Thanks very much for this very helpful advice, I can
now capture the results perfectly.
However, I'm still not getting the right end results
and I think it may be because I'm not weighting the
subgroups properly. Returning to my question below,
could I ask whether the "subgroup sum of weights" is
meant to be incorporated into the decomposition?
Again, many thanks.
Brian
--- "Stephen P. Jenkins" <[email protected]> wrote:
> The new version of -ineqdeco- on SSC is working
> fine, as far as I know. The
> extended set of saved results was provided
> explicitly in order to facilitate
> calculations of Mookherjee & Shorrocks--type
> decompositions of changes of
> inequality in mind.
>
> You were almost there. The problem is in your code.
>
> The line
>
> replace GE92=r(ge0_r) if var==`r'
>
> should be
>
> replace GE92=r(ge0_`r') if var==`r'
>
> ^^^^
> and similarly in later lines as well.
>
> Note also that macro r(levels) is left behind after
> -ineqdeco-, precisely to
> facilitate the sort of post-calculation looping that
> you are doing.
>
> You could also loop over r(levels) directly rather
> than local S, unless of course
> you had already just created local S from r(levels)
> and have simply not shown us
> this.
>
> It is important to use -foreach- as you did, rather
> than -forvalues- (say), because
> -ineqdeco- requires the values of the bygroup
> variable to be non-negative integers
> but they might be non-consecutive integers.
> r(levels) contains the values of the
> bygroup that were actually used in your -ineqdeco,
> by(bygroup)- command.
>
>
> > Date: Fri, 8 Sep 2006 10:20:26 -0700 (PDT)
> > From: Brian Sartene <[email protected]>
> > Subject: st: new ineqdeco - capturing results for
> dynamic decomposition
> >
> > Dear all:
> > I would like to use the new version of Ineqdeco to
> > write a program for the dynamic decomposition of
> the
> > GE(0) measure. however, i have two problems.
> > First, i have trouble retrieving subgroup-level
> > results.
> > i am trying the code given below but it is not
> > capturing the r(ge0_k) results.
> > Second, using the Mookherjee and Shorrocks
> formulation
> > of the decomposition, I can see a need for all the
> > subgroup components except the subgroup sum of
> > weights.
> > I wondered if this is needed for the
> decomposition, or
> > is given for another purpose.
> > Thank you in advance.
> > Regards,
> > Brian
> >
> >
> >
> >
> > egen var=group(comu)
> > levels7 var, local (S)
> > ineqdeco hhed [w=expc] if year==1992, by (var)
> >
> > tempvar GE92 popsh92 relmean92 incsh92 loginc92
> GE03
> > popsh03 relmean03 incsh03 loginc03 weight92
> weight03
> > lnlambda92 lnlambda03
> >
> > foreach r of local S {
> > replace GE92=r(ge0_r) if var==`r'
> > replace popsh92 = r(v_r) if var==`r'
> > replace relmean92=r(lambda_r) if var==`r'
> > replace incsh92=r(theta_r) if var==`r'
> > replace loginc92=r(lgmean_r) if var==`r'
> > replace weight92=r(sumw_r) if var==`r'
> > replace lnlambda92=ln(relmean92) if var==`r'
> > }
>
>
> --
> Professor Stephen P. Jenkins <[email protected]>
> Institute for Social and Economic Research
> University of Essex, Colchester CO4 3SQ, U.K.
> Tel: +44 1206 873374. Fax: +44 1206 873151.
> http://www.iser.essex.ac.uk
> Survival Analysis using Stata:
>
http://www.iser.essex.ac.uk/teaching/degree/stephenj/ec968/
>
> *
> * 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/
>
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
*
* 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/
