# RE: st: RE: More on F test and the translog

 From "Nick Cox" To Subject RE: st: RE: More on F test and the translog Date Thu, 13 Apr 2006 13:52:18 +0100

```I didn't pick that up. Indeed I don't think
you said as much.

If these are externally supplied constants,
then in what sense are they part of your
inferential problem? Perhaps you should
turn this into a Bayesian problem for some
real fun.

Nick
n.j.cox@durham.ac.uk

Gauri Khanna

> Even if I took them in turn the larger probem remains.
> Beta21, Beta31 and
> Beta32 are  not estimated in the translog.

Nick Cox

> >fifth and sixth restrictions put on one line. Try taking
> >them in turn.

Gauri Khanna

> > > More on the F test for homogeniety: Although I was able to
> > > run the first
> > > three test using the test, accumulate command, I was stumped
> > > at the fourth
> > > one(see below). Note I am using a translog production
> > > function with three
> > > inputs
> > >
> > > The restrictions that must be jointly imposed are:
> > >
> > > 1. Beta11+Beta12+Beta13=0
> > > 2. Beta12+Beta22+Beta23=0
> > > 3. Beta13+Beta23+Beta33=0
> > > 4. Beta12=Beta21; Beta13=Beta31, Beta23=Beta32
> > >
> > >
> > > The reason for this is that I when I estimate the translog,
> > > the coefficients
> > > obtained are:
> > > Beta11, Beta12, Beta13, Beta22, Beta23, and Beta33. But the
> > > fourth requires
> > > that I must
> > > have Beta21, Beta31 and Beta32?!
> > >
> > > The equation estimated is:
> > >
> > > LnY= constant+alpha1LnX1+alphaf2LnX2+alpha3LnX3+
> > >         beta11(Lnx1)sq+beta12[Lnx1.Lnx2]+beta12[Lnx1.Lnx3]+
> > >         beta22(Lnx2)sq+beta23[Lnx2.Lnx3]+
> > >         beta33(LnX3)sq
> > >
> > > Sq= squared term.

*
*   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/
```