I am estimating a structural VAR where all my variables are I(1). I took the log differences of each variable and generated the impulse response functions. Is there a way to convert the impulse response functions back to levels?
Asked
Active
Viewed 26 times