<p>Box plot of severe housing deprivation rate, %.</p>

Pearson’s correlation coefficient was used to find linear relationships between pairs of indicators, while a principal component analysis and a cluster analysis were implemented to examine multidimensional structures and patterns. The multivariate analysis was conducted across three time spans: 2015–2019, 2020 (analysed separately as the Covid-19 pandemic disrupted economies) and 2021–2023. The results reveal a general decline in poverty; nevertheless, they also emphasise persistent regional disparities across EU Member States.

Bulgaria, Greece, Romania, and Latvia form the most vulnerable cluster in the final period, whereas several countries in Northern and Western Europe show more stable trends. The results further indicate that multidimensional poverty profiles cannot be explained solely by economic performance or geographical location. In conclusion, the cluster-based approach provides a dynamic perspective on the diverse experiences of poverty in the EU.

The results show the need for tailored policy responses that reflect the specificities of individual EU countries and clusters to promote sustainable and inclusive development and to strengthen resilience to potential future crises.</p></div>