The RATS Software Forum

Dear Tom:
I have used the median target method by Fry and Pagan (2011). However, the point estimate of the impulse response are frequently outside of the confidence bands. Attached is the code and data, would you mind take a look at it and tell me how to address the issue? Thanks a lot!

There's no reason that can't happen. The Fry-Pagan response is the one that comes closest to the median response (in mean square) over the full range of the responses both across horizons and across variables. Your first series has a scale an order of magnitude higher than the others, so the minimizer will be dominated by that---note that it's almost exactly where you would expect it. The example at http://www.estima.com/forum/viewtopic.php?f=8&t=1941 rescales the gaps to take into account the range, but it still has some outside the one standard error bands.