Hello everbody,
i am using Dynare++ to solve a model and want to check for approximation errors. I know there is the build in accuracy check, but i want to do it by hand. Therefore I need to evaulate the policy functions for some sets of state variables. At the moment i am looping over my combinations and use the dynare_simul() function, where i adjust the starting values and set the shocks to zero. I only do 1 period of “simulation” and use the resulting values.
- Is there a way to avoid looping over all combinations without constructing the policy function by myself?
- Am I handling the shocks right? For example I use the productivity level z_t as starting value and set the shock to 0, in my understanding this is equivalent to setting z_t-1 as starting value and using a shock different from zero. (E.g. assume an AR Process for productivity.) I will check this by comparing the results, but i’m interested in the theoretical identity.
Thanks a lot,
Fabian