Praha may share more economically relevant characteristics with high-employment metropolitan regions in other countries than with some lower-performing domestic territories.

Equalisation is neither necessary nor realistic, but persistent structural exclusion is economically costly

A decision-grade assessment should not imply that every European region should converge to an identical employment rate, because demographic structure, sectoral specialisation, urbanisation, education, preferences, commuting and geographic constraints will always produce legitimate regional differences; complete equalisation is neither an economically necessary objective nor a plausible policy benchmark.

The relevant concern is instead the persistence of large, durable and structurally correlated gaps in which low employment coincides with weak productivity, lower investment, outward migration, high inactivity and limited employer density, because such configurations indicate that the region is not simply specialised differently but is operating below its potential and may be caught in a self-reinforcing development trap.

The Commission's Ninth Cohesion Report explicitly uses the concept of a development trap to describe regions experiencing prolonged periods of weak or negative growth, weak productivity growth and low employment creation, while also noting that some transition and less-developed regions began to diverge from the EU average after the 2009 recession. European Commission — Ninth Cohesion Report, Economic Cohesion

The policy objective should therefore be framed as convergence in opportunity and productive capability rather than mechanical equality of headline rates, because the former seeks to remove structural barriers to employment while allowing legitimate regional specialisation and demographic differences to persist.

What the 2015–2025 decade actually proves

The 2015–2025 regional series establishes that the EU experienced a meaningful narrowing of employment-rate dispersion during a period containing the aftermath of the sovereign-debt crisis, the COVID-19 shock, the energy-price shock and major geopolitical disruption, which means that European labour markets demonstrated significant resilience and adaptive capacity; however, the same decade ended with almost one quarter of regions below 74%, an Italian coefficient of variation of 15.1%, and persistent concentrations of low employment in southern and peripheral territories. Eurostat — regional employment disparities, 2015–2025

The correct historical judgment is therefore neither triumphalist nor pessimistic, because the EU has demonstrated that regional convergence is possible, particularly in Central and Eastern Europe and in Member States such as Spain and Portugal where internal employment dispersion has narrowed materially, but the data equally demonstrate that convergence can stall, remain incomplete or bypass specific regions for long periods.

The structural lesson is that integration creates opportunities for convergence without guaranteeing convergence, because the effect of a common market depends on whether regions possess or can build the human capital, institutional quality, infrastructure, business ecosystems and investment capacity required to exploit those opportunities.

Decision-relevant comparative matrix

Territorial system	2025 evidence	Direction of convergence	Structural reading
EU-27 aggregate	Regional CV 9.2%, down from 12.5% in 2015	Clear sigma-convergence	Distribution narrowed, but lower tail remains large
Italy	CV 15.1%; Bolzano 80.0%, Calabria 50.3%	Incomplete and highly uneven	Strongest case of persistent within-state dualism
Spain	CV 6.1%; internal CV fell by at least 40% since 2015	Meaningful convergence	Major progress, but southern/peripheral tail persists
Romania	CV 8.0%; București-Ilfov 81.1%, Sud-Vest Oltenia 61.7%	National catching-up with internal concentration	Capital-led convergence can coexist with domestic divergence
Netherlands	CV ≤ 2.0%; all 12 regions ≥78%	High convergence and high employment	Closest to broad territorial equalisation
Portugal	CV ≤ 2.0%; 7 of 9 regions ≥78%; CV down ≥40% since 2015	Strong internal convergence	Peripheral geography does not preclude broad employment performance
Germany	34 of 38 regions ≥78%; Oberbayern 85.1%	Broad high-employment convergence	Strong regional diffusion despite some metropolitan/deindustrialised exceptions
Greece	11 of 13 regions <74%; national rate 71.0%	Limited equalisation	Weakness geographically broad, not confined to one domestic periphery
France	11 regions <74%, including all 5 outermost regions	Mixed	Metropolitan, deindustrialised and outermost-region mechanisms must be separated

Source: all employment and dispersion values in this matrix come from Eurostat's 2025 regional labour-market release and the associated lfst_r_lmder and regional employment series. Eurostat — Labour market statistics at regional level, 2025

Key judgments

The first judgment is that regional employment convergence is empirically real but incomplete, because the EU-wide population-weighted coefficient of variation fell from 12.5% to 9.2% between 2015 and 2025, while large internal disparities, a persistent sub-74% cluster and frontier gaps exceeding 35 percentage points remain clearly visible in the same official dataset. Eurostat — regional employment disparities, 2015–2025

The second judgment is that convergence has been geographically asymmetric, because Central and Eastern Europe records substantial long-run income convergence and several eastern capital regions now occupy the top of the EU employment ranking, while parts of southern Europe remain concentrated in the lower tail and some regions have struggled to recover the productive trajectory lost after the 2009 crisis. European Commission — Ninth Cohesion Report, Economic Cohesion

The third judgment is that national convergence and regional convergence can move in different directions, as demonstrated by Romania's rapid broader European catching-up alongside an 8.0% internal regional employment coefficient and a 19.4-point difference between București-Ilfov and Sud-Vest Oltenia, while Italy's national average conceals a much larger internal cleavage. Eurostat — Labour market statistics at regional level, 2025

The fourth judgment is that equalisation cannot be inferred from a declining coefficient of variation, because employment-rate convergence can result from participation, migration, changes in inactivity or employment growth without corresponding convergence in productivity, firm density, wages, skills or industrial structure, and those structural variables must therefore be examined independently in the next analytical pillar.

The fifth judgment is that the European policy debate should shift from asking whether cohesion “worked” or “failed” toward identifying which regional mechanisms convert integration and public investment into self-sustaining labour demand, because the coexistence of successful catching-up in Central Europe, narrowing dispersion in Spain and Portugal, broad high employment in Germany and the Netherlands, and persistent weakness in parts of Italy and Greece proves that one uniform causal answer is inconsistent with the official record.

What would change the assessment

The judgment of partial convergence without equalisation would weaken materially if the next five annual regional releases showed not only a continuing decline in the EU-wide coefficient of variation but also sustained and disproportionate improvement in Calabria, Campania, Sicilia, low-employment Greek regions, southern and peripheral Spanish regions and French outermost territories relative to the employment frontier, because that combination would indicate that the lower tail itself was being absorbed rather than merely that the centre of the distribution was compressing.

The judgment would strengthen if the EU employment rate reached or exceeded the 78% 2030 target while a similar number of NUTS-2 regions remained below 74%, because such a result would mean that aggregate progress was being generated principally by regions already close to or above the target rather than through broad territorial catching-up. European Commission — European Pillar of Social Rights Action Plan

The club-convergence hypothesis would gain support if high-employment metropolitan and industrial regions continued converging toward one another across national borders while low-employment southern and peripheral regions remained clustered in a separate range, whereas it would weaken if the latter group began systematically migrating toward the upper distribution independent of national boundaries and historical structural conditions.

---

Open official record

The most important remaining official-record requirement is a harmonised long-run NUTS-2 panel that keeps territorial boundaries sufficiently stable to distinguish genuine economic convergence from classification changes, because NUTS revisions can complicate longitudinal comparison even when the underlying employment series is harmonised; a rigorous econometric test of beta or club convergence should therefore use a boundary-consistent panel rather than simply concatenate annual maps.

A second outstanding requirement is the decomposition of regional employment-rate changes into employment growth, labour-force participation, unemployment, demographic change and migration, because headline convergence alone cannot identify which margin generated the improvement and may therefore overstate or understate convergence in productive capacity.

A third requirement is a fully comparable regional productivity, investment, enterprise-density and human-capital panel for the same 2015–2025 period, which is necessary to determine whether observed employment convergence corresponds to deeper structural convergence or whether labour-market outcomes have narrowed faster than the productive fundamentals underlying them.

The next analytical step is therefore to move from the fact of partial convergence to the mechanism of incomplete equalisation, testing whether endogenous growth, agglomeration, human-capital sorting, firm density and cumulative causation explain why some regions convert European integration into self-reinforcing productive advantage while others remain locked into low-employment equilibria.

From transfers to productive geography

Principal judgment

The central policy problem exposed by the 2025 regional employment map is not whether transfers, cohesion funding or economic integration matter, because the official record shows that they do, but whether they are sufficient to alter the productive structures that generate employment in persistently weak regions, and the balance of evidence indicates that they are not sufficient unless they change the underlying regional production function through higher productivity, stronger firm formation, deeper capital accumulation, better infrastructure, more effective human-capital retention, denser innovation networks and greater access to markets. The European Commission's Ninth Report on Economic, Social and Territorial Cohesion explicitly links regional development performance to productivity, employment and income per head, identifies regions that remain trapped when these dimensions persistently underperform European or national benchmarks, and records that development traps affect regions with very different starting income levels, which means that the problem cannot be reduced to a shortage of fiscal transfers or to initial poverty alone. European Commission — Ninth Report on Economic, Social and Territorial Cohesion European Commission — Economic Cohesion chapter

The appropriate distinction is therefore between redistribution that changes disposable resources and development policy that changes productive capability, because the first can stabilise consumption, reduce poverty, finance public services and prevent social divergence while leaving the local relationship between capital, labour, technology and output largely unchanged, whereas the second must raise the expected return to productive investment or increase the capacity of workers and firms to generate value inside the region. The Commission's RHOMOLO framework for assessing cohesion policy is consistent with precisely this distinction because it models expenditure not merely as an injection of demand but through channels including physical-capital accumulation, human-capital formation and technological progress, while distinguishing short-run demand effects from longer-run supply-side effects. European Commission — Macroeconomic impact of 2021–2027 cohesion-policy programmes

The policy conclusion that follows is narrower than the claim that “transfers do not work,” which would be contradicted by the evidence, because European cohesion policy has supported infrastructure, employment, investment, education, innovation and economic convergence in large parts of Central and Eastern Europe; rather, the defensible conclusion is that transfers produce lasting regional convergence only to the extent that they alter the assets, incentives and institutional conditions from which firms and workers generate future income, while transfers that primarily support current expenditure or compensate for weak private-sector demand can improve welfare without necessarily eliminating the productive gap that caused the transfer dependence in the first place. The Commission itself estimates substantial long-run gains from cohesion policy, including higher output and additional employment, while simultaneously documenting persistent development traps and regional inequalities, demonstrating that redistribution and convergence cannot be treated as synonyms. European Commission — Ninth Cohesion Report: cohesion policy continues to narrow regional gaps

---

A transfer can change income without changing production

The distinction between income support and productive transformation begins with a basic macroeconomic identity that is frequently obscured in regional-policy debates, because a region can experience higher household disposable income without producing an equivalent increase in locally generated value added when income arrives through pensions, social transfers, public-sector wages, intergovernmental redistribution, EU-funded expenditure or remittances; such flows can be economically and socially indispensable while still leaving the private productive base comparatively weak, particularly if a significant share of the additional purchasing power leaks into imports from more productive regions.

This distinction does not imply that demand effects are irrelevant, because additional purchasing power can sustain local services, improve firm revenues, prevent unemployment and support investment under appropriate conditions, yet the multiplier becomes structurally transformative only when stronger demand induces durable expansion of productive capacity rather than remaining dependent on repeated external financing. The European Commission's own modelling of cohesion-policy expenditure distinguishes the initial demand stimulus associated with public spending from persistent supply effects generated through capital formation, skills and technological improvement, which is analytically important because only the second category can continue raising potential output after the original transfer has ended. European Commission — RHOMOLO assessment of 2021–2027 cohesion programmes

The regional implication is that two territories receiving the same nominal fiscal transfer can produce radically different long-run outcomes if one uses the resources to remove transport bottlenecks, raise workforce skills, expand digital infrastructure, improve institutional capacity and crowd in private capital while the other experiences principally a temporary rise in demand, because the first intervention changes the productive environment in which future investment decisions are taken whereas the second largely changes the level of current expenditure.

This mechanism provides a coherent explanation for why persistent fiscal transfers can coexist with persistent employment divergence without proving that the transfers were ineffective, because those flows may have prevented wider social divergence or deeper contraction while failing to generate enough additional productive capacity to move the receiving region toward a self-sustaining higher-employment equilibrium.

---

The relevant object of policy is the regional production function

A productive region converts combinations of labour, physical capital, human capital, infrastructure, technology and organisational knowledge into output, and policy becomes structurally consequential when it changes either the quantity of these inputs or the efficiency with which they interact, which means that the fundamental objective in a persistently low-employment region is not simply to increase expenditure but to raise the productivity and employability of labour while increasing the expected return to locating and expanding firms there.

The relevance of this production-function perspective is especially clear in the European context because cohesion policy itself is officially designed around channels that include productive investment, infrastructure, skills, innovation, digitalisation and territorial connectivity rather than simple redistribution, while the Commission's macroeconomic modelling attributes long-run effects to capital accumulation, human-capital improvements and technological change rather than assuming that public spending permanently raises output simply because it occurs. European Commission — expected macroeconomic impact of cohesion policy 2021–2027

The distinction also clarifies why high employment and high transfers are not opposites, because productive regions can receive substantial public expenditure while remaining highly competitive if that expenditure complements private capital, while weak regions can remain dependent even after large interventions if infrastructure is disconnected from viable productive ecosystems, education does not match labour demand, administrative capacity delays investment, or skilled workers leave before local firms can employ them.

The relevant question for regional policy is therefore not “how much money enters the territory?” but “how much of that expenditure permanently raises the region's capacity to produce competitive goods and services, attract investment, retain workers and generate employment after the programme finishes?”, because those two quantities need not move together.

Agglomeration changes the expected return to location

The strongest reason that regional convergence is not automatic lies in agglomeration economies, because firms and skilled workers derive benefits from locating near other firms and skilled workers through thicker labour markets, specialised suppliers, knowledge spillovers, shared infrastructure, better matching between jobs and workers, larger customer bases and more efficient access to specialised services. The OECD's regional evidence consistently identifies higher productivity in dense urban and metropolitan areas and notes that economic activity located in densely populated urban cores typically records higher productivity and per-capita output, while its more recent analysis of intermediary cities similarly explains that larger settlements tend to exhibit higher income because the clustering of firms, workers and services raises productivity and economic efficiency. OECD — Productivity and Jobs in a Globalised World OECD — Unlocking the Potential of Intermediary Cities for Regional Development

These agglomeration effects create a cumulative mechanism because a region with a sufficiently dense employer base offers workers a greater probability of finding jobs that match their skills, which attracts more workers; the larger specialised workforce then makes the location more attractive to firms, while additional firms support suppliers, professional services, financial institutions, research networks and infrastructure, which in turn raises the productivity advantage of locating there.

The opposite mechanism can operate in a low-density region, because a limited number of high-productivity employers reduces the probability that specialised workers will find appropriate jobs, encouraging migration toward stronger labour markets; once skilled workers leave, potential investors face a thinner labour pool, which reduces the expected return to investment and can further weaken the incentive for firms to locate there.

The process is therefore endogenous because yesterday's productive density helps determine tomorrow's productive density, and this property is fundamentally different from a neoclassical convergence model in which scarce capital should naturally flow toward lower-capital regions where its marginal return is assumed to be higher.

Capital does not automatically flow toward the poorest region

The assumption that weaker regions should eventually catch up because they possess less capital and therefore offer higher marginal returns is too restrictive once modern production depends on complementary assets, because a factory, laboratory, software company or advanced-services firm does not evaluate only wage costs or the physical quantity of capital already present; it also evaluates skills, infrastructure, logistics, regulatory capacity, supplier networks, access to finance, research institutions, digital connectivity, business services and the probability that specialised staff can be recruited and retained.

The OECD reports that regional labour-productivity differences remain extremely large even within the same countries, with productivity in the most productive region averaging nearly twice that of the least productive region across OECD countries, which demonstrates that national institutions alone do not equalise the efficiency with which labour is converted into output. OECD — Productivity and innovation in regions

Such productivity gaps affect investment decisions because a high-productivity region can remain attractive despite higher wages, rents and congestion if those costs are more than offset by superior worker matching, supplier density, innovation capacity and market accessibility, while a low-wage region can remain unattractive when lower labour costs coexist with weaker infrastructure, limited human capital, administrative uncertainty or a narrow supplier ecosystem.

The result is that market integration can intensify rather than neutralise spatial concentration under some conditions, because the removal of barriers allows firms to serve a broader market from a smaller number of highly productive locations, while workers can relocate toward those locations more easily; integration therefore creates opportunities for poorer regions to attract capital, but it simultaneously increases competitive exposure and mobility, which can strengthen leading agglomerations when weaker regions lack sufficient complementary assets.

The Single Market can therefore be both a convergence mechanism and a sorting mechanism

European integration lowers barriers to trade, capital movement and labour mobility, which can promote convergence when lower-cost or rapidly reforming regions attract investment and integrate into continental value chains, as occurred across substantial parts of Central and Eastern Europe; however, the same process can operate as a sorting mechanism when mobile capital and skilled labour select locations offering the highest combined returns rather than the lowest initial income.

This duality helps explain why Central European regions were able to converge rapidly while some southern regions did not, because accession economies situated close to major German and Central European industrial systems could combine lower initial labour costs with improving infrastructure, foreign investment, manufacturing integration, skills and proximity to high-income markets, creating a cumulative productive process rather than a transfer-dependent one. The Commission's Ninth Cohesion Report records that the average GDP per head of Member States joining from 2004 onward rose from approximately 52% of the EU average at accession to nearly 80% two decades later, while their unemployment rate fell from approximately 13% to around 4%, demonstrating that European integration was fully compatible with rapid convergence when productive integration took hold. European Commission — Ninth Cohesion Report: twenty years of enlargement and convergence

This successful eastern convergence makes the persistence of southern underperformance more analytically significant because it removes the possibility of explaining the regional map simply by claiming that the Single Market inevitably concentrates activity in the old European core; Warszawski stołeczny and Praha occupying the top two positions in the 2025 EU NUTS-2 employment ranking show instead that peripheral or formerly lower-income economies can enter the high-employment club when they accumulate the productive conditions necessary to capture investment and skilled labour. Eurostat — Labour market statistics at regional level, 2025

Foreign direct investment follows ecosystems rather than fiscal need

Foreign direct investment is especially important because it can simultaneously deliver capital, managerial know-how, export-market access, technology and supply-chain integration, yet its regional distribution is not determined by the location of the greatest social need; investors select territories according to expected risk-adjusted returns, which means that regions already possessing dense industrial or technological ecosystems can attract additional investment more easily than regions with weaker productive foundations.

The OECD's 2026 work on the interaction between FDI and SME ecosystems in Europe emphasises that the developmental effect of foreign investment depends on linkages with domestic firms, regional innovation systems and the capacity of local SMEs to absorb knowledge and technology, which means that simply attracting an external plant is less transformative than embedding foreign investment in a local supplier and innovation ecosystem. OECD — Connecting FDI and SMEs for Productivity and Innovation in Europe

This distinction is fundamental for cohesion policy because a subsidy that persuades one mobile investor to establish an isolated facility can create employment without necessarily generating broad convergence, whereas investment linked to local suppliers, research institutions, specialised training and logistics can create spillovers that raise the productivity of other firms and improve the region's future ability to attract capital without equivalent subsidies.

The objective should therefore be to transform external investment from an enclave into an ecosystem, because the latter changes productive geography while the former can disappear when the original cost advantage or subsidy expires.

Human capital is both an input and a mobile asset

Human capital occupies a uniquely important position in regional divergence because education raises individual employment probabilities and productivity, but educated workers are themselves mobile and can leave the region that financed their education if local opportunities are insufficient, which means that a weak region can invest successfully in skills yet fail to retain the resulting productive asset.

The Commission's demographic analysis identifies regions at risk of a talent-development trap, where the decline of the working-age population and difficulties in developing, attracting or retaining skilled workers interact with lower GDP per head, wages and household income, creating the possibility that weak labour demand and demographic decline reinforce one another. European Commission — Ninth Cohesion Report, demographic transition

This mechanism is especially relevant to southern Italy, Greece and other lagging regions because higher education alone cannot generate local convergence when graduates subsequently migrate toward Milan, Munich, Amsterdam, Prague, Paris, London or other employment centres; from a national perspective that migration may improve labour allocation, yet from the sending region's perspective it can weaken the human-capital base required for future growth.

The resulting feedback loop is severe because fewer skilled workers reduce the incentive for high-productivity employers to enter, while fewer high-productivity employers strengthen the incentive for skilled workers to leave, making simultaneous intervention on labour demand and human-capital retention more important than education spending considered in isolation.

Demography can transform a growth problem into a capacity problem

Persistent outward migration and population ageing can eventually change the nature of regional underdevelopment because the initial problem of insufficient employment can evolve into a shortage of workers, entrepreneurs, taxpayers and skilled professionals, particularly when young and educated cohorts are disproportionately represented among those leaving.

The Commission's cohesion analysis warns that demographic decline is increasingly concentrated in specific European territories and that regions facing talent-development traps often combine shrinking working-age populations with weaker income and economic performance, which can make convergence more difficult because public-service provision becomes more expensive per resident while the private market supporting firms becomes smaller. European Commission — Demographic transition and talent-development traps

The policy significance is that employment rates alone can become misleading in shrinking regions, because a declining population may mechanically improve the employment ratio if non-employed residents leave faster than jobs disappear, even though productive capacity is weakening; genuine territorial convergence must therefore be assessed jointly through employment, productivity, demographic sustainability and human-capital retention.

Firm density is a missing bridge between employment and productivity

Regional employment ultimately requires employers, and regions with a dense network of firms possess structural advantages because workers have more potential matches, entrepreneurs have more customers and suppliers, specialised business services become viable, failed workers or firms can be reallocated more rapidly and knowledge circulates through formal and informal networks.

The OECD's regional-productivity framework emphasises business dynamism, innovation and productivity growth across tradable and non-tradable sectors, while showing that regional productivity inequality is closely connected to the geography of tradable economic activity and technological upgrading. OECD — Productivity and regional income inequality

This relationship explains why public employment alone cannot substitute indefinitely for weak private-sector density, because a public administration, hospital or school can raise employment and provide essential services while producing fewer cumulative spillovers into export capability, supplier networks and business formation than a diversified productive ecosystem; public investment becomes developmentally strongest when it enables private-sector expansion rather than simply compensating for its absence.

The policy objective is therefore not necessarily to maximise the number of firms, because a region populated by very small low-productivity enterprises can remain weak, but to increase the density of productive, scalable and interconnected firms capable of generating stable labour demand and integrating with national or European markets.

Productivity is the decisive long-run variable

Regional employment can rise temporarily without productivity growth, yet sustained convergence in living standards and fiscal capacity requires workers to generate more value per hour or per person employed, because only productivity growth allows wages, profits and tax revenues to rise together without generating equivalent losses in competitiveness.

The OECD reports that productivity differences across regions remain exceptionally large and that, within countries, labour productivity in the most productive region is on average almost twice that in the least productive region, which shows that spatial inequalities in productive efficiency remain a fundamental feature even where labour law, taxation and national macroeconomic institutions are shared. OECD — Productivity and innovation in regions

The Commission's definition of a regional development trap similarly includes weak productivity growth alongside weak GDP-per-head and employment performance, demonstrating that employment alone is not treated institutionally as sufficient evidence of sustainable convergence. European Commission — Economic Cohesion and regional development traps

The implication for southern Europe is that an increase in low-productivity employment can narrow the headline employment-rate gap without eliminating the productivity gap, while a productive-convergence strategy requires movement into activities capable of sustaining higher output and wages over time.

Agglomeration generates increasing returns that weaken the assumption of automatic sigma-convergence

Traditional convergence intuition often assumes diminishing returns to capital, under which regions with less capital should offer higher marginal returns and therefore attract investment until the gap narrows; agglomeration economies modify this logic because the return to a new investment can increase when it is located near existing complementary capital, skilled workers, suppliers and knowledge institutions.

Under increasing returns to agglomeration, the expected return to the next high-technology firm may be greater in Munich, Utrecht, Prague or Warsaw than in a much poorer region precisely because previous investment has already built the infrastructure, skills, suppliers and networks on which the new firm depends.

This means that low initial capital is not necessarily an advantage and can become a disadvantage when production depends heavily on complementarity, because the absence of one asset lowers the return to the others; constructing infrastructure without skilled labour may have limited effect, training workers without employers can accelerate outmigration, subsidising firms without local suppliers can create enclaves, and financing research without commercialisation capability can fail to create regional employment.

The policy implication is therefore complementarity rather than volume, because development interventions become more effective when infrastructure, skills, innovation, business finance and institutional reforms are sequenced and concentrated sufficiently to move a region across the threshold at which private cumulative investment becomes self-reinforcing.

Capital cities show the power and limits of agglomeration

Recent Commission analysis of European metropolitan productivity reinforces the conclusion that agglomeration has a powerful but uneven effect, because capital cities have generally outperformed other metropolitan regions while secondary cities have often struggled to match their productivity dynamics, indicating that national growth can become disproportionately concentrated in dominant urban centres. European Commission — Capital cities lead, while other cities lag in the EU

The 2025 employment map provides complementary labour-market evidence because many capital regions outperform national averages, with București-Ilfov reaching 81.1% compared with Romania's 69.0% national rate, while Praha ranks second in the entire EU at 86.0% and Warszawski stołeczny first at 86.9%; these cases demonstrate the capacity of metropolitan agglomerations to capture skilled labour, investment and high-productivity activity. Eurostat — Labour market statistics at regional level, 2025

Yet Brussels, Berlin and Vienna show that agglomeration does not guarantee a high resident employment rate, because large cities can simultaneously contain highly productive firms and concentrated inactivity, inequality, housing constraints, migrant-integration challenges and commuter inflows, which means that productivity of the place and employment of residents must remain distinct analytical variables.

This distinction reinforces the argument that regional development should not seek to imitate metropolitan density mechanically, but should identify which agglomeration benefits can be reproduced through networks, transport, digital connectivity, university-industry linkages and stronger intermediary cities.

Infrastructure produces convergence only when it changes effective accessibility

Infrastructure spending is one of the most visible components of cohesion policy, yet its economic value depends less on the physical asset itself than on whether it reduces the effective cost of moving people, goods, energy or information between the region and productive markets.

A motorway that integrates a peripheral industrial district into continental supply chains can raise productivity, expand the labour-market catchment area and attract firms, while a transport project that reduces travel times only marginally or connects locations with weak productive demand can have far smaller long-run effects despite equivalent construction expenditure.

The same principle applies to digital infrastructure, ports, railways and energy networks, because physical connectivity becomes productive capital only when firms and workers can exploit it, making infrastructure complementary to skills, business capacity and institutional competence.

The European Commission's cohesion modelling explicitly includes infrastructure within capital accumulation and allows benefits to propagate across regions through trade and mobility linkages, which is why the impact of spending can extend beyond the territory in which the project is physically constructed. European Commission — RHOMOLO modelling of cohesion-policy effects

Transfers can leak toward stronger regions through trade

A central but often neglected feature of regional redistribution is that the region receiving the fiscal transfer is not necessarily the region capturing the associated production, because households and governments can spend transferred income on goods, services and capital equipment produced elsewhere.

This leakage is not a policy failure in welfare terms because residents still receive goods and services, yet it weakens the assumption that a euro transferred into a low-productivity region automatically generates a euro of regional productive capacity.

Where local supplier density is low, increased demand can disproportionately benefit firms located in stronger regions, thereby raising disposable income in the recipient territory while simultaneously reinforcing production in the core.

This is one reason why regional-policy evaluation should distinguish expenditure incidence from production incidence, because the geographical location of the beneficiary household, infrastructure project or public administration does not necessarily identify the geography in which value added, profits and high-skilled employment are generated.

The same logic applies to procurement and major projects

Large public projects can generate temporary local employment without producing long-run convergence if specialised engineering, technology, finance and management are imported from outside the region and disappear after construction, whereas projects tied to local supplier development, vocational training and research capability can produce lasting spillovers.

The relevant policy metric should therefore extend beyond gross project expenditure to include locally retained value added, supplier upgrading, workforce skill formation, private investment crowded in and productive capacity surviving after the public programme ends.

This approach also reduces the risk of mistaking absorption capacity for development, because a region can successfully spend its allocated resources without producing a proportionate structural change in productivity or employment.

Cohesion policy has produced measurable macroeconomic effects

The evidence does not justify an anti-transfer interpretation because the European Commission estimates substantial effects from cohesion policy and reports that the policy has materially supported convergence, with the Ninth Cohesion Report estimating that programmes can generate around 1.3 million additional jobs in the EU by 2027, while long-run GDP effects are particularly important in less-developed Member States and regions. European Commission — Ninth Cohesion Report policy-impact assessment

These estimates are model-based rather than direct observations and should therefore be interpreted as counterfactual assessments of what output and employment would have been without policy rather than as measured employment counts attributable uniquely to each project, yet they materially weaken any assertion that cohesion spending merely redistributes income without affecting productive capacity.

The correct conclusion is instead that cohesion policy can change productive geography, but its effects vary according to regional starting conditions, project composition, institutional capacity and the complementarity of investments, which explains why large aggregate positive effects can coexist with development traps in particular regions.

Development traps are evidence of failed self-reinforcement

The Commission defines development-trapped regions through persistent underperformance in dimensions including GDP per head, productivity and employment, which provides an institutional formulation of the same cumulative-causation problem identified in endogenous-growth theory, because a region becomes trapped when weak performance reduces its capacity to generate the investment and human capital required to escape weak performance. European Commission — Economic Cohesion and development traps

The concept is particularly important because development traps are not confined to the poorest regions, meaning that middle-income territories can stagnate after exhausting earlier growth models without successfully moving toward innovation-intensive production.

This phenomenon is relevant to older industrial and southern European regions because convergence based initially on low wages, construction, traditional manufacturing, tourism or fiscal transfers can eventually stall when wages rise, demographics deteriorate or competitive pressure increases but innovation capacity and firm sophistication fail to rise correspondingly.

The policy problem therefore changes over time, because the instruments capable of moving a low-income region toward middle income are not necessarily those required to move a middle-income region toward the productivity frontier.

Southern Italy illustrates the difference between income support and productive convergence

Italy provides the most consequential test of this framework because the country combines decades of national redistribution and European regional spending with the EU's highest regional employment-rate dispersion, as the 2025 Eurostat record places Bolzano/Bozen at 80.0%, Calabria at 50.3%, Campania at 50.8% and Sicilia at 51.4% for persons aged 20–64, while Italy's national rate stands at 67.6%. Eurostat — Labour market statistics at regional level, 2025

The persistence of this divide does not prove that transfers worsened southern Italy or achieved nothing, because the relevant counterfactual might have involved substantially lower household income, weaker infrastructure and greater outmigration in their absence; it does demonstrate, however, that redistribution and infrastructure expenditure have not yet generated a productive system capable of equalising southern employment with the strongest northern regions.

The interaction with female employment is especially important because Puglia, Campania, Sicilia, Calabria and Basilicata recorded some of the largest gender employment gaps in the EU in 2025, meaning that increasing productive capacity must be accompanied by policies capable of converting latent female labour supply into actual employment through childcare, transport, incentives, formalisation and employer demand rather than assuming that stronger aggregate GDP will automatically close the participation gap. Eurostat — regional gender employment gaps, 2025

The Italian policy challenge is therefore not simply to transfer more resources from north to south but to raise the expected return to productive activity in the south while simultaneously increasing the effective supply of skilled labour and reducing the incentives for young workers and firms to relocate elsewhere.

France demonstrates why territorial policy must distinguish remoteness from deindustrialisation

France's low-employment regions illustrate two distinct structural problems because the French outermost territories face geographic remoteness, small markets, transport costs and demographic pressures, while parts of metropolitan north-eastern France face the legacy of industrial restructuring in coal, steel and manufacturing.

The same transfer instrument cannot be expected to solve both problems efficiently because an outermost region requires connectivity, human-capital development, public services and economic diversification adapted to exceptional geography, whereas a former industrial region requires technological upgrading, workforce retraining, brownfield redevelopment, innovation and the replacement or transformation of declining tradable sectors.

Eurostat's 2025 map, which places Mayotte at 36.9% and Guyane at 50.0%, while also identifying former industrial regions among Europe's lower-employment territories, demonstrates why the lower tail should not be treated as one policy category merely because the headline employment rate is similar. Eurostat — Labour market statistics at regional level, 2025

Germany demonstrates the value of distributed productive density

Germany provides an important positive comparator because 34 of its 38 NUTS-2 regions had already reached the 78% employment target in 2025, which implies that strong employment is distributed across multiple industrial, metropolitan and medium-sized regional economies rather than concentrated exclusively in one capital or financial centre. Eurostat — regional employment rates, 2025

The German pattern is consistent with a regional structure in which productive capacity is distributed across multiple cities, industrial districts and specialised clusters, which reduces dependence on one dominant metropolitan centre and creates multiple locations capable of retaining skilled workers and attracting investment.

This does not establish that Germany's model can simply be replicated elsewhere, because its industrial history, vocational-training institutions, export specialisation and infrastructure are path-dependent, but it does demonstrate that high employment and geographically distributed productive density can coexist.

The United Kingdom remains analytically relevant precisely because it is outside the EU framework

The United Kingdom should not be included in the EU-27 employment ranking, yet it remains a valuable comparative case because its highly concentrated productivity geography demonstrates that severe regional divergence is not uniquely produced by the euro, EU cohesion rules or the Single Market.

OECD regional evidence has long identified substantial productivity concentration around London and parts of southern England, while weaker productivity persists across several former industrial and peripheral regions, showing that a sovereign currency and independent national fiscal framework do not themselves eliminate spatial agglomeration or regional development traps. OECD — OECD Regional Outlook 2023

The British comparison therefore strengthens rather than weakens the productive-geography argument, because similar spatial polarisation can emerge under different monetary and institutional arrangements when high-productivity activity, skilled labour and investment become geographically concentrated.

Place-based policy is justified when the market failure is spatial

The existence of agglomeration effects provides a theoretical justification for place-based intervention because market outcomes can become path-dependent when firms and workers individually make rational location choices that collectively reinforce already-strong regions, while no single private actor possesses an incentive to finance the full package of infrastructure, skills, supplier development and institutional capacity required to shift the weaker region onto a higher-growth trajectory.

A firm will not normally train an entire regional workforce if competitors can hire the trained workers, a worker will not move to a weak region merely because firms might arrive later, and an infrastructure investor cannot capture all benefits created for surrounding businesses, which means that coordination failures can prevent potentially viable regional clusters from emerging.

Government intervention becomes economically defensible when it coordinates these complementary investments rather than merely subsidising one factor in isolation.

Place-based policy nevertheless carries serious implementation risk

Agglomeration theory does not justify unlimited regional subsidy because governments can misidentify viable sectors, subsidise politically favoured incumbents, create infrastructure with insufficient demand, fragment resources across too many projects or maintain activities that would not survive without permanent support.

A decision-grade policy framework must therefore distinguish building capabilities from protecting existing structures, because the objective should be to increase adaptability, investment attraction and productive complexity rather than freeze the region's historical industrial composition.

The strongest interventions are likely to be those that create assets transferable across firms and sectors, including skills, transport, digital connectivity, research capability, energy infrastructure and administrative competence, because these investments retain value even when individual firms or technologies change.

The appropriate strategy is concentrated complementarity rather than indiscriminate spending

A weak region rarely lacks only one productive input, which means that isolated interventions can fail when other binding constraints remain unchanged; improving a university without creating high-skill jobs can accelerate graduate migration, constructing industrial land without transport or energy capacity can leave sites unused, providing investment grants without skilled labour can produce recruitment constraints, and upgrading transport without strengthening local firms can increase the ability of residents to shop or work elsewhere without generating equivalent local production.

The strategic implication is that regional policy should identify the binding package of complementary constraints and address them jointly enough to change private expectations, because cumulative investment begins only when firms and workers believe that other complementary actors will also remain or arrive.

This creates a strong argument for concentrating resources on fewer coherent regional transformation strategies rather than distributing funding thinly across politically balanced portfolios whose individual components never reach the scale required to alter investment behaviour.

The timing of policy matters because productive geography is path-dependent

Agglomeration advantages compound over time, which means that delayed intervention can become more expensive because the region continues losing firms, skilled workers and fiscal capacity while stronger regions accumulate additional advantages.

The Commission's development-trap analysis is consistent with this interpretation because regions can remain stuck for prolonged periods once weak productivity, employment and income growth become mutually reinforcing, while regions at risk of talent-development traps simultaneously face demographic decline and difficulty retaining skilled labour. European Commission — Economic Cohesion and development traps European Commission — Demographic transition and talent-development traps

The relevant government objective is therefore not simply to maximise annual absorption of funds but to intervene before a region loses enough demographic and productive mass that recovery requires disproportionately larger resources.

A regional production strategy requires different metrics from a transfer strategy

A transfer-oriented evaluation naturally asks how much money was allocated, committed or spent, while a production-oriented evaluation must ask whether private investment rose, labour productivity increased, employment broadened, firms scaled, skilled workers remained, innovation activity expanded and external subsidy dependence declined.

The distinction is fundamental because high absorption can coexist with weak development if the projects do not affect binding productive constraints, whereas lower initial expenditure can generate stronger long-run effects if it changes connectivity, firm behaviour or human-capital retention.

For certified policy evaluation, the relevant dashboard should therefore include at minimum employment rate, productivity per worker, private fixed investment, business births and scale-up rates, tertiary-educated retention, female employment, R&D intensity, export or tradable-sector participation, commuting accessibility and demographic change, with all measures tracked against both national and EU comparators rather than against the region's own previous level alone.

The policy objective is to cross a threshold from externally supported to self-reinforcing growth

The most important conceptual distinction in this chapter is between a region whose income depends heavily on external fiscal support and a region in which external public investment has succeeded in creating conditions for endogenous private accumulation, because only the latter can progressively reduce dependence on compensatory transfers.

A successful development intervention should eventually make subsequent investment easier rather than equally difficult, because the first wave of productive investment creates workers, suppliers, infrastructure usage, knowledge and credibility that increase the attractiveness of the second wave.

When that process does not emerge after repeated programmes, government should not automatically conclude that more of the same spending is required, but should identify which complementarity failed, whether the problem lies in skills, infrastructure, governance, firm capability, access to finance, market connectivity or demographic erosion.

Mechanism matrix: what different interventions can and cannot accomplish

Policy instrument	Immediate effect	Potential long-run production effect	Main failure mode
Household transfers	Higher disposable income and consumption stabilisation	Indirect demand support and human-capital protection	Production leaks outside the region and dependency persists
Public-sector employment	Direct employment and service provision	Human-capital and institutional gains where services raise productivity	Crowding out or fiscal dependence without private-sector expansion
Transport infrastructure	Construction demand and lower travel costs	Larger effective markets, commuting zones and logistics efficiency	Low utilisation or stronger leakage toward external producers
Digital infrastructure	Improved connectivity	Enables remote services, innovation and business scaling	Connectivity without firms or skills produces weak additionality
Education and training	Higher individual human capital	Greater productivity and investment attractiveness	Skilled workers migrate if local demand remains weak
Firm subsidies	Lower investment costs	Potential cluster formation and employment	Enclave investment or subsidy competition without spillovers
R&D and innovation support	Higher research expenditure	Knowledge spillovers and technological upgrading	Research remains disconnected from commercialisation
Childcare and care infrastructure	Higher effective labour supply	Sustained female employment and larger regional workforce	Weak effect if labour demand remains insufficient
Supplier-development programmes	Stronger SME capabilities	Embedding of FDI and deeper regional value chains	Insufficient anchor demand or weak absorptive capacity
Institutional-capacity investment	Faster and more predictable implementation	Lower transaction costs and higher investment credibility	Reform remains procedural without productive impact

The mechanisms represented in this matrix are consistent with the European Commission's distinction between short-run expenditure effects and long-run changes operating through human capital, physical capital and technology, while the OECD evidence on regional productivity and FDI linkages reinforces the importance of firm ecosystems, innovation and local absorptive capacity. European Commission — RHOMOLO cohesion-policy impact framework OECD — Connecting FDI and SMEs for Productivity and Innovation in Europe

Italy, France, Germany and the United Kingdom imply four different policy problems

Italy requires a strategy capable of moving parts of the Mezzogiorno from transfer-supported demand toward stronger private labour demand, with particular emphasis on female employment, human-capital retention, infrastructure execution, supplier density and investment credibility, because the country's 2025 15.1% regional employment dispersion indicates that national macroeconomic improvement alone will not close the internal gap. Eurostat — Labour market statistics at regional level, 2025

France requires differentiated territorial policy because outermost regions and metropolitan former-industrial regions face fundamentally different binding constraints, meaning that uniform redistribution is unlikely to generate equal returns when one problem is extreme remoteness and demographic pressure while another is industrial restructuring and technological transition. Eurostat — regional employment patterns, 2025

Germany demonstrates the strategic value of a more distributed productive geography, because high employment across 34 of 38 NUTS-2 regions reduces dependence on one dominant growth pole and suggests that strong regional employment can coexist with multiple industrial and metropolitan centres rather than a single national agglomeration. Eurostat — Labour market statistics at regional level, 2025

The United Kingdom provides an external control case showing that spatial inequality persists outside the euro and EU cohesion architecture, while London-centred productivity concentration and weaker former-industrial regions demonstrate that sovereign monetary autonomy does not itself neutralise agglomeration; the comparison strengthens the conclusion that productive geography is a deeper structural phenomenon than EU institutional design alone. OECD — Regional Outlook 2023

What integration can accomplish, and what it cannot guarantee

Integration expands market size, lowers cross-border transaction costs, enables labour mobility, facilitates capital movement and allows weaker regions to participate in larger value chains, which creates powerful opportunities for convergence and is consistent with the successful catching-up observed across Central and Eastern Europe.

Integration does not guarantee that mobile factors will locate in weak regions, however, because workers and firms respond to relative returns, meaning that the very freedom of movement intended to improve efficiency can accelerate the movement of talent and capital toward locations where complementary assets already exist.

This is not a contradiction but a defining characteristic of an integrated economy, because efficiency at the European level and territorial convergence at the regional level are not automatically the same objective.

A Single Market can become more productive overall while production becomes spatially concentrated, and the resulting welfare gain can coexist with increasingly difficult political and fiscal questions about how to sustain regions that lose productive population.

Transfers remain necessary even when they are not sufficient

The argument for productive transformation should not be interpreted as an argument against redistribution, because transfers perform functions that markets will not automatically provide, including income insurance, minimum public-service provision, social stability, protection against asymmetric shocks and the maintenance of territorial cohesion.

A weak region that loses fiscal support before private production has expanded sufficiently can enter a deeper contraction rather than a productive adjustment, because falling demand can cause additional business closures, outmigration and deterioration of public services.

The correct sequencing is therefore not “transfers or development” but transfers that protect social cohesion combined with investment capable of reducing the future need for compensatory transfers, which aligns redistribution with productive transformation rather than treating the two as substitutes.

The fiscal criterion should be additionality rather than gross expenditure

Every large regional programme should ultimately be evaluated against an additionality question asking what private investment, productivity, employment or human-capital retention occurred because of the intervention that would not otherwise have occurred, because gross expenditure cannot distinguish between genuinely transformative projects and activities that would have happened without public support.

The Commission's model-based evaluation architecture implicitly addresses this problem by comparing policy scenarios with counterfactual baselines rather than equating funds spent with output created, which is methodologically essential for any serious assessment of cohesion effectiveness. European Commission — macroeconomic modelling of cohesion-policy programmes

The strategic objective should be regional capability, not regional self-sufficiency

Productive convergence does not require every region to contain every industry or become economically self-sufficient, because specialisation and trade are sources of efficiency, while forcing identical sectoral structures across territories would destroy rather than enhance comparative advantage.

The relevant goal is instead to ensure that regions possess enough productive capability, connectivity and human capital to participate in national and European value chains on terms capable of sustaining employment and incomes.

A successful peripheral region may therefore remain specialised and highly open, but it should possess firms and workers capable of capturing sufficient value from its specialisation rather than functioning primarily as a consumption market supported by external fiscal flows.

The deeper policy failure is coordination failure

The strongest interpretation of persistent regional divergence is that weak territories can become trapped not because every individual investment is irrational, but because each investment depends on other investments occurring simultaneously, which creates a coordination problem that private markets can struggle to solve.

Workers will not remain without credible employers, firms will not enter without workers and infrastructure, financiers will not lend without viable firms, universities cannot retain graduates without careers, and local governments cannot finance infrastructure sustainably without a tax base.

A productive regional strategy must therefore coordinate these expectations sufficiently to create a new equilibrium in which private decisions reinforce rather than undermine the development objective.

Key judgments

The principal judgment is that fiscal transfers can reduce welfare disparities without eliminating productive disparities, because household income, public services and consumption can be stabilised even when local productivity, private investment and employer density remain weak; the existence of persistent development traps alongside large-scale cohesion intervention demonstrates why these dimensions must not be conflated. European Commission — Ninth Report on Economic, Social and Territorial Cohesion

The second judgment is that agglomeration creates cumulative advantage, because skilled workers, firms, suppliers, knowledge institutions and infrastructure increase one another's productivity, while the OECD's regional evidence shows exceptionally large within-country productivity differences and systematically stronger productivity in dense urban economies. OECD — Productivity and innovation in regions OECD — Regions and Cities at a Glance 2024

The third judgment is that integration has no single territorial effect, because it can enable rapid convergence where regions possess sufficient complementary assets, as demonstrated by the substantial catching-up of Central and Eastern Europe, while simultaneously strengthening high-productivity agglomerations when workers and capital migrate toward them. European Commission — Ninth Cohesion Report and twenty years of enlargement

The fourth judgment is that human capital cannot be treated independently from labour demand, because educating workers without creating productive employment can increase outward migration, while firms will not invest where the specialised workforce they require is unavailable; successful policy must therefore create complementary labour supply and labour demand rather than maximise either independently. European Commission — Demographic transition and talent-development traps

The fifth judgment is that cohesion policy is capable of changing productive geography, because Commission modelling attributes material employment and output effects to the policy and identifies long-run channels through capital, skills and technology, yet the persistence of regional traps demonstrates that effectiveness varies substantially with project quality, institutional capacity and complementary regional conditions. European Commission — Macroeconomic impact of 2021–2027 cohesion policy

The sixth judgment is that the relevant policy objective is not equal spending per inhabitant but equalisation of productive opportunity, which requires governments to identify the binding regional constraints preventing viable firms and workers from responding to European market opportunities rather than presuming that expenditure volume itself measures structural convergence.

What would change the assessment

The assessment that productive geography rather than transfer volume is the dominant explanation would weaken if regions receiving persistently high fiscal transfers began displaying durable productivity, employment and private-investment convergence without corresponding improvements in human capital, infrastructure, business density, connectivity or institutional capacity, because such evidence would suggest that income support itself can generate much stronger endogenous productive effects than the current mechanism implies.

The assessment would strengthen if regional data showed that areas combining human-capital loss, low productivity and weak firm formation remained persistently in the lower employment tail despite substantial per-capita transfers, while regions receiving comparable or lower transfers but achieving stronger private investment and ecosystem development converged more rapidly.

The agglomeration mechanism would weaken if firm and skilled-worker location became increasingly insensitive to existing productive density, perhaps because remote work, digital production or technological change materially reduced the benefits of geographical clustering, whereas it would strengthen if capital regions and large metropolitan systems continued to capture disproportionate shares of productivity growth and high-skilled employment. OECD — Implications of Remote Working Adoption on Place-Based Policies

Open official record

The most important outstanding evidence gap is a fully harmonised NUTS-2 panel combining EU-LFS employment, regional labour productivity, private investment, firm births and deaths, enterprise scale, FDI, tertiary-educated migration, cohesion expenditure and commuting accessibility over a sufficiently long period, because only such a dataset can identify whether transfer intensity predicts convergence after controlling for initial productive conditions and whether observed employment improvements are accompanied by structural transformation.

A second unresolved issue concerns local fiscal incidence, because expenditure allocated to a region is not identical to value added retained within the region, while procurement, imports and external contractors can move part of the economic benefit elsewhere; a rigorous evaluation therefore requires territorial input-output analysis capable of identifying where production generated by publicly financed expenditure actually occurs.

A third unresolved issue concerns human-capital retention, because educational attainment is measured more systematically than the regional origin and destination of highly skilled workers, yet the distinction between producing graduates and retaining graduates is essential for evaluating whether education expenditure changes local productive capacity or subsidises human capital used elsewhere.

A fourth unresolved issue concerns firm density and scaling, because employment statistics identify whether residents work but do not reveal whether the regional employer base is becoming more diversified, innovative and resilient, which means that business-demography and productivity evidence must be incorporated before concluding that a rising employment rate represents a durable productive transition.

A fifth unresolved issue concerns causal attribution, because the coexistence of transfers and regional convergence does not establish that the transfer caused the convergence, while the coexistence of large transfers and stagnation does not establish that the transfer failed; credible evaluation requires counterfactual modelling or quasi-experimental evidence rather than comparisons of gross expenditure with subsequent outcomes.

Net assessment

The European regional employment map is best understood not as evidence that redistribution failed, but as evidence that redistribution cannot substitute for productive geography, because the deepest regional divergences are reproduced through differences in productivity, skills, firm ecosystems, demographic sustainability, accessibility and cumulative investment attractiveness rather than through household income alone.

Cohesion policy can alter that geography when it raises human capital, infrastructure quality, technological capability and private investment sufficiently to create self-reinforcing productive ecosystems, and the successful catching-up of Central and Eastern Europe demonstrates that this mechanism is real rather than theoretical; however, the persistence of southern Italian, Greek, French outermost and other development-trapped regions shows that integration and transfers do not automatically supply the complementary assets required for convergence. European Commission — Ninth Report on Economic, Social and Territorial Cohesion

The most consequential policy shift is therefore conceptual: regional policy should be assessed not by how much income it redistributes into a territory, but by whether it changes the territory's capacity to generate income after the transfer diminishes, because only that transition transforms cohesion spending from compensation for divergence into an instrument capable of changing the productive map that produced the divergence in the first place.

---

Copyright of debuglies.com - Even partial reproduction of the contents is not permitted without prior authorization Reproduction reserved