[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: 2SLS,fe & two-way error component_hetero&autoco?
Has any one tested for autocorrelation and heteroskedasticity in a 2SLS, fe and two way error component
model? I am working with a three dimensional (firm, analyst and year), large (44th obs) and unbalanced
panel data. The model I am running suffers from endogeneity as well. Therefore, I intend to evaluate
and compare my model base on fixed/random effect model, 2SLS, fe and two way error component
model. I have all the reasons to suspect for autocorrelation and heteroskedasticity in the 2SLS, fe and
two-way error component model. Based on xttest3 and pantest2 tests I have determined that my fixed
effect model suffers from both heteroskedasticity and autocorrelation. But how would I test for these
problems in the 2SLS, fe and two-way error component model? I have tried ivhettest, but it does not
work with the fe.
Any suggestions would be very welcomed.
On 26 May 2004 at 16:36, Clive Nicholas wrote:
> Thanks very much for replying. Just a few notes:
> > If region is a categorical variable, and these are xt data, then there are
> > two possibilities: region modifies the constant term (in which some sort
> > of fe or re model should be used) or region modifies the entire
> > relationship (including the coeff on midch). In the latter case a set of
> > interacted dummies would be used in a fe context, or one could use some
> > sort of random-coefficients model (Hildreth-Houck).
> Of course, I used REGION as an example. In terms of continuous 'third'
> covariates, does the method change? I've been using OLS (when the
> Gauss-Markov assumptions have been satisfied) or FGLS up until now. Most
> of the explantory variables in my models (i.e., net turnout rates and
> party competition) are continuous.
> > I did not respond to the original enquiry since the answer seemed obvious:
> > if there is a third variable that (one suspects) should be in the
> > relationship, and it is measurable, the correct methodology is to include
> > it. After having done so, one may test for its relevance. Techniques such
> > as dealing with proxy issues would only arise if the variable in question
> > is not quantifiable.
> I want to shriek my reply to this, but I'll simply say "I agree with all
> of the above!" That's what I've been doing all along. It was a critical
> query of part of my work that that brought on doubts that I was modelling
> my variables of interest in the correct way.
> CLIVE NICHOLAS |t: 0(044)191 222 5969
> Politics |e: firstname.lastname@example.org
> Newcastle University |http://www.ncl.ac.uk/geps
> * 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/
* For searches and help try: