This may seem like a silly question, but I have trouble understanding the concept.
I am performing regressions where the dependent variables are raw buy-and-hold returns and abnormal buy-and-hold returns, respectively, while controlling for firm characteristics etc.
The point is obviously to study how my independent variables affect returns. When calculating buy-and-hold returns for a period of 12 months for multiple companies using monthly data, should I end up with one value for the whole year/company or monthly values (that add up month by month)? I.e. should my return variables consist of one return value for each company or multiple?
Thank you in advance!