Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Factor Augmented Vector Autoregression in Stata

    Hi, All:

    I am attempting to use the two-step approach to factor-augmented vector autoregression as outlined in the seminal paper by Bernanke and his colleagues.

    Thus far, I have used the factor command to as follows:

    factor (my var list), factors(3) pcf
    rotate
    predict factor1 factor 2 factor 3

    This works fine.

    Next, I have used an svar command using factors 1-3 and the effective funds rate (as my policy instrument). Again - I am all good.


    My questions is this:

    In the FAVAR paper, Bernanke and his colleagues are able to recover impulse response functions for variables used to create the factors. For example, if I have used variables named a,b,c,...x,y,z to create my factors (so I have 26 factor loadings within each of my three factors) and I am using SVAR with those 3 factors in my information set and using the FF rate as my policy instrument, how can I generate the impulse response functions were the FF rate is my impulse variable and then some subset of my variables, say q,r,s, are my response variables?

    It doesn't work to use the 'IRF create' function where I simply plug these variables in as response variables. I've tried this.


  • #2
    bump

    Comment


    • #3
      Anyone can help with this? I want to use FAVAR in Stata as well. Did anyone try it before? Thanks

      Comment

      Working...
      X