I have a panel dataset that looks like this:
| person_id | year | cash |
|---|---|---|
| 222 | 2020q4 | 6,000 |
| 222 | 2021q1 | 7,000 |
| 222 | 2021q2 | 8,000 |
| 321 | 2020q4 | 4,000 |
| 321 | 2021q4 | 11,000 |
| 321 | 2021q2 | 15,000 |
I want to subset the data set for the groups of person_id's that have < 10,000 in 2021q2 and keep all of their previous observations as well, and drop the person_ids who don't have that amount of cash by 2021q2. How can I subset this data by the three variables above?
So in this example I want to keep all observations of person_id 222 but drop all observations of 321.
Appreciate any assistance. Thanks!
CodePudding user response：
There might be an even more succinct way to do this, but I would split it up in these three steps:
clear
input int person_id str6 year int cash
222 "2020q4" 6000
222 "2021q1" 7000
222 "2021q2" 8000
321 "2020q4" 4000
321 "2021q4" 11000
321 "2021q2" 15000
end
*Test if obs has cash>10000 in 2021 q2
gen subset_obs = (cash > 10000 & year == "2021q2")
*By ID, get the max value in subset_obs to copy 1s to all rows for the ID
bysort person_id : egen subset_id = max(subset_obs)
*Keep only IDs with subset_id is 1
keep if subset_id == 1