Calculating growth rate first and last observation

[Max Verheijen]

Dear Statalist,

I want to calculate the total growth rate of a variable over the entire period (2010-2016) however my data set is unbalanced and therefore starting years and/or ending years are different. I was wondering how to calculate the total growth rate by calculating; (last occurrence - first occurrence) / (first occurrence) as i am interested in the growth of this variable over the entire data set.

Thank you for taking the time reading this!

Kind Regards,

Max Verheijen

[Nick Cox]

Let's assume that there are no missing values. Then

bysort id (year) : gen wanted = (whatever[_N] - whatever[1]) / whatever[1]

With missing values present there are other approaches. Here's one:

gen OK = !missing(whatever) 
bysort OK id (year) : gen wanted = (whatever[_N] - whatever[1]) / whatever[1] 
bysort id (wanted) : replace wanted = wanted[1]

You should surely want to keep track too of the number of years concerned

gen OK = !missing(whatever) 
bysort OK id (year) : gen wanted = (whatever[_N] - whatever[1]) / whatever[1] 
by OK id: gen diff = year[_N] - year[1] if OK 
bysort id (wanted) : replace wanted = wanted[1] 
by id: replace diff = diff[1]

The code here is not tested. You didn't give a data example to work on.