What is the effect of an exogenous variable with no innovations in a perfect foresight environment?

Perfect foresight simulations
1

Hi all,

I am trying to understand why these two .mod files give different results.

perfectforesight_3eq.mod, contains a simple NKM. I am imposing a path for inflation, the plots of the state variables are in line with what one would expect from the theory.

However perfectforesight_3eq_GovExp.mod gives completely different dynamics. I am imposing the exact same values for inflation and the models are the same, except that in the second .mod file I introduced government expenditure as an AR(1) shock. I did not impose any path for government expenditures innovations, therefore gov. expenditure is equal to 0 in every period. But for some reason, consumption decreases and slowly reaches the steady state, when in the previous .mod file it goes in the opposite way.

I understand that in a perfect foresight experiment in which agents know that gov. expenditure is zero, the result of both experiments should be the same (also considering that it is c and not y that enters the model equations).

Thank you in advance,
Mario

perfectforesight_3eq_GovExp.mod (848 Bytes)
perfectforesight_3eq.mod (757 Bytes)

2

rplot plots the model including the initial and terminal condition, if necessary. With a lagged endogenous variable, the initial condition will be present, resulting in a plot that is longer by one period. That’s the part you were worrying about. But after that initial (meaningless) period, the simulations are the same, as should be expected.

3

Hello jpfeifer,

Unfortunately that is not the issue I was referring to. When I run both .mod files, I get the following simulated paths. On the left is perfectforesight_3eq.mod, and on the right perfectforesight_3eq_GovExp.mod. As you can see, the evolution is not the same for any of the variables.

Thank you for your help.
Mario

image
image1359×730 46.8 KB

4

Are you using Dynare 5.3?

5

I am using Dynare 4.6.4

6

Please upgrade. I suspect there was a bug.

7

Hi,

You are right, with the new version it now works.

Thank you!
Mario