Spatial Divides of Wealth Inequality and Politics in Switzerland: A Kernel and Optimal Transport Approach

This article investigates how economic, political, spatial, and linguistic factors jointly structure the political geography of Switzerland. Leveraging detailed municipal-level income tax data, we reconstruct wealth distributions and compute different inter-municipality economic distances from Swiss Federal Tax data. These, declined into simple mean and optimal transport distances (Wasserstein), are compared to political distances derived from 381 national referenda (1971--2024). Using a kernel-based approach, multidimensional scaling (MDS), spatial autocorrelation index and multiple regression, we explore the alignment between economic inequality, political behavior, and spatial-linguistic structures. Our results reveal that optimal transport income disparities explain political divides better than do the average wealth, while linguistic and urban--rural divisions remain largely dominant structuring forces of political explanation. This study highlights the potential of MDS visualization and the kernel associations in analyzing spatially embedded political behavior.