Scaled IRF add-in and standard errors
Posted: Wed Jan 25, 2023 10:45 am
Hello everyone,
I was using the Scaled IRF add-in to calculate impulse response functions from a SVAR (non-recursive), I noticed that when I get the standard errors directly from the Eviews command window (Asymptotic), it doesn't match with the standard errors that the add-in gives me (I set a value of 2 in the box "multiplication of standard error"). The response to the shock does match exactly, but the standard errors do not.
Does anyone know if the scaled IRF add-in does a different procedure to calculate standard errors? or is there an error? Which of the two standard errors should I trust?
Thanks in advance
I was using the Scaled IRF add-in to calculate impulse response functions from a SVAR (non-recursive), I noticed that when I get the standard errors directly from the Eviews command window (Asymptotic), it doesn't match with the standard errors that the add-in gives me (I set a value of 2 in the box "multiplication of standard error"). The response to the shock does match exactly, but the standard errors do not.
Does anyone know if the scaled IRF add-in does a different procedure to calculate standard errors? or is there an error? Which of the two standard errors should I trust?
Thanks in advance