using csdid when data is observed every five years

[Gene Korku]

Hello all,

I would like to estimate the effect of a policy that is implemented in a staggered manner using csdid. My data is collected every five years as shown below.

* Example generated by -dataex-. For more info, type help dataex
clear
input int year
2007
2012
2017
2022
end

I would like to know if csdid can be used in this case especially if the policy is implemented in the years 2010, 2013, and 2018 across different counties. How do I assign my gvar in this case.
I have a balanced panel of counties.

[Gene Korku]

FernandoRios additionally, I'd like to know how this would work under jwdid.

[FernandoRios]

That won’t work
you need the treatment periods to coincide with the observed periods.
for something like this you may instead do the regressions by hand ( you create all dummies as needed)

[Gene Korku]

That won’t work
you need the treatment periods to coincide with the observed periods.
for something like this you may instead do the regressions by hand ( you create all dummies as needed)

If I understand what you are saying, then I can't for example define a treated year 2010 to be 2012, 2013 to be 2015 and 2018 as 2022?

[FernandoRios]

You certainly can but not sure how to interpret that

[Gene Korku]

You certainly can but not sure how to interpret that

I have seen some papers with data collected over similar intervals use csdid/did_multiplegt to create event-studies so I thought something like that would be feasible in my case. That notwithstanding going by your suggestion of doing this manually by hand how do i go about it.

This is how the gvar(ban_date) looks like originally before i convert to the adjacent observable year

           |             ban_date
      Year |      2010       2013       2018 |     Total
-----------+---------------------------------+----------
      2002 |        86         68         11 |       165
      2007 |        87         67         12 |       166
      2012 |        88         71         13 |       172
      2017 |        88         71         12 |       171
      2022 |        88         70         11 |       169
-----------+---------------------------------+----------
     Total |       437        347         59 |       843


After changing the date

          |             ban_date
      Year |      2012       2017       2022 |     Total
-----------+---------------------------------+----------
      2002 |        86         68         11 |       165
      2007 |        87         67         12 |       166
      2012 |        88         71         13 |       172
      2017 |        88         71         12 |       171
      2022 |        88         70         11 |       169
-----------+---------------------------------+----------
     Total |       437        347         59 |       843