Generating a Time to Treat variable for an Event Plot with Multiple treatments starting at Different times

[Anupam Ghosh]

Hello Statalisters,

I am in dire need of help as I am unable to generate the time_to_treat variable needed in most event study packages. My dataset has two treatments Major and Minor and I have created the respective years of those treatments, I now need to generate the time variable relative to those treatments (Year_Minor and Year_Major)

input int year long fips float(Year_Minor Year_Major)
1991 1001 . .
1992 1001 . .
1993 1001 . .
1994 1001 . .
1995 1001 . .
1996 1001 . .
1997 1001 . .
1998 1001 . .
1999 1001 . .
2001 1001 . .
2002 1001 . .
2003 1001 . .
2004 1001 . .
2005 1001 . .
2008 1001 . .
2010 1001 . .
2014 1001 . .
2015 1001 . .
2016 1001 . .
2017 1001 . .
2018 1001 . .
2019 1001 . .
2020 1001 . .
2021 1001 . .
2022 1001 . .
2023 1001 . .
1991 1003 . .
1992 1003 . .
1993 1003 . .
1994 1003 . .
1995 1003 1995 .
1996 1003 . .
1997 1003 1997 .
1998 1003 1998 .
1999 1003 . .
2001 1003 . .
2002 1003 . .
2003 1003 . .
2004 1003 . 2004
2005 1003 2005 .
2008 1003 . .
2010 1003 . .
2014 1003 . .
2015 1003 . .
2016 1003 . .
2017 1003 . .
2018 1003 2018 .
2019 1003 . .
2020 1003 . 2020
2021 1003 . .
2022 1003 . .
2023 1003 . .
1991 1005 . .
1992 1005 . .
1993 1005 . .
1994 1005 . .
1995 1005 . .
1996 1005 . .
1997 1005 . .
1998 1005 . .
1999 1005 . .
2001 1005 . .
2002 1005 . .
2003 1005 . .
2004 1005 . .
2005 1005 . .
2008 1005 . .
2010 1005 . .
2014 1005 . .
2015 1005 . .
2016 1005 . .
2017 1005 . .
2018 1005 2018 .
2019 1005 . .
2020 1005 . .
2021 1005 . .
2022 1005 . .
2023 1005 . .
1991 1007 . .
1992 1007 . .
1993 1007 . .
1994 1007 . .
1995 1007 . .
1996 1007 . .
1997 1007 . .
1998 1007 . .
1999 1007 . .
2001 1007 . .
2002 1007 . .
2003 1007 . .
2004 1007 . .
2005 1007 . .
2008 1007 . .
2010 1007 . .
2014 1007 . .
2015 1007 . .
2016 1007 . .
2017 1007 . .
2018 1007 . .
2019 1007 . .

[George Ford]

egen ttminor = max(Year_Minor), by(fips)
egen ttmajor = max(Year_Major), by(fips)

g t2tminor = year-ttminor
g t2tmajor = year-ttmajor

[Anupam Ghosh]

Hi George,

Thank you very much for responding. However, this does not work for my design. I am trying to generate two separate event plots one for a major hurricane and another for a minor hurricane. Thus, I had created the years of the specific treatments and hence, I need to get the relative time_to_treat from these years of treatment. Also in my data, a county is treated multiple times.

bys fips year: gen Year_Minor = year if Wind>=64 & Wind<=95, before(year)
bys fips year: gen Year_Major = year if Wind>95 & Wind<999, before(year)

input int year double fips float(Year_Minor Year_Major Wind)
1990 1001 . . .
1991 1001 . . 2.2893946
1992 1001 . . 17.231062
1993 1001 . . 1.446508
1994 1001 . . 29.7827
1995 1001 . . 59.5928
1996 1001 . . 4.4682536
1997 1001 . . 30.071825
1998 1001 . . 26.45237
1999 1001 . . 6.222265
2000 1001 . . .
2001 1001 . . 33.63532
2002 1001 . . 19.51485
2003 1001 . . 3.527311
2004 1001 . . 51.00212
2005 1001 . . 43.84644
2006 1001 . . .
2007 1001 . . .
2008 1001 . . 11.092205
2009 1001 . . .
2010 1001 . . 1.9921166
2015 1001 . . 4.912673
2016 1001 . . 22.574387
2017 1001 . . 30.829926
2018 1001 . . 40.05725
1990 1003 . . .
1991 1003 . . 1.6038173
1992 1003 . . 22.37851
1993 1003 . . .8866239
1994 1003 . . 44.78434
1995 1003 1995 . 73.828476
1996 1003 . . 2.9656675
1997 1003 1997 . 66.42067
1998 1003 1998 . 66.62316
1999 1003 . . 4.2566943
2000 1003 . . .
2001 1003 . . 42.00578
2002 1003 . . 34.52595
2003 1003 . . 7.246689
2004 1003 . 2004 112.5204
2005 1003 2005 . 79.35527
2006 1003 . . .
2007 1003 . . .
2008 1003 . . 27.58461
2009 1003 . . .
2010 1003 . . 2.527151
2014 1003 . . 3.516019
2015 1003 . . 4.6265264
2016 1003 . . 20.51896
2017 1003 . . 53.46472
2018 1003 2018 . 65.19765
1990 1005 . . .
1991 1005 . . 3.594553
1992 1005 . . 11.187775
1993 1005 . . 1.9936098
1994 1005 . . 34.939507
1995 1005 . . 59.0539
1996 1005 . . 6.938338
1997 1005 . . 28.01658
1998 1005 . . 35.86039
1999 1005 . . 9.723162
2000 1005 . . .
2001 1005 . . 40.92657
2002 1005 . . 14.635206
2003 1005 . . 3.12525
2004 1005 . . 37.109432
2005 1005 . . 41.45851
2006 1005 . . .
2007 1005 . . .
2008 1005 . . 10.149692
2009 1005 . . .
2010 1005 . . 1.482621
2015 1005 . . 3.6058176
2016 1005 . . 36.78129
2017 1005 . . 27.29396
2018 1005 2018 . 71.08755
1991 1007 . . 1.780672
1992 1007 . . 23.25992
1994 1007 . . 23.457073
1995 1007 . . 40.64491
1996 1007 . . 3.896007
1997 1007 . . 26.58225
1998 1007 . . 25.45125
2000 1007 . . .
2002 1007 . . 23.321436
2003 1007 . . 3.597654
2004 1007 . . 47.12711
2005 1007 . . 36.923332
2006 1007 . . .
2007 1007 . . .
2008 1007 . . 11.34131
2009 1007 . . .
2010 1007 . . 2.429665
2015 1007 . . 6.089266
2016 1007 . . 16.46585
2017 1007 . . 34.35956
2018 1007 . . 27.97218
1990 1009 . . .
1991 1009 . . 2.0355842
1992 1009 . . 24.81802

[Clyde Schechter]

gsort fips -year
by fips: gen minor_num = sum(!missing(Year_Minor))
by fips: gen major_num = sum(!missing(Year_Major))
gsort fips minor_num -year
by fips minor_num: gen tt_minor = Year_Minor[1] - year
gsort fips major_num -year
by fips major_num: gen tt_major = Year_Major[1] - year
sort fips year
drop minor_num major_num