Say production Y = exp(A)*K^alpha.
I know the standard timing in Dynare is for an exogenous shock is:
A = rho*A(-1) + shock
But I want to use:
A(+1) = rho*A + shock(+1)
such that I get the policy function in terms of A. Is this possible? This would really make my life a lot easier
(My actual model involves learning, where I use interpolation between policy functions derived from Dynare, which makes it fuzzy to use the first solution).
No, this is not possible. What you mean with
is that you actually want to consider A as exogenous and not A(-1) and shock. That is not a structure that Dynare allows. But maybe DSGE_mod/SGU_2004/SGU_2004.mod at master · JohannesPfeifer/DSGE_mod · GitHub
is helpful.
Thanks Johannes, I saw those SGU files, but rho=0 is the special case where the timing does not matter.
And yes, I meant a policy function where A_t is a state variable, so A is treated as exogenous (with A(+1) = rho*A + shock(+1)).
It’s a pity there is not even a trick to get that directly from Dynare. Of course you can back out that formulation from the Dynare policy functions, but in my case it gets ugly.
Sorry, but I don’t get it. Dynare’s timing is in a way that (-1) indicates a state variable. The reason is that at time 0, you cannot change something that has timing (-1). When you say A_t is a state variable, I don’t see what that is supposed to mean, unless it is purely exogenous.