Dear All, I have been asked the following question. The data is
Code:
* Example generated by -dataex-. To install: ssc install dataex
clear
input long company_id int year float(cash2 asset2 roa2) int event_year byte post
35 2007        .        .         .    . .
35 2008        .        .         .    . .
35 2009        .        .         .    . .
35 2010        .        .         .    . .
35 2011        .        .         .    . .
35 2012        .        .         .    . .
35 2013        .        .         .    . .
35 2014 .1127009 21.90875  .0570275 2014 0
35 2015 .0645279 22.20465  .0548749 2014 1
35 2016 .0473218 22.51477   .032463 2014 1
35 2017 .0453601 22.81886  .0356166 2014 1
37 2007        .        .         .    . .
37 2008        .        .         .    . .
37 2009        .        .         .    . .
37 2010        .        .         .    . .
37 2011        .        .         .    . .
37 2012        .        .         .    . .
37 2013        .        .         .    . .
37 2014        .        .         .    . .
37 2015 .2241614 22.24493 -.1642308 2011 1
37 2016 .2352649 22.19659  .2031146 2011 1
37 2017 .3559973 21.78238  .0031499 2011 1
38 2007        .        .         .    . .
38 2008        .        .         .    . .
38 2009        .        .         .    . .
38 2010        .        .         .    . .
38 2011 .0292621 20.26197  .0088071 2011 0
38 2012 .0302361 20.22923  .0187552 2011 1
38 2013 .0280254 20.33895  .0174681 2011 1
38 2014 .0518462 20.34243 -.0743601 2011 1
38 2015 .1417207  20.4286  .0267013 2011 1
38 2016        .        .         .    . .
38 2017        .        .         .    . .
39 2007        .        .         .    . .
39 2008        .        .         .    . .
39 2009        .        .         .    . .
39 2010        .        .         .    . .
39 2011        .        .         .    . .
39 2012 .0828916 24.86628   .041901 2011 1
39 2013 .0657115 25.00831  .0464264 2011 1
39 2014 .0417811 25.01401  .0375714 2011 1
39 2015 .0420292 25.01401  .0275726 2011 1
39 2016 .0382863 25.01401  .0064241 2011 1
39 2017 .0280802 25.01401  .0319415 2011 1
end
  1. The purpose is to keep `company_id' with both pre- and post-event data.
  2. The event occurred in `event_year' (suppose that the effect of the policy begins in the next year).
  3. Taking company_id=35 for example, the pre period is year=2014 and the post period is year=2015 (of course, including other 2016, ... years), both with data on cash2, asset2, and roa2, and so on. Thus, we want to keep this company_id for later analysis.
  4. Taking another company_id=37 for instance, the pre period is year=2011 and the post period is year=2012 (of course, including other 2013, ... years). There is no data (all missing data) in the pre period. Thus, we want to drop this company_id.
  5. Any suggestion is appreciated.