Dear all,

I have the following problem and would be very happy if you could help me out!

I downloaded the Execucomp dataset and would like to use the methodology from Bertrand & Schoar, 2003 (--> analysis of manager fixed effects), thus, I need to restrict the "attention to the subset of firms for which at least one specific top executive can be observed in at least one other firm". Ideally, I would end up with a dataset that shows the CEO Mr. NiceGuy at firm A from 1992 to 1996 and at firm B from 1996/1997 to 2019, CEO Mr. OtherGuy at firm C from 1993-1995 and at firm D from 1995/1996 to 2003 and at firm E from 2003/2004 to 2019 and so forth...

My idea was to generate a new variable X that indicates the number of unique firms within the EXECID group (manager ID). Thus, for EXECID == 50
  • for the years 1992-1998 (with gvkey = 1573), value 1
  • and for years 2016 - 2019 (with gvkey = 178310), value 2

Array


Afterwards, I would have managed to find the "highest" number within each EXECID / X group and generate a new variable Y with the specific value. In my example above, the value would be 2 for EXECID == 50.

Then, I could drop Y if < 2 (i.e. 1) and end up with my sample.

Unfortunately, I do not fully understand how to generate my X variable.

Thus, I appreciate your help and insight!

Best, L