Joint modelling of dyadic and monadic count outcomes: an application to modelling forced migration flows

Abstract

The aim of this study is to explore the adoption of a joint modelling framework for dealing with dyadic and monadic count outcomes with excess zeros simultaneously via a common latent structure. As a case study, we consider the problem of identifying the different push and pull factors of cross-border forced migration and internal displacement. We consider a full panel data analysis and estimate a random effects joint hurdle model following the Bayesian paradigm; the resultant posterior is approximated through the integrated nested Laplace approximation.