I am not quite sure I understand your output data: both tables with cumulative returns start 50 days after the first date, but the values are different.
One way to calculate the cumulative returns for many variables could be to use a two-dimensional array:
data want50;
set have;
array returns50 (0:50;1:163) 8 _temporary_;
array returns(*) c1-c163;
array cumulative(*) 8 cr50_c1-cr50_c163;
retain cr50_c1-cr50_C163;
do _I_=1 to dim(returns);
returns50(mod(_N_-1,51),_I_)=returns(_I_);
cumulative(_I_)+returns(_I_);
cumulative(_I_)+-returns50(mod(_N_,51),_I_));
end;
if _N_>=50;
drop _I_ c1-c163;
run;
So, apparently we are adding the day's returns and then subtracting the next day's returns from the cumulative return. How does that work? It works because the next day has not been read yet, so if there is anything there, it is the the return 50 days before that is subtracted. Note that first index of the two-dimensional array is zero-based, so that actual number of elements is 51*163 - as the return that we need to subtract is before the 50-day period.
I renamed your output variables so that they could be named for an array.
To calculate the 200 days cumulative return, change the number 50 to 200, and the number 51 to 201.
