[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
RE: st: binary endogenous variable with panel data (call for GMM)
People can vote for what they like -- naturally. And
I'm sure lots of Stata users want this stuff and
that StataCorp is listening seriously.
However, if anyone wants to start assuming or presuming
that Stata is primarily an econometrics program,
that is surely incorrect and divisive.
> I want to join the call for a GMM routine in Stata... this is a big
> gap from other software (e.g. LIMDEP). Models like Chamberlain's
> approach to panel data are not easy to estimate because of this... The
> econometrics literature has shifted in the last years away from ML and
> is moving towards methods with less assumptions, among which GMM.
> Hopefully Stata won't lag behind on this.
> On 11/23/05, Pierre Azoulay <email@example.com> wrote:
> > I believe the answer is no. The newey method only works with
> > continuously distributed endogenous variables.
> > In general, estimating non-linear panel data models with endogeneity
> > requires GMM estimation. The idea is to treat the derivative of the
> > log likelihood function as implicitly defining a moment condition.
> > Unfortunately, Stata does not have a GMM routine at this point. It's
> > really one of the only critical gaps in what is otherwise a
> > package. I have a similar problem with count panel data
> models, and I
> > am going to have to do all these analyses in TSP.
> > Unfortunately, TSP does not handle weights, and TSP commands do not
> > have an easy to use cluster() option either.
> > Dr. Gould, Save us!
> > Pierre
> > Newey, Whitney K., "Efficient Estimation of Limited
> Dependent Variable
> > Models with Endogenous Explanatory Variables" Journal of
> > 36(3) November 1987, pp. 231-50.
* For searches and help try: