Hi,

First, I am sorry for not being able to find a better title than this. But my problem is a little more complicated. The data example below shows directors (with unique director ids) belonging to firms (with unique firm ids). Fiscal years are denoted by fyear while tsize, ind, city, and perf are size tercile the company belongs to, its industry code, city its headquarters are located at, and its past performance, respectively. I am trying to find the determinants of director hire. Specifically, I want to know how director's positions at other firms in the last year influence the probability of his selection in the current company (assuming the director is newly hired). For this reason, I need to compute whether, in the previous year, the director has served in a company of same size etc. I need to do that separately for industry and city. For performance, I have to calculate the average performance of all companies the director served on in the previous year.

Here's the example dataset.

Code:
* Example generated by -dataex-. To install: ssc install dataex
clear
input int(gvkey fyear directorid) byte(tsize ind) str4 city float perf
1001 2000 12345 2 12 "Hel"    .2
1001 2000   123 2 12 "Hel"    .3
1001 2000  4531 2 12 "Hel"    .1
1002 2000    98 3 15 "Tamp"   .1
1002 2000 12345 3 15 "Tamp"  .05
1002 2000   123 3 15 "Tamp"  .04
1003 2000   100 1  9 "Hel"   .12
1003 2000    98 1  9 "Hel"   .02
1003 2000   109 1  9 "Hel"  -.01
1001 2001 12345 2 12 "Hel"    .2
1001 2001   123 2 12 "Hel"    .3
1001 2001  4531 2 12 "Hel"    .1
1002 2001    98 3 15 "Tamp"   .1
1002 2001 12345 3 15 "Tamp"  .05
1002 2001   123 3 15 "Tamp"  .04
1003 2001   100 1  9 "Hel"   .12
1003 2001    98 1  9 "Hel"   .02
1003 2001   109 1  9 "Hel"  -.01
end