We contribute to the non-life experience ratemaking literature by introducing a computationally efficient approximation algorithm for the Bayesian premium in models with dynamic random effect, where the risk of a policyholder is governed by an individual process of unobserved heterogeneity....