Hi, I am working with a model of the form : R(t+1) = Summation ( B(ti) * F(ti) ) + ewhere, R(t+1) is a N X T matrix containing returns of N instruments for T time periodsF(ti) are 'i' sets of N X T matrices ,i.e, each F(t) is a N X T matrix containing factor values of N instruments for T time periodsB(ti) are 'i' sets of N X T matrices ,i.e, each F(t) is a N X T matrix containing factor coefficients of N instruments for T time periodsAlso, For each N , Summation [ B(ti) ]= 1 The objective is to find out which out of the 'i' Factors best explain the forward returns.Any help regarding how to set this up in excel/Eviews will be much appeciated.Thanks

You want to use the Fama-MacBeth method. I'm unclear from your description about which of your matrices holds your "factor exposures", but the procedure is to estimate a series of weighted cross-sectional regressions of the returns on the standardized factor exposures, one time period at a time. The coefficients are then returns to 'factor portfolios'. Loop over time, running a cross-sectional regression in each time period, collect the coefficients in a TxK matrix (where K is the number of factors you're working with), and then analyze the significance of the estimated coefficients. See Grinold & Khan pp72-75, and chapter 12, "Information Analysis" for more specific discussion of this. Not sure how you would do this efficiently in Excel, and I haven't used eViews in over a decade. Preferably, you need something like Matlab or R with good numerical capabilities and with which you can do some basic programming.

Thanks Skip, I will try this out.