i.year command ran for Fixed effects regression but results dropped

[Ketty Chih]

Dear all,
I ran the fixed effects regression and my code is as follows:
xtset d_id monthlydate1 xi :xtreg AR lncaptial lnfeerate lnage i.year if d_sample3==1&t>=-12&t<=12,fe vce(robust)
The code of "year" is :
gen year=cond(y>=1999&y<=2004,1,0)
replace year=2 if y>=2005&y<=2010
replace year=3 if y>=2011&y<=2015
replace year=4 if y>=2016&y<=2020
The d_id is the name of my sample and the monthlydate1 is including the year+month (e.g.2014m1)

Please tell me why the variables "_Iyear_2 _Iyear_3 _Iyear_4" are being dropped.
Thank you.

[Carlo Lazzaro]

Ketty:
please share am example/excerpt of your dara via -dataex-. Thanks.
Please note tha the -xi:- prefix is redundant.
Eventually, it is probably more appropriate to remove social network links from your post. Thanks.

[Ketty Chih]

Carlo Lazzaro ,
Thanks for your reminding, but i can't find the button to edit my post.
The prefix of -xi:- is that i want to tell i did that code and i missed to press the enter.

[Carlo Lazzaro]

Ketty:
screenshots are impossible to elaborate on and, as such, wisely discouraged by the FAQ.
To post an exampe/excerpt of your data, please use -dataex- or use the #-shaped toggle from the advanced editor (A) toolbar.

this is the result of putting your data, code, whatever within CODE bounds

That said:
1) I'm unclear with your explanation cocnerning -xi:- prefix;
2) please post an example/excerpt of your data containing -year- variable, too. Thanks.

[Carlo Lazzaro]

Ketty:
-xi:- prefix does not add anything to your code (as it is the case for all the canned Stata commands).

[Ketty Chih]

Carlo,
year is :

gen year=cond(y>=1999&y<=2004,1,0)
replace year=2 if y>=2005&y<=2010
replace year=3 if y>=2011&y<=2015
replace year=4 if y>=2016&y<=2020

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(d_id d_sample3 AR lnage lncaptial lnfeerate year)
1 .          .         .         .          . 3
1 .          .         . 14.027865 -3.6119184 3
1 1          .         0 13.711934 -2.3330443 3
1 1   .5803861  .6931472   13.6827 -2.4304185 3
1 1   3.675469 1.0986123  13.70369  -2.488915 3
1 1  -2.410889 1.3862944 13.707518 -2.5510464 3
1 1 -.13151123  1.609438  13.67305  -2.453408 3
1 1   -3.23541 1.7917595 13.729343 -2.5383074 3
1 1   1.254812   1.94591 13.778533  -2.525729 3
1 1    .131283 2.0794415 13.879187   -2.56395 3
end

I'm new with stata, Thanks for your patience.
Let me know if i miss any information

[Carlo Lazzaro]

Ketty:
as per you data example, replicating your issue is unfeasible.
As you provide one year only, the best approach is going OLS mimicking an -fe- specification by plugging in the -panelid- (that does not vary; as such, it refers to an unique panel only).
That said:

. reg AR lncaptial lnfeerate lnage i.d_id i.year
note: 1.d_id omitted because of collinearity
note: 3.year omitted because of collinearity

      Source |       SS           df       MS      Number of obs   =         7
-------------+----------------------------------   F(3, 3)         =      0.63
       Model |  12.2251122         3  4.07503739   Prob > F        =    0.6448
    Residual |  19.5075174         3  6.50250581   R-squared       =    0.3853
-------------+----------------------------------   Adj R-squared   =   -0.2295
       Total |  31.7326296         6   5.2887716   Root MSE        =      2.55

------------------------------------------------------------------------------
          AR |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
   lncaptial |   23.71878   22.57403     1.05   0.371    -48.12186    95.55942
   lnfeerate |   30.85099   32.07198     0.96   0.407    -71.21638    132.9184
       lnage |  -1.547011   3.368463    -0.46   0.677    -12.26696     9.17294
      1.d_id |          0  (omitted)
      3.year |          0  (omitted)
       _cons |  -246.1281   285.3071    -0.86   0.452    -1154.103    661.8465
------------------------------------------------------------------------------

.

-d_id- and -year- are perfectly collinear and are therefore kicked out from the calculation.

[Ketty Chih]

Carlo Lazzaro :
Thanks you for your help! Really appreciate what you've done.