Page 1 of 1

A general question: more complicated, worse?

PostPosted: Fri Jul 24, 2015 11:41 pm
by superztt
When I have more variables and more conditions characterized by equations, it's harder to get steady state and satisfy the Blanchard Kahn condition. The model running in dynare becomes very delicate. For example, real price or nominal price, with adjustment cost or not.

Is my feeling right or there've got to be something wrong when I make the model more complicated and it's nothing to do with the number of variables or equations?

Re: A general question: more complicated, worse?

PostPosted: Sun Jul 26, 2015 2:18 pm
by jpfeifer
Steady state finding becomes harder as you need to provide more good starting values (except for linearized models of course). Similarly for Blanchard-Kahn, the more equations you enter the more scope for mistakes and your part. Complexity is not a problem for the computer but for the human user. Thus, it is always recommended to start with a working core model and then subsequently add features.