Dear Ingo,
You will find an interesting paper on MA unit root process here:
http://ideas.repec.org/a/cup/etheor/v14y1998i3p326-38.html
and an introduction to ARIMA models here
http://econ.la.psu.edu/~hbierens/EasyRegTours/ARIMA.HTM
Briefly, your time series will exhibit a unit root or will be stationary but
not both!
The (Augmented) Dickey-Fuller test will help you to test if their is a unit
root in your time series. If after on differenciation your time serie is
still stationary, it is of order 1, after two, it is of order 2 etc...
Regards.
---
Sylvain BARTHELEMY, Senior Economist
Thierry Apoteker Consultant
http://www.tac-financial.com | http://barth.netliberte.org
-----Message d'origine-----
De : The ox-users list is aimed at all Ox users
[mailto:[log in to unmask]]De la part de Ingo Klein
Envoyé : mercredi 19 mars 2003 12:24
À : [log in to unmask]
Objet : MA Unit Root Process
Dear fellow Econometricians,
After differencing a stationary variable according to ADF, I have found some
problems with the new variables. Apparently, this is caused by something
called MA Unit Root Process, however, I am unable to find any literature on
this. Does anyone know what de facto implications such effect would have on
a linear regression? And, maybe even more important, how to test for it?
Best,
Ingo Klein
Economics and Finance
University Of Durham
England
|