Hello Joel,
I am not completely sure that my calculation of the Lagrange M test
for autocorrelation is ok. I have several doubts: for example if I
want to test for remaining serial correlation after using the LDV,
should I inlude the LDV in the second regression as well (i.e.,I mean
the auxiliary regression of the residuals on the lagged residuals and
the explanatory variables)? Should country dummies for FE be included
in these tests? And so on...
I'm sorry I can't help much. I am trying to teach this to myself now.
Best,
Carlos
On Mon, Jun 8, 2009 at 10:01 AM, Joel Miller<[email protected]> wrote:
> Carlos,
>
> Did you conclude that your hand-calculated LaGrange test was
> appropriate? Did you find serial autocorrelation?
>
> I'm grappling with the same problem. I seem to have serial correlation
> using xtpcse with a lag (according to your hand calculated method) and
> trying to figure out how to deal with it.
>
> Thanks
>
> Joel M.
>
> On Thu, May 28, 2009 at 3:45 AM, Martin Weiss<[email protected]> wrote:
>>
>> <>
>>
>> -xtserial- is user written and obtained via
>>
>> *************
>> findit xtserial
>> *************
>>
>> A look at xtserial.ado shows you the inner workings of the command, which
>> are exceptionally easy to understand. Also note
>> http://www.stata.com/support/faqs/stat/panel.html
>>
>> BTW, could you explain what "TSCS" means?
>>
>> HTH
>> Martin
>>
>>
>> -----Ursprüngliche Nachricht-----
>> Von: [email protected]
>> [mailto:[email protected]] Im Auftrag von Carlos
>> Rodriguez
>> Gesendet: Donnerstag, 28. Mai 2009 02:28
>> An: [email protected]
>> Betreff: st: Test for serial correlation in TSCS data: pooled Lagrange test
>> vs. xtserial command
>>
>> Dear listserv participants,
>>
>> I wish to test for serial correaltion in my TSCS data. I have first
>> done it by hand (not sure if I did it OK -please, see 1 below) but
>> have now found out about xtserial command in Stata10. I'd appreciate
>> your clarifications on the difference (if any) between the xtserial
>> command and 1)
>>
>> 1) I have tested for serial correlation by saving the residuals from
>> my regression and then running a regression of the residuals on the
>> lagged residuals and all of variables from the original regression. I
>> have then multiplied the N by the R-squared from this auxiliary
>> regression and checked the Chi-square value with 1 degree of freedom.
>> (not sure if this is right way to do it, I am grateful for your help)
>>
>> 2) could I use the xtserial command instead of 1)?
>>
>> Thanks very much in advance.
>> Carlos Rodriguez
>> *
>> * 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/
>>
>
> *
> * 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/
