[Gretl-devel] --no-fd-corr in system

Sven Schreiber svetosch at gmx.net
Tue Jul 7 12:34:10 EDT 2015


Am 06.07.2015 um 15:26 schrieb fdiiorio at unina.it:

> I noticed some problem using --no-df-corr in system using gretl 10.10.1
> (built date 2015-04-04)

Hold on a second...

>
> estimate "Kl1" method=ols
...
>               coefficient   std. error   t-ratio    p-value
>    ---------------------------------------------------------
>    const      16.2366       1.30270      12.46     5.62e-010 ***
>    P           0.192934     0.0912102     2.115    0.0495    **
...
> but the std. error reported above refer to an estimation with NO DF
> correction
> (see Calzolari, 2012, MPRA Paper No. 64415, pag.34,
> http://mpra.ub.uni-muenchen.de/64415/)

Are you sure? (I haven't checked the paper.)

What I see in this case --where gretl is supposed to apply a 
d.o.f.-correction as described in the guide -- is a bigger standard 
error (e.g. 1.30) than below without a correction:

>
> estimate "Kl1" method=ols --no-df-corr

>               coefficient   std. error   t-ratio    p-value
>    ---------------------------------------------------------
>    const      16.2366       1.17208      13.85     1.09e-010 ***
>    P           0.192934     0.0820650     2.351    0.0310    **

But this is right way, no? T-K instead of T in the denominator means 
higher variance estimates.

So I think either there is some misunderstanding, or perhaps the paper 
is wrong (in this detail).

cheers,
sven




More information about the Gretl-devel mailing list