1. The history contained in the preceding ten chapters covers but a limited number of English writers. At the time of writing, a short list of economists was drawn up in advance, which enumerated the thinkers at that time generally regarded as the leaders in the development of the old English political economy. It was stated at the outset that the purpose was to review the opinions of these writers only. An endeavor was made to define this purpose as being to make an intensive rather than an extensive study of the history of English theory. By an extensive study we should mean the effort to discover writers who have made important contributions to the thought the history of which is being traced, but have been previously unrecognized or insufficiently accredited. Such examinations into the history of English economic literature will probably in the immediate future result in important alterations in what might be called our accepted lists of chief writers. It will be found that thinkers now supposed to have expressed the doctrines of their time with the greatest clearness and power were in some cases surpassed in these points by writers at present almost or quite forgotten. Or it will be found perhaps that new ideas—such for instance as the conception of marginal utility—really originated earlier than at present supposed.

A brief time elapsed between the writing of the tenth and the eleventh chapters of this essay. But in this period there has been published a discussion of the work of some earlier British economists, which establishes beyond a doubt that the marginal utility theory of value was conceived and formulated in most excellent fashion by an English writer as early as 1833, two decades before Gossen and a generation before Menger, Jevons and Walras. The English writer was W. F. Lloyd, Professor of Political Economy at Oxford. A full description of Lloyd’s theory, with citations showing the excellence of its statement, appeared in The Economic Journal, over the signature of Professor Seligman, of Columbia University.[196] Professor Seligman has called attention to still another writer of great consequence because of his contributions to the theory of value. This is Mountifort Longfield who, as Professor Seligman states, in 1833 gave an able exposition of what is now the modern doctrine of marginal demand and marginal cost.[197]

Avoiding the larger task involved in the extensive study of the history of the theory of value in England, the present monograph has endeavored to interpret the labor theory as it passed through the minds of some nine economists beginning with Smith and ending with Cairnes. The history is that of a subjective-cost philosophy of value and the difficulties of its application to explain the facts of industry. According to this philosophy, or this ultimate explanation, utility is a condition essential to the existence of value, but cost or difficulty of attainment is the essence of value. The idea was elucidated by a variety of figures of speech. Utility was conceived as a sort of resting-ground for value, but the height of value upon this ground, the value as an amount, was held to be determined by or measured by cost of production. Thus Ricardo wrote to J. B. Say:

Karl Marx was accustomed to speak of value as “a congelation of human labor,” and to speak of a useful object, or an object made useful in the process of production, as a sort of receptacle for value.

In his involved “philosophical” account of value, as we termed it, Adam Smith taught that the value, or “real worth,” of a good is measured equally well by the amount of labor which it has cost to produce, or by the labor which it can command in exchange. When, however, Adam Smith turns his attention to the proximate principles of value in the actual competitive market, we find him confessing that the theory which he first developed applies without modification only to a primitive state of society, without land rent and interest on capital. In this primitive state the amount of labor which a commodity costs determines the amount of labor which it can command in exchange. Under the conditions of advanced society, the rent of land and the “profits of stock” must come out of the exchange value of the product, and the labor cost of the latter, which is paid for by wages, no longer determines its value.[199] If we take advantage of modern terminology, and throw Adam Smith’s theory into our own words, we make its precise significance clearer. It means virtually that the exchange value of a good in the fully developed social economy is determined by its entrepreneur’s or money cost of production, so far as it is determined by cost at all. Competition must be perfect to enable cost to determine actual values. Entrepreneur’s cost is composed of expenditures for wages of the labor, rent of the land, and “profits” of the capital necessary for production. The labor cost of producing the commodity determines only the amount of the wages cost to the entrepreneur. The other elements helping to make the total of entrepreneur’s cost are not determined by the labor cost of producing the commodity. Thus the exchange value of a good is determined by its entrepreneur’s cost, but this latter is not determined by labor cost, and consequently the exchange value of the good is not determined by labor cost. The existence of rent and interest destroys the regulation of exchange values by labor cost. It must be kept in mind that all this is very much more explicit than what Adam Smith said. It signifies that the labor-cost philosophy of value cannot be true—perhaps Smith would say only that it is imperfect or not precisely true—because it is in conflict with a more certain empirical law of value, namely, the law of entrepreneur’s cost.

Ricardo adopted the labor-cost philosophy of value virtually as a premise, and the most important parts of his reasoning on the subject are concerned with removing or minimizing the empirical difficulties with this philosophy.[200] In the end he admitted that interest is an element in that form of cost which exercises the most direct and compelling influence on exchange values. He made concessions, which pursued to their logical outcome, signify that the existence of interest throws the exchange values of goods out of proportion to their pure labor costs of production. We have seen how he put this concession in a peculiar and misleading manner.[201] As for rent of land, Ricardo hastens to repudiate Adam Smith’s admission that it also is a source of difficulty to the labor theory. He gets rid of rent by explaining that the exchange values of commodities are regulated by the quantity of labor required for production on the least fertile land in use, or the quantity required to produce the most expensive portions of the supplies. In modern phraseology, he urged that it is not the actual labor cost of a good but its marginal labor cost, which regulates its value. Ricardo himself used that fatally ambiguous formula, “rent does not enter into price.” Having removed rent by making the value determining cost marginal, and having minimized the difficulty of interest, Ricardo proceeded as if the labor theory were, for the sake of argument, intact.

The chief purport of the work of Malthus was, first, to deny Ricardo’s right to disregard the interest difficulty and, second, to reaffirm Adam Smith’s opinion that land rent also throws values out of relation to labor costs. Turning now to Senior, who was the next writer to suggest a worthy and new idea, we find that in the view of this economist the existence of land rent and interest as elements in entrepreneur’s cost is fatal to the labor theory. But Senior explained interest as the reward for abstinence, just as wages are the reward for labor. In his view labor and abstinence are independent, co-ordinate elements in subjective cost.[202] However, the more important idea that can be extracted from Senior’s reasonings is that wages as actually paid are not in proportion to the quantities of labor engaged in different employments. For in his view all skilled labor gains a wage which really includes a rent to “scarce natural talents.”[203]

Summing up the results of Senior’s argument for ourselves (since he himself did not make the present applications of his doctrines), the labor theory requires that the entrepreneur’s costs of commodities should be in proportion to their labor costs, but entrepreneur’s costs are out of proportion to labor costs, not only because they include rent of land and interest on capital, but because the very payments of wages themselves may be out of proportion to the comparative amounts of labor employed and remunerated. When we say that entrepreneur’s costs are out of proportion to labor costs, we do not mean that they are in excess of wages cost, though they are this, but that for commodity A to cost the same quantity of labor as commodity B, is not a sign that the two commodities have the same entrepreneur’s costs. In other words, relative entrepreneur’s costs are not determined by relative labor costs.

John Stuart Mill took Ricardo’s view of land rent and of interest, but took Senior’s view that skilled labor occasions the entrepreneur an expense out of proportion to the quantity of labor remunerated. That is to say, he held that the higher wages of skill do not represent a higher labor cost. This opinion he adopted without at the same time taking up Senior’s particular use of the term, “rent to talents.” As we concluded at the end of Chapter IX, it may be said that Mill followed Ricardo more closely than any other of his predecessors on the question of the relation of labor to value. Turning to Cairnes we note that this writer gave one sentence in his book to the doctrine of land rent, and in this he acquiesced in the judgment of Ricardo and J. S. Mill. But Cairnes adopted a position with reference to interest identical with that assumed by Senior, and his theory of “non-competing groups” merely emphasizes the claim that the amount of wages paid in different employments is not a test of the quantities of labor employed.

Taking the whole period covered in this history, we see that a goodly number of points of criticism were raised against the pure labor-cost theory. The reader may have noted that all these points were implicit in the work of Senior, and in his alone. Does the entrepreneur’s payment of land rent, of interest on capital, and his payment of a superior wage to skilled labor (a wage out of proportion to the disutility of skilled labor), make impossible the theory that labor cost regulates value? Is it not possible that labor cost may be conceived in some way, perhaps as “marginal cost,” or “social marginal disutility,” such that the exchange values of the products of industry can be shown to depend upon the labor cost of these products? In the following pages the writer will try to give an answer to this question, so far as it lies in his power. It will be indispensable to bear with a considerable number of preliminaries. The ultimate relation of cost (in any or all of its forms) to value, cannot be discussed with any success, unless the parties to the discussion have reached some understanding as to the relation of utility to value and as to the meaning of other proposed laws of value than cost laws. Merely a moderate acquaintance with the contemporary literature of economic theory gives complete assurance that the necessary understanding just mentioned cannot be presumed to exist, but must be established as carefully as possible.

2. Historically, there have been two distinct conceptions of cost—at least in English political economy—namely, (1) labor cost, and (2) entrepreneur’s cost. Logically, there are two elementary forms of cost, (1) pain cost, and (2) potentiality cost. The first consists in the human discomforts or undesirable feelings incidental to the production of wealth, whether the disutility of labor, or that of “abstinence” or “waiting.” This is also frequently called “true,” “real,” or “subjective” cost. The terms “true” and “real” are hardly commendable, for the other elementary form of cost is quite as real as this. The word “subjective” is genuinely distinctive, but probably less so than the generic term “pain.”

The second elementary form of cost is that emphasized chiefly by the Austrian writers. In the making or acquisition of economic products, certain scarce agents are either destroyed, or else for the time being employed in a way that excludes their employment in the production of other goods. We may say that the production of any given good involves the destruction of certain productive agents, or, permitting a convenient liberty of expression, of the uses of agents from which other goods might have been produced. Thus with the emergence of one good the possibility of some other good is excluded. Professors von Wieser and Böhm-Bawerk, who have written the best explanations of the relation of this form of cost to value, have suggested no distinctive name for it. The reason appears to be that they consider this as the only form of cost which requires especial explanation, since, as they hold, it is the only kind of cost which can be correlated with value. In an article in the Quarterly Journal of Economics,[204] Professor D. S. Green has suggested the term “opportunity cost” for the Austrian conception. Professor H. J. Davenport, having in mind the same conception, refers to it as “sacrifice cost.”[205] Professor Heinrich Dietzel speaks of the same cost as “Nutzeneinbusse.”[206] This concept has also in various places been designated “social cost.” Though this cost is social in a certain comparatively irrelevant sense in which the other kind of cost is not social, the term “social cost” is not a good one because it lacks distinctiveness. There is nothing intrinsically social about potentiality cost, nor would the other elementary form of cost, that is, pain cost, be best designated “individual cost.” Both forms would appear in an isolated individual economy, for instance in our fictitious but highly useful Crusoe economy. “Sacrifice cost” is ambiguous, because the word “sacrifice” is used as often as not, though perhaps improperly, to signify the discomfort or pain endured in production. A term is needed to distinguish a form of cost from pain cost. “Potentiality” cost appeals to the writer as being a term somewhat better than “opportunity” cost. When certain common production goods, capable of being turned to the making of more than one kind of thing, are used up to make a given thing, their potentiality to make an alternative thing is destroyed. The potentiality cost of a commodity is measured by the highest other value that might have been obtained if this commodity had not been produced from the productive agents entering into it.

If it be correct that pain cost and potentiality cost are the two elementary kinds of cost, what is the relation of the two historical forms of cost, labor and entrepreneur’s cost, to these? Just as the word labor stands for two distinct things, toil and productive power, so may “labor cost” mean either pain cost or potentiality cost. If the term labor cost is used without a qualifying adjective or explanatory phrase, it would naturally call up in most minds the conception of pain cost. But labor force is the most disposable of all productive agencies, and when the productive power of labor (what the entrepreneur buys) is used up in the production of a given commodity, we have a perfect example of potentiality cost. As we have seen, Adam Smith used labor at different times, both in the sense of toil or disutility and in the sense of productive power or potential commodity. Ricardo, to the best of the writer’s knowledge, said nothing to indicate definitively which of these concepts he designated by the term labor. In the writer’s judgment the presumption is that by the labor cost which regulates the exchange value of commodities, Ricardo meant what we call pain cost. At one place, Ricardo said “difficulty of attainment” is the true measure of value. It seems almost assured that this must mean “pain cost.” When cost is conceived as the ultimate essence of value, the cost will almost certainly be pain cost or “real cost.” It is interesting to note that Professor Dietzel, in arguing that the Ricardian labor theory of value is perfectly reconcilable with all that holds good in the utility theory, states that labor cost must be conceived solely as “Nutzeneinbusse,” i. e., utility-sacrifice, or potentiality cost. When the labor theory is founded upon the conception of labor as toil (“Unlust”), he considers it to be “built upon sand.”[207]

We see that it is possible to mean either pain or potentiality cost by the words “labor cost.” Labor as “pain” and labor as productive power are not the same thing but the first is incident to the second. It remains to consider the relation of entrepreneur’s expense to the two elementary forms of cost. Torrens desired to exclude the money outlays of the entrepreneur in interest charges from the money cost of production of a good. That is, he maintained that what he called “profits,” the chief constituent of which was interest, is no part of cost of production. This view was never adopted by any subsequent economist of weight. The very simple reasons why it is indefensible were mentioned in the chapter on Torrens. Conceding then that interest is a part of entrepreneur’s cost, the relation of the latter to “pain cost” can be stated in a few words. The total “pain cost” of any article, which is produced by entrepreneurs, finds its remuneration in those payments which go to make up cost from the view-point of the entrepreneur. The point to be held fast, a point already emphasized, is that the subjective costs of goods so produced can influence their exchange values only by way of influencing their entrepreneur’s costs.

The relation of entrepreneur’s expense to potentiality cost is less simple and familiar than the foregoing. It will best be taken up in a subsequent section after we have endeavored to state the gist of the utility theory of value. We may conclude the present discussion of cost concepts by noting that there are several ways of reckoning or analyzing entrepreneur’s cost. (1) Adam Smith’s method, adopted by Malthus, is set forth in the following definite words written by the latter: “The cost of producing any commodity is made up of all the wages, all the profits, and all the rent which ... are necessary to bring that particular commodity to market in the quantity required.”[208] (2) Perhaps the most approved modern method of analyzing the elements in entrepreneur’s cost is merely into wages and interest. In this case rent paid for the use of land is treated in the same way as rent paid for buildings or for machinery or power. (3) The most direct treatment of entrepreneur’s cost defines it shortly as including the prices of all the productive agencies used up in the making of the product, or as the value of raw material, machinery, and labor power “entering into” the product. It is always necessary to explain immediately that some production goods are in no sense consumed in the making of the product. Such are the land and buildings. Some production goods are consumed only in very small part in the making of a single product. The total money cost of a product is according to one system divided into “prime cost” and “establishment cost.” The former includes the prices of all those elements which are entirely used up in making the product. The latter includes the product’s due share of the money cost of the rest of the establishment, worn to a certain extent or occupied for a time in its making. Numerous practical formulæ are in use to aid in the difficult problem of imputing to the product its due share in the various general charges. A second necessary explanation connected with the third form of calculating entrepreneur’s cost is that, since the prices of the production goods have to be paid before the product is finished, each price must be increased by interest for the time of its advance, to give the complete cost of the product. In order to state the relation of entrepreneur’s expense to potentiality cost, it is not necessary to discuss in full the mutual relations of the various modes of calculating entrepreneur’s cost. Such a discussion would, it seems to the writer, involve us in the theory of interest and indeed in the entire theory of distribution. Fortunately it is sufficient for our present ends to point out that all forms of analyzing entrepreneur’s cost must be based upon the third, or, as we might call it, the practical method. If it is desired to reduce the cost of a product to wages and interest, the practical cost of the article as defined above will have to be ascertained as a starting-point. We will be content in a later section to trace the connection between potentiality cost and the practical form of entrepreneur’s cost.

3. It is impossible to give the term value any one meaning. The word is so ambiguous that Jevons advocated its abandonment. The central thought of the value concept seems to be capacity to excite desire, but there are two grand kinds of economic value which are best designated, (1) exchange value and (2) esteem value.[209] The exchange value of an article has often been defined as a ratio—or specifically, as the ratio in which the unit of measure of the article exchanges for a multiple, or fraction, of the unit of measure of any other determinate thing[210]—and again as the quantity of that other thing for which it exchanges. But both of these conceptions are involved in metaphysical difficulties which make them impossible to employ in actual reasoning, and all writers are accustomed to make assertions about exchange value which are not in the least true of these, their purely formal, definitions. Walsh has shown most clearly that the only conception of exchange value free from difficulty is that of power in exchange. Exchange value is the power in a thing by means of which its owner is enabled to command other things possessing a similar power.[211] This power is measured objectively by the physical quantity of some other thing selected for the purpose, but the value is not that quantity, though speaking elliptically we may say the value of a coat is twenty bushels of wheat or twenty dollars. The exchanging power in our daily thought is always and properly referred to the thing and not to the man or owner. When the thing is gone the power is gone. The purchasing power of a twenty-dollar piece does not reside in the holder.[212] It goes without saying that this power can exist in an article only when there is another thing for which it can be exchanged and when there are men to effect the exchanging. It is not implied that articles of commerce have the power to exchange themselves in market places without human intervention. But the purchasing power resides in the article; it is always referred to the article in our thought, and it leads to nonsensical results to run counter to our commonest and most practical forms of thought and endeavor to locate it elsewhere.

The end of the theory of value is to explain the causes which govern exchange values, or practically, market prices. Other conceptions of value than exchange value derive their just importance in political economy only from the aid they may render in this explanation. This defines the place of the conception of esteem value. A finished explanation of exchange values is impossible without a theory of esteem values, just as an ultimate explanation of the counterbalancing powers of different objects in the scales is vain without the general law of gravitation. The power of an apple to counterbalance two eggs might have been thought of as a matter purely relative between apples and eggs, before the general conception was framed of the common attraction exerted by the earth upon both apples and eggs. Exchange values were once stated to be “purely relative.” The objectionable point in the statement lies in the adverb “purely.” For exchange values are derived from esteem values very much as weights, or powers to counterbalance, are derived from the earth’s attraction. That is, these two derivations are generally similar. Here as elsewhere analogies can be pressed too far. The derivation of the exchange values of different goods in the social market from the esteem values put upon these goods by the consumers, or purchasers in the society, is a process indefinitely more difficult to explain than the relative weights of ponderables.[213] Given the gravity of separate objects it is but a step to explain their relative weights, but given the esteem values of different economic commodities it is a long road to follow the process through which these determine the exchange values of the same commodities. If esteem value be the gravity behind exchange value, the case is rendered complex in that each good possesses a separate esteem value for every different consumer whereas there is but one earth to attract each object whose relative weight we desire to explain.

In the judgment of the writer, the best definition of esteem value is “the significance (Bedeutung) which concrete goods attain in our estimation when we realize that we are dependent upon them for the satisfaction of some want.” This is a loose translation of the definition formulated by Carl Menger in 1871.[214] The individual good attains value not simply when it is capable of affording us satisfaction, but when it conditions the satisfaction. Goods existing in superfluity give satisfactions but do not condition them, hence any unit of such goods possesses no value. The removal or destruction of a unit occasions the loss of no satisfaction. Menger’s definition was a triumph of theory in stating the relation of value to human satisfaction and to utility.

The law of marginal utility is but a corollary of the principle involved in this definition. The utility of a good is its power to afford satisfaction.[215] When goods occur in stocks of like units the phenomenon of “marginal” utility emerges. As the stock of such goods to be used by a consumer within a given time is increased, the satisfaction afforded by each successive unit declines.[216] The actual utility of each successive increment is lower than the actual utility of the preceding increment. The actual utility of the last or marginal increment is the “marginal” utility of any of the increments. The reason why the value of any such increment is determined at the height of its “marginal” utility is only because any one increment conditions merely the satisfaction afforded by the last or marginal increment. Remove or destroy any increment and rationally only the satisfaction of the marginal increment will be given up. In effect any increment is the marginal one. Thus the law of marginal utility is not the fundamental law of value. Menger’s definition contains this fundamental law and gives a universal principle of value. The theorem that value depends upon marginal utility is merely a deduction from this fundamental principle, and is of limited scope, since it applies only where there are goods in stocks.[217]

An absolutely essential point to be kept in mind is that the value of an object is not derived from the sacrifice made to obtain it. On the contrary we make the sacrifice because the object has this value. The value is first, the sacrifice second. The only means of estimating how much sacrifice or discomfort we can afford to undergo to obtain an object is by judging its value to us previously to and independently of the sacrifice. If the labor cost, say, determined the value, we could expend labor cost regardlessly in producing any objects whatsoever. But this is just what we cannot do. We must have a care when we expend labor. A care for what? For the value of the result. The value of the object is derived from the satisfaction which it can afford, but it is attributed to the object only when it is the indispensable condition of that satisfaction.

When a particular object conditions some satisfaction of ours, it possesses a superior power over our welfare to that possessed by a good which, while it affords us satisfaction, does not condition it. It would be a waste of energy to spend it in producing or conserving the thing of inferior power. The ultimate reason why our minds instinctively assign a superior importance to the things of higher power over our welfare is that such instincts contribute to our fitness in the struggle for existence. The things of lower rank possess merely utility. The things of higher rank possess a superior power which we recognize and distinguish as value. Lack of this distinction would lead to a waste of energy in the modifications which we effect in our environment.[218]

4. Let us consider now the essential features of the utility theory of value. According to this doctrine the sole source of value is human satisfaction. The thought is well conveyed, though the expression may be somewhat loose, by saying that the only inherently valuable thing is satisfaction, and that the value of satisfaction is passed out to, or attributed to, any external object upon which the satisfaction is dependent. Human satisfactions are quantities, for they are capable of being more or less. But they are inexact and wavering quantities. Similarly, the esteem values of goods, derived from the satisfactions conditioned by these goods, are quantitative, but are incapable of exact or constant numerical expression. The exchange values of goods are exact and relatively constant quantities. Nevertheless, exchange values are assuredly founded upon esteem values. The description of the process by which these wavering and more or less uncertain esteem values determine the exact and definite exchange values of goods, is perhaps the most difficult part of the utility theory. But without this part—commonly called the “theory of price”—the utility doctrine remains a mere torso.

Böhm-Bawerk’s theory of price is the only attempt that has been made to complete the utility theory in this direction, which is at once well known and authoritative.[219] Though the fundamental lines of this explanation are correct, it is in externals at least defective in two noteworthy respects. In the first place, one of its principal assumptions is untrue to the typical conditions of exchange under the division of labor. This assumption consists in supposing that sellers’ subjective valuations are effective factors in determining price in the market of an organic society. In Böhm-Bawerk’s theory, the reader will recall, the miniature but supposedly typical market is assumed to consist of a number of sellers owning horses and a number of buyers desiring them. The first step in the argument is to assign a money expression to the esteem or “subjective” value of a horse to each buyer and each seller. Then if these money expressions, or “price equivalents,” are higher for some one buyer than for some one seller an exchange of money for a horse is possible between the two. If buyer B values a horse at $45 and seller S at $40, B can afford to part with, and S can afford to take for a horse any sum of money over $40 and under $45. By considering all the price equivalents of sellers on one side and of buyers on the other side, Böhm-Bawerk shows us that for every given combination of such figures in a market there is a certain definite number of sales possible, and these sales must take place at a price fixed between the price equivalents of the last buyer and the last seller. In other words, the market price will be fixed between the money valuations set upon a horse by the “marginal pair.” The region so delimited by the marginal pairs becomes narrower as the number of buyers and sellers entering the market increases. Thus, in a large market the price is virtually determined to a point. The difficulty with this theory is that under the division of labor, sellers make products for the market, in view of the market price, and make them in numbers and keep them in stocks far in excess of their own needs. Under the division of labor, the lowest price at which a seller will part with a commodity is not set by the marginal utility or subjective value of the commodity to him. A theory of price applicable to the modern market must not employ the subjective valuations of sellers as a factor in price determination.

In the second place, Böhm-Bawerk’s theory of price is misleading, since it obscures the fact that the Austrian theory of value always assumes the supply of the goods whose value it is to explain. The Austrian writers themselves teach us that the value of a good depends upon the supply of it. The theory of marginal utility explains very well why an increase of supply lowers value or a decrease raises value. But if there should be any cause which limits or regulates the supply of goods with reference to their value, by some kind of an adjustment to value, this cause would be both a cause and a regulator (or at least a part regulator) of value. Cost of production in some ultimate form is by many writers supposed to be such a cause. The human “pain cost” of producing goods is of equal importance in theory with the human pleasure gain had from utilizing the goods. The value of an addition to the stock of a given sort of goods always furnishes a motive for the increase of the stock. Any cause which limits the supply at a certain point in the face of this human desire for the increase is a cause of value. True, it is a cause more remote than utility, but still a cause of value. Since the Austrian writers virtually ignore the cause or causes governing supply (and thus governing or helping to govern marginal utility itself), the doctrine of price which they advance ought to rest openly and squarely upon the assumption that the supply of the good is taken for granted. In Böhm-Bawerk’s theory of price, the total supply of horses in the miniature market is simply assumed, not accounted for.[220] If horses were more plentiful in this market the sellers’ price equivalent would be lowered and the market price would turn out lower. It is a fair criticism that Böhm-Bawerk obscures the important point of the dependence of price upon supply, by assigning sellers and buyers an arbitrary series of money valuations as the very first step in his argument.

The “theory of price,” as the Austrians call that part of their theory which traces the connection between consumers’ “subjective” values and market exchange values, must begin with a clear recognition that the pure utility theory of value assumes outright the extent of the supplies of all goods. Let us, then, inquire first how the price of a given supply of consumption goods is determined.[221] If a certain supply of some consumption good is presented for sale in the social market, there is theoretically some one price at which just this amount of goods can be sold. Following Professor Marshall, we may call this the “social demand price.” At a higher price, only a less supply could be disposed of. At a lower price more could be sold. The competition of buyers ultimately prevents this lower price being set. The dependence of the social demand price of a given supply of goods upon the esteem value of these goods to the consumers in the social market may be traced as follows:

(1) From esteem values to price equivalents. The sum of money which a consumer would pay for the addition of an article to his possessions rather than go without that article is called its price equivalent. The price equivalent must not be confused with the price he may pay actually to buy the article, namely, the market price. A piano may have a market price of $600, but have a price equivalent of $1,000 to A. If the market price of pianos were to ascend to $1,000, A would still purchase one. But more than $1,000 he would not pay. When A assigns to a piano a price equivalent of $1,000, this sum of money, of course, has significance merely as the representative of the indefinite variety of other goods which A supposes to be within the command of $1,000. Thus A’s ability to think out price equivalents depends upon the already existing exchange value of money, or, in other words, upon prices themselves. If the market prices of carpets, carriages, wines and other things were different from what they are, the amount of other things commanded by $1,000 would be altered, and assuredly A’s price equivalent for a piano would change. Thus the price equivalent of one thing can be named only in view of previously existing scales of market prices of other things. A consumer comes to form a conception of the significance of a unit of money to his welfare. Into the question of the inner nature of this conception we cannot afford here to push our inquiry. Without this conception he could not set price equivalents. It is merely a matter of experience that in fact consumers do set price equivalents. The sole possible explanation of the fact that when a monopoly raises the price of a consumption good the sales of it decline, is that some buyers have been excluded because the price asked for has passed above their price equivalents.[222] The worth of money to a consumer depends upon the extent of his money income. Thus the price equivalent set upon a good by any consumer depends (1) upon the esteem value of that good to him, and (2) upon the extent of his money income. Given a consumer’s money income, his price equivalents for various articles will be determined by, and be in proportion to, the esteem values those articles have for him.

(2) From price equivalents to market-price. The price at which a given supply of a certain good can be sold to a body of consumers is a resultant from their individual price equivalents for this good. It is in no sense an average or mean of these price equivalents.[223] This is best understood by imagining the supply of the good offered in the market to be increased by one unit. This extra unit must be added to some person’s stock. It will go normally to the person who will pay the most for it, but the price of the unit will have to be lowered sufficiently to bring it down to this person’s price equivalent for a unit. In the open market, then, the prices of all units will have to be lowered to the level of the price of this unit. This illustrates the way in which the social demand price of a given supply of goods is determined at some individual price equivalent. The price of a given supply is determined at the point of a marginal individual price equivalent.

In the last section it was asserted that unless a good is possessed in a plurality of units, that is, in a stock, by the individual consumer, its value will not be determined by marginal utility. The Austrian writers have made this perfectly clear, but there are innumerable places in the literature which has sprung up about the Austrian theory, either expounding or criticizing it, where the value of such a good as a piano or a furnace is said to depend on marginal utility. Let us suppose that no person possesses more than one piano. In this case, properly speaking there is nothing marginal about the value-determining utility of a piano. There are, however, two methods in vogue of discovering an alleged marginal utility in such a single unit commodity. The first is to point out that a piano may serve several uses. For instance, it may be used to produce music and also as an ornamental piece of furniture. It is then suggested that one of these uses is greater or less than the other and is marginal. Some suggest by implication or directly that it is the least use to which a piano is put which determines its value to the owner. If it should be suggested in reply that a piano might be used to conceal a discolored place in the wall, which could equally well be done by a two-dollar screen, the probable reply would be that it is only the least use to which the piano can be put economically which determines its value. Even so acute a writer as Smart[224] is guilty of this perversion. When it is in the pursuit of such margins, the mind is far adrift from the true logic of the utility theory. When a piano is actually used to cover a piece of wall, this is assuredly an “economic” use of the article. This use does not exclude or hamper any of its other uses. It is true no one would pay $600 for a piano merely to cover a bad piece of wall, but very likely few would pay that sum for any one use of the piano. The truth is, the value of a piano to its user depends upon the sum of its uses to him. The value of the piano measures the total amount of satisfaction conditioned upon its possession. When goods are used in stocks any one unit conditions only the satisfaction had from the last unit. Thus only does marginal utility—the actual total utility of the marginal increment—determine value. It is quite futile to attempt to distinguish between the different uses of a unit commodity and arrange these in a descending scale and choose a marginal use. When the unit is taken away all these uses would be sacrificed. Professor Dietzel, an undiscerning critic and imitator combined of the Austrians, has stated that it is the “highest use,” the use on the upper margin, which determines the value of a unit commodity.[225]

A second method of discovering a marginal utility for a piano is to conceive of the utility of a piano to that possessor who has just been able to afford the price as the marginal utility of pianos. All men pay the same price for a given grade of piano, but the rich men have much higher price equivalents than the poor. If the supply of pianos to be sold in a given social market be increased, the price will fall. This fall is interpreted as being caused by a decline in the “marginal utility” of pianos. There is no justification for this logic in the utility theory. It is not possible to compare the satisfactions had from pianos by different persons. It is not possible to imagine the pianos in society arranged in a series, the pianos of highest utility being those held by the persons who could afford to pay most and so on. The price of a piano depends upon the marginal price equivalent of a piano, but neither the exchange value nor the esteem value of a piano depends upon marginal utility.

To conclude, all goods derive their exchange values from the esteem values placed upon them by consumers. The exchange value of a good in money is determined in a marginal manner by the price equivalents set upon the good by the consumers. Since the extent of a consumer’s money income helps determine the price equivalents placed upon all articles by him, it is impossible to show that these price equivalents depend solely upon esteem values. But it is still proper to say that the esteem value of a good is the sole source of its exchange value. A consumer will assign no price equivalent to a good unless it possess esteem value, and when he does assign a price equivalent, it will be precisely in proportion to the esteem value of the good as compared with other goods which he values.

The price of a commodity is a definite and fairly stable quantity, e. g., the price of an oil-stove is $4.50. Is it possible that this definite price can be said to be determined by the utility of oil-stoves to consumers? Consumers cannot reduce their estimates of the utility of articles directly to figures. But, nevertheless, a consumer can determine upon a sum of money whose general purchasing power he considers approximately equivalent to the value of an oil-stove to him. The value of the oil-stove is a wavering quantity, but having struck a money estimate on the basis of that value, taking it for what it is at the instant of the decision, this money sum is a definite something that will be carried in mind as such though the value is indefinite and wavering. Given these definite “price equivalents” the definite market price is a resultant from them. The market price of a good is a sort of social institution and has the momentum or stability of such an institution. Being once determined, it will not waver as do the numberless individual estimates of “esteem value” upon which it is founded.

5. If a commodity fetching a definite and exact price, as for instance an oil-stove selling for $4.50, is produced under competitive conditions, the apparent and proximate reason why the article has this particular price is because it costs its manufacturer about this sum of money to produce it. Putting aside the complications due to the fact that competition frequently takes place between firms producing at different costs, the commonest law of exchange value, stated in the usual language, is that price is “determined by” entrepreneur’s cost of production. Whether entrepreneur’s cost is reckoned in terms of wages, interest, and rent; wages and interest alone; or in terms merely of the prices of all the production goods “entering into” the product, the law of entrepreneur’s cost, as stated above, reduces itself to the proposition that the exchange value of production goods “determines” the exchange value of products. For all forms of calculating entrepreneur’s costs are based on the simple, practical, or first method of reckoning costs, as the prices of labor, raw material, machinery, power, etc.[226]

There is seeming antagonism between the law of entrepreneur’s cost and the utility theory of value. For, according to the latter, the value of production goods is derived solely from the value of their products. Value originates in human satisfaction, flows out to those consumption goods upon which the satisfaction is immediately dependent (i. e., by which it is conditioned), from these flows out to the production goods upon which the consumption goods as products are dependent for their existence. From these on, the flow of value continues to those production goods which are still farther removed, and so on, rank by rank, until unproduced agents are reached. Put in other words, raw material, machinery and similar goods, have value solely because the entrepreneur can afford to pay for them, and this he can do solely because his products have value. If value conduction runs from product to production goods, how can value determination run in the reverse direction, from production goods to product? If the stream runs from the spring, we know that the volume of the spring must determine the volume of the stream and not the contrary. But does not the law of entrepreneur’s cost assert that the value of production goods determines the value of products? The solution of the enigma of this apparent conflict of the utility theory with the great empirical law of cost (in its common form of statement) is one of the most interesting and important products of the acute thinking of the Austrian economists.

The difficulty exists solely because many single production goods, such for example as iron and wood, or pre-eminently common labor, enter into a variety of products. When several various products are related to one another by reason of the fact that a common production good enters into all of them, we may, following von Wieser, call them “cognate products.” If cognate products A, B, and C are made in part from the common production good P, P will derive its value from the values of A, B, and C, but the value of P itself will have a peculiar reactionary effect, yet to be described, upon the value of these products taken individually. This reaction is the phenomenon really at the foundation of the law of entrepreneur’s cost. For an instant permit a supposition quite contrary to fact. Suppose that A, B, and C are made entirely from P so that no other production good enters into them. Then the exchange values of A, B, and C will be derived solely from the “marginal” utility of these products, but their exchange values are peculiar in this, that they will be related, will be adjusted, each to the others by reason of their common origin in P. If a unit of P entering into A attains in that form a higher value that when entering into B or C, then more A’s will be made from P and less B’s and C’s. The increase in the supply of A’s will decrease their value by decreasing the marginal utility of A’s.[227] This process keeps the values of A’s, B’s, and C’s in a mutual adjustment. If we carelessly confine our view to a part of this process of adjustment, we see what appears to be a determination of the value of A, the product, by the value of P, the production good. If the value of A is out of adjustment with its cost in P, and then the adjustment is effected, it is the value of A which seems to move to that of P. The value of P may seem to be the independently determining factor. As a matter of fact, in the first place, the value of A moves only when its marginal utility has been altered by a change of its supply, and in the second place, when it does move toward the value of P, that of P also moves toward it. It depends on the relative importance of A’s in comparison with all of the rest of the products of P, how far the value of A moves and how far the value of P moves in their mutual adjustment. If it is discovered that the two stars in a “double star” are approaching each other, we easily conceive that both take part in the moving. But when an apple falls to the earth it is more difficult to realize that the earth also falls toward the apple, moving its due proportion of the distance between them. In the same way, when we see that the value of some relatively unimportant product remains equal to its cost of production, we are inclined to state the case as one of pure determination of value by cost. But this “determination” is but a part, viewed by itself, of a larger process by which the supplies, and consequently the values, of cognate products are being adjusted to one another. The whole truth is that the value of production goods is determined by the value of their products. Because of the existence of great common production goods of manifold productive uses, we have the peculiar reaction of the value of production goods on the value of products, a part of which process is described in the law of entrepreneur’s cost.

The case of real life is more complex than the one considered above in that a plurality of common production goods always enters into a product. In this more complex case we find an entirely new problem, which rejoices in the name of the “imputation of the productive contribution,” but the explanation of the law of entrepreneur’s cost remains exactly the same in principle as that given for the artificially simplified case. When a product is made from several production goods, as a carriage from wood, iron, leather and labor, the value which by anticipation the production goods derive from the product is divided among them. The share of value of each production good is called its productive contribution. The different solutions of the problem of imputing the productive contribution already offered in the literature of value and distribution, will not be discussed here.[228] One principle, however, will be taken for granted. It will be assumed that the productive contribution of any kind of production good in the value of a particular kind of product varies inversely with the supply of this production good which is turned to the making of that kind of product. If product A is made from production goods P, Q and R, and the amount of P used in producing A’s is increased—with or without an increase in the amount of Q’s and R’s used[229]—the productive contribution of a unit of P in the value of A’s will be decreased.

Considering now the reaction of the value of production goods on the value of particular products, we may represent the more complex case, corresponding to real life, by supposing products A, B and C to be made from production goods P and Q, P and R, and P and S respectively. In this case the products are cognate only by reason of their relationship through P. The production goods Q, R and S are not common. Here as before the supplies and values of the products A, B and C are in a relation of mutual dependence. The dependence is, however, not so close as in the first artificially simplified case considered above. In the present case it is not the values of A, B and C themselves which are brought to an equilibrium, but it is merely the productive contributions of P in the values of A, B and C that must reach an equality. If a unit of P obtains a higher productive contribution in A than in B and C, more P will be put to the making of A’s and less to the making of B’s and C’s, until a unit of P attains the same productive contribution in each of these products. If the amount of P put to the making of A’s is increased, the supply of A’s will be increased and A’s will decline in value. But the decline in the value of A’s caused in this manner will fall entirely upon the productive contribution of P. The decline of value takes place merely to bring the contribution of P’s in A’s to an equilibrium with the contribution of P’s in B and C. Abstract as the foregoing formulæ are, they are nevertheless real. If entrepreneurs were not able to ascertain, at least approximately, the productive contributions of the various production goods entering into the product which they manufacture, they would be unable to tell either how much of each productive factor they can afford to buy or what price they can pay for it.

When we call to mind the fact that in actual industry most production goods are themselves products, and that into the majority of final products nearly all the great common production goods enter, we realize the stupendous complexity of the relationships of cognate products in actual life. It is no wonder that ordinarily a whole half of the process by which the values of all these fellow products are brought to mutual adjustment escapes our notice. Pig iron derives its value from a thousand and one kinds of products. When the value of one of these alone is being brought into adjustment with the value of pig iron, the mass is all on the side of the pig iron, if we may so express it. In this movement the single product is seeking a value-equilibrium with all the vast multitude of other products of pig-iron. It seems itself to effect all the adjusting. As a matter of fact, it contributes its due share to the determination of the value of the raw iron. Thus far, we may safely affirm, the difficulty which, at first sight, the law of entrepreneur’s cost seems to present to the utility theory has been quite surmounted.

It is a matter of some interest to define as exactly as possible the relation of potentiality cost to entrepreneur’s cost. When an entrepreneur, in making his product, uses a production good capable of other applications, by other entrepreneurs, his product is made at the expense of potentiality cost. Whether in the isolated or in the social economy, there is waste whenever a good is produced at a potentiality cost which is higher than the value of the good itself. A greater value will be sacrificed to obtain a less. In the isolated economy Crusoe will easily guard himself against this form of waste. In the social economy, the competition of entrepreneurs for the supplies of production goods prevents the same form of waste. Competition being granted, nowhere will there be found an entrepreneur who uses up production goods to make a value less than the highest value that these goods could produce elsewhere. For otherwise the entrepreneurs located at the other points of higher return would be able to command the production good for their purposes, and the possibility of profits would furnish them with the motive to bid for it. If entrepreneurs’ costs always consisted solely in the value of production goods capable of manifold applications, we could say that the potentiality cost of a product determined its entrepreneur’s cost. For the former cost consists in the highest other values that these production goods may be made to produce elsewhere, and the entrepreneur will normally have to pay that value for them and no more. His necessary outlay for a product is thus regulated by its potentiality cost.

But sometimes in the making of a product certain valuable production goods may be used whose employment does not involve potentiality cost. These are, of course, production goods capable of being used in this one product alone. A production good may be capable of making but one kind of product and yet receive from that product a certain share of value as its productive contribution. A mineral spring may be so situated as to be such a production good. A mine is a perfect example. Here a new question confronts us. Does or does not an entrepreneur’s outlay in the value of such a production good constitute a part of entrepreneur’s cost? This is solely a question as to how we choose to define entrepreneur’s cost. It may be defined either way. But in the event that we define this cost to include outlays for single-use production goods, it will no longer be possible to assert that potentiality cost governs entrepreneur’s cost wholly and in all cases. Let us give an illustration of the question. If the bottled water of a mineral spring can sell for ten cents in a neighboring city, and it costs five cents for the bottle and labor and two cents for transportation, is or is not the three cents per bottle which remains as the rent (“price-determined surplus”) to the spring a part of the entrepreneur’s cost of producing bottled mineral water? If the vender of the water did not own the spring, he would be inclined to reckon the rent paid for it to its owner as a part of his money costs. But economists are agreed that the distinction between costs and surpluses does not hinge on relations of legal ownership. If the producer of the bottled water owned the spring, he would merely pay the rent of it to himself. To the present writer it seems that entrepreneur’s cost may be defined either to be coextensive with potentiality cost, or to exceed this cost by the inclusion of “price-determined” rents,[230] provided a consistent usage be maintained. In the one case, entrepreneur’s cost is determined by potentiality cost; in the other case it is principally determined by potentiality cost.[231]

In accounting for the value of consumption goods the Austrian theory takes their supply for granted. In the same way, when the Austrians come to their explanation of the law of cost they take the supplies of production goods for granted. If the supply of pig iron brought every year to the iron market be increased, the supply of the products of iron will be increased and the exchange values of these products will fall. The exchange value of pig iron will fall in consequence.[232] At this point a question—a criticism in behalf of the cost theories of value—naturally suggests itself. Is not supply ultimately regulated by cost of production in some form, and is not cost of production thus either the ultimate regulator of value itself, or at least a joint regulator with utility? We have suggested here the famous question of the “reconciliation” of the cost and the utility theories of value. It is certain that the only form of cost which can exercise ultimate control over the supply of any produced good is what we have called by the generic name of “pain cost.” The potentiality cost of a product is measured in the value of the production goods entering it. But this value itself depends on the supply of these production goods. The ultimate cost regulator of the value of both the products and the production goods cannot be potentiality cost. The influence of potentiality cost causes the supply of the individual kind of product merely to be adjusted to the supplies of its cognate products. But potentiality cost has no influence whatever over the total supply of the production goods or the absolute supply of the total mass of cognate products. To appeal to a simple illustration, if a flow or stream of some production good be supposed to divide into several branches as it proceeds, each branch representing one of the several cognate products of that good, the influence of potentiality cost may determine the relative volumes of the different product-streams, but only pain cost—if any cost at all—can influence the volume of the parent stream, and thus govern the absolute volume of all the branches. Entrepreneur’s cost also, most obviously, fails as an ultimate regulator of supply. This cost is but the proximate agency through which the two elementary forms of cost exert their influences upon the relative and absolute supplies of products. Undoubtedly the recognition that pain cost is the only form of cost capable of exerting any ultimate control of value, helps to suggest that it be called “real cost” or “true cost.” Professor Marshall, for instance, analyzes cost into two forms, (1) real, and (2) money costs.

6. There is to-day a large following for the doctrine that cost and utility are joint and equal regulators of value. Professor Alfred Marshall, for instance, states that “we might as reasonably dispute whether it is the upper or the under blade of a pair of scissors that cuts a piece of paper, as whether value is governed by utility or cost of production.”[233] Historically, there have been developed by economists two distinct and apparently antagonistic theories professing to afford ultimate explanations of value; the earlier or cost and the later or utility theory. But if it can be shown that in reality cost and utility are but joint regulators of value, recent thinkers hold this equivalent to a demonstration that the two apparently hostile doctrines are after all but the two parts of a larger harmonious whole. It is, therefore, maintained by many that the two opposed schools of value merely failed to take a broad enough view of the problem. Thus we have ever before us the interesting question of the reconciliation of the cost and utility theories. If it be desired to effect a fundamental reconciliation, what appears to be the most propitious starting point is found in the theory of the “marginal or final equivalence of utility and disutility,” a doctrine which originates purely as a theory of “subjective” or esteem value in contrast with exchange value. The first writer to give this theory a definite formulation was probably Gossen, but it was J. B. Clark’s later but entirely independent statement of the same idea which was first to bring it to the notice of economists generally.[234]

Professor Clark develops a theory that the ultimate standard of the value of a good is the “effective social disutility” cost of its acquisition. Thus he presents a theory of distinctively social valuation.[235] But we also find as a part of his doctrines a theory of purely individual valuations. Professor Clark distinguishes clearly between these two, and in fact develops the theory of social from the theory of individual valuation. In drawing the reader’s attention to these doctrines, it will suit our purposes best to emphasize as much as possible the distinction between value in the individual economy and in the social economy. We may, therefore, adopt the device of “Crusoe economics,” and consider the relation of subjective cost to value in the isolated individual economy. In the theory of value, the pursuit of this plan does not involve a waste of time, but on the contrary it is an excellent measure for which there is precedent at some point or other in the writings of nearly all theorists. The plan corresponds to the artifice of the primitive society of hunters and fishers so frequently used and also abused by the classical economists. It is in a peculiar degree a device of what Roscher called the “idealistic method.” We are, of course, now dealing with a problem entirely outside the possible sphere of the “historical method.” We find in the Crusoe economy the prototype, as it were, of many a complex value relation in advanced social conditions, and an appreciation of certain simple and highly generalized principles true of this economy may greatly facilitate our understanding of the difficult subject of cost and value under real conditions. But the strongest reason of all for considering consciously and explicitly the case of Crusoe by himself, is because there are many examples in the literature of value where certain doctrines are laid down ostensibly as universal principles, which are in reality conditioned upon the unconscious assumption of Crusoe conditions. The case of Crusoe should be discussed if only to show what principles are limited to the conditions of his economy.

Crusoe might possess various articles of value which cost him no labor, or cost him an entirely negligible amount, such as certain scarce fruits; but the major part of Crusoe’s wealth, let us suppose, is produced, and is freely reproducible, by his labor. The amount of labor power which he expends upon his island is variable at his pleasure within wide limits. If he choose to work only a few hours a day through the year, he will produce only a certain limited amount of the most useful things. If he add more hours per day, he will produce more of the old kinds of goods and also other different articles, luxuries as contrasted with necessities, but at any rate things of lower utility. Thus as Crusoe increases the hours of labor power put forth per day, he finds that there is a decline in the additional utility produced by each successive increment of labor. On the other hand, he finds that the pain cost or disutility of labor increases as he toils longer and longer. Crusoe will, for his average day, work until the increasing disutility of labor comes to an equality in his judgment with the decreasing utility of the things being produced. It would not be rational for him to stop at an earlier point, for then further labor would produce him a means of satisfaction greater than the “pain” of the labor itself. Nor would he labor beyond this point so that the pain would exceed the pleasure gain. Thus the utility produced by, and the disutility of, the final increment of labor in the working-day counterbalance, or we have the “marginal equivalence of utility and disutility.”

The most important part of Professor Clark’s teaching is that the disutility of labor expended in producing goods is the ultimate standard of their value. To establish this thesis, it is necessary for him to show that the “effective” disutility of an increment of labor is always the actual disutility of the final increment, and that the “effective” utility produced by any increment is the actual utility produced by the final increment. If Crusoe works ten hours a day, any hour of the day’s labor will have the same effective disutility as the tenth hour. Thus if it costs one hour of labor to produce the article A, the pain cost of A is always in effect the disutility of the final or tenth hour of labor, no matter in what part of the day the good A happens actually to be produced. For if Crusoe should decide to go without this article in order to avoid the hour of labor which it costs, his day would then consist of nine hours spent upon the rest of his products, and the effect would be to save himself the disutility of the tenth or last hour.[236] By shortening the working-day an hour he cannot do otherwise than save himself from suffering this marginal disutility.

In a similar way it can be shown that the effective utility of a product A produced by the first hour of labor is the same as the actual utility of the product B produced by the tenth hour. For if A were to be lost, destroyed, or traded away, another A could be produced in its stead through the sacrifice of a B, by turning the tenth hour from B to the production of an A. In effect, then, upon the possession of A is dependent merely the utility of B. The utility of a good is defined as its power to afford satisfaction, but the “effective utility” of a good as conceived by Professor Clark is, we may say, the power of that good over a man’s satisfactions, taking into consideration the adjustments he may make in his productive activities in case of the loss of that good. If he produces this same good over again, and instead goes without some other good, in effect he foregoes the utility of the latter, and this is precisely the effective utility of the first good.[237] Thus, in Professor Clark’s view, the utility produced by the last increment of a man’s labor affords a unit for measuring the effective utility to him of any and all the freely producible products of his labor. But it has already been shown that the disutility of this last increment of labor is equal to the utility produced by it. Thus terminal disutility becomes also an available unit of “subjective” or esteem value, and Professor Clark adopts this as the “ultimate standard of value” upon the ground that pain is a more convenient measure than pleasure.[238] The theory signifies that the subjective value to the isolated producer of any good whose supply depends on labor can be most conveniently measured by the labor cost of that good. To quote: