Policy & Transition Function

Hello all,

I have a question relating to outputs produced by Dynare.

  • After running a .mod file, I see on the command window a table of policy transition function, from which we can derive the approximated solution for example:
    k = constant + ak(-1) + ba(-1) + c*e (1)
    with a, b, c are elements in the table

  • however, there is another output like matrices of coefficient of A and B in the policy & transition functions in the state-space form, like
    Y=As(-1) + Be (2)

What makes me confused is that, why in (1) we have constant term, and in (2) we dont have while both of them are policy/transition functions?

Thanks,

In the second case the variables are in deviations to the steady state, so there is no constant.

Best,
Stéphane.

Thank Stephane, I got it.

Regarding the coefficient matrices A & B in state-space form, I have one more question.

As I read, they store the coefficients in the “decision rule” order: static variables --> backward-looking variables --> forward-looking variables. That is all I can found.

But in case that my model have several backward-looking variables, then how can I see the order of coefficients between these backward-looking variables? I need to know the order of variables exactly to use these coefficients to do simulations.

Thank you,

If you cannot sort this out with the information provided in the reference manual here:

http://www.dynare.org/manual/Stochastic-solution-and-simulation.html#Typology-and-ordering-of-variables

you really should use the simult_ routine, as I aready suggested you here:

Best,
Stéphane.

1 Like

Oh I am so sorry I bookmarked the link of the manual to read on this, but then totally forget…
I got it now thanks to the manual: “The DR-order is the following: static variables appear first, then purely backward variables, then mixed variables, and finally purely forward variables. Inside each category, variables are arranged according to the declaration order.

Thank you,