Capital asset pricing models revisited: Evidence from errors in variables