It is probably a silly question but please assume the following. I gather hedge funds returns and I want to do a regression against few dependent variables (FF factors, etc.). In essence, I get a panel data (multiple years and multiple hedge funds returns with multiple dependent variables. What I do is I merge those hedge funds returns based on some characteristic to get a simple column named "Hedge funds returns" for that characteristic (essentially an average hedge funds returns for that particular month with that particular characteristic). Something like:
Date Commodity_HF_Returns SP500 SMB HML ... 2018-08-31 -0.56 -3.35 0.85 -2.64 2018-09-30 -0.23 0.54 -1.34 2.33 ...
I started with multiple hedge funds that invest in commodities but now it is an average. My question is: If I want to run OLS regression at this point, is it simple OLS or PooledOLS estimator that must be used?