Gini coefficient for population level data

[Flavien Coukan]

Hello Everyone,

I have a population-level dataset that breaks down the proportion of income share by various sociodemographic variables (age, gender, and ethnicity) and I would like to calculate the GINI coefficient to compare pre- and post-COVID19 pandemic to check whether the inequity gap has widened.
I know Stata doesn't have its own command to calculate the Gini coefficient and I looked into the various users' commands, but none seem to match my dataset specifications: the command glcurve can plot the Lorenz curves, but does not provide the Gini coefficient, and the command lorenz can only be used with individual-level data, etc...

Any advice on how I could get the Gini coefficient for population-level data?
I'll take any help/suggestions!

Thank you,
Flav

[Justin Niakamal]

You can look into rifvar (Stata Journal) and dstat (ssc)

. sysuse nlsw88, clear
(NLSW, 1988 extract)

. 
. 
. dstat (gini) wage

Summary statistics                Number of obs   =      2,246

--------------------------------------------------------------
        wage | Coefficient  Std. err.     [95% conf. interval]
-------------+------------------------------------------------
        gini |   .3325258    .006698      .3193909    .3456607
--------------------------------------------------------------

. di r(table)[1,1]
.33252581

. 
. egen Gini = rifvar(wage), gini

. sum Gini

    Variable |        Obs        Mean    Std. dev.       Min        Max
-------------+---------------------------------------------------------
        Gini |      2,246    .3325258    .3174303   .1413221   2.501671

[Flavien Coukan]

Thank you Justin, however your suggestion only seems to apply to individual-level data and the data I am working with is aggregated at the population level, unless I'm mistaken?

[FernandoRios]

Can you describe what kind of data you have
per a snapshot
if you have just aggregates, look for the paper “a tale of two Gini’s”
Also this paper https://www.mdpi.com/2227-7390/9/20/2551