Gini coefficient

[fereshte yousefzade]

Hello
Excuse ، I have the income data and the number of people in each household and the weight of households. How can I calculate the Gini coefficient?
please guide me
Thank you

[Stephen Jenkins]

search gini should yield lots of hits. Examples of programs on SSC include ineqdeco, ineqdec0, and svylorenz. Their help files should be informative regarding useage.

[fereshte yousefzade]

Thank you
I have DASp but I do not know the calculation steps, especially with the decalcification method

[Nick Cox]

On DASP I guess that the advice in previous threads still applies. https://www.stata.com/statalist/arch.../msg00932.html

My own bias is that packages for which you need to register are semi-private. If I can't look at the code or help without registering I back off. Support may be available from the authors or from people who have registered.

[John Riveros]

Thank you
I have DASp but I do not know the calculation steps, especially with the decalcification method

if it helps on anything, this is the direct link of the help of ineqdeco which has a technical details section http://fmwww.bc.edu/repec/bocode/i/ineqdeco.html . So you can check how the calculation was perform.

[fereshte yousefzade]

Thank you all friends

[fereshte yousefzade]

Hello
Excuse me ask the question again.
I have to calculate the Gini coefficient for the provinces of Iran in the following way
First, I will divide household income by the number of people in each household and earn per capita income
Then multiply the per capita income in every household's weight
In the next step, I need to divide the new data into ten
Then calculate the Gini coefficient for these deciles
What code should I enter and what package to install?

[fereshte yousefzade]

Who does not help me?

[Stephen Jenkins]

Please read the Forum FAQ, especially the sections about posting clear questions and "bumping"!

Suppose you have the following variables for each household: y (income), size (# persons in household), wgt (household weight), province (identifer, numeric variable)

Here is some pseudo-code

ge ypc = y/size  // per capita household income
ineqdeco ypc [aw = wgt], by(province)   // inequality of distribution of per capita household income among households
ge iwgt = size * wgt
ineqdeco ypc [aw = iwgt], by(province)   // inequality of distribution of per capita household income among individuals

I am assuming that you wish to examine inequality within and between provinces, not decile groups as your last message suggested. (I can't imagine why you would want to do the latter.)

But if you really do wish to examine inequality within and between decile groups, then use something like

xtile decilegp = ypc [aw = iwgt], nq(10)   // decile groups of individuals
ineqdeco ypc [aw = iwgt], by(decilegp)

Over to you now. If you wish to ask follow-up questions, please ensure that you have read the Forum FAQ, and the help files for all the program modules cited, before you post again.

[fereshte yousefzade]

Excuse me for repeated questions

Thank you very much.

[fereshte yousefzade]

Hi I want to calculate Gini coefficient with classification method for Iran.I have studied in Persian papers that Gini coefficient is calculated without and with classification method.In without classification method this formula is used:
G=1-∑(Yi+1 + Yi ) (Xi+1 - Xi ) (1)

Y_i: comparative accumulated frequency cost for families with _i class or row.
X_i: family’s comparative accumulated frequency.
With this method I know how to calculate Gini coefficient through excel or stata. But in classification method after arranging families per capita income ascendingly data is divided to equal groups then it is determined that each one of groups what percentage of society’s income allocated to themselves and this formula is used:
G = 1- 1/N ∑(Yi+1 + Yi) (2)
In this formula 1/N is classification’s ratio( if decile is used 1/10)
Iran’s statistics institute state that with classification method Gini coefficient calculated this way:
Each family’s income(Y) divided to family’s size so that per capita income is obtained.
P_Y = Y/size
Then per capita income is multiplied in weight
W_Y = P_Y * weight
Then after arranging W_Y ascendingly data is divided to equal groups(10 group). It is determined that each one of groups what percentage of society’s income allocate to themselves and through formula (2) Gini coefficient is calculated. I have done this method with excel but with stata after decile clustering I don’t know the code to enter. Previously I have asked this question and I have seen other friend’s questions but I didn’t get any results. I’ll be grateful if you help me.
(my point of calculating Gini coefficient isn’t for each decile but calculating the whole country’s Gini coefficient with classification method)

[Stephen Jenkins]

You appear (now) to want to calculate the Gini from grouped data. (Why you might do that if you have access to unit record data is unclear). You should search the internet for relevant articles and formulae. Publicly accessible sources I found quickly include: https://www3.nd.edu/~wbrooks/GiniNotes.pdf ; https://www.une.edu.au/__data/assets/pdf_file/0011/15986/emetwp00-1.pdf . More generally, the nature of the calculation depends on what information is available for each of the groups (e.g. do you have an estimate of each group's mean income)

You might also check out von Hippel et al.’s ‘robust Pareto midpoint estimation’ (RPME) on SSC: search rpme

[fereshte yousefzade]

thank you very much