search

mboeck11 / BGVAR

Toolbox for the estimation of Bayesian Global Vector Autoregressions in R.
27 stars 20 forks source link

Unit and scaling in irf #15

Closed 88312 closed 2 months ago

88312 commented 1 year ago

Hi, I'm wondering if the unit in the irf function could be clarified. The tutorial (vignettes/examples.html) says it adjusts the scale of an irf by "shockinfo_chol$scale<--100" "for a negative 100 bp shock applied to US short term interest rates:" However, the stir data in eerData, which is used in the tutorial, are not in bp. Also, the s.e. of the short term interest rates does not happen to be 1 bp. So, it seems that the scaling is not correct. The responses of the variables in "Responses of US country model" figure looks indeed too big for a 100 bp shock.

I guess the unit of a shock in the irf function is the same as the data. Is this correct?

mboeck11 commented 2 months ago

Thank you for pointing this out - 100 basis points correspond here to -1 not -100! We've corrected this in the new version.