comparing WLS regressions

[Rutger]

Hi all,

I have a question regarding WLS regressions. I have performed two WLS regression for two subsamples (the first consist domestic mergers and acquisitions (D) and the second consist of cross-border mergers and acquisitions (CB) ) where the weights (using options Weighted LS) are determined by the square root of the number of observations (n^0.5) of each subsample. Next, I regressed the following equation to determine whether CB significantly differs from D;

(CB-D) c SMB HML (rm-rf)

I tried to give the weights for the two subsamples as follows nCB^0.5-nD^0.5 However, as the sample size is equal the coefficient should be the same as the difference between the two individual subsamples but this is not the case :(

I hope someone is able to help me out.....

[Rutger]

Does someone know ho to give the weights for a regression which compares two individual wls regressions where the weights in the individual regressions are determined by the square root of the number of observations??