[Gretl-users] long memory

[Riccardo (Jack) Lucchetti]

On Mon, December 4, 2006 10:33, javier garcia enriquez wrote:
> Hi;
>
>   In my present investigation we are interested in see the posibility of
> fractional integration in some time series. When Gretl do the peridogram, it
> also show the GPH and the Whittle estimation and, in brackets, appears a
> value we don´t know what means, because below just appears a test for
> fractional integration, when the null hypothesis is that the integration
> order is zero. Then... ¿what is that value? (In the following example we
> have accentuate in yellow)
>
>   Contraste de integración fraccional (Geweke, Porter-Hudak)
>     Orden de integración estimado = 0,605383 (0,146662)
>     Estadístico de contraste: t(9) = 4,12773, con valor p  0,0026
>
>   Estimador local de Whittle (T = 101, m = 14)
>     Orden de integración estimado = 0,591367 (0,133631)
>     Estadístico de contraste: z = 4,42538, con valor p  0,0000
>
>   Thanks
>   Javi

That's the asymptotic standard error: 0.605383/0.146662 = 4.12773 (with a bit
of rounding).


Riccardo (Jack) Lucchetti
Dipartimento di Economia
Facoltà di Economia "G. Fuà"
Ancona