Random effects are particularly useful in insurance studies, to capture residual heterogeneity or to induce cross-sectional and/or serial dependence, opening hence the door to many applications ...