I try to estimate a new keynesian model with unit root inflation target. I wonder whether I have to stationarize my model w.r.t inflation target. As far as I know, I should, but I have already writtten a model with stationary inflation target, inflation target is AR(1) with coefficient smaller than 1. When I put the AR coefficient equal to 1, dynare still solves the model. Does this mean that I dont need to stationarize the model and can I estimate the model without stationarizing it?
See Adolfson et al. 2005 wp 179 of bank of Sweden. There they mention that they choose inflation target as AR(1), but they also try non stationary by setting coef to 1 and they say the data favours (with respect to marginal likelihood) an AR(1) process. See page 25 bottom last paragraph… there you’ll find it
Hi,
Thank you for your reply. Although, as you said, taking inflation target as AR(1) process is enough, I still want to see the implications of unit root inflation target. Therefore I wonder whether I have to stationarize my model w.r.t inflation target or not. As I said before, when I put the AR coefficient equal to 1, dynare still solves the model. Does this mean that I dont need to stationarize the model and can I estimate the model without stationarizing it?