to the randomness of profits, rather than to errors in the measurements of K, which would imply a negative bias in p', this too could not explain away the observed negative as, since again one would expect bias /3' in this case too, to be less than one in absolute value. Thus, the observed negative cannot be explained as the consequence of using a bad cost of capital variable and therefore can be taken as an indication of greater capital-schooling (skill) complementarity.4