Revisiting loop over parameters

Steady state
1

Dear all,
I am struggling with implementing loop over parameters in the new version of Dynare (I’m running 5.1).
I have written a plain vanilla RBC .mod file, which by hand runs for a range of a certain parameter.
When I try to write an .m file looping over this same range, I cannot get any results.
I hope there is an easy fix for an issue that I’m overlooking as I’m still getting used to the new syntax.
Thank you!
RBC266.mod (2.5 KB)
RBCParameterLoop.m (781 Bytes)

2

I am getting an error about steady state computations. You may need to analytically compute the steady state.

3

Of course. Here are the two updated files. By itself, the RBC266Analytical.mod file runs for the lower and upper ranges of betta (0.8 and 0.98) over which I am trying to loop, but I cannot get the loop file to run.
Thank you!
RBC266Analytical.mod (2.9 KB)
RBCParameterLoop.m (791 Bytes)

4

Use steady_state_model for the analytical steady state:
RBC266Analytical.mod (2.9 KB)

5

Thank you so much, the loop runs now!
For future reference, is it true that the only way to be able to run a loop over parameters is to have an analytical model solution in a _steadystate.m file? The reason I’m asking is that I am working on a much larger model where the only way to get a solution is to use fsolve on a subsystem of the model equations, which I cannot possibly solve by hand. It seems that I cannot call on fsolve from within _steadystate.m file.

6
  1. No, loops do not require analytical steady states, but loops often requires updating the initial values, which you did not do. Hence the problem.
  2. You can use fsolve in a _steadystate.m-file. That’s the main purpose of those files.
7

Got it, thank you for taking time to clarify the issue. For the sake of conversation completeness, and for anyone else reading it, here’s the other option (without _steadystate.m). In my .mod file, after the model and shocks blocks, I used

steady_state_model
analytical expressions for each endogenous variable
end

Then I was able to use an external .m file with a loop over parameters as attached above.
P.S. the reason for the hassle is that I am trying to teach a basic RBC model and its solution in Dynare in an undergraduate course, and the fewer files I need to explain, the better (I think).