Well in that whole DASP manual, they never discuss how they get the
standard errors. If that were a Stata Journal submission, it would
probably be returned with a request for Methods and Formulas section.
It is not mentioned in their huge analytical manual, either --
although of course that's a very useful summary of all things in
distributional economics, with all the right references. And
unfortunately I've seen more than once that something produced and
promoted by the World Bank eventually gets rebutted for not being
technically/statistically/econometrically solid -- maybe for the
simple reason that it does not have to pass peer review and be
presented at half a dozen seminars with critical audience to appear as
a working paper (and it does not help much that the number of people
who understand both economics of distribution and analysis with
complex survey data is miserably small). So if I were using this
package I would at least make sure that it produces correct standard
errors by cross-checking with other procedures.
Of course producing standard errors is a useful practice. But even
computing the standard error of a difference of two say Gini
coefficients or FGT indices between two surveys may not be that easy
if the samples are not perfectly independent.
Just a word of caution.
On 12/5/08, Abdel Rahmen El Lahga <[email protected]> wrote:
> you can use the DASP package developed by A. Araar and J-Y Duclos
> available at http://132.203.59.36/DASP/index.html
> you can estimate various poverty inequality measure with their s.e and
> taking into account the sampling design.
> Concerning the remark of Stas Kolenikov I note that std-err of
> poverty measure are required by most journal and policy maker to
> assess the effect of sampling variability on poverty and inequlity
> measure. This is a usuel practice in the profession
--
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
