Divorce Dynamics and Attitudes Toward Divorce in Southern Europe: A Comparative Analysis of Greece, Italy, Spain, and France

Introduction:
Divorce has risen across Europe since the 1970s, yet Southern European countries have long exhibited lower levels than many Western settings. This study compares divortiality patterns and attitudes toward divorce in Greece, Italy, Spain, and France, examining how institutional context, cohort change, and values interact. France is included not as a normative benchmark nor as a proxy for “Western Europe,” but as a theoretically relevant contrast case with a distinct historical and institutional trajectory.

Methods:
We combine (i) official divorce and marriage statistics by marriage duration from the European Divorce Observatory (EDO), the Hellenic Statistical Authority (ELSTAT), Eurostat, the Instituto Nacional de Estadística – Spain (INE), the Institut National de la Statistique et des Études Économiques – France (INSEE), and the Istituto Nazionale di Statistica – Italy (ISTAT) to compute the Total Divorce Rate (TDR) and median duration at divorce (France 1952–2016; Italy 1971–2022; Greece 1972–2023; Spain 1981–2022); and (ii) individual-level data from the European Values Study 2017, accessed via the Longitudinal and Demographic Survey (LADS, 2018) (pooled N=8,494). Unconditional acceptance is defined as reporting that divorce “always justifies” (10/10). Analyses include descriptive statistics, chi-square tests, binary logistic regression, and tree-based machine learning models (Random Forest, XGBoost, LightGBM, CatBoost) with feature-importance and SHAP diagnostics.

Results:
Since 2010, divorce affected about 46% of marriages in France and 49% in Spain, compared with 27% in Italy and 28% in Greece. Unconditional acceptance is highest in France (36.5%) and Spain (32.9%) and lower in Greece (25.5%) and Italy (23.4%). Acceptance declines with age (≈1.1% lower odds per year) and is lower among men, the religious, and the married or widowed, while higher among those viewing marriage as outdated. Random Forest achieved high recall (0.87), and both RF and XGBoost identify age and religiosity as dominant predictors.

Conclusions:
Divorce behaviour and attitudes are shaped by cohort dynamics and institutional change, supporting a reciprocal values–behaviour interpretation of family change across Southern Europe and the French contrast case.