Large values problem

Hi,

1.I can run this model but the results do not sound right, I would be happy if you check it and tell me if there is a problem.
(Some figures do not look standard and also and it seems that there is something wrong with VARIANCE DECOMPOSITION SIMULATING ONE SHOCK AT A TIME )

2.When analyse results, which one is more important, the result of correlation between variables or policy and transition functions? (In case of discovering the relationship between two variables)

3.can I omit the “SP” equation and still examine its relationship with other variables or it would be invalid?(SP is a crucial variable)

aliye.mod (3.1 KB)

  1. What are the issues you spot?
  2. Correlations simply reflect the policy functions, but are often easier to interpret.
  3. If it’s crucial, you can typically not drop it. However, in your model SP seems to be exogenous, so it’s your choice if you want to set it to 0 in all periods.

Thank you,
for example look at M figure, it seems that it has equilibrium problem.


a

Why do you think that?

Should not M get back to zero when the shock ends?

You cannot judge whether a variable is mean-reverting after just 40 periods.

Thank you.

How can I expand periods in Dynare?

Set e.g. periods=500.

Thank you.

but I meant for figures.

Sorry, my mistake. I meant
irf=500.

Thanks,
so if after for example 500 or 1000 periods my figures do not go back to zero it means my model has a problem?

Not necessarily. Some models have variables that feature unit roots. An example is the price level in the New Keynesian model. But if you expect your model to be reverting to the steady state, then having a variable not return after a thousand periods or so is a problem.

Thank you.