Public Finance

THE SHIFTING AND INCIDENCE OF TAXATION
*40

BOOK III, CHAPTER V

§ 1. Up to the present we have avoided all but the most incidental mention of the difficult problems connected with the real as opposed to the apparent pressure of taxation. This course has the great convenience of allowing an acquaintance to be made with the leading features and general guiding rules of the system, free from the complications that are inherent in any discussion of the question of incidence. The omission must now be remedied: we have to consider the nature and consequences of the series of processes usually known as shifting or repercussion of taxation, and to study the effects produced by them on the financial organisation. A correct solution of the problem is indispensable for full knowledge of the subject. Our judgments on every part of the tax system will be affected by our theory of incidence. Take,
e.g. the question of justice. How can we say that any given arrangement of taxation is fair unless we know its real, not merely its apparent, incidence? The extent and limits of the shifting of taxation are elements in estimating the expediency of exempting the minimum of subsistence, of imposing a progressive income-tax, or of taxing articles of consumption. Instead of confining our attention to surface appearances, we must examine the underlying conditions, and estimate in their entirety the effects of fiscal regulations.

This complicated investigation will occupy the present chapter. We shall seek to establish the general laws of repercussion and their most important results, reserving the treatment of the several taxes for the appropriate place.
*41

Here a question as to the proper use of terms arises. Most writers distinguish between the ‘incidence’ and the ‘effects’ of taxation, employing the former to denote the falling of the actual burden, and reserving the latter for the various economic results. This is the position of Professors Seligman and Adams. But the latter at least appears to limit the burden to the actual payment to the State.
*42 It seems more in accordance with language and principle to treat the loss incurred by the citizen as coming under the head of incidence even though the State does not obtain revenue. The burden of many taxes is greater than their yield, but it is hardly admissible to cut up this burden into two parts, one, measured by state receipts, belonging to the topic of incidence, while the balance is treated as the ‘effect’ of taxation. Mr. Cannan proposes the heroic measure of discarding the term ‘incidence’ altogether.
*43 But the expression is far too well established, and also far too convenient to be thus summarily abandoned.

§ 2. Popular discussion of financial matters has always given a large place to this special topic. The real incidence of tithes, of import duties, and of local rates has been hotly debated at many a dinner-table and in many a tavern, and very positive conclusions have been reached in entire ignorance of the grave difficulties that surround any attempt to determine accurately these and similar points. It takes some training to see that confident decisions as to the division of rates between landlord and tenant, or of duties between producer and consumer, cannot be made in a ready and off-hand way. Such is, however, the case. The complications are too great; the subtle modes in which pressure applied at one point is diffused over a wider area are too hard to be followed without a clear appreciation of the general conditions and a careful use of the slippery instrument of abstract deduction. In dealing with the problems of incidence we are at that part of finance that touches most closely on economic theory in its hardest form.

Scientific students have long recognised the fact, and the earliest efforts of financial inquiry have been directed to the question of the incidence of taxation. In some instances it was apparent that duties temporarily paid by the seller of a commodity were only advanced by him, to be obtained later on from the buyer in the form of increased price.
*44 The extension of this result to all cases of taxation on producers or dealers was so plausible that it became an accepted doctrine of practical finance before passing into a theoretical form. This particular position dealt only with readily observed facts, and was confined to outward and apparent effects of taxation. A far more important step was made when the fruitful idea of a single source from which all taxes must, in the last resort, come suggested itself. The doctrine found a definite expression in Locke’s statement ‘that taxes however contrived, and out of whose hands soever immediately taken, do in a country where their great fund is inland for the most part terminate upon land.’
*45 Put forward tentatively by Locke in connexion with his controversy as to interest, the conception of land as the true source of taxation was made the basis of a practical plan by Vanderlint,
*46 and more fully developed by the Physiocrats as a part of their view of the ‘net product.’

This earliest scientific theory of shifting rested on certain general assumptions, some of which we have already noticed.
*47 They are (1) that wages are at the minimum point of subsistance, (2) that taxation of profits will drive capital out of industry, and by thus reducing the supply raise the rate to its former point, the tax excluded, and (3) that expenditure for consumption is simply the employment of income, and that increased prices through taxation will compel a compensatory increase of money wages and profits. By aid of these propositions it was easy to establish that taxation of wages or profits or their outlay must be passed on to some other class in society. The gains from commerce and professional avocations were brought under the same principle by the ingenious argument that they also were necessary to secure the continuance of the particular trade or profession.
*48

We thus obtain a rigorously complete theory of incidence accompanied by a description of the mode in which the shifting is carried out from the points of initial pressure to the ultimate resting-place. The equalising agency of competition and the necessity for normal wages and profits are the forces that push the burden of taxation on to the landowner’s revenue. This conception of society as a mechanism in which strains were distributed in obedience to general laws, quite independent of legislation or intention, was thoroughly grasped by the physiocratic school, and was applied by them to the incidence of taxation as well as to economic distribution in general.
*49

The contrast so often noticed between Adam Smith and his French contemporaries, appears clearly in his treatment of the question of incidence.
*50 The sharp and definite theory that regards all taxation except that on rent as necessarily shifted is changed into the broader doctrine that transference may or may not take place according to circumstances, and may fall on any one of the three constituents of income. In the application of these general positions several qualifications are introduced. Taxes on wages are always, Adam Smith holds, transferred, partly to the consumer in higher price, partly to the landlord in lower rent. The employer must have his ordinary profit, and he recoups himself for his larger wages’ bill by increasing his sale price, or, if a farmer, by giving his landlord a smaller amount of produce. The share of profit known as employers’ gain is also unamenable to taxation, as being merely the necessary reward of the entrepreneur.
*51 Interest, though capable of bearing some of the public charges, is difficult to estimate, and its root, capital, is apt to emigrate when placed under exceptional charges. Taxes on necessary articles of consumption tend to raise the wages of labour, and therefore are, like direct wages taxes, passed on to the consumer or the landlord. A house-tax tends to fall partly on the occupier, partly on the ground-landlord, the builder in the long run always receiving his normal profits.

The result of the inquiries on incidence in the
Wealth of Nations is a modification of the
Économistes’ view. Though the landlord is still the chief bearer of public charges which are shifted on to him from various points, while his special burdens are not transferable, he is not the sole bearer: the capitalist has to contribute a share, and the vague class of consumers has to pay on the taxed forms of expenditure which may come from rent, profit, or even (in the case of taxes on luxuries) wages. The landlord has, nevertheless, as Falck remarks, the ‘lion’s share of the payment of taxes,’ and therefore in part Adam Smith occupies the physiocratic position.
*52

The title of Ricardo’s treatise marks taxation as one of its subjects, and it may be said that the space devoted to that topic is altogether occupied with the question of incidence. Adam Smith’s positions are corrected in the light of the newer theories of population and rent. In fact, Ricardo’s doctrine of taxation is a development of his theory of distribution. Notwithstanding the generally loose form of his writings, there is an amount of precision about his statements as to the movement of taxes that has made him the leading representative of economic orthodoxy on these points. Reduced to a definite form his views are represented in the following propositions, resting, it must be noted, on the assumptions of (1) self-interest as the motive power of action, and (2) the complete mobility—at least within the same country—of labour and capital.

(1) A tax on economic rent is not transferable; it remains on the landlord. (2) A tax on the wages of ordinary labour is transferred to the employer, and is in reality a tax on profits. (3) A tax on profits in general is not transferable, and must remain on the capitalist; but (4) a tax on the profits of a particular employment will be transferred to the consumers of its product. (5) Taxes on commodities paid by the producer are passed on to the consumer, as in the case of taxes on particular profits. (6) In the case of commodities consumed by the labourers there is a further shifting from the consuming labourers to the capitalists who employ them.
*53

The main outlines of this theory of incidence reappear in J. S. Mill’s
Principles, with some not unimportant amendments. For example, the higher classes of wage-earners are admitted as possible bearers of taxation. In their case a tax on wages may or may not be shifted. The results as to tithes and profits are somewhat altered, and greater emphasis is laid on the tendency of profits to a minimum. But these are special points: speaking broadly, there is no part of Mill’s work which so fully deserves the description ‘a readable Ricardo’
*54 as that which deals with taxation.
*55

§ 3. The very general adoption of the Ricardian theory in its developed form as the sole and exclusive scientific doctrine makes it advisable to note some of the objections that prevent us from accepting it as a complete interpretation of the phenomena of incidence.
*56 Some of these criticisms have been forcibly urged by Cliffe Leslie and Held, but they may be put in a more general form. The first weak point in Ricardo’s position is his ambiguous treatment of consumption and consumers. In his general scheme of distribution there is no place for the consumers as a class, but we often find him asserting that a given tax does fall on ‘the consumer.’ How are we to explain this apparent discrepancy? The most natural answer is that landlords and capitalists make up this class, the labourers being normally outside it, as their consumption is a part of the expenses of production. This explanation is not completely satisfactory, for it is plain that all landlords and capitalists are not affected by particular taxes falling on consumers, and yet no criterion for distinction is suggested. The necessity for studying the forms of expenditure as a department of economics seems clear from this consideration. Besides the pressure that falls on the primary divisions of income, there is the additional one on the employment of that income, and differences in its employment produce differences in pressure. A doctrine of incidence that is confined to the receipt of income without regard to its expenditure is so far defective. A second objection to the theory is its dependence on a few unduly simplified conditions. Social forces are regarded as definite and precise in their action, and very positive statements are made on the strength of this insecure foundation. Thus taxes on wages and on labourers’ necessaries are regarded as being always transferred to the capitalist, a proposition true only on the assumption that wages are at the minimum, and that any change in them will at once act, and act proportionately, on population. In the same way the equality of profits and the complete dependence of rent on the margin of cultivation are made to support very sweeping propositions as to the incidence of taxation. If we allow that the economic forces of population, of competition in regard to the employment of capital, and the movement of cultivation are not quite so regular in their action, the deductions made from them must be qualified. Thirdly, the theory exaggerates the simplicity not merely of economic forces, but also of the forms of taxation. Taxes on rent, on profits, or on wages are not all the same, and the particular mode of imposition often affects the incidence. There is need for much care and discrimination in using those results of deduction that depend on the identification of so many different taxes. Finally, far too little notice is taken of the actual facts and of the unavoidable limitations in the application of theoretical principles. The orthodox theory of incidence professes to explain what will happen in the long run, ‘but taxes,’ as Leslie well remarked, ‘are paid immediately under the real conditions of life and out of the actual wages and profits or other funds of individuals, not out of hypotheses or abstractions in the minds of economists.’
*57 Knowledge of what will happen when the limit is reached is, no doubt, desirable, but what takes place during the process of adjustment should also be noticed.

The existence of these imperfections does not destroy the great service of the doctrine as a preliminary or introductory inquiry (
Vorstudium). Without some such attempt the intricacies of incidence could never be explained, and it is by expansion and correction of the Ricardian procedure that advance can best be made in the explanation of these problems. As an intellectual exercise the abstract theory of the shifting of taxation has a high disciplinary value. The root-error of its followers lay in taking a part for the whole.

§ 4. The course of development in the preceding theories is clear enough. From the first suggestion of Locke to the compact exposition of J. S. Mill there can be traced a series of connecting links and alterations in consequence of wider knowledge. Adam Smith has the French theory constantly in mind, as Ricardo in turn has the ideas of the
Wealth of Nations. All have in common the recognition of certain points on which the pressure of taxation ultimately rests, and all, it may be added, suggest the wisdom of adapting legislation to the conditions of incidence in order to secure a fair distribution, or at least to prevent unnecessary waste through friction.

Another group of doctrines has a quite different tendency. In place of investigating the complicated shiftings that settle the ultimate incidence, it either denies the possibility of ascertaining them, or assumes that they bring about a general diffusion of the burden over the whole society. The natural conclusion is that the particular forms of taxation are altogether immaterial, as, whatever be the immediate charges, the burden is finally distributed in an equitable, or inequitable, manner.

The first scientific statement of this view is ascribed to N. F. Canard, whose essay, attacking the theory that all taxation must fall on the owners of land, obtained a prize from the French Academy. The gist of his argument is that there is surplus product in labour and commerce as well as land, and that taxation falls on all of these ‘net products.’ The process of diffusion is carried out by exchange, buyer and seller in each transaction dividing the amount of tax imposed, and at every fresh exchange a division of the part of taxation transferred takes place until ultimately the charge is spread over all the parties concerned. Extending this conception to the whole society, taxation comes to be regarded as after a time diffused equably among all its members. The qualification as to time is important, for the process of diffusion is not complete at first; consequently old taxes are the best, and all new taxes, or even changes in existing ones, are to be deprecated as disturbing the beautiful and harmonious distribution which relieves the legislator of any trouble respecting the apparent merits or demerits of existing taxes.
*58

The comfortable nature of this theory has made it a popular one. Without adopting Canard’s peculiar explanation of the mode of diffusion, Thiers asserts the general diffusion of the public charges; Stein, from a still different point of view, reaches what is practically the same result. In his opinion the whole theory of shifting is an error arising from imperfect comprehension of the real nature of the process, which in reality contains two different parts. For, first, a tax is a part of the cost of production similar to the expense of raw materials or labour, and like other expenses must enter into price, and taxation is through this medium ‘diffused from one to another’ until it extends to all. Again, from a higher point of view, the portion of product paid in taxes is a surplus product, the result of the services of the state administration, which pays back at least what it receives. The conception of incidence of taxation has to be replaced by that of the ‘production of taxes.’
*59

This theory has also had a good deal of vogue in England amongst members of Parliament and officials, but is often held along with other and inconsistent theories. Thus, Sir E. W. Hamilton, after discussing at considerable length the incidence of certain taxes, finally comforts himself with the reflection that ‘perhaps there is more truth than is popularly supposed in the optimistic theory of general diffusion, which is that “taxes equate and diffuse themselves, and if levied with certainty and uniformity they will, by a diffusion and repercussion, reach and burden all property with unerring certainty and equality.” ‘
*60 Sir R. Giffen expresses his full agreement ‘in the opinion that all old taxation tends to become equally diffused over the whole community.’
*61 So Lord Avebury makes the same quotation as Sir E. W. Hamilton, and approves of it in similar terms.
*62 This does not hinder him from asserting that ‘the Commissioners make out a very strong case for further relief to real property, especially after the additional burden thrown on it by Sir W. Harcourt’s budget.’
*63

Closely allied to belief in the theory of diffusion is the disposition to regard the problem of incidence as insoluble, and at the same time to treat all questions of taxation as if it were non-existent.
*64 To deny that the incidence of taxation is discoverable seems to be the first step towards believing that it is unimportant.
*65

No lengthy criticism of the negative theory of incidence is needed.
*66 Facts speak for themselves: if the incidence of the public burdens be really so settled that legislative action has no effect, how comes it to pass that some forms of taxation are much more oppressive than others? It is impossible to escape entirely the weight of a load by judicious arrangement of it, but it is quite feasible to diminish the fatigue it produces. Canard’s doctrine is contrary to experience, and is not established by abstract reasoning. There can be no doubt that taxes are not always a part of cost of production, but in any case the real question is whether they can always be shifted by the immediate payer, and to this the answer must be a decided negative. The desire to pass on the burden may be universal, but the capacity to do so is limited. Even in the special case of taxes on commodities it is not always open to the producer to shift the duty to the consumer. As regards other taxes, the very idea of cost of production is quite inapplicable.

§ 5. In proceeding to examine more closely the conditions of our problem, it is well to remember that the total denial of the existence of shifting and the assertion of its universal existence are both unfounded. That,
e.g. taxes on commodities are sometimes transferred by the immediate payers is an obvious fact. No one can believe that the distillers bear the whole burden of the English spirit duties. On the other hand, it is just as incredible that a landlord could entirely shift a tax on rent to his tenant or any other class, or that the payers of income tax could completely relieve themselves at the expense of others. The existence and the extent of the process of transference must depend largely on the conditions of the case, and it is these conditions that a general theory of incidence has to consider and explain.

The course of transference may be in different directions, and according to its starting-point and direction it is necessary to distinguish. Where the movement is from the producer to the consumer, or, more generally, from the seller to the buyer, there is ‘forward shifting’; where it is from the buyer to the seller, there is ‘backward shifting’; where the process of shifting affects more than two parties, it leads to ‘diffused incidence.’
*67

The simplest instance of shifting is, as more than once mentioned, that in which the producer of a commodity passes on the charge in increased price to the consumer. A closer examination of this familiar case will suggest some important conditions. Why does the buyer submit to this additional charge? An increase in price tends to reduce demand, and will not the falling off bring about a return to the old level? The usual reply would be that the dealer or producer had been obtaining normal profits before the imposition of the tax, and that without an increased return sufficient to compensate for the new charge he would not, or could not, continue in the business. The doctrine of average profits resulting from the effective mobility of capital is thus the foundation of the proposition that taxes on commodities levied on the producer are shifted to the consumer. The reason for the proposition also shows its limitations; wherever an industry is yielding such exceptional gains that taxation will not reduce them below the supposed normal level, the motive for abandoning the employment not being present, the force that produces shifting will not be in operation. It may therefore be allowed that, so far as producers’ gains are at all of the nature of monopoly, taxation will specially affect them. But here a further qualification is presented. Exceptional gains may be made by some producers, but not by others; in fact in nearly every industry some of those engaged in it can barely hold their ground. This unfortunate class must, on the increase of taxation, either raise their price or leave their business; if they can succeed in the former attempt, their successful competitors will gain by it, and shift their charge to the consumer; if they fail, the margin of pure profit is raised, and the burden will remain on the producers. It is possible, and indeed likely, that the actual result may be a compromise. Some of the weakest producers may be driven out, but the price may also be somewhat raised, in which case there will be division of the charge. Therefore the true conclusion is, that when there is complete mobility taxation will be shifted from producer to consumer. Again, the possibility of shifting taxation of the kind under notice does not depend simply on the elasticity of supply; the effect that changes in price will produce on demand must be considered. Taxation imposed on a necessary article, or one which forms a very small part of the total outlay of the consumer, will, since demand is inelastic, be more likely to pass on at once to the consumer than if the commodity belonged to that large intermediate class, the demand for which is speedily checked or increased by an upward or downward movement of price.
*68 Again, if increased expenditure has to be devoted to taxed articles, less remains to be applied in the purchase of other goods; the consequent reduction may lower prices in other industries from which withdrawal is not economically expedient, and accordingly diffuse the indirect incidence of the tax to a different set of producers. Further, it must be remarked that as all industrial processes are really complex, it is quite possible that a tax may not affect the holder of floating capital who is ready to seek other investments, but may fall entirely on the owners of land, or specialised capital suitable for the production of the article. Both land and fixed capital are indeed capable of different uses, but the alternative ones are necessarily less profitable than that in which they are actually engaged. Hindrances to mobility are hindrances to shifting of taxation. The very application of a tax of itself produces disturbing effects. Additional capital has to be employed, restrictions, which mean the sacrifice of time and money, come into force, both tending to reduce the number of producers and to concentrate industry. The production approaches to a qualified monopoly, and thus the weight of taxation falls, so far as actual receipts are concerned, on the consumers, with a further loss to the small producers excluded from the business.

The case of a strict monopoly is of sufficient theoretical importance to receive some special notice. Starting from the admitted fact that the normal monopolist endeavours to make his net return as high as possible, it follows that a tax on the commodity that he produces will, by increasing his expenses of production, tend to reduce his net receipts, but whether the whole or even the greater part of the tax will be borne by him or by the consumers will depend on (
a) the conditions of demand, and also (
b) on those of supply. If a slight rise of price seriously checks consumption, or, in other words, if the demand is elastic, the monopolist suffers more than in the case of inelastic demand. Again (
b) if the condition of diminishing return operates, the tax may, and probably will, be in part compensated by the cheaper production of the marginal portion of the reduced supply. The condition of increasing return makes a tax more oppressive, since the cost will rise with contraction of supply.

But the theoretical conception of a pure monopoly is of little direct service in dealing with the question of incidence; for in very few cases is a monopoly strictly so called to be found. There is in truth rather a number of instances of limited or qualified monopolies, arising in part from natural, in part from legal limitations. Both monopoly and competition have to be considered, and in particular the interaction of these opposed conditions as well as the effect of financial changes in readjusting these areas.
*69

The precise method of taxation employed will have an important influence; whether the duty be imposed at an early stage or allowed to lie over till the article is ready for the consumer; whether the measure adopted is supposed capacity of production or actual product are very material circumstances in deciding the exact incidence.

Thus the apparently simple case of taxation of commodities appears to be really surrounded with complications that need close and careful study. The same questions would arise if the tax were levied directly from the consumer; there would be the possibility of a backward shifting, just as there is of failure of the forward one. In fact, as the position is sometimes explained, there is a struggle between producer and consumer, each striving to throw the loss on the other, and much will depend on the relative strength of the parties. As a rule producers are a smaller and better organised class, and therefore have the chances in their layout, though where they possess any differential gain, this advantage is lost to them. Fresh increases of taxation are passed on to the consumer more readily than reductions are restored to him. This element of friction has another effect. Small additional amounts of taxation are not easily shifted; a few pence on or off the gallon of spirits cannot directly influence retail price. The initial shifting always implies an effort, which, however, very readily takes place in industries accustomed to alter prices as the various expenses of production change. Additional taxation and a rise in the price of hops are events of exactly the same kind to the brewer, and their final result is distributed in the same way.

The diffusion of the burden may be still more complicated. In modern society products pass through the hands of several distinct classes before reaching the consumer, and the struggle of buyer and seller will be repeated at each separate stage. The existence of monopoly or of some form of limitation at any point may prevent the shifting passing any further. An economically strong intermediate group may throw a charge back to the producer, send it forward to the consumer, or divide it between both.

The foregoing analysis of the actions and reactions that may accompany or result from the imposition of a tax on a commodity shows the general conditions that are influential.
*70 They are (1) the presence or absence of mobility; in the former case, the normal shifting to the consumer will take place; in the latter it is retarded: (2) the law of demand for the particular commodity; on this depends very much the extent to which there will be a reflection of the burden either back to the producer or to other industries: (3) the presence or absence of monopoly: (4) the method of taxation as affecting the preceding conditions: (5) the organisation of the industry and its division, and (6) the amount of taxation. In regard to this last circumstance, it may happen that additional taxation will increase the force of competition. The new element may be just the last thing wanting to break up the existing settled conditions. This will be easily understood by considering the effect of successive very small additions to the duty on a given article. Each of these will tend to remain on the payer, but as soon as the additions are sensible, or easily distributed, the shifting movement will begin to act. Even in the case of the most rigid and gainful monopoly, the producer must, if taxation be carried sufficiently far, either pass on the weight or abandon the undertaking.

§ 6. The comparatively easy case of a tax on goods enables us to perceive the general character of the changes in incidence produced by the process of shifting. We have now to deal with the more important and interesting question of the movements of incidence in respect to the incomes of the different economic classes. The whole tendency of modern economic science has been in the direction of emphasising the fundamental similarity between the departments of exchange and distribution.
*71 Rent, wages, interest, and employers’ earnings are exhibited as the prices of the respective services of land, labour, capital, and business ability. Might we not say that a tax on any of these commodities would be amenable to the same reasoning as that already applied to material goods, the consumers of each of the factors of production being those other factors that need its co-operation? This mode of treatment is, we believe, unsuitable, owing to the distinguishing pecularities of the shares in distribution. Their production is not in the same form or subject to the same conditions as that of ordinary commodities. Nor is the nature of demand the same in respect to them. The attempt to bring commodities and services under a common heading seems to be an undue straining of the analogy that undoubtedly exists. A better mode of dealing with the question is rather to consider it in the light of the theory of distribution, while availing ourselves of whatever is applicable in the case of taxation of material commodities.

There is no need for attempting here to re-state the economic theory of distribution. The work of Ricardo has been filled in and placed in closer relation to actual conditions by the ablest workers of the past and present generations,
*72 who have carefully elaborated the originally fragmentary doctrines on the subject. The main conditions affecting changes in distribution must, however, be noticed; for the effect of taxation is plainly a deduction from the total produce—
i.e. so much loss to be re-distributed among the parties concerned.
*73

Assuming competition, the main circumstances on which the amount of rent depends are the position of the margin of cultivation, and the several qualities of land that lie above it. Change either the worst land in cultivation, or the relation to it of the superior soils and the quantity of rent will be altered.
*74 In estimating the incidence of a tax on rent, its effect on these conditions is the first consideration. The usual way of showing that a tax on rent cannot be shifted is to point out that it does not affect that particular land that pays no rent, and consequently leaves the determination of the total amount, including the tax, as before. Ricardo and some of his nearest disciples differed as to the incidence of tithes or proportional taxes on the raw produce of land. The former maintained that such a tax must fall on the consumer, since in the case of produce from the worst land in cultivation there was no rent on which it could be placed, and it was the yield from this land that determined price; as the cultivator would need his average profit, the shifting to the consumer was necessary. Senior and McCulloch, on the other hand, held that the rise of price would check demand, and therefore by changing the position of the margin in an upward direction would reduce rent.
*75Without discussing the special point at issue, which belongs to the group of land taxes,
*76 we see that the criterion used by both is the effect on the general condition of agricultural industry. That on the hypothesis of perfect competition a tax on rent must remain on the payers is an indisputable truth, but for the cases of actual taxation it is important to bear in mind that economic rent is mixed up with other elements. The investment of capital in land yields a return in many instances indistinguishable from economic rent, but at the same time it is ‘really the profits of the landlord’s stock.’
*77 So far as no discrimination can be made between these components, the incidence of a tax will fall to some extent on the return to capital, and, if sufficient to discourage its investments, will tend to be passed on to the consumers of agricultural products, since land of inferior natural quality must be cultivated in order to supply the required quantity.

The opposite cases of taxes imposed elsewhere falling on rent is much more probable. The class of differential gains of which rent is one very conspicuous instance is peculiarly liable to be affected by taxation. The influence of competition is, speaking generally, effective in distributing special burdens on a particular industry; but where special gains have been obtained an equivalent tax is the restoration, not the destruction, of equality. This is the kernel of truth in the Physiocrats’ belief, and on it their exaggerated doctrine was based. No kind of actual tax can be imagined which might not under certain conditions diminish the fund that goes to the landowner. Wages, interest, employers receipts, duties on goods, or on acts, all supply such examples, and they all accomplish their effect by operating on the margin of cultivation in the widest sense. The complicated working of the tax-system is very well shown by this circumstance. It is, as we discovered, very difficult to single out differential gains for exclusive or extra taxation, but the ordinary agencies of economic life are tending to that object, though of course in a very limited and imperfect way. They strike alike the earned and the unearned increment, the investment profitable through the foresight of the prudent employer and the lucky chance of the rash speculator, the rents of careful and improving as well as of inattentive and tyrannical landlords.

§ 7. Taxation of the capitalist’s share of the national income gives rise to more difficult problems than those connected with rent. Between the doctrine of Turgot, that a tax on profits is always, and that of Ricardo, that it is never, shifted,
*78 we have to take an intermediate position. A general tax on interest, as it affects all employments equally, would appear certain to remain on the payers. The mobility of capital cannot here, so long as we confine our attention to a single country, have any effect. Where the tax does not extend to capital invested abroad it is evident that it would discourage home investments and lead to the emigration of wealth to other places. ‘The proprietor of stock,’ as Adam Smith tells us, ‘is properly a citizen of the world, and is not necessarily attached to any particular country.’
*79 Even within a limited area another feature of capital will affect the incidence of special taxes imposed on it. Unlike land, it can be indefinitely increased by human foresight and providence, having as a chief inducement the return to be obtained by investment. Taxation on interest lowers that return, and is therefore a direct discouragement to saving.
*80 So far as it is effectual, the diminution in the supply of available capital tends to raise the rate of interest and transfer the incidence to the consumers of capital,
i.e. the other factors in distribution, and as rent is not likely to be much affected, in reality to the producers, including both employers and labourers. How far the check to production will show itself in a higher cost of production and therefore fall on the consumer is not easily determinable; if there were to be a substantial check to the investment of capital this would be a probable result, causing a diffusion of the incidence, some of it returning to the capitalist in his capacity of consumer.

For most purposes of economic reasoning there is an advantage in neglecting the differences between the different forms of capital and dealing solely with the characteristics common to all. But in handling the problem of incidence, it is necessary to see that there are two broad classes of reproductive wealth, the one free and capable of being turned in any direction, the other fixed in some particular industry. It is primarily to the former that the arguments from the mobility of capital apply. Capital once invested, the difficulty of withdrawing it places the possessor for the time being in the same position as the landlord. A tax on fixed capital would thus seem to resemble in its effects a tax on rent, and to be equally untransferable. One instance—that of land improvements—has been already discussed, and in considering it we saw that the mode of relief to the capitalist was simply by reducing future investments. The single tax on fixed capital in the sense used by Menier and his followers would be at first a heavy burden on the owners of those forms of wealth, that would show itself later on in reduced investment and retarded production. Free capital, if separately taxed, has much readier modes of escape. Employment outside the particular tax area makes it very difficult, even if the law enacts it, to enforce collection; consequently the chance of placing an effective tax on movable capital is much reduced by both economical and technical circumstances.

The chief condition, then, on which the incidence of a tax on interest depends is its effect on the accumulation and investment of capital, including its action on the saving propensities of the inhabitants and their disposition to move their wealth to escape taxation. If the rate of interest is determined by what Jevons calls the ‘final utility’ of capital, it is plain that the possibility of shifting the tax will depend on the effect produced on this margin of investment. If it is forced up the weight will be transferred from the recipients of interest to that intermediate class which gains by the cheapness of capital.
*81 A tax on the returns of fixed capital will at first rest on the payers, and only be transferred with difficulty, but it will ultimately, when the old supply is sufficiently contracted, come under the same influences.
*82

Mixed up with the interest of capital in Ricardo’s treatment of taxes is that element of profit variously described as ‘wages of superintendence,’ ‘earnings of management,’ or ’employers’ gain.’ It has, however, strong claims to separate treatment. The profit of the entrepreneur has some points of resemblance with wages, as it has others with rent, and we must therefore be prepared to find that the movement of taxation is different in its case from that of ordinary interest. The analogy of rent would lead us to believe that a tax on the gains of the employer could not be transferred, since there would be no opportunity for escape on the part of the immediate bearers. A tax on this very indeterminate element of the gross profit of business would, however, be certain in practice to trench in some degree on the other constituents. It is almost impossible to avoid levying such a tax on that minimum ’employer’s return’ which is sometimes regarded as equivalent to ‘no profits.’ The struggling marginal producer will then need an increased price in order to recoup himself for the tax, and unless he obtains it will have to yield to the pressure of the ‘last straw,’ and therefore abandon his business. Taxation of this kind would operate somewhat as taxation of commodities. It may be urged that when the gains of all industries are taxed there is no reason for the weaker employers giving up business. They can, however, pass down to the class of labourers, as others by taxation may be hindered from leaving it.
*83 The effect on the marginal employer appears as the condition determining the shifting of taxes on the employer’s gain.

This share of national income may also suffer through the operation of taxes on commodities. When such a duty is not transferred to the consumer the burden is likely to fall on the differential element in profits; the tax has to be paid without the compensation of a rise in price, and there is no way of shifting the burden, unless in the case of raw materials, where rent may be curtailed. It is quite in accordance with the analogous case of rent that taxes should be shifted to the peculiar gains of the employer. It is, besides, possible that a tax on interest may be transferred to profit in the limited sense. When the rate of interest is raised, as we have seen that it may be, by taxation, the employer has to pay dearer for his borrowed capital, and, so far as what he works with is his own, loses on one hand what he gains with the other. On the whole we may confidently say that the broad and simple statement that taxes on profits fall on the capitalist, who can in no wise transfer them to others, requires to be very much limited before it can be accepted as correct. We must separate the two essentially different elements of interest and employer’s gain, and recognise that while the one is affected by changes in the point of final utility of capital, the other is connected with the opportunity for profitable industrial effort.

§ 8. If the older theories on the subject of incidence assumed too hastily that rent and profits had to bear their own immediate burdens and under certain conditions those of others, they made amends when dealing with wages. The transference of taxes on this part of revenue was asserted in the most positive manner.
*84 The landlord, the capitalist, the consumer might all be affected by a tax on wages, but the labourer was always exempt from contributing to the requirements of the State. This immunity was believed to extend to the higher kinds of wages and salaries, since they had a fixed relation to the ordinary rate.
*85

The historical explanation of this belief is afforded by the evolution of the system of hired labour from the earlier condition of serfdom. The slave as an instrument of production received what was needed for his maintenance; any reduction in its amount would reduce his efficiency. Taxation was paid altogether out of his master’s income, it did not concern the living machines engaged in the creation of wealth. This conception survived in the earlier period of free labour, and gained support from the doctrine of a ‘natural’ rate of wages common to the French and English economists. Any reduction in the rate would, it was held, act on population, and by diminishing its number restore the former real reward of the labourer. In spite of occasional concessions, such was the opinion of Turgot, of Adam Smith, and of Ricardo,
*86 and, given the premise, the conclusion was sound enough. It is also true that in both the France of the eighteenth century and the England of the Napoleonic war wages did seem to have touched the subsistence point, and to give a direct verification of the economical doctrine. But, strange to say, no notice was taken of the fact that one of the causes of this deplorable situation was the heavy and unequal pressure of taxation. The French peasant and the English labourer were the greatest sufferers by the fiscal systems under which they lived, and financial reform was one of the means of their relief.

No account of the incidence of a tax on wages can be satisfactory that does not fully recognise the existence of varying standards of comfort, even among the lowest unskilled labourers at different times and places. Beside and above the physical minimum, there is what Mill calls a ‘moral’ minimum. The conditions must be exceptionally unfortunate that do not allow the labourer something above mere subsistence, and, when that minimum point is exceeded, there is something on which taxation may fall. To estimate the incidence of taxation we must know its effect on the standard of life. If that is maintained unaltered there will be a transference of the tax to the capitalist or employer; if it is lowered the labourer bears the tax himself. This consideration applies to each industrial grade, but it is evident that the higher the usual scale is placed the less is the probability that it will be readjusted to suit taxation. When a group of labourers possesses a monopoly, it, in common with all holders of differential gains, has no power to throw off the burdens imposed on it, and, as most skilled labourers have more or less special advantage, the shifting of taxes is in their case beset with difficulty.

Thus, as in the case of rent, interest, and profit, we find that the ultimate incidence of a tax on wages will vary according to the special conditions under which it is imposed; and of these the most important are the effect on the usual standard of living, and, so far as the higher kinds of wages are concerned, the extent to which their receivers are privileged through natural or artificial causes. Peculiar gains of labour are just as much at the mercy of fresh taxation as any other differential advantage. The process of shifting requires the actual exertion of force to carry it out, and those forces can only be the agencies that work through supply and demand. If the same supply of labour of any particular kind is forthcoming with an unchanged demand, then direct taxation of labour will not be transferred. The great difficulty of adjusting the supply of labour is a reason for believing that any shifting of taxes imposed on it must be a slow and uncertain process.

A similar conclusion applies to the case of taxes on the labourer’s consumption. We do not find that duties on food produce higher wages; they only bring the starvation point nearer, as the history of the English Corn Laws shows. When Ricardo argued that taxes on articles of the labourer’s consumption are exactly the same as a tax on profits he assumed far too rigid a connexion between the cost of living and the supply of labour. A tax on the food of animals used in production would increase cost, because food so given is regulated to secure efficiency, but the labourer seeks to procure the best terms for himself. The element of free contract present in the latter case entirely alters the position. For completeness of statement it is desirable to add that a great deal of wages is really the return on capital invested in the education of the workers, but in reality this does not produce as much practical effect as might be expected. A tax on wages, unaccompanied by an equivalent tax on the yield of material capital, would apparently discourage expenditure in the formation of personal or immaterial capital, and turn it towards the production of goods. This check to the supply of trained workers would tend to raise the price of their services, and shift the pressure to the consumers of the goods produced by them or to the employers. In practice the calculations of parents and others who make the investment in the education of the young are not so carefully worked out as to be influenced by the existence of a tax on the wages of the higher employments. Still, even with the actual imperfect estimates, some effect would probably be traceable if the tax were a heavy one. The necessary expenses of living in a suitable way and the cost of training are the two agencies that give some justification for Adam Smith’s doctrine of a balance not to be disturbed by taxation between the different employments.

§ 9. Our examination of the general conditions that help to determine the true incidence of taxes on the different constituents of income, though necessarily brief, at least makes it plain that the theoretical explanation of the subject is not the simple process sometimes imagined. The movement of a given tax is not invariably in the same direction: its course will be guided by the surrounding circumstances. Without knowing what these are we cannot tell the direction, much less the precise extent, of its incidence. To pretend to say where,
e.g. a tax on profits will fall, without possessing further data, is as vain as to seek to determine the space traversed by a moving body whose initial velocity and period of movement are both unknown.

The difficulty of estimating the incidence of taxation is increased by the complementary alterations that take place in the economic system. A change in rent implies changes in the amount and probably in the relations of the other shares in distribution; a rise or fall in the price of one article leads to other changes of price, and we may therefore expect that even in the most precise and determinate cases of incidence some additional diffused effect will be produced.
*87

At the best, and after the exercise of the utmost care, there will remain some obscurity as to the exact extent to which shifting takes place, owing chiefly to the difficulty of employing statistical verification.
*88 Deduction from general propositions cannot overcome this obstacle, and special vigilance is therefore necessary to avoid errors arising from the want of a check such as the process of verification provides. The earlier theories are so many warnings of the danger of hasty deductions from insufficient premises.

But, subject to these cautions, the use of the theory is by no means slight. We may not be able to give confident answers to general questions on the subject, but in dealing with particular instances we shall have the advantage of knowing what conditions we ought to notice and what effects we may reasonably look for. So understood, the theory of incidence is an indispensable part of financial doctrine.