[Magashi Joseph]

Hi Atta,
Sorry I didnt respond to your previous email.
So couple of points.
1. rifhdreg does exactly what the PDF you sent me does. So, if you would like to replicate their tables you simply run the regression, as I previously suggested. Something like:

rifhdreg rchealth female black age c.age#c.age, rif(eindex(rchealth) lb(0) ub(1) )

My understanding from Heckleys et al paper is that the coefficients you get from this regression is what they use and interpret as decomposition. The first paper that you cite does exactly the same thing.
2. if what you are more interested is a Oaxaca-Blinder decomposition, that can be easily done using the command oaxaca_rif (also part of "rif")

gen age_sq=age*age
oaxaca_rif rchealth black age age_sq, rif(eindex(rchealth) lb(0) ub(1) ) by(female)
** note that oaxaca_rif does not work with factor notation.

or you can use the more advanced strategy (reweighted analysis).

oaxaca_rif rchealth black age age_sq, rif(eindex(rchealth) lb(0) ub(1) ) by(female) rwlogit(black age age_sq)

3. for the interpretation, you can follow the examples I have on my paper. The working paper has examples for the analysis of inequality in the US using CPS data, and the newer version has an example using the "oaxaca" toy data.

4. For further examples on how to use the syntax, please see the examples within each command. Most of them are executable, so you should be able to replicate them.
HTH
Fernando

Dear Fernando

I have tried to run regression using rifireg and rifhdreg as follows

rifireg stunting child_sex mobirdity media_access householdsize mother_working,lb(0) ub(1) eindex rankv(wealth_index)
rifhdreg stunting child_sex mobirdity media_access householdsize mother_working , rif(eindex( wealth_index ) lb(0) ub(1))

Results obtained on the coefficient are not the same although the size does not differ much. Is this okay or am I making mistakes in the estimation?
See the coefficients of the obtained estimations

Estimates for rifireg

rifEI	Coef.	Std. Err.	t	P>t	[95% Conf.	Interval]
child_sex	-.0127779	.022257	-0.57	0.566	-.0564067	.0308509
mobirdity	-.0834324	.0259076	-3.22	0.001	-.1342171	-.0326476
media_access	-.0601567	.0277468	-2.17	0.030	-.1145468	-.0057665
householdsize	.0140704	.0026327	5.34	0.000	.0089096	.0192311
mother_working	.0233143	.0294883	0.79	0.429	-.0344895	.0811181
_cons	-.1861618	.0419054	-4.44	0.000	-.268306	-.1040177

Estimates for rifhdreg

stunting	Coef.	Std. Err.	t	P>t	[95% Conf.	Interval]
child_sex	-.0164388	.0222152	-0.74	0.459	-.0599857	.0271081
mobirdity	-.0897227	.0258589	-3.47	0.001	-.1404121	-.0390333
media_access	-.0581846	.0276947	-2.10	0.036	-.1124726	-.0038966
householdsize	.0149052	.0026278	5.67	0.000	.0097541	.0200563
mother_working	.030088	.0294329	1.02	0.307	-.0276073	.0877832
_cons	-.1993676	.0418267	-4.77	0.000	-.2813575	-.1173777

With regards,
Magashi

[FernandoRios]

hi Magashi
There is no mistake in what you are doing.
When I was working on the command, I encounter the same "phenomenon" And the reason for why this happens is because all the rank dependent indices require some sorting of the data, and sorting is done at random when there are ties. (see help sort)
On my command, to be able to reproduce the same results, I add the option "seed" which makes it reproducible.
HTH
Fernando

[Magashi Joseph]

Hello FernandoRios
Much appreciated for the quick response

[Magashi Joseph]

Dear FernandoRios
I have read your paper you published on RIF regression and decomposition published by The Stata Journal at https://www.stata-journal.com/articl...article=st0588 and I have one question:
The interpretation you made on page 79 on RIF decomposition especially when you say "Inequality among women is between 11.5% to almost 48% higher than for men" was referring to which coefficient to be specific on Table 5 (Page 78)?. I addition to that you made interpretation on the detailed decomposition effect. Is it possible to make interpretation by looking at the coefficient as you did on the RIF regression on page 66 (First and second paragraph)?.

Thank you in advance
Magashi.

[FernandoRios]

Hi Magashi
Since i was comparing various measures of inequality, I used a comparable metric across groups.
For example, if you look at the iqe ratio Men IQR was 0.95 and women 1.06. So women's inequality is 11.6 (=(1.06/.95 - 1)*100) larger than men. The same calculation is used for other measures.

It is possible to make such interpretations, but those need to be made with care.
1. you need stronger assumptions to make interpretation of the detailed decomposition components.
2. The detailed decomposition are easier to interpret if you put the result in the context of previously estimated regressions. That would be equivalent to presenting rif-regressions similar to table 3, but for each group.

HTH

[Magashi Joseph]

Thank you FernandoRios for this clarification
One more question with regard to the second point you highlighted on Interpretation. How do I put the result in the context of previously estimated regressions (for the oaxaca_rif decomposition) so that I can make I can make interpretation for each group as you suggested?

With thanks
Magashi

[FernandoRios]

similar to what you would do with a standard OB decomposition.
You need the regressions for both groups, and the mean characteristics for both. WIth that you can make a better interpretation of the detailed components.

[Magashi Joseph]

Noted and thank you FernandoRios

[atta asif]

Hi respected FarnandoRios
hopw you are doing well

i need guidelines regarding Bayesian RIF regression and Bayesian RIF decomposition analysis for inequality measurement. so could you please guide about the software that may be helpul for data analysis. one article which described these methods but did not discuss the software that have used (file attached).

your guidance if any might be very helpful for me just as previous.
best wished
asif

[FernandoRios]

Hi Atta
I didn't see any attached paper. In any case, I have not worked with Bayesian econometrics, so your best option may be to contact directly to the authors of the paper.
They may be able to provide you with details on their implementation.
Fernando

[atta asif]

many thanks for response

[atta asif]

meanwhile look at the attached document please

[FernandoRios]

Thank you for the reference and paper

[atta asif]

respected Farnando how are you?

i have come to know a prefix command "bayes" in stata. however upon checking i have come to know that your well-renowned command about "rifhdreg" and Oaxca_rif" did not support it.can you please guide me about, how i can corporate it.

i shall be very thankful

[FernandoRios]

Hi atta
unfortunately, I haven’t work with the Hayes prefix before, but I ll put that on my agenda for a future update
best regards