On Tue, 16 Oct 2012, Sven Schreiber wrote: > (I don't know if there is some predefined test for seasonality in gretl > -- in the case of OLS models one way might be to just add seasonal > dummies and do an F-test for their significance.)
Fourier stuff is not even remotely as popular as seasonal dummies, but I personally much prefer it for monthly (or higher-frequency) data and yes, it's just an F-test. Example: <hansl> open bjg.gdt series years = time/12 series ann1 = sin(2*pi*years) series ann2 = cos(2*pi*years) series sem1 = sin(4*pi*years) series sem2 = cos(4*pi*years) series qua1 = sin(8*pi*years) series qua2 = cos(8*pi*years) list SEAS = ann1 ann2 sem1 sem2 qua1 qua2 ols lg const time --robust add SEAS # rejection -> seasonality </hansl> -------------------------------------------------- Riccardo (Jack) Lucchetti Dipartimento di Economia Università Politecnica delle Marche (formerly known as Università di Ancona) r.lucchetti(a)univpm.it http://www2.econ.univpm.it/servizi/hpp/lucchetti --------------------------------------------------
