Bayesian IRF Matching - Inflation vs Price Index


Thank you so much for integrating BIRF Matching into Dynare. One of the IRFs I would like to match is CPI. The model obviously does not have CPI but it has inflation. Normally, this is not an issue because I would simply cumsum(inflation) to get IRF for CPI. I was wondering whether Dynare IRF matching has a similar feature, where we can modify the IRF variable.


Unless there is a trend, you can simply define the CPI via its link to inflation. Dynare should then take care of the rest.

Thanks so much for your swift response. Could you be more specific? The model has a trend but the shock does not influence it. What I would like to do is to match data IRF of CPI to model IRF of cumsum(inflation). How can I do that?


In your model, you can define


and the simply use the IRF of P. That only does not work if you have trend inflation.

Sorry: CPI is log(P) itself. In the model I have Pi = P/P(-1).

Impulse response to log(P) is equivalent to impulse response to cumsum(log(Pi)). I just want to make a modification after the model generates IRFs to match the data.

Then define