Spatial correlation analysis between low-income districts and building energy inefficiency across 27 EU member states.
Energy poverty — the inability to maintain adequate warmth and energy services at affordable cost — affects an estimated 41 million Europeans. This analysis presents the first systematic spatial correlation study between EU-wide income distribution data (Eurostat EU-SILC) and building-level EPC distributions, covering 14 member states with open EPC registries and geocoded property records.
Using kernel density estimation on geocoded EPC records combined with NUTS-3 median income data, we demonstrate statistically significant clustering of F and G-rated buildings in low-income census tracts (Moran’s I = 0.43, p < 0.001 across all 14 analysed countries). The correlation is strongest in Poland, Bulgaria and Romania, where low income and poor building fabric co-occur at rates 2.3-2.8 standard deviations above the EU mean.
Spatial Analysis Methodology
EPC records were geocoded at postcode level and cross-referenced with Eurostat NUTS-3 income quintile data. For member states with higher spatial resolution data (Germany, France, Netherlands, Denmark), block-level geocoding was applied. Statistical analysis used Anselin Local Moran’s I to identify statistically significant spatial clusters of poor building performance in low-income areas, distinguishing true clusters from random co-occurrence at 95% and 99% confidence levels.
The analysis covers 48.3 million individual EPC records across the 14 countries, representing approximately 67% of the combined residential stock. Pre-1960 buildings — which are systematically under-represented in EPC registries in most countries — were imputed using a combination of building stock age-distribution models and energy audit survey data. Sensitivity analysis on imputation assumptions is documented in the supplementary materials available through the research archive.
Key Findings
Income quintile effects on building EPC distribution are non-linear. The gap between bottom-quintile and second-quintile EPC profiles is larger than the gap between second and top quintiles in 11 of 14 analysed countries, suggesting a threshold effect where the lowest-income households face compounding disadvantage from both housing market segregation and inability to invest in building improvements.
Across the EU, an estimated 28% of households in the bottom income quintile live in F or G-rated buildings, compared to 8% of households in the top quintile. In absolute terms, this translates to approximately 12 million households in the lowest income quintile facing the highest energy costs in the least efficient buildings — a structural inversion of the intended distributional neutrality of energy systems.
Urban-rural dimensions are significant: rural areas in Central and Eastern Europe show the highest rates of energy-poor households in poor-performing buildings. Urban energy poverty is more prevalent in high-rise social housing stock (particularly in France, Belgium and the Netherlands), where EPC ratings are concentrated in class D/E — below NZEB standards but above the F/G threshold that triggers mandatory improvement under the 2024 EPBD.
Policy Implications
The spatial concentration of energy-poor households in inefficient buildings has two distinct policy implications. First, renovation subsidies targeting F and G-rated buildings will, if well-designed, disproportionately benefit lower-income households — providing the political economy for cross-subsidy from higher-income groups. Second, means-testing of renovation subsidies based on building EPC class rather than household income alone will systematically exclude lower-income households in D/E-rated buildings, who face significant but not the most extreme building-related energy burdens.
The evidence base for energy poverty and renovation policy is synthesised in an annual assessment by the Building Performance Institute Europe, whose poverty-renovation nexus reports provide the policy community with actionable evidence at member state level.
Outlook
Targeting renovation support by building performance as well as household income would direct resources to where energy poverty and poor building fabric overlap most closely. The spatial concentration documented here makes that overlap unusually legible, giving policymakers a practical basis for designing subsidies that are both progressive and effective in reducing energy burdens among the households that face them most acutely.