generate retirement age

[Mads Oernfeldt]

Hey all,

I am using panel data and have a full population dataset. I can observe an individuel over time, their age and their retirement status (dummy variable). I would like to generate a variable over their retirement age by ID.

I tried the following code:

by ID_nr: gen retirement_age = age if retirement == 1
by ID_nr: replace retirement_age = min(retirement_age)

However, it does unfortunately not work.

Can anyone help?

Thank you in advance!

[Øyvind Snilsberg]

gen wanted = age if retirement
bys id (wanted): replace wanted = wanted[1]

[Mads Oernfeldt]

Thanks! It works

[William Lisowski]

To show why the code in post #1 almost worked

by ID_nr: gen ra = age if retirement == 1 
by ID_nr: egen retirement_age = min(ra)
drop ra

The min() function on the generate command finds the minimum of its arguments in a single observation; the egen min() function finds the minimum of its single argument n across observations.

This can be shortened to

by ID_nr: egen retirement_age = min( cond(retirement==1, age, .) )

which finds the minimum value of the expression that evaluates to age if the individual is retired and a missing value otherwise.