Hello everyone.
I’m trying to extend a New Keynesian model with hand-to-mouth consumers for monetary/fiscal mix analysis. One of the assumption of the model is different taxation for the two types of consumers, which creates some issues for model determinacy. Firstly, I used two fiscal rules allowing for different tax responses and the model is correctly closed, generating IRF in monetary regime. Switching to the fiscal regime (allowing for zero responses of taxes to government spending and deficit) two fiscal rules of this type suffer of collinearity issue and so rank failure. Using one fiscal rule for aggregate taxes may solve the problem but I don’t know how to close the model and correctly determine taxes for the two types of consumers.
Thank you for the attention and time.
The mod file is attached here
tesigali3.mod (4.5 KB)
tesipar.m (1.7 KB)
Why are there steady_state operators everywhere? That may be problematic and may cause collinearities.
Sorry I didn’t specify that I made a try with a log linearized government budget constraint, so it implies a normalization for the debt-to-gdp ratio in steady state. My problem is that with that specification the model works, the nominal debt decreases for the increase in inflation and the real debt to gdp ratio goes up; instead with a linear GBC there is rank failure.
I don’t understand. The only difference between linearization and log-linearization is a division by the steady state of the variable. Correctly done, that should not change anything.
Thank you professor.
Why instead the non-linear version of that budget constraint is not able to correctly account for FTPL dynamics?
There must be an implementation error.
I’m trying to catch it. Given what you have seen from the code, could it be a problem on household euler equation or on the Phillips Curve?
The parameterization I used is standard for a fiscal-led regime
I still don’t entirely understand the problem. What works in your model and what does not? Or which exact modification is problematic?
Sorry for that.
The problem is the simulation of a news shock in government spending in a fiscal-led regime: including a non linear budget constraint the model suffers of rank failure when I don’t allow tax responses to debt in the fiscal rule, while in the monetary-led regime the problem disappear. Turning to the linearized version of the GBC, the model works well in the fiscal regime, instead the responses in the monetary one are quite weird, for example taxes move at the impact of the news shock, something which is not likely to occur or in general with the other specification of the GBC it does not happen.
But again, if you enter the same GBC once in its linearized and once in its non-linear form, the results must be identical at first order. If that is not the case, there must be a mistake in at least one of them.