The study’s central innovation is methodological. Rather than asking whether a single indicator such as unemployment or GDP predicts populist voting — the approach that has dominated the so-called geography of discontent literature — Refisch used k-means cluster analysis to group regions according to eleven socioeconomic indicators simultaneously. These included GDP per capita and its long-term growth since 1993, employment change since 1995, the share and change of industrial jobs, tertiary education levels, net migration between 2011 and 2021, the proportion of residents aged 65 and older, and population density. All indicators were standardized within each country, so a region’s score reflects its position relative to the national average rather than absolute European rankings. The algorithm then partitioned the regions into six statistically robust types, each representing a distinct configuration of economic, demographic, and social conditions.

The six region types read like a taxonomy of contemporary Europe. Cluster 1, the capitals and high performers, contains the most prosperous regions in each country, with employment growth of 36 percent between 1995 and 2021, the highest GDP growth rates, and an influx of young, educated migrants. Cluster 2 comprises deindustrialized metropolises — densely populated, affluent, and highly educated urban centers such as Paris, Barcelona, Cologne, and Katowice that nonetheless experienced the largest decline in industrial employment of any cluster, nearly 10 percent. Cluster 3, labeled modern industry, is the manufacturing heartland: these regions hold the highest industrial job share at 30 percent of employment and match the capitals in GDP growth, spanning southern Germany, the Warsaw region, southern and eastern Czechia, western Romania, and much of Lombardy.

The remaining three clusters capture Europe’s more troubled geographies. Cluster 4, gaining ground, includes regions on an upward trajectory, with 28 percent job growth since 1995, rising education levels, and positive migration balances — coastal France, southeastern Spain, western Lithuania, and parts of the Netherlands and Sweden. Cluster 5, aging, emigration, and job loss, is the only cluster that saw an absolute decline in employment over the study period; these are the most rural and sparsely populated regions, where more than a quarter of residents are over 65 and the young and skilled keep leaving. Northwestern Spain, rural eastern Germany, western Pomerania, and central France fall here. Cluster 6, weak economy, contains regions still struggling after industrial decline, with the lowest GDP per capita and the weakest growth since 1993 — Sicily, eastern Poland, much of Estonia, and central and eastern Slovakia among them.

Against this regional backdrop, Refisch combined the most recent national parliamentary election results available up to March 2025, weighting each country equally so that no state’s many regions could dominate the picture. The results are striking in their systematicity. Radical-right parties achieve their strongest support in Cluster 5, the structurally disadvantaged regions, where their average vote share runs 0.91 standard deviations above the overall mean. But they also overperform in Cluster 3, the growing industrial regions, and in Cluster 6, the weak-economy regions, while underperforming in the urban clusters 1 and 2. Green and radical-left parties show nearly the mirror image: their strongest support comes in the deindustrialized metropolises of Cluster 2, with above-average results in the capitals, and their weakest in industrial and weak-economy regions.

Perhaps the most dramatic single finding concerns the social democrats. Once the standard-bearers of the industrial working class, social-democratic parties now register their lowest vote shares precisely in the modern industrial regions of Cluster 3 — a striking 1.77 standard deviations below their mean, the most pronounced effect observed for any party family in the study. Their strongest support, by contrast, comes in the disadvantaged rural regions of Cluster 6 and Cluster 5. Conservative parties occupy the opposite pole: they peak in the industrial regions and suffer their worst results in the progressive, deindustrialized metropolises. Liberal parties show a clear urban focus, peaking in the capitals and weakest in the aging, emigrating regions, while agrarian parties — present in only six countries — predictably perform best in the rural clusters.

The radical-right success in prosperous industrial regions poses a genuine puzzle for the geography of discontent thesis, which interprets anti-system voting as revenge for deindustrialization and regional decline. Cluster 3 regions are not left behind; they are growing, competitive, and integrated into global value chains. Refisch offers two complementary interpretations. The first is compositional: these regions have relatively few highly educated residents and many manufacturing workers, groups that individual-level studies consistently show are more likely to support right-wing populist parties. The second involves status anxiety rather than actual decline. Previous research has found that many radical-right voters have not experienced downward mobility themselves but perceive their status as threatened. Global competition puts sustained pressure on European industry, and workers in even successful industrial regions may fear future loss — a dynamic that transforms economic optimism on paper into electoral anxiety at the ballot box.

The broader implication is that spatial differentiation is not a radical-right specialty. Every party family in the study exhibited pronounced regional over- or underperformance across the six region types, which supports the argument, advanced by political scientists Rod Ford and Will Jennings, that place deserves recognition as a political cleavage in its own right. In the classic cleavage theory of Seymour Martin Lipset and Stein Rokkan, territorial divides such as center–periphery conflicts structured entire party systems; the new findings suggest such structuring is alive and well, but operating through contemporary configurations of deindustrialization, demographic aging, educational sorting, and selective migration. The continued outflow of educated, cosmopolitan voters from rural and peripheral regions reinforces these patterns, deepening the divergence between metropolitan and non-metropolitan electorates with each electoral cycle.

The study is not without limitations, which the author acknowledges candidly. Comparable data on unemployment and poverty were unavailable for all countries, omitting an important dimension of regional disadvantage, and public service provision — increasingly recognized as politically salient — could not be included. Some successful parties with hybrid profiles, such as ANO 2011 in Czechia, the Dawn of Nemunas in Lithuania, and the Five Star Movement in Italy, were assigned to a no-family category. Sensitivity analyses confirmed that the main patterns hold across Europe, though results for smaller party families are more sensitive to the exclusion of individual countries, and the Christian-democratic results are disproportionately shaped by Germany, where the Bavaria-based CSU dominates regions classified as modern industry.

Still, the methodological lesson is clear: electoral behavior responds to multidimensional regional configurations, not to isolated variables, and typologies that capture those configurations reveal spatial structures that conventional regressions can miss. Refisch also points to a normative shift the findings invite. Rather than asking only under what conditions voters reject democratic norms, researchers and policymakers should equally investigate which regional conditions foster support for pro-democratic parties — a question that becomes urgent as Europe’s internal divides continue to widen and as the geography of party competition increasingly maps onto the geography of opportunity itself.

Subject of Research: The relationship between regional socioeconomic development types and electoral support for party families across Europe

Article Title: Electoral outcomes of party families across regional development types in Europe

Article References: Refisch, M. (2026). Electoral outcomes of party families across regional development types in Europe. Geographica Helvetica, 81(3), 435-452. https://doi.org/10.5194/gh-81-435-2026

Image Credits: AI Generated

DOI: 10.5194/gh-81-435-2026

Keywords: electoral geography, regional inequality, radical-right parties, k-means cluster analysis, deindustrialization, party families, geography of discontent, Europe, voting behavior, regional typology, social democracy, urban-rural divide