After many frustrated days without any result to calculate the cumulative abnormal return, I need some help to know how to code the car. As you can see below, I added a picture of the situation. I want to capture the CAR t+4 : t+60. Or in other words, the starting point to calculate the car (AR) is 4 trading days (data) after the earnings announcement (anndats_act) till 60 trading days after the earnings announcement. I already merged the databases that I needed, and created a sample document based on one company with multiple earnings announcement. In this case 2015, and 2016. Otherwise, if I am testing the code, the run will take forever, if it is based on million observations

cusip8 = company code
anndats_act = earnings announcement
data = trading days
AR = Abnormal returns

For example, 24feb2015 is the earnings annoucement (anndats_act) for a certain company (cusip8). I want to capture the cumulative abnormal return (AR) from 27feb2015 (4 trading days after the earnings announcemen) till 60 trading days after earnings announcement. I want to capture the car for multiple companies that contain multiple annual earnings announcements. So, I want to create a new variable like gen car t+4 : t+60 = sum(AR) , by (cusip8 & anndats_act t+4 t+60).

I look forward to hear your suggestions how I can write the code to capture the car t+4 : t+60 for each company and unique earnings announcement.
Array