which is best fitted to promote, or least opposed to, this great end, though it may not press quite equally on different orders of society, is to be preferred to a more equal but otherwise less advantageous tax. The distinguishing characteristic of the best tax is, not that it is most nearly proportioned to the means of individuals, but that it is easily assessed and collected, and is, at the same time, most conducive, all things considered, to the public interests.1

Far from ignoring all ethical considerations, as Bastable suggests, this is a distinct recognition of an ethical principle more definite and more fundamental than any which Bastable himself recognizes in his discussion, or shows any signs of being aware of.

Leaving out of consideration for the present all benefits which the levying and collecting of a tax may confer, such as the suppression of an undesirable industry or the deepening of the taxpayer's interest in the affairs of the state, let us turn our attention to the sacrifices involved. There is, of course, to be considered the direct sacrifice on the part of him who pays a tax. Having his income curtailed by the amount of the tax, his power to consume, or to enjoy the use of wealth, is correspondingly reduced. This form of sacrifice is the most prominent, and has, naturally enough, generally appealed most strongly to writers on taxation. But there is also another form of sacrifice quite as important and fully as worthy of attention. Any tax which represses a desirable industry or form of activity not only imposes a sacrifice on him who pays it, but also upon those who are deprived of the services or the products of the repressed industry. Taxes should therefore be apportioned in such a way as to impose the smallest sum total of sacrifice of these two kinds.

While it is essential that both forms of sacrifice should be considered before reaching any final conclusion as to the best system of taxation, nevertheless the preliminary discussion may be facilitated by first considering them separately. If one were to consider only the first and more direct form of sacrifice, with a view to determining how the total sacrifice of this kind could be reduced to a minimum, he would be driven to conclude in favor of a highly progressive rate of taxation on incomes, with a somewhat

1 Taxation and the Funding System, London, 1845, p. 19.

2 Public Finance, London and N.Y., 1895, p. 314.

higher rate on incomes derived from more permanent sources, such as secure investments, than upon incomes from insecure sources, such as salaries. From the gross income which comes to him in the form of a salary, the receiver must make certain deductions in the way of insurance premiums, e.g., to provide for the future, before he is on a level, in point of well being, with one whose net income comes to him from a permanent investment. The man with a salary of five thousand dollars would be no better off than another with an income of four thousand from a permanent investment, if the former would have to spend one thousand dollars of his salary in life insurance premiums in order to provide as well for his family as the latter's family would be provided for by the investment itself. Under these conditions, the sacrifice involved in the payment of an equal amount to the state would be equal, though the nominal incomes are unequal.

Leaving such matters out of consideration, a highly progressive rate of taxation would be necessary in order to secure the minimum of sacrifice, and for the following reasons. In the first place, a dollar is worth less, generally speaking, to a man with a large income than to a man with a small income, and a dollar taken from the former imposes a smaller sacrifice than a dollar taken from the latter. Moreover, if after the first dollar is taken from the first man, his income is still greater than that of the second man, the taking of a second dollar will occasion him less sacrifice than would the taking of a first dollar from the second man; so that if only two dollars were to be raised, they should both be taken from the first man. Applying this principle rigorously, we should continue taxing the largest income until it was reduced to such a level that the last dollar of the remaining income was worth as much to its owner as the last dollar of the next largest income is worth to its owner, and then only should we begin to tax the latter at all. Then the two should be taxed until they were reduced to a similar level with respect to the third largest, before the third largest is taxed at all, and so on until a sufficient revenue is raised.1

1 For a fuller discussion of this point, see an article by the present writer on "The Ethical Basis of Distribution and its Application to Taxation," in The Annals of the American Academy of Political and Social Science, July, 1895.

Such an application of the principle involves the assumption that wants are equal, which, though obviously not true, approximates more nearly to the truth than any other working assumption that could possibly be invented. Since the state must collect a revenue, it must have some definite assumption upon which it can proceed. The question is not, therefore, whether men's wants are equal, but whether there is any rule of inequality of wants upon which the apportionment of taxes could be made with a nearer approximation to the truth. If there be such a rule, it has not yet been discovered. To assume, for example, that the man whose income is greater than five thousand has correspondingly greater wants than the man whose income is less than five thousand, would be obviously unsafe, because there are even chances that the opposite would be true. Where the chances are even on both sides, it is safer to assume equality. Of a given number of men of the same age and the same general standard of health (by way of illustration) it is obviously untrue to assume that they will all live the same number of years, yet it is nearer the truth to assume that than any other definite workable principle. Consequently the life-insurance company acts justly when it assumes that they will live the same number of years, and apportions their premiums accordingly.

This in no way ignores the fact that wants expand with the opportunity of gratifying them. This objection, however, could only apply at the time when the tax was first imposed. At such a time it would doubtless be true that the five thousandth dollar taken from a man with an income of ten thousand would occasion him a greater sacrifice than the taking of the first dollar from an income of five thousand dollars would occasion its owner. But the reasons for this are twofold. In the first place, by taxing the first man so heavily the state would be depriving him of so many things which he was accustomed to enjoying that by the time the five thousandth dollar was reached, the taking of each particular dollar would be keenly felt. The last dollar of his remaining income would represent a greater utility to him than would the last dollar of the five thousand dollar income to its owner. In the second place, by taxing the second man so lightly as compared with his present taxes, the state would be allowing

him to consume some things to which he had not become accustomed. The taking of the particular dollar in question would not involve a very high sacrifice, for the reason that it would deprive him only of some enjoyment which had not yet entered into his standard of living. But both these reasons would disappear after the new tax had been in operation for a generation, or long enough to bring the standards of living of the two men to the same level.

Drastic as this method of taxation would be, yet, the writer contends, this is the method which would be logically forced upon us if we should adopt the utilitarian test, and should, in applying it, have regard only to the direct sacrifice on the part of those who pay the taxes, ignoring the indirect forms of sacrifice which a system of taxation inevitably imposes. J. S. Mill, who advocated equality of sacrifice as the rule of justice in taxation, was guilty of faulty reasoning on this point, doubtless because he had not made the analysis which subsequent writers have made into the nature of wants and their satisfaction. He was too good a utilitarian to advocate equality of sacrifice if he did not believe that it would involve the least sacrifice on the whole. This is shown by the following quotation, the italics of which are mine.

For what reason ought equality to be the rule in matters of taxation? For the reason that it ought to be so in all affairs of government. As a government ought to make no distinction of persons or classes in the strength of their claims on it, whatever sacrifice it requires from them should be made to bear as nearly as possible with the same pressure upon all, which, it must be observed, is the mode by which the least sacrifice is occasioned on the whole. If any one bears less than his fair share of the burden, some other person must suffer more than his share, and the alleviation to the one is not, caeteris paribus, so great a good to him, as the increased pressure upon another is an evil.1

The last proposition in the above quotation would be true only of persons whose incomes were approximately equal. If A's income is twice as great as B's, or, to state it more accurately, if A's income were enough greater than B's so that a dollar is worth half as much to A as it is to B, then equality of sacrifice would be

1 Principles of Political Economy, bk. v, ch. ii, sec. 2.

secured by making A pay twice as many dollars as B: by collecting $100, for example, from A and $50 from B. But the last dollar of A's remaining income would still be worth less to A than the last dollar of B's remaining income is worth to B: and the last dollar taken from A would occasion him less sacrifice than the last dollar taken from B has occasioned him. Then by taking more than $100, say $110, from A, and less than $50, say $40, from B the same revenue would be raised with a smaller sum total of sacrifice, for the gain to B by this change would be greater than the loss to A. This will appear at once to any one who at all understands the principle of marginal utility. The only conclusion one can draw is that the least sum total of direct sacrifice is secured, not by equality of sacrifice, but by equality of marginal sacrifice. Equality of marginal sacrifice would be secured by so apportioning taxes that, as a general rule, the last dollar collected from one man should impose the same sacrifice as the last dollar collected from any other man, though the total amount collected from each man might impose very unequal total sacrifices.

We are now in a position to test the validity of the minor premise in the argument on page 70: viz., if each individual would voluntarily contribute in proportion to his ability, the whole burden of taxation could be most easily borne i.e., with the minimum of sacrifice. If one's ability is assumed to be measured by one's income, real and potential, and to vary with that income, then the minimum of sacrifice would not be secured by each one's paying according to his ability. If the rich would volunteer to pay more than in proportion to their ability, allowing the poor to pay less than in proportion to their ability, the burden would be more easily borne-i.e., with less sacrifice than if all should pay proportionally. As a statement of individual obligation, even, the faculty theory is untenable, unless modified and defined more rigidly than has yet been done. From the strictly utilitarian standpoint, the individual who measures his obligation to society by his total income is less to be commended than the individual who determines whether he has fulfilled his social obligations by considering, not how much he has given, but how much he has left. The latter type of individual