if we run the regression over all the stocks i e if the regression is run successive 620194

In practice, to estimate an asset”s alpha and beta, the following linear regression is usually estimated:

r_j – r_ft = a_j + ß_j(r_mt – r_ft)+ ?_j

where r_j is the asset return, r_ft the risk-free rate, and r_mt the market return.

1. If the CAPM accurately describes the data one is using, should the estimated alpha be zero?
2. If the estimated alpha is statistically different from zero, is the CAPM true or not?
3. If the asset return is that of a mutual fund, and if the alpha is positive, does it mean that the mutual fund beat the market (i.e., outperformed the market)?
4. If we run the regression over all the stocks (i.e., if the regression is run successively for each of the available stocks), what can we say about the average alpha, beta, and idiosyncratic risk?