problem of simulation with permanent shock

[threetom]

Hello all!

When I use the code in Chapter 3 of the User Guide, which is a deterministic case with permanent shock (as in page 26 of the User Guide), I find an interesting thing: when I change the time notation of the .mod file (for example, in the code, I change equation (a) and (b) to (a)-1 and (b)-1), then the simulation results are not the same. For example, the two paths of consumption are:

c c2
0.707986 0.707986
0.708271 0.711416
0.708615 0.711602
0.708941 0.711778
0.709251 0.711945
0.709545 0.712103
0.709825 0.712254
0.71009 0.712397

So, I am confused, with the same model, but just with different time notation of the same equation, we may get different result. I also tried another code, the difference is much bigger than this one.

Why this may happen? Which one of the result is more reliable? What can I do to handle this problem?

Would anyone mind to help me to handle these problems?

Thank you, very much!

Threetom

[MichelJuillard]

The Dynare timing convention is there for Dynare to understand the user's model. If you change timing in an equation, in most cases, you change the model. Don't forget that Dynare puts implicit expectation conditional on period t in front of each equation member. So a variable in period (t+1) doesn't have the same meaning as a variable in period (t)

Best

Michel