Difficulties associated with deriving AX = X as a necessary condition
from Ay + c being an unbiased estimator of X beta are discussed in ter
ms of the linear model y similar to (X beta, V). We demonstrate two in
stances in which AX = X and c = 0 are not necessary conditions but the
n argue that these can be ignored in practice. Excluding these possibi
lities leads to AX = X and c = 0 being necessary.