Factor Augmented Auto-Regression for Stata, any tips?

[adriano donofrio]

Dearest,

I've been looking for a while for a Stata procedure to implement a Factor Augmented Auto-Regression (FAVAR) model. Any idea where to find it or any suggestion about how to implement this?

More generally, it seems to me that when it comes to dynamic factor modelling Stata has only the dfactor command which computes the base-version of this tecnique.

Hope you can help me,

Thank you

[adriano donofrio]

bump

[George Ford]

Can't find anything either.

Don't know much about it, but it looks like you use PCA to create factors and then use those factors in a VAR model and proceed as usual. That's not hard to code, unless there are some details to worry about.

It looks like there an intermediate step where you regress the slow moving factor on Y and use the residuals as the variable for the VAR. Also easy.

[George Ford]

I suspect you could follow this even if you don't know R:

https://jbduarte.com/blog/time%20series/r/favar/2020/04/24/FAVAR-Replication.html