Sunday, February 19, 2006
Re: st: Using dwstat after regress and dfuller
On Fri, 17 Feb 2006, Coffey, Seamus wrote:
> Can someone explain why dwstat returns different values (in Stata 8.0) after > the dfuller and regress commands for what are the equivalent regressions? > > The results are shown below and you can see that the d value is 1.996121 > after dfuller and is 1.314681 after regress. > > I know the d statistic is not appropriate for an autogregressive model but > surely the computed value should be the same. > > Thanks >
The DW statistic reported after -regress- is correct.
-dfuller- uses a temporary variable to represent the trend in the regression it runs, and, being temporary, that variable is not known to -dwstat-. Therefore, the residuals used by -dwstat- are incorrect.
In short, you cannot use -dwstat- after -dfuller-. You'd need to refit the Dickey-Fuller regression using -regress- before calling -dwstat-.
In a future adofile update, we will fix -dwstat- to spot such conditions and exit with an error message if necessary.
-- Brian Poi -- firstname.lastname@example.org
* * 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/
Links to this post: