Compute yearly returns by monthly returns given unfavorable data layout

[Jannic Cutura]

Dear All,
I am currently facing the following problem and was wondering whether anyone could help. Suppose you have data in the following format:

id     time         ret
1  01jan2000    -5.3
1  01feb2000     4.5
...
1  01dec2000    5.6
1  01jan2001    -5.3
...
1  01dec2001    5.6
2  01jan2000    -5.3
2  01feb2000     4.5
...
2  01dec2000    5.6
2  01jan2001    -5.3
2  01feb2001    -5.3
2  01jun2001    -5.3
2  01jul2001    -5.3
2  01dec2002    5.6
2  01jan2002    -5.3
2  01feb2002    -5.3
....
2  01dec2002    -5.3
....

So essentially the problem is, that there can be missing months, are a series end for example in jun. I would like to "ignore" this problem. So if the series end in june, just use the frist six months to compute the yearly returns. I guess the following approach needs to be taken:

1. create a new time variable that is easier to work with (convert it on monthly first I guess?)
2. and then, "by id year" compute yearly returns

Can anyone point me towards useful commands for this task?

Thanks in advance!

[Carole J. Wilson]

Assuming your time variable is a string, first read help datetimes to change your string time variable to a variable with a Stata date format.

Something like:

gen stata_time=date(time, "DMY")
format stata_time %td

You could create a year varible from this

gen year=year(stata_date)

Now you can calculate yearly returns with:

bysort id year: egen yr_ret=total(ret)