Balanced panel data with consecutive observations and distinct starting point

[Stan Hissink]

Hi all,

Using Stata 15.1, I am trying to keep consecutive observations per gvkey close to the announcement date of a transaction. Announcementdummy=1 is the closest observation to the transaction and this is my distinct starting point. Using the below codes, I still end up with an unbalanced panel. What can be improved?

Thanks in advance.

* Example generated by -dataex-. To install: ssc install dataex
clear
input long gvkey byte(announcementdummy _spell _seq) float length
1082 1 1 1  8
1082 2 1 2  8
1082 3 1 3  8
1082 4 1 4  8
1382 1 1 1  9
1382 2 1 2  9
1382 3 1 3  9
1382 4 1 4  9
2290 1 1 1  9
2290 2 1 2  9
2290 3 1 3  9
2290 4 1 4  9
2589 1 1 1  6
2589 2 1 2  6
2589 3 1 3  6
2589 4 1 4  6
2597 1 1 1  5
2597 2 1 2  5
2597 3 1 3  5
2597 4 1 4  5
3116 1 1 1  7
3116 2 1 2  7
3116 3 1 3  7
3116 4 1 4  7
3169 1 1 1  5
3169 2 1 2  5
3169 3 1 3  5
3169 4 1 4  5
3310 1 1 1 15
3310 2 1 2 15
3310 3 1 3 15
3310 4 1 4 15
3497 1 1 1  8
3497 2 1 2  8
3497 3 1 3  8
3497 4 1 4  8
3612 1 1 1  9
3612 2 1 2  9
3612 3 1 3  9
3612 4 1 4  9
end


. tsset gvkey announcementdummy
       panel variable:  gvkey (unbalanced)
        time variable:  announcemen~y, 1 to 22, but with gaps
                delta:  1 unit

. xtset gvkey announcementdummy
       panel variable:  gvkey (unbalanced)
        time variable:  announcemen~y, 1 to 22, but with gaps
                delta:  1 unit

. tsspell, f(L.announcementdummy==.)
warning: data contain gaps; see help on tsspell

. egen length = max(_seq), by(gvkey _spell) 

. keep if length >=4
(3,155 observations deleted)

. keep if inrange(announcementdummy, 1, 4)
(6,898 observations deleted)

. tab announcementdummy

announcemen |
    t dummy |      Freq.     Percent        Cum.
------------+-----------------------------------
          1 |      1,670       24.80       24.80
          2 |      1,681       24.96       49.76
          3 |      1,688       25.06       74.82
          4 |      1,696       25.18      100.00
------------+-----------------------------------
      Total |      6,735      100.00

[Stan Hissink]

I already figured it out. Was plain simple. Had to use a different order of coding.

tsset gvkey announcementdummy
xtset gvkey announcementdummy
keep if inrange(announcementdummy, 1, 4)
tsspell, f(L.announcementdummy==.)
egen length = max(_seq), by(gvkey _spell) 
keep if length >=4