About This Event
Generalized linear models have been mainstream for quite a few years. However, the reflection of credibility in such models is not necessarily a settled issue. Hierarchical models (also known as fixed and random effects models, or linear mixed models) are an established technique to address this. Under certain conditions, they will generate parameter estimates that are identical to the Bühlmann-Straub credibility estimates that have been part of actuarial practice for decades.