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

st: White noise tests


From   "Rune Logstein" <Rune.Logstein@nfh.uit.no>
To   <statalist@hsphsun2.harvard.edu>, "Rune Logstein" <Rune.Logstein@nfh.uit.no>
Subject   st: White noise tests
Date   Wed, 15 Jun 2005 10:10:43 +0200

Hello 
I have a problem. This is about how to understand the differences
between the white noise tests Bartlett's periodogram based white noise
test and portmanteau white noise test.

I have time series data with approximately 32500 observations. (The time
series is really a panel data but for simplicity I treat the data as
ordinary time series). The time series are daily observations

After I have regressed the dependent variable on a lot of endogenous
variables and specified an ARMA(4.1) structure for the residuals I test
for white noise of the residuals with the tests mentioned above.

Bartlett's test says that the residuals are white noise (0.4 / 0.99) But
the Portmanteu say that the residuals are not white noise (57.9 /
0.03)

Wich test result should I belive in?
Why the contradictory result? Does the negative conclusion from the
portmanteu test stem from the big sample size?


Best regards
Rune Logstein



*
*   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