Dear all,

I am trying to obtain rolling "Bayesian" coefficients using a panel dataset with 50-quarter timespan for 1700 firms. In other words, I want to derive firm-specific betas that are updated every quarter by regressing depvar on indepvar,

Code:
xtset firmid quarter
I did get simple rolling regression coefficients by running the following (Please correct me if this is incorrect):
Code:
rolling _b, window(10) recursive saving(coefficients1, replace): reg depvar indepvar
What I actually want to run is something like this:
Code:
rolling _b, window(10) recursive saving(coefficients2, replace): bayes: regress depvar indepvar
However, it seems that bayes is not supposed to be used in a "rolling" command, since I see the following error:

Code:
bayes may not be used in this context
Therefore, I would like to know if there is a way to do the Bayesian rolling estimation. Any suggestions are welcomed. Thank you in advance!