Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: ARMA modell, lagged dependent variables and autocorrelation


From   "Rune Logstein" <Rune.Logstein@nfh.uit.no>
To   <statalist@hsphsun2.harvard.edu>, <Statalist@hsph.harvard.edu>
Subject   st: ARMA modell, lagged dependent variables and autocorrelation
Date   Wed, 25 May 2005 13:45:27 +0200

Hello everyone at the statalist

I want to regress a model with some lagged endogenous variables as
explanatory variables (1. and 2. lag) and I have manually constructed
these lagged variables.

In addition I want to correct for autocorrelation of second order in the
error terms. For this purpose I regress an ARMA model where I allow for
a second order autocorrelation in the error terms. 

By this I hope that I can regress the model without having some
autocorrelation in the error terms which is said to cause bias and
inconsistency in the presence of lagged variables. 

But in this ARMA regression I cant do the durbin h test ("durbina") that
tests for autocorrelation in the error terms in presence of lagged
endogenous variables. I can only do the durbin watson test that is said
to be biased in the presence of lagged variables. Nevertheless the
durbin watson test in this case is very very close to 2.0 (1.999)

Do the very happy result from the durbin watson test suggest that the
autocorrelation is removed anyway??

How to do the durbin h test in the ARMA regression??

Does anyone know?? 



Rune Logstein
norway


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index