Regarding the Cholesky decomposition and bootstrap

[mandarpriya]

Sir,
I really appreciate for the library as it has helped me a lot in understanding the GK 2015 paper. I really like the closeness of the code to the GK paper, which has been very useful as for me the matlab is very difficult. i wished to know about how to use the bootstrapping and Cholesky decompostion with bootstrapping as an extension to the code. In this regard i was going through the svars package which does id.chol and wild.boot for cholesky and boostrapping. So can they be used with code.

Thanks & Regards
Mandar

[angusmoore]

Hi Mandar, sorry for my very slow reply, have been on parental leave for the past 6 months and not keeping up to date with github.

Short answer is: no, the bootstrapping in svars won't work with the code here, as that package (from a quick skim of the code) expects the results from the identification to be in a particular format, which this code won't align to. It may well be possible to rejig it in to a format that works with svars, but skimming the svars bootstrapping code it looks hard to do that.

[mandarpriya]

Thanks so Sir. I am sorry for the late response. I agree to what you say as i see that your approach is the for the frist model of Gertler-Karadi (Proxy SVAR model). I have used another way , by changing the approach and was able to use the svars library functionality for the same.