I apologize for not making my statement clear about non-
reproducibility of the standard errors using sqreg. Please see the
example below:
My dependent variable is diff, and my independent variables are
diff3, trend and lmis_. I have 129 subjects identified by the
variable id. So I run two identical models with different sequencing
of the independent variables using the latest updated Stata 8, and
find that although the coefficients are consistent, the standard
errors are different (see below). I would be eager to send you my
dataset if you would like to have a closer look into the matter.
Thanks.
. quietly tab id,gen(dummy)
. drop dummy1
. set seed 1234567
. xi:sqreg diff dummy* diff3 i.trend lmis_
i.trend _Itrend_0-5 (naturally coded; _Itrend_0
omitted)
(fitting base model)
(bootstrapping ....................)
Simultaneous quantile regression Number of obs
= 207
bootstrap(20) SEs .50 Pseudo R2
= 0.6683
----------------------------------------------------------------------
--------
| Bootstrap
diff | Coef. Std. Err. t P>|t| [95% Conf.
Interval]
-------------+--------------------------------------------------------
--------
q50 |
dummy2 | -33.11931 20.62797 -1.61 0.113 -
74.22142 7.98281
dummy3 | -29.4156 18.75973 -1.57 0.121 -66.79518
7.963973
.
.
.
dummy128 | -40.79279 23.68435 -1.72 0.089 -87.98489
6.399318
dummy129 | -16.64872 14.72689 -1.13 0.262 -45.99268
12.69524
diff3 | 14.18931 12.47035 1.14 0.259 -10.65841
39.03704
_Itrend_4 | -9.091088 11.34316 -0.80 0.425 -31.69284
13.51066
_Itrend_5 | -.6516196 13.12833 -0.05 0.961 -26.81038
25.50714
lmis_ | -1.73158 1.449246 -1.19 0.236 -4.619266
1.156106
_cons | 57.45794 21.34711 2.69 0.009 14.9229
99.99298
----------------------------------------------------------------------
--------
. set seed 1234567
. xi:sqreg diff diff3 i.trend lmis_ dummy*
i.trend _Itrend_0-5 (naturally coded; _Itrend_0
omitted)
(fitting base model)
(bootstrapping ....................)
Simultaneous quantile regression Number of obs
= 207
bootstrap(20) SEs .50 Pseudo R2
= 0.6683
----------------------------------------------------------------------
--------
| Bootstrap
diff | Coef. Std. Err. t P>|t| [95% Conf.
Interval]
-------------+--------------------------------------------------------
--------
q50 |
diff3 | 14.18931 8.877925 1.60 0.114 -3.500338
31.87897
_Itrend_4 | -9.091087 8.446916 -1.08 0.285 -
25.92194 7.73976
_Itrend_5 | -.6516195 9.521289 -0.07 0.946 -19.6232
18.31996
lmis_ | -1.73158 1.102537 -1.57 0.121 -
3.928434 .4652733
dummy2 | -33.11931 22.73976 -1.46 0.149 -78.42926
12.19065
dummy3 | -29.4156 18.20915 -1.62 0.110 -65.69812
6.866916
.
.
.
dummy128 | -40.79279 20.23759 -2.02 0.047 -
81.11707 -.468505
dummy129 | -16.64872 20.00269 -0.83 0.408 -56.50496
23.20751
_cons | 57.45794 24.24653 2.37 0.020 9.145678
105.7702
----------------------------------------------------------------------
--------
--- In [email protected], smerryman@k... wrote:
>
> ----- Original Message -----
> From: mehryar_karim <akarim@t...>
> Date: Thursday, July 29, 2004 11:49 am
> Subject: Re: st: Is there a fixed effect quantile regression in
STATA?
>
> > I achieved the fixed effect results using dummy variables for
> > subjects in the sqreg model. My major issue was the
reproducibility
> > of the standard errors even after using the set seed command
before
> > implementing `sqreg'. I think there is a bug in the `sqreg'
command.
> > I'm not sure if my response was helpful.
>
> Could you please expand on the issue of the non-reproducibility of
the standard errors? It seems to work for me (results below)
>
> As to Bo's (bo@m...) original question on fixed effects quantile
regression -- you would have to generate dummy variables and include
them in the regression. I believe this would be interpreted as a
pure-location shift. This seems to how it is done in the applied
literature (see, for example, "A Quantile Regression Analysis of the
Cross Section of Stock Market Returns" by Michelle L. Barnesa1 and
Anthony W. Hughesb
(http://www.bos.frb.org/economic/wp/wp2002/wp022.pdf) who use time
dummies to control time specific effects).
>
> You might also find useful Roger Koenker's paper "Quantile
Regression for Longitudinal Data" (
http://www.econ.uiuc.edu/~roger/research/panel/long.pdf )
>
> In it he writes (page 3):
> "What role should the a_i's play? Generally, the a_i's would be
intended to
> capture some individual specific source of variability,
or 'unobserved heterogeneity,'
> that was not adequately controlled for by other covariates in the
model. For example,
> in a study of the effect of a dietary intervention on blood
pressure, it would be
> desirable to estimate departures from individuals' idiosyncratic
levels. If the number
> of observations m_i were large for each individual then we might
even hope to estimate
> a distributional shift a_i(t) for each individual. This would
certainly be useful for
> groups of individuals: a distributional shift for men versus women,
or for blacks
> versus whites. However, in most applications the m_i, the number of
observations on
> each individual, would be relatively modest and then it is quite
unrealistic to attempt
> to estimate a t-dependent, distributional, individual effect. At
best we may be able to
> estimate an individual specific location-shift effect, and even
this may strain credulity."
>
> Hope this helps,
> Scott
>
> --------------------------------------------------------------------
------------------
>
>
> . set seed 123
>
> . sqreg price mpg fore, q(.1 .9)
> (fitting base model)
> (bootstrapping ....................)
>
> Simultaneous quantile regression Number of obs
= 74
> bootstrap(20) SEs .10 Pseudo R2
= 0.0878
> .90 Pseudo R2
= 0.2546
>
> --------------------------------------------------------------------
----------
> | Bootstrap
> price | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
> -------------+------------------------------------------------------
----------
> q10 |
> mpg | -71.42857 46.74642 -1.53 0.131 -
164.6383 21.78114
> foreign | 591.8571 508.9618 1.16 0.249 -
422.9839 1606.698
> _cons | 5370.429 979.445 5.48 0.000
3417.471 7323.386
> -------------+------------------------------------------------------
----------
> q90 |
> mpg | -348 82.17086 -4.24 0.000 -
511.844 -184.156
> foreign | 1654 961.6789 1.72 0.090 -
263.5332 3571.533
> _cons | 16257 1973.42 8.24 0.000
12322.11 20191.89
> --------------------------------------------------------------------
----------
>
> . set seed 123
>
> . sqreg price mpg fore, q(.1 .9)
> (fitting base model)
> (bootstrapping ....................)
>
> Simultaneous quantile regression Number of obs
= 74
> bootstrap(20) SEs .10 Pseudo R2
= 0.0878
> .90 Pseudo R2
= 0.2546
>
> --------------------------------------------------------------------
----------
> | Bootstrap
> price | Coef. Std. Err. t P>|t| [95%
Conf. Interval]
> -------------+------------------------------------------------------
----------
> q10 |
> mpg | -71.42857 46.74642 -1.53 0.131 -
164.6383 21.78114
> foreign | 591.8571 508.9618 1.16 0.249 -
422.9839 1606.698
> _cons | 5370.429 979.445 5.48 0.000
3417.471 7323.386
> -------------+------------------------------------------------------
----------
> q90 |
> mpg | -348 82.17086 -4.24 0.000 -
511.844 -184.156
> foreign | 1654 961.6789 1.72 0.090 -
263.5332 3571.533
> _cons | 16257 1973.42 8.24 0.000
12322.11 20191.89
> --------------------------------------------------------------------
----------
>
>
>
> *
> * 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/
*
* 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/
