Hi everyone, I am solving the model running perfect foresight, and I want to figure out the reduced form and see the state variables in order to specify their initial values in histval block and create a scenario with the model. How can I do that technically?
thank you! I am attaching the code below.
endo_nonlinear_loss_LR.mod (58.5 KB)
The model_info command can be used to display the state variables.
Thank you Mr. Pfeifer
I tried using model_info and also the command
states = M_.endo_names(M_.state_var)
to identify the state variables. However, the output also includes auxiliary and multiplier variables, and I am not sure how to treat them or what values should be assigned to them. I also ran a stochastic simulation of the model and obtained the policy and transition functions. Since the rows of these matrices correspond to state variables, I assumed this might be another way to identify them—is this correct?
Apart from identifying the state variables, I am also unsure how many lagged values need to be provided. When I run the command M_.maximum_lag, it indicates that the maximum lag in the model is 1. Does this mean that I only need to set values for period (0) for each state variable, or should I also provide values for period (-1)?
My third question concerns the homotopy issue. When I provide values for too many state variables, Dynare switches to a homotopy procedure. However, if I do not provide these values, the simulations become unreasonable. What would be the recommended approach in this case to avoid triggering homotopy while still obtaining meaningful simulations?
And also one thing, If I have histval block, should I replace steady_state_model block with Initval block since initval provides terminal values in this case?