Energy accounting and final uses of energy: statistical inconsistencies and open problems