Re: [Gretl-users] Vintage Data and fcast command
Am 22.09.2015 um 16:22 schrieb Francesco Raffaele P.:
My model is a blue, it's a simple OLS regression like Y: X Z W
X Z W are standard data (latest aviable) but Y variable uses realtime data.
aha, ok.
So there's a way i can run fcast --rolling command with dependent
variable changing at every new step.
To be honest, I wasn't really aware of the "--rolling" option of fcast.
(Reading its doc ["fixed starting point"] I would say it's a bit of a
misnomer, because rolling is typically meant to be fixed sample-length,
while with a fixed starting point perhaps "sequential" is a more common
term. But that's a different issue.)
So, no, you cannot do that with fcast. Perhaps in the longer term "fcast
--rolling" should also get this additional feature to work on realtime
data, but it's not trivial in terms of the syntax.
As I indicated, I still think you have to write your own loop, and at
each step re-define your Y variable accordingly. Then do fcast without
any special option.
cheers,
sven