Re: st: RE: lrtest and wald or chi2 statistics

[oliazim <oliazim@yandex.ru>]

Hi Konstantaras,

Thank you very much for your great help.
The only problem that i have the data that not normally distributed and have autocorrelation (stock prices, inflation) and gmm assumes variance-covariance matrix that allows both heteroscedasticity and autocorrelation.
Do you know how the best to fit xtgee model that will allow for these fitures?

(Or maybe, somehow i can withdraw the parameters from the estimated gmm models (unrestricted and restricted) and
compare them by lrtest?)

Thank you in advance,

Olga

12.09.10, 19:40, "K Konstantaras" <dinokon@otenet.gr>:

> The reason why you get this is because gmm does not involve maximum
>  likelihood estimation. Try user written qic command for xtgee, which
>  modifies fit statistics for quasi-likelihood estimation. Since you employ
>  panel data, xtgee may fit a wide variety of dependent variables and error
>  correlations.
>  
>  I believe by: ssc install qic , you can download it.
>  
>  Good luck
>  
>  Konstantinos K.
>  
>  -----Original Message-----
>  From: owner-statalist@hsphsun2.harvard.edu
>  [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of oliazim
>  Sent: Sunday, September 12, 2010 6:00 PM
>  To: statalist@hsphsun2.harvard.edu
>  Subject: st: lrtest and wald or chi2 statistics
>  
>  Dear Statalist members,
>  I need to compare 2 models: Conditional factor model for stock prices and
>  Unconditional CAMP that only imploy World index as a benchmark and obtain
>  Wald test chi2 statistics to compare errors of two models and prove the
>  importance of additional factors. 
>  When i calculate parameters with gmm model along with panel data for 20
>  countries in Stata 11:
>   
>  ". gmm (Conditional:
>  indexprem_c-{a0}-worldindex*{b1}-worldindex*cpi1*{b_cpi}-worldindex*prr1*{b_
>  prr} -worldindex*woil_price1*{b_oil}), 
>  instruments(Conditional: worldindex cpi1 prr1 woil_price1) wmatrix(hac
>  bartlett 1)"
>  and 
>  ". gmm (Unconditional: indexprem_c-{a0}-worldindex*{b1}),
>  instruments(Unconditional: worldindex) wmatrix(hac bartlett 1)"
>  
>  the program does not give me Wald or Chi2 statistic. After, when i try to
>  make lrtest it gives me the error:
>  
>  "lrtest ( Conditional) ( Unconditional)
>  Conditional does not contain scalar e ( l l )"
>  
>  How can i get it or i need to change my models?
>  Olga
>  *
>  *   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/