Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Autocorrelation(testparm or wntstmvq?)


From   Robson Glasscock <glasscockrc@vcu.edu>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Autocorrelation(testparm or wntstmvq?)
Date   Sat, 11 Jun 2011 16:59:26 -0400

Hi Katia,
The model you estimated attempts to explain current variation in your
dependent variable (rt) using 4 lags of rt. The regression output
tells you the impact/significance of each individual lag (l.rt- l4rt)
on the current value of rt. The testparm command you wrote uses a Wald
test to test the null hypothesis that beta l.rt= beta l2.rt= beta
l3.rt= beta l4.rt= 0. Note that in your model you don't need to use
the testparm command to test that all of your parameters are jointly
equal to zero because the standard output already gives this to you in
the form of the overall F/LR test in the upper right-hand corner of
the output.

Testing for autocorrelation in residuals of time-series models is
different. The presence of autocorrelation in the residuals of our
"best" model tells us that we haven't modeled the process perfectly.
There is still some systematic variation in the error terms, but
knowing that autocorrelation exists via the test still won't tell us
what, exactly, the autocorrelation is caused by.

best,
Robson Glasscock
On Fri, Jun 10, 2011 at 9:53 AM, Katia Bobulova
<katia.bobulova@googlemail.com> wrote:
> Dear All,
>
> I would like to test the autocorrelation between rt,rt-1 and so on.
>
> I typed this command:
>
> reg rt L(1/4).rt
> testparm L.rt L2.rt L3.rt L4.rt
>
> However, I found in the book "Alaysis of Financial Time Series", pag.
> 27 that I can test jointly that several autocorrelations of rt are
> zero with the potmanteau test.
>
> The command in stata is:  wntstmvq.
>
> However, all the exmaples that I found related to this command refer
> to autocorrelations in the residuals. Is it correct to do something
> like this, to test instead the autocorrelation in the resturns?:
>
> wntstmvq bq
>
> Are testparm and wntstmvq two different ways to test the same thing?
>
> I am a little bit confused on which one should I use in my case. Any
> help would be really appreciated.
>
> Katia
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
>

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   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   |   Site index