Municipal waste generation and domestic extraction in the European Union: Does women's political representation matter? Evidence from a KRLS machine learning approach

This study examines the nonlinear and heterogeneous association between municipal waste generation (MWG) and domestic extraction (DE) across 27 European Union economies over 2012–2024, with particular attention to women's political representation (WSP). Using Kernel-Based Regularized Least Squares (KRLS), the analysis estimates observation-specific marginal associations without imposing a predetermined functional form and examines whether WSP conditions the MWG–DE relationship. The baseline results indicate a negative average association between MWG and DE, alongside substantial heterogeneity across country-year observations. However, the estimated MWG association becomes statistically insignificant under the per-capita specification, while excluding economic reliance on domestic extraction (DE/DMC) strengthens the negative association. These findings suggest that the magnitude and statistical significance of the MWG–DE relationship depend partly on model specification. WSP does not exhibit a stable independent association with DE across specifications, although the cross-partial results indicate that it conditions the MWG–DE relationship, with substantial variation in the direction and magnitude of this moderation. GDP shows the most consistent positive association with DE across the examined specifications. The positive DE/DMC relationship requires caution because DE enters directly into the ratio. Two-way fixed effects, panel-corrected standard errors (PCSE), LASSO, System GMM, out-of-sample validation, permutation placebo tests, and country-level sensitivity analyses provide complementary evidence, although the robustness results do not establish a uniformly stable MWG association or resolve potential endogeneity. Overall, the findings characterize the MWG–DE relationship as heterogeneous and context dependent, highlighting the relevance of economic structure, resource dependence, and institutional conditions. The results support differentiated circular-economy strategies while remaining explicitly associational rather than causal.