[Gretl-users] Re: Hansen's Superior Predictive Ability (SPA) test
Am 05.11.2019 um 01:50 schrieb Burak Korkusuz:
Hi,
Does anyone know is it possible to use the SPA test in gretl?
Diebold-Mariano (DM), Giacomini-White (GW) etc. tests are available,
but as far as I know, the SPA test (relatively new and more popular)
is not available. Is there any chance to apply that test with codes in
gretl?
Hi, that depends what you mean by "possible". There is no existing
implementation that I am aware of. But it certainly is possible to
program it in gretl/hansl.
But, speaking as a co-author of the FEP package, I quite like the model
confidence set (MCS) approach, also by PR Hansen (and others). We do not
have that, either, but what makes you think that SPA is more popular?
And BTW, I don't think it is more popular than DM, so I guess you mean
by comparison with GW - do you have evidence for that statement? (It's
an honest question.)
cheers
sven
Attachments:
- attachment.htm (text/html — 1.6 KB)