D-vine Copula Quantile Regression for a Multidimensional Water Expenditures Analysis: Social and Regional Impacts

Water expenditures variable, as the primary indicator of water consumption, is primordial to analysing and designing water policies. Numerous studies have analyzed water expenditures by considering various factors, including social, spatial, and climate variables predominantly relying on modeling tools often perceived as “black boxes”. In this study, we adopt a different approach, employing D-vine copula quantile regression to scrutinize water expenditures. This method has been proven to be efficient in predicting quantiles for different areas, especially when the normality assumption is inappropriate. Indeed, vine copulas offer the flexibility to select different marginal distributions and a variety of dependence structures. An illustration of the proposed methodology is applied to water consumption in Morocco. The results show different relationships between water expenditures and a set of determinant factors (geographical Area, Sex, and Educational level of the householder), they also show the quantiles’ variability at the regional scale.