Problem about calculating robust QMLE standard errors
Posted: Fri Nov 06, 2009 2:50 am
Dear all
I’m confusing about the robust QMLE standard errors of ARCH model estimated by EViews (4.0)
According to the User’s guide, robust QMLE standard errors could be calculated from the gradient (or score) and Hessian matrix.
Because I have to calculated the robust QMLE standard errors manually.
But I can’t get the same figure as Eviews provided.
I tried to compare the figure from other statistic software (ex. Gretl, which is a free statistic software)
The formula of robust QMLE standard errors of Eviews and Gretl are the same.
I can got the same robust QMLE standard errors of ARCH model manually with Gretl but not Eviews
Please help me to figure out what’s different in QMLE between Eviews and Gretl.
Thanks a lot for help.
I’m confusing about the robust QMLE standard errors of ARCH model estimated by EViews (4.0)
According to the User’s guide, robust QMLE standard errors could be calculated from the gradient (or score) and Hessian matrix.
Because I have to calculated the robust QMLE standard errors manually.
But I can’t get the same figure as Eviews provided.
I tried to compare the figure from other statistic software (ex. Gretl, which is a free statistic software)
The formula of robust QMLE standard errors of Eviews and Gretl are the same.
I can got the same robust QMLE standard errors of ARCH model manually with Gretl but not Eviews
Please help me to figure out what’s different in QMLE between Eviews and Gretl.
Thanks a lot for help.