Hi, I am trying to estimate an unconditional quantile regression with with individual fixed effects as described in Rios-Avila and Maroto. Indeed, I follow their code closely:
I have a subsample of the data. time2 is the squared version of time.
When I estimate try to estimate the model for y2 at the 10th quantile, none of the parameters are computed
:
And I receive the following error right below it.
When I ignore remove y2 from the loop, I receive the insufficient observations warning:
When I remove y3 from the loop, the results for y4 at quantile 10 is essentially the same as the results for y2 at quantile 10 (i.e., all the parameters are zero).
I encounter similar issues when I don't use fixed effects (i.e., omit "abs(id)"). The loop runs until after I estimate the results for y2 at the 90th quantile. After those results are displayed, I get the error below:
When I omit y3 from the loop, the loop is able to run without any warning/error messages. However, some of the results for specific quantile-y combinations display zeros for all estimated parameters.
I've looked through prior posts and have not come up with a solution
I would appreciate any insight or guidance. FernandoRios any idea what's going on?
Code:
global m4vars time1 time2 c.time1#c.group2 c.time1#c.group3 c.time2#c.group2 c.time2#c.group3 controlvar
global dvars y1 y2 y3 y4
foreach i in $dvars {
rifhdreg `i' $m4vars , rif(q(10)) abs(id) keepsingletons
outreg2 using qreg_`i'_results.doc, replace ctitle(10th) label dec(2) title(Results for `i')
rifhdreg `i' $m4vars , rif(q(25)) abs(id) keepsingletons
outreg2 using qreg_`i'_results.doc, append ctitle(25th) label dec(2) title(Results for `i')
rifhdreg `i' $m4vars , rif(q(50)) abs(id) keepsingletons
outreg2 using qreg_`i'_results.doc, append ctitle(50th) label dec(2) title(Results for `i')
rifhdreg `i' $m4vars , rif(q(75)) abs(id) keepsingletons
outreg2 using qreg_`i'_results.doc, append ctitle(75th) label dec(2) title(Results for `i')
rifhdreg `i' $m4vars , rif(q(90)) abs(id) keepsingletons
outreg2 using qreg_`i'_results.doc, append ctitle(90th) label dec(2) title(Results for `i')
}
When I estimate try to estimate the model for y2 at the 10th quantile, none of the parameters are computed
:
And I receive the following error right below it.
When I ignore remove y2 from the loop, I receive the insufficient observations warning:
When I remove y3 from the loop, the results for y4 at quantile 10 is essentially the same as the results for y2 at quantile 10 (i.e., all the parameters are zero).
I encounter similar issues when I don't use fixed effects (i.e., omit "abs(id)"). The loop runs until after I estimate the results for y2 at the 90th quantile. After those results are displayed, I get the error below:
When I omit y3 from the loop, the loop is able to run without any warning/error messages. However, some of the results for specific quantile-y combinations display zeros for all estimated parameters.
I've looked through prior posts and have not come up with a solution
I would appreciate any insight or guidance. FernandoRios any idea what's going on?
Comment