Dear statalister,
I have a question concerning an issue that has been raised before but has
not been answered in statalist (at least I havent found an answer in the
archive): xtpcse and fixed effects.
I have a dataset of max. 10 states and about T=15 years. I want to conduct a FE estimation and have detected serial correlation in the disturbances. I am aware of xtregar, fe, but for three reasons I am interested in a prais-winsten, fe
estimation:
(1) xtregar, fe causes the loss of one time period and Prais-Winsten does
not.
(2) xtregar does not allow to account for heteroscedasticity but xtpcse does.
(3) I would like to compare results from both estimation techniques.
Stata offers a panel version of prais in xtpcse, but does not offer a FE option.
My question is whether the correct way of estimating a FE prais-winsten estimation is to run xtpcse on the data including dummy variables (or,
alternatively to run xtpcse on deviations from means or orthogonal deviations). After having read XT pcse I do not see any reason against such a procedure but I am not sure because XT pcse is rather silent on implementing FE.
Thank you very much!
Gerhard
--
mailto: [email protected]
+49 351 451 9603 (privat)
+49 176 6002 6783 (mobil)
+49 351 463 39722 (TUD)Der GMX SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen!
Ideal f�r Modem und ISDN: http://www.gmx.net/de/go/smartsurfer
*
* 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/
