Erhanfadli Azrai schrieb: > Hello everyone. > > Can the command 'system' and 'estimate' be used for structural var > equation? or alternatively, could anyone point me to a manual or > write-up to do svar using gretl?
Unless there have been secret additions recently, SVARs are not implemented yet in gretl. To my knowledge there are no SVAR guides for gretl. > I've read somewhere that i can do this in JMulti, but i prefer to code > in gretl. Yes Jmulti is great for SVARs. With respect to coding this yourself in gretl: I'd say it's reasonable if you only have just-identified shocks. However, you're likely to want to do more, so I really would recommend Jmulti. What you could do is to port the SVAR code from Jmulti to gretl, that would actually be a great thing to do. The jmulti backend code is written in Gauss and is open source, you can browse it on sourceforge just like the gretl source, see jmulti.de for details. (The Gauss runtime engine is obviously not open source, and one of the problems is that it can happen that you get different results on Linux and Windows, it happened to me...) cheers, sven
