Hi,
I am working with a 2-country model with complete asset markets and full risk-sharing. Dynare tells me the model has 3 unit roots. However, when I plot IRFs they all return to zero within 40-50 periods (I have persistent shocks), and when I simulate the model for as much as 100,000 periods, all variables remain stationary. I am wondering how I can "detect" which variables give rise to the unit roots, so that I can eventually solve the problem. Any hints? Thanks in advance.