Principles of Economics by Alfred Marshall (1890)

Book Five: General Relations of Demand, Supply and Value

Chapter 8, Marginal Costs in Relation to Values. General Principles

1. This Chapter and the three following are given to a study of the marginal costs of products in relation to the values of those products on the one hand, and on the other hand to the values of the land, machinery, and other appliances used in making them. The study relates to normal conditions and long period results. This fact must ever be borne in mind. The market value of anything may be much above or much below the normal cost of production and the marginal costs of a particular producer at any time may stand in no close relation to marginal costs under normal conditions. (1)

It was indicated at the end of Chapter VI that no one part of the problem can be isolated from the rest. There are comparatively few things the demand for which is not greatly affected by the demand for other things to the usefulness of which they contribute; and it may even be said that the demand for the majority of articles of commerce is not direct but is derived from the demand for those commodities to the making of which they contribute, as materials or as implements. And again this demand, because it is so derived, is largely dependent on the supply of other things which will work with them in making those commodities. And again the supply of anything available for use in making any commodity is apt to be greatly influenced by the demand for that thing derived from its uses in making other commodities: and so on. These inter-relations can be and must be ignored in rapid and popular discussions on the business affairs of the world. But no study that makes any claim to thoroughness can escape from a close investigation of them. This requires many things to be borne in mind at the same time: and for that reason economics can never become a simple science. (2)

The contribution which this group of chapters aims at making covers little ground: but that ground is difficult and we shall need to work over it carefully, and from more than one point of view; for it is thickly strewn with pitfalls and stumbling blocks. It deals primarily with the earnings of land, machinery, and other material agents of production. Its main argument applies to the earnings of human beings; but they are influenced by some causes which do not Affect the earnings of material agents of production: and the matter in hand is sufficiently difficult without further complicating it by side issues.

2. Let us begin by recalling the action of the principle of substitution. In the modern world nearly all the means of production pass through the hands of employers and other business men, who specialize themselves in organizing the economic forces of the population. Each of them chooses in every case those factors of production which seem best for his purpose. And the sum of the prices which he pays for those factors which he uses is, as a rule, less than the sum of the prices which he would have to pay for any other set of factors which could be substituted for them: for, whenever it appears that this is not the case, he will, as a rule, set to work to substitute the less expensive arrangement or process. (3)

This statement is in close harmony with such common sayings of every-day life, as that "everything tends to find its own level," that "most men earn just about what they are worth," that "if one man can earn twice as much as another, that shows that his work is worth twice as much," that "machinery will displace manual labour whenever it can do the work cheaper." The principle does not indeed without hindrance. It may be restricted by custom or act law, by professional etiquette or trade-union regulation: it may be weakened by want of enterprise, or it may be softened by a generous unwillingness to part with old associates. But it never ceases to act, and it permeates all the economic adjustments of the modern world.

Thus there are some kinds of field work for which horsePower is clearly more suitable than steam-power, and vice versa. If we may now suppose that there have been no great recent improvements in horse or steam machinery, and that therefore the experience of the past has enabled farmers gradually to apply the law of substitution; then, on this supposition the application of steam-power will have been pushed just so far that any further use of it in the place of horse-power would bring no net advantage. There will however remain a margin on which they could be indifferently applied (as Jevons would have said); and on that margin the net efficiency of either in adding to the money value of the total product will be proportionate to the cost of applying it. (4)

Similarly, if there are two methods of obtaining the same result, one by skilled and the other by unskilled labour, that one will be adopted which is the more efficient in proportion to its cost. There will be a margin on which either will be indifferently applied. (5) On that line the efficiency of each will be in proportion to the price paid for it, account being taken of the special circumstances of different districts and of different workshops in the same district. In other words, the wages of skilled and unskilled labour will bear to one another the same ratio that their efficiencies do at the margin of indifference.

Again, there will be a rivalry between hand-power and machine-power similar to that between two different kinds of hand-power or two different kinds of machine-power. Thus hand-power has the advantage for some operations, as, for instance, for weeding out valuable crops that have an irregular growth; horse-power in its turn has a cleat advantage for weeding an ordinary turnip field; and the application of each of them will be pushed in each district till any further use of it would bring no net advantage there. On the margin of indifference between hand-power and horse-power their prices must be proportionate to their efficiency; and thus the influence of substitution will tend to establish a direct relation between the wages of labour and the price that has to be paid for horse-power.

3. As a rule many kinds of labour, of raw material, of machinery and other plant, and of business organization, both internal and external, go to the production of a commodity: and the advantages of economic freedom are never more strikingly manifest than when a business man endowed with genius is trying experiments, at his own risk, to see whether some new method, or combination of old methods, will be more efficient than the old. Every business man indeed, according to his energy and ability, is constantly endeavouring to obtain a notion of the relative efficiency of every agent of production that he employs; as well as of others that might possibly be substituted for some of them. He estimates as best he can how much net product (i.e. net addition to the value of his total product) will be caused by a certain extra use of any one agent; net that is after deducting for any extra expenses that may be indirectly caused by the change, and adding for any incidental savings. He endeavours to employ each agent up to that margin at which its net product would no longer exceed the price he would have to pay for it. He works generally by trained instinct rather than formal calculation; but his processes are substantially similar to those indicated in our study of derived demand; and, from another point of view, they may be described as those which might be reaped by a complex and refined system of book-keeping by double entry. (6)

We have already followed some simple estimates of this sort. We have noticed, for instance, how the proportion of hops and malt in ale can be varied, how the extra price which can be got for ale by increasing the quantity of hops in it is a representative of the causes which govern the demand price for hops. Assuming that no further trouble or expense of any kind is involved by this additional use of hops, and that the expediency of using this extra amount is doubtful, the extra value thus given to the ale is the marginal net product of the hops of which we are in search. In this case, as in most others, the net product is an improvement in quality or a general contribution to the value of the product; it is not a definite part of the produce which can be separated from the rest. But in exceptional instances that can be done. (7)

4. The notion of the marginal employment of any agent of production implies a possible tendency to diminishing return from its increased employment.

Excessive applications of any means to the attainment of any end are indeed sure to yield diminishing. returns in every branch of business; and, one may say, in all the affairs of life. We may take some additional examples of a principle that has already been illustrated. (8) In the manufacture of sewing machines some parts may well be made of cast iron; for others a common kind of steel will suffice; there are yet others for which a specially expensive steel-compound is needed; and all parts should be finished off more or less smoothly, so that the machine may work easily. Now if any one devoted a disproportionate care and expense to th selection of materials for the less important uses, it might truly be said that that expenditure was yielding a rapidly diminishing return; and that he would have done better to give some of it to making his machines work smoothly, or even to producing more machines: and the case might be even worse if he devoted an excessive expenditure to mere brilliancy of finish, and put low grade metal to work for which a higher grade was needed.

This consideration seems at first to simplify economic problems; but on the contrary it is a chief source of difficulty and confusion. For though there is some analogy between all these various tendencies to diminishing return, they yet are not identical. Thus the diminishing return which arises from an ill-proportioned application of the various agents of production into a particular task has little in common with that broad tendency to the pressure of a crowded and growing population on the means of subsistence. The great classical Law of Diminishing Return has its chief application, not to any one particular crop, but to all the chief food crops. It takes for granted that farmers raise, as a rule, those crops for which their land and other resources are best adapted, account being taken of the relative demands for the several crops; and that they distribute their resources appropriately between different routes. It does not attribute to them unlimited intelligence and wisdom, but it assumes that, taking one with another, they have shown a reasonable amount of care and discretion in the distribution of these resources. It refers to a country the whole land of which is already in the hands of active business men, who can supplement their own capital by loans from banks wherever they can show it is likely to be well applied; and asserts that an increase in the total amount of capital applied to agriculture in that country will yield diminishing returns of produce in general. This statement is akin to, but yet quite distinct from, the statement that if any farmer makes a bad distribution of his resources between different plans of cultivation, he will get a markedly diminishing return from those elements of expenditure which he has driven to excess.

For instance, in any given case, there is a certain proportion between the amounts which may with best advantage be spent on ploughing and harrowing, or manuring. There might be some differences of opinion on the matter, but only within narrow limits. An inexperienced person who ploughed many times over land, which was already in fairly good mechanical condition, while he gave it little or none of the manure which it was craving, would be generally condemned as having so over applied ploughing as to make it yield a rapidly diminishing return. But this result of the misapplication of resources has no very close connection with the tendency of agriculture in an old country to yield a diminishing return to a general increase of resources well applied in cultivation: and indeed exactly parallel cases can be found of a diminishing return to particular resources when applied in undue proportion, even in industries which yield an increasing return to increased applications of capital and labour when appropriately distributed. (9)

5. The part played by the net product at of production in the modern doctrine of Distribution is to be misunderstood. In particular many able writers have supposed that it represents the marginal use of a thing as governing the value of the whole. It is not so; the doctrine says we must go to the margin to study the action of those forces which govern the value of the whole: and that is a very different affair. Of course the withdrawal of (say) iron from any of its necessary uses would have just the same influence on its value as its withdrawal from its marginal uses; in the same way as the pressure in a boiler for cooking under high pressure would be affected by the escape of any other steam just as it would by the escape of the steam in one of the safety valves: but in fact the steam does not escape except through the safety valves. In like manner iron, or any other agent of production, is not (under ordinary circumstances) thrown out of use except at points at which its use yields no clear surplus of profit; that is, it is thrown out from its marginal uses only.

Again, the finger of an automatic weighing machine determines, in the sense of indicating, the weight sought for. So the escape of steam from a safety valve, governed by a spring representing a pressure of a hundred pounds to the square inch, determines the pressure of steam in the boiler, in the sense of indicating that it has reached a hundred pounds to the inch. The pressure is caused by the heat; the spring in the valve governs the pressure by yielding and letting out some of the steam when its amount is so great, at the existing heat, as to overbear the resistance of the spring.

Similarly, with regard to machinery and other appliances of production made by man, there is a margin through which additional supplies come in after overcoming the resistance of a spring, called "cost of production," For when the supply of those appliances is so small relatively to the demand that the earnings expected from new supplies are more than sufficient to yield normal interest (or profits, if earnings of management are reckoned in) on their cost of production, besides allowing for depreciation, etc., then the valve opens, and the new supplies come in. When the earnings are less than this, the valve remains shut: and as anyhow the existing supply is always in process of slow destruction by use and the lapse of time, the supply is always shrinking when the valve is closed. The valve is that part of the machinery by which the general relations of demand and supply govern value. But marginal uses do not govern value; because they, together with value, are themselves governed by those general relations.

6. Thus, so long as the resources of an individual producer are in the form of general purchasing power, he will push every investment up to the margin at which he no longer expects from it a higher net return than he could get by investing in some other material, or machine, or advertisement, or in the hire of some additional labour every investment will, as it were, be driven up to a valve which offers to it a resistance equal to its own expanding force. If he invests in material or in labour, that is soon embodied in some saleable product: the sale replenishes his fluid capital, and that again is invested up to the margin at which any further investment would yield a return so diminished as not to be profitable.

But if he invests in land, or in a durable building or machine, the return which he gets from his investment may vary widely from his expectation. It will be governed by the market for his products, which may change its character largely through new inventions, changes in fashion, etc., during the life of a machine, to say nothing of the perpetual life of land. The incomes which he thus may derive from investments in land and in machinery differ from his individual point of view mainly in the longer life of the land. But in regard to production in general, a dominant difference between the two lies in the fact that the supply of land is fixed (though in a new country, the supply of land utilized in man's service may be increased); while the supply of machines may be increased without limit. And this difference reacts on the individual producer. For if no great new invention renders his machines obsolete, while there is a steady demand for the things made by them, they will be constantly on sale at about their cost of production; and his machines will generally yield him normal profits on that cost of production, with deductions corresponding to their wear and tear.

Thus the rate of interest is a ratio: and the two things which it connects are both sums of money. So long as capital is "free," and the sum of money or general purchasing power over which it gives command is known, the net money income, expected to be derived from it, can be represented at once as bearing a given ratio (four or five or ten per cent) to that sum. But when the free capital has been invested in a particular thing, its money value cannot as a rule be ascertained except by capitalizing the net income which it will yield: and therefore the causes which govern it are likely to be akin in a greater or less degree to those which govern rents.

We are thus brought to the central doctrine of this part of economics, viz.: — "That which is rightly regarded as interest on 'free' or 'floating' capital, or on new investments of capital, is more properly treated as a sort of rent — a Quasi-rent — On old investments of capital. And there is no sharp line of division between floating capital and that which has been 'sunk' for a special branch of production, nor between new and old investments of capital; each group shades into the other gradually. And thus even the rent of land is seen, not as a thing by itself, but as the leading species of a large genus; though indeed it has peculiarities of its own which are of vital importance from the point of view of theory as well as of practice." (10)


1. Numerous objections have been urged against the important place assigned to marginal costs in modern analysis. But it will be found that most of them rely on arguments, in which statements referring to normal conditions and normal value are controverted by statements relating to abnormal or particular conditions.

2. The reader is referred to the footnote on p. 393 with special reference to the compressed mathematical version of the central problem of value which begins in Note XIV in the Mathematical Appendix and culminates in Note XXI.

3. Compare V, III, section 3; and V, IV, sections 3, 4; and Note XIV in the Mathematical Appendix.

4. This margin will vary with local circumstances, as well as with the habits, inclinations, and resources of individual farmers. The difficulty of applying steam machinery in small fields and on rugged ground is overcome more generally in those districts in which labour is scarce than in those in which it is plentiful; especially if, as is probable, coal is cheaper, and the feed of horses dearer in the former than the latter.

5. Skilled manual labour being generally used for special orders and for things of which not many are required of the same pattern; and unskilled labour aided by specialized machinery being used for others. The two methods are to be seen side by side on similar work in every large workshop: but the position of the line between them will vary a little from one workshop to another.

6. The changes, which he desires, may be such as could only be made on a large scale; as for instance the substitution of steam-power for hand-power in a certain factory; and in that case there would be a certain element of uncertainty and risk in the change. Such breaches of continuity are however inevitable both in production and consumption if we regard the action of single individuals. But as there is a continuous demand in a large market for hats and watches and wedding cakes, though no individual buys many of them (see III, III, section 5), so there will always be trades in which small businesses are most economically conducted without steam power, and larger businesses with; while businesses of intermediate size are on the margin. Again, even in large establishments in which steam is already in use, there will always be some things done by hand-power which are done by steam power elsewhere; and so on .

7. See p. 387, and Mathematical Note XVI. See also other illustrations in, VI, VII. A further illustration of the relation between the wages of the marginal shepherd, and the net product of his labour will be worked out in detail in VI, I, section 7 .

8. See V, IV, section 4; see also the note on von Thünen, below, p. 523.

9. See above IV, iii, section 8; and Carver, Distribution of Wealth, ch. ii, and above footnotes on pp. 319, 320. Mr J. A. Hobson is a vigorous and suggestive writer on the realistic and social sides of economics: but, as a critic of Ricardian doctrines, he is perhaps apt to underrate the difficulty of the problems which he discusses. He argues that if the marginal application of any agent of production be curtailed, that will so disorganize production that every other agent will be working to less effect than before; and that therefore the total resulting loss will include not only the true marginal product of that agent, but also a part of the products due to the other agents: but he appears to have overlooked the following points: — (1) There are forces constantly at work tending so to readjust the distribution of resources between their different uses, that any maladjustment will be arrested before it has gone far: and the argument does not profess to apply to exceptional cases of violent maladjustment. (2) When the adjustment is such as to give the best results, a slight change in the proportions in which they are applied diminishes the efficiency of that adjustment by a quantity which is very small relatively to that change in technical language it is of "the second order of smalls" — ; and it may therefore be neglected relatively to that change. (In pure mathematical phrase, efficiency being regarded as a function of the proportions of the agents; when the efficiency is at its maximum, its differential coefficient with regard to any one of these proportions is zero.) A grave error would therefore have been involved, if any allowance had been made for those elements which Mr Hobson asserts to have been overlooked. (3) In economics, as in physics, changes are generally continuous. Convulsive changes may indeed occur, but they must be dealt with separately: and an illustration drawn from a convulsive change can throw no true light on the processes of normal steady evolution. In the particular problem before us, this precaution is of special importance: for a violent check to the supply of any one agent of production, may easily render the work of all other agents practically useless; and therefore it may inflict a loss out of all proportion to the harm done by a small check to the supply of that agent when applied up to that margin, at which there was doubt whether the extra net product due to a small additional application of it would be remunerative. The study of changes in complex quantitative relations is often vitiated by a neglect of this consideration, to which Mr Hobson seems to be prone; as indeed is instanced by his remarks on a "marginal shepherd" in The Industrial System, p. i 10. See Professor Edgeworth's masterly analyses of the two instances mentioned in this note, Quarterly Journal of Economics, 1904, p. 167; and Scientia, 1910, pp. 95-100.

10. This statement is reproduced from the Preface to the first edition of the present volume.

Contents | next chapter | Political Economy Archive