I am currently working on a DSGE model and I am facing two problems:
I’m not sure whether my timing is correct. I thought the convention is that in the current period households choose K, which is thought of as next periods capital stock, and take K(-1) as given, whereas firms choose K(-1) today. However, if I model it like this, the shocks seem to enter the model with a delay such that my IRFs look a bit funny. But if I let the firm choose K (instead K(-1)) as well, the IRFs look fine and the problem seems to be solved. So is this simply the way to do it? (–> See model_smoothIRFs.mod + outputhelperfiscal_smooth.m & steadystatehelperfiscal_smooth.m for the steady state computation)
Also, I am having troubles calibrating the parameters to match some empirical facts. I do not want to estimate the model on data using Bayesian estimation techniques or something like that, I simply want to choose the parameters such that my model satisfies a few empirical facts.
In more detail, I have two sectors (official and underground) in latter of which agents and firms do not have to pay taxes. Now I want to match some targets, e.g. that the underground sector is about 27% of total production (official + underground production). The way I approached so far is I declared some parameters as endogenous variables and introduced extra equations in the model and steady_state_model sections specifying my targets (e.g. “Yu/Y = 0.27” for the 27% underground ratio). Unfortunately, though, it always returns complex steady states now, even if I don’t specify the actual empirical targets but the exact values computed in steady state when I don’t endogenize the parameters… (–> see model_calibration.mod + outputhelperfiscal_calibration.m & steadystatehelperfiscal_calibration.m for the steady state computation)
Any form of help would be much appreciated!
Thanks a lot in advance!