.-
help for ^sepov^	Dean Jolliffe, Anastassia Semykina
.-

Robust standard errors for the FGT poverty indices
--------------------------------------------------

^sepov^ ^varlist^ [weight] [^if^ exp] [^in^ range]^,^ ^p^ovline^(^varname|#^)^ [sepov_options]

where sepov_options are
		^a^lfa^(^#^) str^ata^(^varname^) psu(^varname^) fpc(^varname^)
		by(^varlist^)^ {^com^plete|^av^ailable} ^sub^pop(varname)
		^srs^subpop ^nolab^el ^l^evel^(^#^) ci deff deft meff meft obs size^


^sepov^ allows ^pweight^s and ^iweight^s; see help @weights@.

Warning:  Use of ^if^ or ^in^ restrictions will not produce correct variance
          estimates for subpopulations in many cases.  To compare estimates
          for subpopulations, use the ^by()^ or ^subpop()^ options.


Description
-----------
^sepov^ provides estimates of the Foster-Greer-Thorbecke (FGT)
class of poverty indices, as well as estimates of the sampling
variance of these indices. The FGT indices are:

Palpha    = 1/n SUM I( y < z)[(z-y)/z]

where n is the sample size, z is the poverty line, y is some measure
of individual or household welfare, I is the indicator function
which takes the value of one if the condition is true and zero
otherwise, and subscripts have been dropped for simplicity.
By default sepov reports the headcount (P0), poverty-gap (P1),
and squared poverty-gap indices (P2), but the user may request
any variant of the FGT by specifying alpha.

If no sample design information is specified, the reported standard
errors are accurate if the design is simple random. For other
designs it is necessary to specify the strata and primary sampling
units.

Options
-------

^p^ovline^(^varname|#^)^ specifies the poverty line, which can be either a scalar
    or a variable. By accepting variables, ^sepov^ allows for spatial variation
    in the poverty line.

^strata()^, ^psu()^, and ^fpc()^ are described in ^stset^; see help @stset@.

^by(^varlist^)^, ^subpop(^varname^)^, ^nolabel^, {^complete^|^available^}, ^level(^#^)^,
^ci^, ^deff^, ^meff^, ^obs^ and ^size^ are described in ^svymean^; see @svymean@.

^a^lfa^(^#^)^ specifies the type of poverty index. If alfa is not specified then,
    by default, sepov reports the headcount, poverty-gap and squared
    poverty-gap indices.


Examples
--------

	. ^sepov pcexp, p(100)^
	. ^sepov pcexp, p(z) psu(clust) strata(strata)^
	. ^sepov pcexp [w=weight], p(z) a(3) by(race) psu(clust) strata(strata)^


Also see
--------

 Manual:  ^[R] svymean^
On-line:  help for @ci@, @svy@, @svymean@ @svydes@, @svylc@, @svyreg@, @svyset@, @svytest@