Proc MCMC for hierarchical tweedie model

[dschmidt]

I'm looking to model some insurance claims, and based on a few quick skims of papers, the tweedie distributions seem like a good candidate - a poisson mixture of gammas. Is anyone aware of methods/studies/resources which do hierarchical modeling of tweedies? If it were another distribution, I would use proc mcmc, but I've read that there's no closed form for the distribution, so I can't use the "General" option to specify the distribution.

What I'm really interested in, though, is examining a sampling distribution. So I could consider a normal hierarchical model based on a fixed group size if there's no alternatives. But ideally I'd like to be able to generalize for different group sizes. 

Thanks!

[pink_poodle]

There is an example of modeling automobile insurance claims with Proc Hpgenselect here:

https://support.sas.com/rnd/app/stat/examples/hpgenselect/hpgenselect.htm