Piecewise function not holding when estimating IRF

Stochastic simulations
1

Hello there,

I’m estimating a model with a government policy rule in which the government intervenes with a wage subsidy.

The rule states the government will spend Gt = max(0, C_bar - Ct). However, when I estimate the impulse response function, it says that the government responds to a positive economic shock with negative spending. I am trying to get it such that the government spendings nothing in response to a positive shock and only responds in response to a negative shock.

I’ve attached my .mod file and an image of the IRFs that are confusing.

Any help would be much appreciated.

Kind Regards,
Henry
IRF.fig (40.9 KB)
wagesubsidy.mod (2.8 KB)
wagesubsidysimul.m (1.6 KB)

2

You cannot use perturbation solutions for non-differentiable functions. You could try to use OccBin.