[Gretl-users] Re: Identification of SVARs using heteroskedasticity

Thanks for the information. I will take a look at that, though I do not use R. You are in principle right, but according to my understanding of the literature in macro, it is by much more probable to have breaks in the variance - covariance matrix, which suggest that this way to identify shocks is quite natural, assuming structural parameters are stable (if not, Bacchiocchi and Fanelli op. cit.). Robustness is an issue, in many more cases in my opinion. In fact, a finding should be considered robust if different methods yield similar results, qualitatively at least. In any case, the aforementioned identification method should have several applications in finance. AB & SVECM: fiscal policy SVARs imposing deficit stationarity for example, though I am sure other people would find other uses. This way IRFs cannot diverge to unsustainable deficits. Thanks again and goodnight, Andreas