Fixed effect regression with some conditions

[Saad Al]

Hello Stata experts,

I am using the following data:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(IVOL DUM FirmID)

   -.012846745 0 1
    .003010829 0 1
    -.01994809 1 1
     -.0102941 1 1
   -.016133793 1 1
    .003362658 1 1
    .010100462 0 1
   -.005357542 1 1
     .01985686 1 1
    -.02125267 0 1
   -.013251625 0 2
   -.011405208 0 2
    .005397665 0 2
   -.016385317 0 2
    -.02082585 0 2
    .004482734 0 2
   -.008927207 0 2
-1.3124484e-06 1 2
   -.002564483 0 2
    -.02041283 0 2
   -.021711256 0 3
    .019218445 0 3
    -.01553059 0 3
   -.001207368 0 3
   -.020130247 1 3
   -.015929932 1 3
   -.012757675 1 3
    .002185969 1 3
    -.01975868 1 3
   -.010493663 1 3
   -.021384994 1 4
   -.016919166 1 4
   -.016161574 0 4
    .006752182 0 4
   -.009183821 0 4
   -.022988187 0 4
    .005700997 0 4
   -.016790474 0 4
   -.010007303 1 4
   -.013928052 0 4
  
end

to run this regression:

xtreg IVOL DUM, fe

where DUM is basically a dummy variable and FirmID is the panel firm ID.

So, the observations of variable DUM ==1 vary from one firm to another. We have firms with no observations for DUM==1 and other firms with greater observations of DUM==1. So what I would like to do is to run the above regression if the observations of variable DUM==1 is greater than the average or mean of observations for DUM==1 for all firms. Is there a way to get to the mean of observations for DUM==1 for all firms without doing something like:

by FirmID: sum DUM if DUM==1

because this will calculate the observations for DUM==1 for each firm and then I will have to calculate the mean observations for all firms manually.
Once we do this I need to run the above regression for firms with observations of DUM==1 greater than the mean observations for all firms.

Thank you

[Carlo Lazzaro]

Saad:
are you looking for something along the following lines?

bysort FirmID: egen flag_1=mean(DUM)
egen flag_2=mean(DUM)
list FirmID DUM IVOL flag_1 flag_2

     +--------------------------------------------+
     | FirmID   DUM        IVOL   flag_1   flag_2 |
     |--------------------------------------------|
  1. |      1     0   -.0128467       .6       .4 |
  2. |      1     0    .0030108       .6       .4 |
  3. |      1     1   -.0199481       .6       .4 |
  4. |      1     1   -.0102941       .6       .4 |
  5. |      1     1   -.0161338       .6       .4 |
     |--------------------------------------------|
  6. |      1     1    .0033627       .6       .4 |
  7. |      1     0    .0101005       .6       .4 |
  8. |      1     1   -.0053575       .6       .4 |
  9. |      1     1    .0198569       .6       .4 |
 10. |      1     0   -.0212527       .6       .4 |
     |--------------------------------------------|
 11. |      2     0   -.0132516       .1       .4 |
 12. |      2     0   -.0114052       .1       .4 |
 13. |      2     0    .0053977       .1       .4 |
 14. |      2     0   -.0163853       .1       .4 |
 15. |      2     0   -.0208258       .1       .4 |
     |--------------------------------------------|
 16. |      2     0    .0044827       .1       .4 |
 17. |      2     0   -.0089272       .1       .4 |
 18. |      2     1   -1.31e-06       .1       .4 |
 19. |      2     0   -.0025645       .1       .4 |
 20. |      2     0   -.0204128       .1       .4 |
     |--------------------------------------------|
 21. |      3     0   -.0217113       .6       .4 |
 22. |      3     0    .0192184       .6       .4 |
 23. |      3     0   -.0155306       .6       .4 |
 24. |      3     0   -.0012074       .6       .4 |
 25. |      3     1   -.0201302       .6       .4 |
     |--------------------------------------------|
 26. |      3     1   -.0159299       .6       .4 |
 27. |      3     1   -.0127577       .6       .4 |
 28. |      3     1     .002186       .6       .4 |
 29. |      3     1   -.0197587       .6       .4 |
 30. |      3     1   -.0104937       .6       .4 |
     |--------------------------------------------|
 31. |      4     1    -.021385       .3       .4 |
 32. |      4     1   -.0169192       .3       .4 |
 33. |      4     0   -.0161616       .3       .4 |
 34. |      4     0    .0067522       .3       .4 |
 35. |      4     0   -.0091838       .3       .4 |
     |--------------------------------------------|
 36. |      4     0   -.0229882       .3       .4 |
 37. |      4     0     .005701       .3       .4 |
 38. |      4     0   -.0167905       .3       .4 |
 39. |      4     1   -.0100073       .3       .4 |
 40. |      4     0   -.0139281       .3       .4 |
     +--------------------------------------------+

xtreg IVOL DUM if flag_1>flag_2, fe

Fixed-effects (within) regression               Number of obs     =         20
Group variable: FirmID                          Number of groups  =          2

R-sq:                                           Obs per group:
     within  = 0.0214                                         min =         10
     between =      .                                         avg =       10.0
     overall = 0.0207                                         max =         10

                                                F(1,17)           =       0.37
corr(u_i, Xb)  = 0.0000                         Prob > F          =     0.5504

------------------------------------------------------------------------------
        IVOL |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         DUM |  -.0037558   .0061638    -0.61   0.550    -.0167603    .0092486
       _cons |  -.0050274   .0047744    -1.05   0.307    -.0151006    .0050458
-------------+----------------------------------------------------------------
     sigma_u |  .00329603
     sigma_e |  .01350418
         rho |  .05622311   (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(1, 17) = 0.60                       Prob > F = 0.4508