# multi factor equity model exposures not as expected

I'm researching an equity multi factor model.

It contains three factors, say A, B & C. The factors are weighted as such,

         60%        40%
(70% A + 30% B) + C


I am running a back test on this model. When running the model I constrain the risk factors (momentum, beta & size etc) to have a limited exposure. So ideally most of the return should be explained by my model. Looking at the exposures of my factors A, B & C in the last 12 months the exposure of B is much larger than A. I am trying to understand why this might be but not sure where to start?

-
Could you provide more detail? E.g., is the data real or simulated? How do you "constrain" the factors? –  James Sep 12 '14 at 19:06