Let us suppose that the ideal by which all systems are to be tested is that the total taxation shall impose a slightly progressive burden upon all incomes.

1 See Bastable, pp. 388, 389, for the same idea ; also Ross, " A new Canon of Taxation," Political Science Quarterly, VII. This is no very new canon. Adam Smith said : " Taxation should retard as little as possible the growth of wealth."

Then it is necessary to examine all the taxes with their various shiftings and to see how the total shifting affects the final result.1 It is a fundamental principle, too often overlooked in the discussion of incidence, and one that cannot be too frequently restated, that the possibility of shifting depends largely upon the relation of the tax inquestion to the other parts of the system. Let us use an old illustration : the tax on mortgages is in California shifted to the mortgagor because there are other investments for capital that escape taxation. If every possible channel into which capital might go led to the payment of a similar tax, it would not be so possible as it now is to shift this tax. It is not quite true that the tax would not be shifted at all, but it certainly could not be so universally shifted. We shall see later when it may be shifted. If we had a tax system so arranged as to fall, in the first instance, upon all parts of each individual's income, — an unattainable ideal, — there could be but little shifting. But when a part of the nation's wealth is exempt, taxes upon all wealth that might be transferred into this exempt form are peculiarly liable to be shifted.

1 There have been many theories which have undertaken to explain the final incidence of all taxation on some other plan than that of an examination of all the different taxes. Professor Seligman discusses ten different classes of theories in regard to incidence of which only two require this method. See Shifting and Incidence of Taxation.