[R] Functions for QUAIDS and nonlinear SUR?

[Arne Henningsen]

On Sat, Jan 9, 2010 at 1:21 AM, Werner W. <pensterfuzzer at yahoo.de> wrote:
> I would like to estimate a quadratic almost ideal demand system in R which is
> estimated usually by nonlinear seemingly unrelated regression. But there is no
> such function in R yet

The "systemfit" package has the function nlsystemfit() for estimating
systems of non-linear equations, e.g. by non-linear SUR. However, in
contrast to the systemfit() function for estimating systems of linear
equations, the function nlsystemfit() is still under development and
has convergence problems rather often. So, I cannot recommend using
nlsystemfit() for an important analysis. :-(

> but it is readily available in STATA (nlsur), see B. Poi (2008): Demand-system
> estimation: Update, Stata Journal 8(4).
> Now I am thinking, what is quicker learning to "program" STATA which seems
> not really comfortable for programming or implement the method in R which
> might be above my head in terms of econometrics.

You do not have to start from scratch but you could improve the
nlsystemfit() function, e.g. by implementing analytical gradients of
the objective function -- and I could assist you with this. If you are
interested in improving nlsystemfit(), please apply at R-Forge [1] for
getting write access to systemfit's SVN repository.

[1] http://r-forge.r-project.org/projects/systemfit/

/Arne

-- 
Arne Henningsen
http://www.arne-henningsen.name