Page 1 of 1

Scaling impulse responses and changing confidence intervals

Posted: Thu Jul 11, 2019 12:15 pm
by heer0
Hi guys,

I am using EViews 10 (Univ. ed.) and I estimated impulse response functions from standarized data in different VAR and FAVAR models, which all include three different (non-standardized) monetary policy variables.

I next want to manipulate the impulse responses in two different ways:

1) Scaling shocks (e.g., to the central bank's interest rate) such that they represent a one-period 25 bps increase

Am I right in assuming that in order to scale the response to policy variable y_i, I would need to multiply the response values and the associated confidence intervals by a factor equal to 0.25/stdv(y_i)?

2) Changing the coverage of the confidence intervals

Using the command

Code: Select all

, I derived impulse response functions along with the relevant standard errors. Am I correct in assuming that I would just need to multiply the retrieved standard errors by a factor equal to 1.96 in order to retrieve the upper (and by implication lower) bands of a 95% confidence interval?

Your help will be much appreciated. Thanks in advance!