Forecasting based on random intercepts models requires imputation of the individual permanent effects to the simulated individuals. When these individuals enter the simulation with a history of past outcomes this involves sampling from conditional distributions, which might be unfeasible. I...