Hello, all Statalists!
I want to measure the effect of rail station opening in certain city on the average wage of the citizens of this city.
In particular, my dataset is a panel data for years 2004-2018.
I explore the opening of rail stations in cities (874, 7700, 9200) on the same day in October 2016.
Thus,
I constructed a very simple difference in differences model using city fixed-effects:
Here is a subset of my dataset:
Following my model specification, I run this fixed-effect regression:
As you may see, in addition to the omission of year==2004 (as a "base value"), Stata also omits year==2018.
Why does Stata also omits year==2018?
Many Thanks!
I want to measure the effect of rail station opening in certain city on the average wage of the citizens of this city.
In particular, my dataset is a panel data for years 2004-2018.
I explore the opening of rail stations in cities (874, 7700, 9200) on the same day in October 2016.
Thus,
Code:
* define treatment and control groups gen treat = 0 replace treat = 1 if (code == 9200) | (code == 7700) | (code == 874) * define post treatment period gen post = (year>=2017) * define interaction term gen treatXpost = treat*post
I constructed a very simple difference in differences model using city fixed-effects:
Here is a subset of my dataset:
Code:
* Example generated by -dataex-. For more info, type help dataex clear input int(city_code year) float(treat post treatXpost lwage) 70 2004 0 0 0 8.533657 70 2005 0 0 0 8.631414 70 2006 0 0 0 8.483016 70 2007 0 0 0 8.698013 70 2008 0 0 0 8.740017 70 2009 0 0 0 8.751474 70 2010 0 0 0 8.804475 70 2011 0 0 0 8.856803 70 2012 0 0 0 8.868977 70 2013 0 0 0 8.919186 70 2014 0 0 0 8.940891 70 2015 0 0 0 8.986822 70 2016 0 0 0 8.995661 70 2017 0 1 0 9.055323 70 2018 0 1 0 9.083416 681 2004 0 0 0 9.053803 681 2005 0 0 0 9.129347 681 2006 0 0 0 9.109082 681 2007 0 0 0 9.180603 681 2008 0 0 0 9.203316 681 2009 0 0 0 9.230437 681 2010 0 0 0 9.308193 681 2011 0 0 0 9.353747 681 2012 0 0 0 9.350189 681 2013 0 0 0 9.359019 681 2014 0 0 0 9.380083 681 2015 0 0 0 9.431081 681 2016 0 0 0 9.435082 681 2017 0 1 0 9.481436 681 2018 0 1 0 9.494165 874 2004 1 0 0 8.462525 874 2005 1 0 0 8.474703 874 2006 1 0 0 8.467373 874 2007 1 0 0 8.560444 874 2008 1 0 0 8.61214 874 2009 1 0 0 8.570734 874 2010 1 0 0 8.630343 874 2011 1 0 0 8.691818 874 2012 1 0 0 8.723882 874 2013 1 0 0 8.748781 874 2014 1 0 0 8.77323 874 2015 1 0 0 8.84376 874 2016 1 0 0 8.853379 874 2017 1 1 1 8.923191 874 2018 1 1 1 8.947025 6800 2004 0 0 0 8.624252 6800 2005 0 0 0 8.65504 6800 2006 0 0 0 8.589328 6800 2007 0 0 0 8.743372 6800 2008 0 0 0 8.77106 6800 2009 0 0 0 8.776938 6800 2010 0 0 0 8.8459215 6800 2011 0 0 0 8.888481 6800 2012 0 0 0 8.908694 6800 2013 0 0 0 8.9449415 6800 2014 0 0 0 8.971956 6800 2015 0 0 0 9.024974 6800 2016 0 0 0 9.041449 6800 2017 0 1 0 9.098739 6800 2018 0 1 0 9.11603 7700 2004 1 0 0 8.530702 7700 2005 1 0 0 8.4593525 7700 2006 1 0 0 8.497194 7700 2007 1 0 0 8.620832 7700 2008 1 0 0 8.650325 7700 2009 1 0 0 8.647519 7700 2010 1 0 0 8.683724 7700 2011 1 0 0 8.750999 7700 2012 1 0 0 8.791486 7700 2013 1 0 0 8.829812 7700 2014 1 0 0 8.853808 7700 2015 1 0 0 8.915432 7700 2016 1 0 0 8.907613 7700 2017 1 1 1 8.967887 7700 2018 1 1 1 8.994669 9200 2004 1 0 0 8.431853 9200 2005 1 0 0 8.405591 9200 2006 1 0 0 8.416267 9200 2007 1 0 0 8.531096 9200 2008 1 0 0 8.570354 9200 2009 1 0 0 8.569026 9200 2010 1 0 0 8.614864 9200 2011 1 0 0 8.68169 9200 2012 1 0 0 8.7194805 9200 2013 1 0 0 8.752423 9200 2014 1 0 0 8.782783 9200 2015 1 0 0 8.852236 9200 2016 1 0 0 8.852808 9200 2017 1 1 1 8.927181 9200 2018 1 1 1 8.950792 end
Code:
. xtset city_code
Panel variable: city_code (balanced)
. xtreg lwage treat post treatXpost i.year, fe
note: treat omitted because of collinearity.
note: 2018.year omitted because of collinearity.
Fixed-effects (within) regression Number of obs = 90
Group variable: city_code Number of groups = 6
R-squared: Obs per group:
Within = 0.9819 min = 15
Between = 0.4846 avg = 15.0
Overall = 0.3526 max = 15
F(15,69) = 248.97
corr(u_i, Xb) = -0.0094 Prob > F = 0.0000
------------------------------------------------------------------------------
lwage | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
treat | 0 (omitted)
post | .4752808 .016468 28.86 0.000 .4424281 .5081335
treatXpost | .0325399 .015584 2.09 0.040 .0014508 .0636291
|
year |
2005 | .0197757 .0145079 1.36 0.177 -.0091667 .0487181
2006 | -.0124221 .0145079 -0.86 0.395 -.0413645 .0165203
2007 | .1162613 .0145079 8.01 0.000 .0873189 .1452037
2008 | .1517364 .0145079 10.46 0.000 .122794 .1806788
2009 | .151556 .0145079 10.45 0.000 .1226136 .1804984
2010 | .2084548 .0145079 14.37 0.000 .1795123 .2373972
2011 | .2644577 .0145079 18.23 0.000 .2355153 .2934001
2012 | .2876525 .0145079 19.83 0.000 .2587101 .3165949
2013 | .3195616 .0145079 22.03 0.000 .2906192 .3485041
2014 | .3443263 .0145079 23.73 0.000 .3153839 .3732688
2015 | .4029185 .0145079 27.77 0.000 .3739761 .4318609
2016 | .4081998 .0145079 28.14 0.000 .3792574 .4371422
2017 | -.0220569 .0145079 -1.52 0.133 -.0509993 .0068855
2018 | 0 (omitted)
|
_cons | 8.606132 .0102586 838.92 0.000 8.585667 8.626598
-------------+----------------------------------------------------------------
sigma_u | .23607787
sigma_e | .02512838
rho | .98879722 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(5, 69) = 1237.14 Prob > F = 0.0000
Why does Stata also omits year==2018?
Many Thanks!
Comment