{smcl}
{* 01Sept2006}{...}
help for {hi:gdecomp}
{hline}
{title:Decomposition of outcome differentials after nonlinear models}
{title:Syntax}
{p 8 16 2}
{cmd:gdecomp} {it:groupvar} [, {it: options} ] {cmd: :} {it:estimation_command}
{p_end}
{p 4 4 2} where
{p 8 16 2} {it:groupvar} specifies a binary (numeric) variable identifying the two groups;
{p 8 16 2} {it:estimation_command} (see help {help estcom}) should begin with the
{it:logit}, {it:logistit}, {it:logistit}, {it:probit}, {it:poisson}, or {it:nbreg};
{p 8 16 2}
{it:options} are
{p_end}
{p 16 16 2}
{cmdab:dxw:eight}{cmd:(}{it:high}|{it:low}{cmd:)} {cmdab:rev:erse} {cmd:eform}
{cmdab:l:evel:(}{it:#}{cmd:)} {cmdab:nohead:er} {cmd:nocoef}
{p_end}
{p 16 16 2}
{cmdab:d:ummies(}{it:varlist_1} [{cmd:\} {it:varlist_2} ..]{cmd:)}
{p_end}
{title:Description}
{p 4 4 2} {cmd:gdecomp} implements a generalized Blinder-Oaxaca decomposition which applies to categorical
and count outcomes (and parallel to this, to nonlinear regression models). First, {it: estimation_command} is
estimated in the two groups of {it: groupvar}. Then the observed difference in the
dependent variable of {it: estimation_command} between the groups defined by {it:groupvar} is decomposed
into three parts: (1) a part due to differences in endowments (labeled by E),
and (2) a part due to differences in {it: marginal effects} and finally
(3) a part due to difference in baseline predictions or constants (labeled by U).
See the {cmd: Methods and formulas} section below.
{p 4 4 2} Typed without arguments, {cmd:gdecomp} replays the estimation results.
{cmd:gdecomp} shares all features of estimation commands; see help {help estcom} for details.
{p 4 4 2} Before using {cmd:gdecomp}, please install the latest version of {help margeff}.
(The latest version is 2.0.1, dated 15 Septermber 2006).
See other packages carrying out Blinder-Oaxaca decompositions at the bottom of this help file.
{title:Options}
{p 4 8 2}{cmdab:dxw:eight}{cmd:(}{it:high}|{it:low}{cmd:)} affects the calculation of the endowment effect.
If {cmdab:dxw:eight}{cmd:(}{it:high}{cmd:)} is specified then differences in endowments are evaluated at the
high-outcome regression line. If {cmdab:dxw:eight}{cmd:(}{it:low}{cmd:)} is specified
then differences in endowments are weighted with the marginal effects from the low-outome group.
The default is {cmdab:dxw:eight}{cmd:(}{it:high}{cmd:)}.
{p 4 8 2} {cmdab:rev:erse} tells {cmd:gdecomp} that the group with the lower average of the outcome variable
should be treated as the high-outcome group. By default, {cmd:gdecomp} defines the low-outcome group to be
the group with the largest observed mean of the outcome variable. The default behavior generalizes the idea
that average earnings are higher in the high-outcome group. The {cmdab:rev:erse} option makes sense and should be used
only if high value of the outcome variable indicate outcomes that are "negatively" valued (or, outcomes decreasing
subjective utility). Do not use this option if large categories of the outcome variable record high salaries or
being in the labor force; use this option if large categories of the outcome variable record being unemployed.
{p 4 8 2}{cmd:eform} tells {cmd:gdecomp} that the dependent variable is the natural logarithm of the outcome variable,
so that correct marginal effects (changes in the exponential of the linear prediction) can be calculated.
This option is useful if the dependent variable is the logarithm of wage.
{cmd: Warning:} with this option, you do not request the results to be displayed in exponentiated form.
{p 4 8 2}{cmd:level(}{it:#}{cmd:)} specifies the confidence level, in
percent terms, for the confidence intervals of the computed statistics;
see help {help level}.
{p 4 8 2} {cmd:noheader} suppresses the display of overall and variable-level decomposition results.
{p 4 8 2}{cmd:nocoef} suppresses the display of the decomposition results for the variables,
and forces {cmd:gdecomp} to display the E, C and U components (without respective standard errors).
.
{p 4 8 2}
{cmd:dummies(}{it:varlist_1} [{cmd:\} {it:varlist_2} ... ]{cmd:)}
modifies the calculation of marginal effects for dummy variables. Here,
{it:varlist_1} [{cmd:\} {it:varlist_2} ... ] are lists of dummy variables, where all dummies of a list indicate
different categories of the same underlying categorical variable. Let {it:xvar} be a categorical variables
with {it:K}+1 ({it:K}>1) categories. In this case, not xvar, but {it:K} dummies - say, D1, ..., DK - are included
in the regression model. The estimated marginal effects for these {it:K} dummies
may be misleading (see an example in the help file {help margeff}). The correct result is obtained if
one specifies the {cmd:dummies(}D*{cmd:)} option.
{p_end}
{title:Methods and Formulas}
{p 4 4 2} Let y1 and y0 be the means of the dependent variable Y in the high-outcome
and the low-outcome groups, respectively (thus y1>y0). Let {bf:x}1 and {bf:x}0 the
row vectors of the means of the explanatory variables X1,...,Xk, and {bf:m}1 and
{bf:m}0 the column vectors of the marginal effects in groups 1 and 0, and {it:a}1 and {it:a}0
the baseline predictions in groups 1 and 0.
{p 4 4 2} If the {cmdab:dxw:eight(}{it:high}|{it:low}{cmd:)} option is omitted
or {cmdab:dxw:eight(}{it:high}{cmd:)} is specified, then the raw differential
y1-y2 is approximated as
{p 8 8 2} y1-y0 = ({bf:x}1-{bf:x}0){bf:m}1 + {bf:x}0({bf:m}1-{bf:m}0) + {it:a}1-{it:a}0
{p 4 4 2} If, however, the {cmdab:dxw:eight(}{it:low}{cmd:)} option is specified, then the raw differential
y1-y2 is approximated as
{p 8 8 2} y1-y0 = ({bf:x}1-{bf:x}0){bf:m}0 + {bf:x}0({bf:m}1-{bf:m}0) + {it:a}1-{it:a}0
{p 4 4 2} Whatever method is chosen, the first part on the right-hand side is the endowments effect (E),
and the second part on the right-hand side is the coefficient effect (C), and the third part is the difference
due to differences in "constants" (unexplained part, U).
{title:Author}
{p}Tamás Bartus, Corvinus University, Budapest, tamas.bartus@uni-corvinus.hu
{title:Also see}
{p 0 19}On-line: help for {help gdecomp} {help decompose} {help oaxaca} if installed {p_end}