Joint models of multivariate longitudinal outcomes and discrete survival data with INLA: An application to credit repayment behaviour

Survival models with time-varying covariates (TVCs) are widely used in the literature on credit risk prediction. However, when these endogenous, inclusion procedure has been limited to practices such as lagging variables or treating them exogenous. That leads possible biased estimators (depending strength of exogeneity assumption) and a lack prediction framework that consolidates joint evolution survival process endogenous TVCs. The use is suitable approach for handling endogeneity, however, it comes at high computational cost. We propose model bivariate TVCs discrete data using integrated nested Laplace approximation (INLA). illustrate implementation via simulations build full-prepayment consumer loans. also methodology individual method more accurate approximations than comparable approaches. evidence superiority over traditional an out-of-sample out-of-time analysis.