Capturing and representing the multidimensionality of welfare through structural equation modeling and a goal-based composite indicator with multiple constraints

This study examines the concept of multidimensional welfare, which encompasses multiple aspects across various dimensions. The study introduces an approach that combines ranking normalization, structural equation modeling, and a new method for constructing composite indicators. This approach enables improved differentiation of welfare levels, confirming the multidimensional nature of welfare and representing it through a readily understandable unidimensional measure. This innovative approach fills a gap in methodologies by considering the interrelationships between dimensions, avoiding aggregating dimensions that carry little information into the composite indicator, ensuring the composite indicator's multidimensionality, and avoiding making it predominantly explained by a single dimension. Other advantages of this approach include a rigorous explanation of the conceptual framework of multidimensional welfare, avoiding the assignment of equal weights to the dimensions due to the lack of a clear and consistent weighting scheme, and providing transparency in the objective definition of dimension weights. The results indicate that government efforts to provide social services and protection are insufficient to improve welfare levels in the poorest municipalities. Governments should not allocate resources solely to social assistance and protection; instead, they should generate employment and income opportunities and promote digital inclusion, leisure, culture, and sports. In addition to contributing to the welfare literature and informing the formulation of more effective social policies, this study advances the composite indicators literature by offering an innovative weighting scheme that ensures conceptual compatibility and preserves the composite's multidimensionality.