[Andrew Musau]

Can you clarify what the july and the tag variables should be indicating? In the data from #6, why is it assigning a 1 to the tag for line 5 of your list (Alabama policyid 6)? I don't see two policies that meet the maxduration criteria listed for 2008.

For me, july and the tag still assign the same values (1 and 0 respectively) to all of my observations.

The maxduration code correctly identifies policyid "6" as the policy to keep for Alabama in 2008.

I thought the tag variable should select the policy with July as the active month only for case where there is more than one policyid meeting the maxdurtion criteria for a state year. See below data example. This is what my clunky coded drop_flag does (which, as you previously noted, is not a good way to do it).
[/CODE]

"july" in my code is an indicator of whether a policy was active in the month of July. It distributes the 0 or 1 to all observations corresponding to a particular policy. "tag" identifies whether there is a policy active in the month of July, just in case that there are multiple policies that are tied for maximum number of months active. So in my code, it is the combination of july and tag that defines what observations are to be dropped. I highlighted two potential issues with your code in #8 (please reread). If these two issues do not arise, the codes should give you the same results. But my code is "safer" in that it accounts for the issued raised in #8. One more thing about july, as the policies may span over multiple years, we need to add a year restriction. If you are noticing a specific issue with the code, be explicit about what it is.

* Example generated by -dataex-. For more info, type help dataex
clear
input str14 state int policydatebegin float policydateend
"Mississippi"  13880 14244
"Mississippi"  14426 14517
"Mississippi"  14518 14975
"Mississippi"  15035 15917
"Mississippi"  16223 16952
"Mississippi"  17257 17470
"Mississippi"  17532 17805
"Mississippi"  17898 18382
"Mississippi"  18383 19935
"Mississippi"  20089 21185
"Pennsylvania" 13849 16860
"Pennsylvania" 16861 17074
"Pennsylvania" 17167 17622
"Pennsylvania" 17714 17805
"Pennsylvania" 18079 18109
"Pennsylvania" 18232 19448
"Pennsylvania" 19540 22780
"Virginia"     13880 14244
"Virginia"     14426 14975
"Virginia"     15127 15340
"Virginia"     15341 16252
"Virginia"     16253 17074
"Virginia"     17348 17439
"Virginia"     17440 17805
"Virginia"     18079 18352
"Virginia"     18353 19723
"Virginia"     19724 21730
end
format %td policydatebegin
format %td policydateend



*CHECK IF STATE POLICYDATEBEGIN POLICYDATEEND UNIQUELY IDENTIFY AN OBSERVATION
isid state policydatebegin policydateend
*IF #1 PASSES
egen policyid= group(state policydatebegin policydatebegin)
g startym= ym(year(policydatebegin), month(policydatebegin))
g endym=  ym(year(policydateend), month(policydateend))
expand endym-startym+1
bys state policyid (startym): g yearmonth= startym[1]+_n-1
format *ym yearmonth %tm
g year= year(dofm(yearmonth))
bys policyid year: g duration=_N
bys state year: egen maxduration= max(duration)
keep if duration==maxduration
sort state policyid yearmonth
bysort state policyid year: egen july= max(month(dofm(yearmonth))==7)
bysort state policyid (year): keep if _n==1
bysort state year (july): gen tag=  july[_N] & !july[1]
drop if !july & tag
l state policyid policydatebegin policydateend year yearmonth july tag, sepby(state year)

Res.:

. l state policyid policydatebegin policydateend year yearmonth july tag, sepby(state year)

     +--------------------------------------------------------------------------------+
     |        state   policyid   policyd~n   policyd~d   year   yearmo~h   july   tag |
     |--------------------------------------------------------------------------------|
  1. |  Mississippi          1   01jan1998   31dec1998   1998     1998m1      1     0 |
     |--------------------------------------------------------------------------------|
  2. |  Mississippi          2   01jul1999   30sep1999   1999     1999m7      1     1 |
     |--------------------------------------------------------------------------------|
  3. |  Mississippi          4   01mar2001   31jul2003   2001     2001m3      1     0 |
     |--------------------------------------------------------------------------------|
  4. |  Mississippi          5   01jun2004   31may2006   2004     2004m6      1     0 |
     |--------------------------------------------------------------------------------|
  5. |  Mississippi          6   01apr2007   31oct2007   2007     2007m4      1     0 |
     |--------------------------------------------------------------------------------|
  6. |  Mississippi          7   01jan2008   30sep2008   2008     2008m1      1     0 |
     |--------------------------------------------------------------------------------|
  7. |  Mississippi          8   01jan2009   30apr2010   2009     2009m1      1     0 |
     |--------------------------------------------------------------------------------|
  8. |  Mississippi          9   01may2010   31jul2014   2010     2010m5      1     0 |
     |--------------------------------------------------------------------------------|
  9. |  Mississippi         10   01jan2015   01jan2018   2015     2015m1      1     0 |
     |--------------------------------------------------------------------------------|
 10. | Pennsylvania         11   01dec1997   28feb2006   1997    1997m12      0     0 |
     |--------------------------------------------------------------------------------|
 11. | Pennsylvania         12   01mar2006   30sep2006   2006     2006m3      1     0 |
     |--------------------------------------------------------------------------------|
 12. | Pennsylvania         13   01jan2007   31mar2008   2007     2007m1      1     0 |
     |--------------------------------------------------------------------------------|
 13. | Pennsylvania         14   01jul2008   30sep2008   2008     2008m7      1     0 |
     |--------------------------------------------------------------------------------|
 14. | Pennsylvania         15   01jul2009   31jul2009   2009     2009m7      1     1 |
     |--------------------------------------------------------------------------------|
 15. | Pennsylvania         17   01jul2013   15may2022   2013     2013m7      1     0 |
     |--------------------------------------------------------------------------------|
 16. |     Virginia         18   01jan1998   31dec1998   1998     1998m1      1     0 |
     |--------------------------------------------------------------------------------|
 17. |     Virginia         19   01jul1999   31dec2000   1999     1999m7      1     0 |
     |--------------------------------------------------------------------------------|
 18. |     Virginia         20   01jun2001   31dec2001   2001     2001m6      1     0 |
     |--------------------------------------------------------------------------------|
 19. |     Virginia         21   01jan2002   30jun2004   2002     2002m1      1     0 |
     |--------------------------------------------------------------------------------|
 20. |     Virginia         22   01jul2004   30sep2006   2004     2004m7      1     0 |
     |--------------------------------------------------------------------------------|
 21. |     Virginia         23   01jul2007   30sep2007   2007     2007m7      1     1 |
     |--------------------------------------------------------------------------------|
 22. |     Virginia         25   01jul2009   31mar2010   2009     2009m7      1     0 |
     |--------------------------------------------------------------------------------|
 23. |     Virginia         26   01apr2010   31dec2013   2010     2010m4      1     0 |
     |--------------------------------------------------------------------------------|
 24. |     Virginia         27   01jan2014   30jun2019   2014     2014m1      1     0 |
     +--------------------------------------------------------------------------------+

.

[KaLeigh White]

Thank you. I got it to work to this time! I tried to be as explicit as possible before, but I was not certain where I was having an issue. The code ran without error but I could not replicate the results you achieved with it. However, I was able to get there with this second data example. Thank you for your time and help.

KaLeigh