Question on: near-0 eigenvalue, and B-K condition

[LittleFish]

A quick question: In order to resolve a near-0 eigenvalue and to meet Blanchard-Kahn condition, should I change initial values, historical values, or parameters, - Or should I change some part of the model equations? Thanks.

[jpfeifer]

I'm afraid the answer to your question won't help you. If the model is wrong (or entered incorrectly), change the model equations.If the model is correct, but your parameters are in the indeterminacy/no solution region, change the parameters. Initial values have nothing to do with the problem if Dynare is able to find the (unique or correct) steady state.

[LittleFish]

I'm afraid the answer to your question won't help you. If the model is wrong (or entered incorrectly), change the model equations.If the model is correct, but your parameters are in the indeterminacy/no solution region, change the parameters. Initial values have nothing to do with the problem if Dynare is able to find the (unique or correct) steady state.

Hi jpfeifer! Thank you for your reply. Dynare is able to find my steady states. But if I include 'check' after the 'steady' command in my code, Dynare would say that 1 eigenvalue is near 0.

[jpfeifer]

For the BK conditions, 0 is not the issue, but 1. Zero is an issue if there is trouble with the rank condition.

[LittleFish]

For the BK conditions, 0 is not the issue, but 1. Zero is an issue if there is trouble with the rank condition.

Thanks, jpfeifer. My steady state is checked to be okay. But when I try to do stochastic simulation, Matlab says that "Blanchard Kahn conditions are not satisfied: no stable equilibrium". So what should I do with the Dynare code in order to satisfy the B-K condition? Thanks.

Little Fish @ Washington DC