30
Stata Technical Bulletin
STB-48
The poverty measures estimated by poverty are computed as follows:
1 n / ~ ʌ. ∖ «
-ÏJ *<I<
Foster-Greer-Thorbecke class:
Headcount ratio: 0 = FGT(Q)
|
Poverty gap ratio: |
pgr = FGT(1) |
|
Income gap ratio: | |
|
Aggregate poverty gap: |
aPg = Σ"=1 - yt) wi1i |
|
Watts index: |
1 n watts = — V (ln(z) - ln(j∕i)) wiIi IN δ— |
|
Clark-Hemming-Ulph class: |
⅛∑(>-(?/)^ |
|
Thon index: |
2 n tll°n = z(ΛT+1)ΛT Σ (jv + 1 - - ⅜) wΛ |
|
Takayama index: |
tak = 1 + ɪ - 2 ∑i=1 (jv + 1 ~ Wi 1Vi1i + zβ~Ii}'} N _ ∑j^=ιNwi(yiIi + z(l - Ii')') _ |
|
Sen index: |
2 n sen = z(5+1)AT Σ (5 + 1 - r⅛) (z - wi1i |
In the Foster-Greer-Thorbecke class, along with FGT(Q) and FGT(I), poverty computes FGT(<a) with a =Q.5, 1.5,
2, 2.5, 3, 3.5, 4, 4.5, 5. In the Clark-Hemming-Ulph class, poverty computes GHU(∕3) with β =Q.1, Q.25, Q.5, Q.75, Q.9.
Syntax
poverty varname Weeihht] [if exp] [in range] [, line(#) gen(,newvarname) select-options]
aweights and fweights are allowed.
Options
line(#) specifies the value of the poverty line. If # is set to —1, the poverty line is computed as half the median of varname.
If # is set to —2, it is computed as two-thirds the median of varname. Default is —1.
newvaewvaename) creates the new variable newvarname and sets it to 1 for all observations identified as poor (i.e., observations
for which varname is below the specified poverty line) and Q for observations identified as non-poor. neevarname is set to
missing for observations with missing varname or not included by the if in statements.
select-options are options used to select the indices to be computed. It can be any of the following (multiple selections are
allowed, see examples below):