[Indrani Manna]

Also Sir, how do I interpret the result from the jacknife correction. Is the table we get at the end refer to some particular quantile? Median regression?

[Joao Santos Silva]

By default, you get the median, but you can choose other quantiles.

Best wishes,

Joao

[Ibra Haidar]

Joao Santos Silva
Dear Professor Silva;

I am using a panel of 178 firms over 2007-2016 and over 2008-2013. I am considering the use of xtqreg as my data is very sensitive to outliers. However, as you may have noticed n/T is really big. So, I wonder if using the Jackknife will correct any bias created by using xtqreg. If so, I wonder I you would generously share it with me as I believe I cannot send you and email because of account restriction.

Thanks a lot for your time and contribution.
Best wishes,

Ibra

[Joao Santos Silva]

Dear Ibra Haidar,

It should indeed help. I have an example of how to do that in the research page of my webpage.

Best wishes,

Joao

[Ibra Haidar]

Thanks a lot Prof. Santos Silva
For interested people, I found it in the publications section of the Prof's webpage as an example of the paper "Quantiles via Moments".

[merve kaya]

Dear Joao,

Does panel quantile regression model give every cross section's coefficient statistic for each quantile degree (.10, .20. ..... , .90) like xtpmg full model?

[merve kaya]

Dear Indrani Manna,

I suggest you compute the standard errors by using clustered bootstrap, as illustrated here.

Also, note that the warning suggests that you may need to make your model more flexible.

Best wishes,

Joao

Dear Silva,

I tried to use your code(above).
I got some results, but I couldn't understand which results show me the robust clustered quantile regression? The last table I got gives us the result of what? And then how can I estimate the wald test?
Thank you very much.

. clear

. *(7 variables, 506 observations pasted into data editor)
. xtset idcode
variable idcode not found
r(111);

. xtset idcode
panel variable: idcode (balanced)

. by idcode: egen c=count(idcode)
not sorted
r(5);
. set seed 123
. g s=runiform()<0.5
. xtwreg yer lyer latm lmzn lkpss, ls q(0.10(0.40)0.90)
command xtwreg is unrecognized
r(199);
. xtqreg yer lyer latm lmzn lkpss, ls q(0.10(0.40)0.90)

MM-QR regression results
Number of obs = 506

Location parameters
(Std. Err. adjusted for 23 clusters in idcode)

Robust
yer Coef. Std. Err. t P>t [95% Conf. Interval]

lyer .9325496 .0932261 10.00 0.000 .7392105 1.125889
latm .0710216 .0310097 2.29 0.032 .0067114 .1353318
lmzn .0031011 .0328224 0.09 0.926 -.0649684 .0711706
lkpss -1.474689 1.287192 -1.15 0.264 -4.144161 1.194784
_cons 157.4729 212.1304 0.74 0.466 -282.4586 597.4043

Scale parameters
(Std. Err. adjusted for 23 clusters in idcode)

Robust
Coef. Std. Err. t P>t [95% Conf. Interval]

lyer .0965352 .0349899 2.76 0.011 .0239705 .1690998
latm .0387408 .0127606 3.04 0.006 .012277 .0652047
lmzn -.0299535 .0415497 -0.72 0.479 -.1161222 .0562152
lkpss 4.784316 1.605389 2.98 0.007 1.454942 8.113689
_cons -291.243 129.3061 -2.25 0.035 -559.4075 -23.07852

WARNING: 5.1383399% of the fitted values of the scale function are not positive
.1 Quantile regression
Coef. Std. Err. z P>z [95% Conf. Interval]
lyer .7807878 .4765032 1.64 0.101 -.1531413 1.714717
latm .0101176 .3347972 0.03 0.976 -.6460729 .6663081
lmzn .0501906 .4181658 0.12 0.904 -.7693993 .8697805
lkpss -8.996051 18.9618 -0.47 0.635 -46.16049 28.16839

.5 Quantile regression

Coef. Std. Err. z P>z [95% Conf. Interval]

lyer .9464727 .1676911 5.64 0.000 .6178042 1.275141
latm .0766091 .1182141 0.65 0.517 -.1550863 .3083045
lmzn -.001219 .1477751 -0.01 0.993 -.290853 .2884149
lkpss -.7846541 6.687959 -0.12 0.907 -13.89281 12.3235

9 Quantile regression
Coef. Std. Err. z P>z [95% Conf. Interval]
lyer 1.083462 .2247761 4.82 0.000 .6429088 1.524015
latm .1315846 .1567294 0.84 0.401 -.1755994 .4387687
lmzn -.0437248 .1953783 -0.22 0.823 -.4266592 .3392095
lkpss 6.004572 8.892183 0.68 0.500 -11.42379 23.43293

. qui bs, cluster(idcode) rep(20): yer lyer latm lmzn lkpss
yer command not found
r(111);

. qui bs, cluster(idcode) rep(20): xtqreg yer lyer latm lmzn lkpss
WARNING: 5.1383399% of the fitted values of the scale function are not positive

. mat V=e(V).
. matrix b=e(b_location)
. matrix g=e(b_scale)
. matrix Q50=e(q)
. scalar N=e(N)
. qui xtqreg yer lyer latm lmzn lkpss
WARNING: 5.1383399% of the fitted values of the scale function are not positive

. scalar N0=e(N)
. matrix Q050=e(q)
. matrix g0=e(b_scale)

. qui xtqreg yer lyer latm lmzn lkpss
WARNING: 5.1383399% of the fitted values of the scale function are not positive

. scalar N1=e(N)
. matrix Q150=e(q)
. matrix g1=e(b_scale)
. mat bjk=b+(2*g-g1*(N1/N)-g0*(N0/N))*(2*Q50-Q150*(N1/N)-Q050*(N0/N))

. ereturn post bjk V
. ereturn display

Coef. Std. Err. z P>z [95% Conf. Interval]

lyer .9325496 .0758899 12.29 0.000 .7838082 1.081291
latm .0710216 .0412574 1.72 0.085 -.0098414 .1518846
lmzn .0031011 .0299044 0.10 0.917 -.0555104 .0617126
lkpss -1.474689 .9124915 -1.62 0.106 -3.263139 .3137619


.

[Joao Santos Silva]

Dear merve kaya,

Your post is rather difficult to read, but it appears that the final table gives you the estimated for the median with clustered standard errors. To better understand what you are doing, please read the original paper before using the estimator.

Best wishes,

Joao

[Funsho Obakemi]

my dependent variable is continuous from 1 to 7. can i use quantile regression. Which one do i interpret among 0.25, 0.5 and 0.75. Do i choose on, or interpret the three? pls i need clarification
Thanks

[Funsho Obakemi]

Greetings to all.
I have a panel data with 106 countries for the period 2007-2020. my dependent variable is continuous but within 1-7. i intend to use quantile regression or system GMM. Which one do i interpret among 0.25, 0.5 and 0.75. Do i choose on, or interpret the three? pls i need clarification
Thanks

[Funsho Obakemi]

Greetings to all.
I have a panel data with 106 countries for the period 2007-2020. my dependent variable is continuous but within 1-7. i intend to use quantile regression or system GMM. I am estimating at 0.25, 0.5 and 0.75, Which one do i interpret among 0.25, 0.5 and 0.75. Do i choose to interpret one, or interpret the three? pls i need clarification
Thanks