# st: Too small standard errors with QREG?

 From "Jeannette Wicks-Lim" To Subject st: Too small standard errors with QREG? Date Thu, 10 Mar 2005 11:09:48 -0500

Hi,
I've estimated a quantile regression (QREG) which requires an iterative process to minimize the sum of the absolute residuals. Something strange happens with my results: the standard errors are extremly small (note that the CI are points rather than ranges...). Does anyone have any thoughts about why this might be happening? My model does have in it many variables (lots of interaction terms) but I don't hink that should be a problem given the large sample size.

Thanks in advance for any help!

Below I've pasted some of the output:

.9 Quantile regression Number of obs = 422027
Raw sum of deviations 66889.66 (about 2.5118887)
Min sum of deviations 64818.5 Pseudo R2 = 0.0310

------------------------------------------------------------------------------
lnrwage | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lnpmin1r | .1971322 1.53e-15 . 0.000 .1971322 .1971322
_Iyear_78 | -.004403 3.95e-16 . .000 -.004403 -.004403
_Iyear_79 | .0034332 4.48e-16 . 0.000 .0034332 .0034332
_Iyear_80 | -.0193586 5.30e-16 . 0.000 -.0193586 -.0193586
_Iyear_81 | .0143617 3.49e-16 . 0.000 .0143617 .0143617
_Iyear_83 | .0150885 3.69e-16 . 0.000 .0150885 .0150885
_Iyear_84 | .0203746 4.08e-16 . 0.000 .0203746 .0203746
_Iyear_85 | .0300213 4.44e-16 . 0.000 .0300213 .0300213
_Iyear_86 | .0464875 4.63e-16 . 0.000 .0464875 .0464875
_Iyear_87 | .0385665 5.01e-16 . 0.000 .0385665 .0385665
_Iyear_88 | .0299613 5.33e-16 . 0.000 .0299613 .0299613
_Iyear_89 | .0426883 5.65e-16 . 0.000 .0426883 .0426883
_Iyear_90 | .0202615 5.32e-16 . 0.000 .0202615 .0202615
_Iyear_91 | .000379 4.67e-16 . 0.000 .000379 .000379
_Iyear_92 | .000173 4.68e-16 . 0.000 .000173 .000173
_Iyear_93 | .00088 5.00e-16 . 0.000 .00088 .00088

Jeannette Wicks-Lim
Department of Economics
University of Massachusetts, Amherst
(413) 577-0820
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/