Risk, Uncertainty, and Profit

Frank H. Knight
Knight, Frank H.
Display paragraphs in this book containing:
First Pub. Date
Boston, MA: Hart, Schaffner & Marx; Houghton Mifflin Co.
Pub. Date
1st edition. Based on award-winning dissertation essay.

Part II, Chapter IV

Joint Production and Capitalization


The present chapter will bring a greater semblance of reality into the imaginary, highly simplified economic system partially constructed above. Many of the features of everyday life abstracted for simplification can now be introduced in succession and their relations and bearings separately studied. In this way we shall ultimately determine what is necessary to perfect competition and what is not. It will be found that most of the simplifying assumptions hitherto made can be dropped without destroying the conditions necessary to a perfect equilibrium in which costs and values are identical throughout. So long as we adhere to the fundamental condition already emphasized, that men know exactly what they are doing, that no uncertainty is present, other elements of reality hitherto abstracted merely complicate the process of adjustment without changing the character of the result. Their elimination has served the necessary end of simplifying the study of the fundamentals of economic behavior and made possible the separate study of these complicating considerations themselves, which we shall now undertake.


The first step in this further development of the imaginary social structure is to examine the nature and bearings of organized production. Hitherto our society has been arbitrarily restricted to the unorganized or individual creation of goods; there has been only "primary" division of labor, through the exchange of products. We now turn to consider "secondary" division of labor, or division of occupations within the separate industries, the coöperation of a large number of persons in the making of a single product. This added element in the situation gives us two serious new problems, though closely related; first, the mechanism of the actual organization of productive groups through free contract alone, and, second, the division of a joint product among the individuals making different kinds of contributions to its production. The latter is the familiar problem of "imputation" (Zurechnung) or "distribution" in the technical sense.


Practically speaking, we are now turning to the second general problem of economics as it is met with in the real world. For methodological reasons we have, indeed, found it necessary to discuss a society in which specialized production takes place, but not joint production. In reality, of course, production is joint, practically without exception. The subject for discussion now is, therefore, the general principles of social organization under free exchange where given resources are used (in the production of goods) for the satisfaction of given wants (and under given conditions as to available methods of technical organization, etc.). It is the problem of the "static state." In order to keep the problems of the organization of production and the division of the product as simple as possible and to introduce complicating factors one at a time, no other changes are now to be made in the arbitrary specifications of the system we are studying. In regard to production particularly, we assume the absolutely continuous creation of the complete article and its immediate exchange and consumption when complete, and the absence of productive "property" in the ordinary sense.*34 That is, there are to be no material productive agents which are not either superabundant, and therefore free, or else rigidly attached to the persons of their owners, and no way is to be open either to increase the productive efficiency of person or thing or to decrease it through use. The only change now introduced in the conditions of our problem is that at least a large part of the commodities produced and consumed in our society are to be made by groups of individuals, performing a number of different kinds of productive work. It is not necessary that every individual perform a unique function; rather let it be typically true that considerable numbers perform the same sort of work and that there are gradations of similarity in the different tasks.*35


The possibility of an automatic organization of production through free agreements between individuals depends upon a technological principle governing joint production and not hitherto introduced. This new axiom is as fundamental to economic thought and process as the principle of choice or diminishing utility, and very similar to it in statement. It is the principle of the variation of proportions in the factors of production, already long famous under the name of "diminishing returns," though its clear and approximately accurate formulation in general terms is a relatively recent achievement. This new law is a generalization from the facts of physical nature as the former is a generalization from the facts of human nature. Like the other, and all other "laws," it is an approximation, and its approximateness must be kept in mind in making practical applications of conclusions resting on it as a premise. Like the other great axioms in economics, it is purely a principle of relativity, dealing with proportions only. In this respect the current statements of the principle are generally less misleading than in the case of diminishing utility, there being less temptation to give it an absolutistic interpretation. It does seem strange, however, that it took economists so long (nearly a century) to recognize the inherent reversibility of a change in proportions and to draw the obvious inferences from the fact. We may observe finally that the new principle is much "truer"; i.e., more universally and accurately in conformity with the facts, more dependable, than its psychological counterpart.


In many other respects, also, there is similarity between the two fundamental principles of proportionality, the psychological law of diminishing utility and the technological one of diminishing returns. A formal and accurate statement of either presupposes continuous divisibility of the variable element, which is not true to fact in a particular case, but which does hold good with practical accuracy in a large market. In both cases divisibility breaks down completely (in an individual case) for minimum amounts. As there is a definite minimum quantity of any consumption good required to give it any significance, so there are limits to the proportions of productivity agencies which will yield any effect at all. As to minima in the case of consumption goods in the different sense of minimum amounts necessary to life, this, though commonly assumed, is ordinarily not true. It is only under very special circumstances that any particular commodity, as the market defines and differentiates commodities (and this is the only sound or relevant method), is indispensable.


In the case of both the law of diminishing utility and that of diminishing returns, also, there are maxima to be taken into account beyond which the good or agency ceases to enter into problems of conduct at all, becoming a "free good"—better called a potential good, as we have seen. The correct procedure is of course to treat superabundant elements in production as we did those in consumption; i.e., to take them absolutely for granted and ignore them completely. Only the "possibility" of a situation arising in which a thing would not be superabundant can give it significance or lead to its being consciously considered in any way.


In discussing the principle of diminishing returns a special difficulty arises in the confusion of varying proportions in a combination with changes in the absolute size of the combination as a whole. These things must imperatively be kept separate; in the writer's opinion more error has arisen over this point than any other single matter in distributive theory. If the amounts of all elements in a combination were freely variable without limit and the product also continuously divisible, it is evident that one size of combination would be precisely similar in its workings to any other similarly composed. But under this condition the tendency to monopoly in the production of every good would be unimpeded. For the competitive system to work, it is necessary to postulate that the conditions as to divisibility of factors are such that the bargaining unit of any one factor is quite small in relation to the total stock of agencies which more or less effectively compete with that unit, and also that an establishment of relatively small size in proportion to the industry as a whole is more efficient than a larger one. Under these conditions the first effect of competition must be to bring all the plants within an industry to the most economical size, and leave a sufficient number in operation to compete effectively for the productive agencies which all use.*36


The principle of diminishing returns in its now current form runs somewhat as follows: As successive increments of any one agency are added to fixed amounts of other agencies in a combination, the physical product of the combination will increase, but after a certain point the output will increase in less proportion than that of the agency in question and will ultimately decrease absolutely.*37 A more general formulation, emphasizing the reference to proportionality in contrast with absolute size, and the reversibility of the law, might run as follows: When the proportion of agencies in a combination is continuously varied over a very wide range, there is generally a first stage in which the product per unit of either agency increases; then a stage in which the product per unit of the relatively increased agency decreases and the product per unit of the relatively decreased agency increases; and finally a third stage in which the product relative to either agency decreases. Since either agency may be the increasing and the other the decreasing one, the first and third stages are identical in meaning.*38


It is requisite for an intelligent organization of production and a determinate division of the produce among the factors by competitive price forces that not merely the product increase in less ratio than the factor, but that equal arithmetic increments of factor yield decreasing increments of product. These two principles have entirely different meanings, of course, but they are badly confused in many statements of the theory of diminishing returns. The second can, however, be deduced from the first, which follows from the very nature of an economic situation, as shown below. The relations of the various elements in the problem can best be shown by reference to a graph. In the accompanying figure, the horizontal Graph IV.1.  Click to enlarge in new window.or X distances represent quantities of the single variable productive factor in a combination, and the vertical or Y distances, the corresponding total physical output of the group. In graphic terms the point where diminishing returns begin is the point (3) where this curve becomes tangent to a straight line through the origin. Less than this proportion of the variable agent cannot intelligently be employed even if it is free, for the output could be increased by discarding a portion of the other factors, if no more of the variable one could be obtained at a uniform price. It is true, necessarily and a priori, that there is such a point on the curve, that for less amounts the product increases in greater ratio than the factor. That is, for any point on the curve between this point (3) and the intersection of the curve with the X axis the tangent must cut the X axis positively. Now, if below this point (3) the tangent to the curve cuts the positive X axis, if at this point it passes through the origin and beyond this point it cuts the positive Y axis, then manifestly the curve is concave downward at the point in question. And this is the graphic condition representing decreasing increments of product. It seems reasonable to assume that the same condition (concavity downward) holds from point 3 to the maximum point (4), but this is not demonstrable a priori. If it is untrue for a certain stage in this interval between points 3 and 4 over the whole field of industry, as represented by the dotted line in the figure, there is indeterminateness in the competitive situation in that interval and to that extent, but this is a rather incredible supposition.


It is immaterial what shape the curve has below point 3 so long as its tangent always cuts the X axis. No doubt in any one industry the curve will show stages of increasing returns interspersed with stages of decreasing returns, and various proportions of combination of the factors are wise and stable.*39


If men are supposed to know what they are doing there is no occasion for discussing the first and third stages at all. The boundaries of the second stage represent extreme limits where one agency or the other becomes a free good and passes out of consideration altogether. Beyond this point the product is absolutely diminished by increasing one agency or the other, as the case may be, which is an absurdity. The identity in meaning of the first and the third stages is evident; the first stage when passing in one direction is the third when reading the data in the opposite order. It is a mere matter of the arrangement of results, not of the results themselves. Beyond the limits of the stage of "decreasing returns," therefore, or under circumstances where the law did not hold, there could not exist an "economic" situation. Unless the return per unit of any agency does decrease it is not productive at all; its use adds nothing to the output of the combination. If we imagine increasing returns the agency is negatively productive. This fact has been recognized in the case of land in the common statement that additional land would never be taken up until diminishing returns set in on that*40 already in use.


The facts of variability in the proportions of agencies in the productive organization, and of the variation of the yield relative to the different agencies in accordance with the principle of diminishing returns not merely make possible the economic organization of society through free contract, but in their absence the whole question of organization would be meaningless; there would be no such problem. Unless there were open for use various combinations of various productivities, with the possibility of comparing them, there would be no question of using any one arrangement rather than any other. Organization is called for, is possible, and is carried out only through the fact that the separate contributions of separate agencies to a joint product can be identified. The organization through free contract under competition is possible and real and effective in so far as such a system tends to give to the owner of each agency the separate contribution of that agency. Modern society is organized through the association of control over productive agencies with the right to their yield. Only because the income is greater where the product is larger is such organization possible at all. In the absence of a law connecting distributive share with effective contribution our social system would be no system, but chaos. It is, therefore, inappropriate for economists to argue as to whether the separation of contributions to a joint product can or cannot be made; it is made; it is our business to explain the mechanism by which it is accomplished.


The business man does find out how much different agencies or units of productive power are worth to the productive process or he could not carry on his business. It is obvious that the business man, in bidding for the use of separate agencies, must think in terms of the added contributions of added units,—in technical economic parlance the "marginal" product,—and it is demonstrable that when the units are sufficiently small the sum of the separate, specific contribution of all the agencies exhausts the total joint product.*41


It is to be observed that when a new productive unit is added to a productive combination the technical law of diminishing returns does not fully describe the variation in the output. In consequence of this law alone, the added physical product of similar agencies will rise in the position from which the one in question is withdrawn and fall in that into which it moves.*42 But in addition, since the transfer decreases the total output of the commodity from whose production the agency is withdrawn, and increases the output of the industry into which it is moved, the price of the former will rise and of the latter fall relatively. In an organized free exchange society, producers naturally estimate product in terms of its exchange value and not of its physical magnitude. The variations in physical contribution and in the value of that contribution when an addition of any kind of agency is made, work in the same direction and must be added to give the total decrease in the value product. We shall call the aggregate variation by the name of diminishing value productivity or simply diminishing productivity, which must always be distinguished from the diminishing physical returns.*43


It is unnecessary to introduce into our society any factors or agencies other than labor in order to study the mechanism of imputation. Groups of individuals more or less specialized to and specializing in different productive functions in the making of the same commodity represent in principle all that is involved in the coöperation of agencies of whatever difference in nature. We may, therefore, refer to these different functionaries as types of agencies, or indeed as "factors" of production, though we shall presently find reasons for avoiding this term, on account of its misleading connotations. When the conditions of a "static" society—i.e., given conditions of the production and consumption of goods—are correctly laid down, there is, as we have seen, no room for property in any sense which differentiates it from productive capacities inherent in the person of the owner.*44


This matter will be discussed at greater length as we proceed. Let it merely be understood at this point that any class or group of agencies, or "factor" of production to which we refer, is formed on the basis of the physical facts and includes those things which are actually interchangeable one with another in the production process. If we speak of "factors" at all, there will thus be not three, but a quite indefinitely large number of them.*45


As a matter of fact, a great deal of unnecessary mystification has been thrown around the problem of imputation. It is merely a case of joint demand, and the same situation is common in the case of consumption goods. There is really no more mystery or special difficulty about separating the demand for labor or any particular kind of labor, due to the fact that it is not employed alone, than there is about constructing a separate demand curve for butter, which is always consumed along with other commodities. The principle of variable proportions is the key to the solution in both cases. Commodities always used together and always in the same proportions would not be separate commodities, as far as consumption is concerned, but parts of one commodity, though they might still be valued separately if the conditions of production were distinct.


Keeping in mind the above facts and the simplified conditions under which we are working, it is not difficult to picture the actual mechanism of the organization. Let us begin as in the last chapter with a random adjustment and follow through the successive readjustments to the equilibrium condition. Suppose that groups of producers are formed by guess in any chance way, the product of each group as a whole being determined in the manner already described and its division among the members of the group arranged on any basis whatever. It is evident that the desire of every individual to better himself will lead at once to three sorts of inquiries. First, each person will endeavor to ascertain his own value to the group of which he is a member and compare it with the share which he is receiving; and second, he will similarly inquire what he might be worth to other groups. Third, as a member of a group each individual will interest himself in the value to the group of other individuals in it and in the value which individuals outside it would have if they could be procured for his group. As a result, (1) remunerations will rapidly be readjusted toward the values which the individuals contribute to the output of the groups with which they work, and (2) all individuals will gravitate toward those groups in which they can make the largest contributions to output. Any individual receiving from his group more than he is worth will be released or have his remuneration reduced. Any individual receiving less than he is worth will be able to secure his full value,*46 since we have specified conditions under which perfect competition will exist between the groups.


All productive groups would thus compete among themselves for the services of actual and potential members, and the individuals in the society would compete for positions in the group in a manner quite analogous to the existing order of things. The standard of what a group could afford to pay for a man is clearly the amount which he enables it to produce more than it would produce without him. In the final adjustment the individual's contribution to the income of the group is his contribution to the income of society as a whole, which he is under pressure to make as large as possible by placing himself in the position where he is really most effective.*47 The tendency of a competitive organization is, therefore, toward that ideal adjustment familiar in the literature of laissez-faire. In the final adjustment the organization could not be changed without bringing uncompensated losses, and the total produce would be divided among all claimants by giving each his added product.*48


The conditions precedent to this theoretical result are indeed abstract; but they are the conditions of perfect competition, and they are the conditions which actual society more or less closely approaches. It is important both to understand free competition because society does approach it more or less closely as an ideal, and to be fully aware of the artificiality of the conditions necessary to realize it perfectly.


Another way of formulating the condition of equlibrium is to view the adjustment as a continual repricing of productive services. This process would be more closely analogous to the process by which the prices of consumption goods are determined. We can think of each producer or group as being in the market with a certain amount of money to spend for productive power in the abstract. At the price level established at any moment those productive agencies will, of course, be purchased which make the largest price contribution to product for a given price outlay. But since the amounts of all agencies in existence are fixed, competition will quickly force a readjustment of prices to that point at which equal price amounts of all agencies make equal price contributions to product, just as in the former case equal price amounts of all goods must represent "equal utilities" to all consumers. The organization of the productive system as a whole is in fact quite analogous to that of the expenditure of income. Productive agencies are now the given resources of which the best use is to be made by distributing them so as to secure equality of remuneration for similar units in all employments. In the organization as a whole, the two principles combine. The money income may be omitted, as an instrumental intermediary, and the result stated by saying that the real resources of society tend to be so distributed among all employments that similar physical units everywhere make contributions psychically equivalent to all persons in the system in a position to choose between them.


It will now be in order to notice the more important objections which have been made to the productivity theory of distribution, though many or all of them have already been answered and probably would not be made against the form of the theory presented above. To begin with, let us insist on the complete separation of the theory of distribution proper from certain sweeping moral and social dogmas, which have been deduced from it. Professor J. B. Clark, the leading American exponent of the theory, is partly responsible for this confusion, through a few unguarded paragraphs in "The Distribution of Wealth."*49 The illegitimacy of these ethical deductions has been well argued, however, by Professor Carver,*50 another expositor of the theory, as well as by Professor J. M. Clark in defending the theory itself.*51 We may, therefore, pass over the strictures of those writers who do not like social implications which the theory does not have, which include a considerable part of the criticism of Professors Davenport*52 and Adriance;*53 we shall take up briefly the question of the ethical aspects of the competitive system in chapter VI.


Against the productivity theory itself an old and common criticism is that well stated by Wieser,*54 who attempts to refute Menger's presentation of it, and substantially the same line of attack has been followed more recently by Hobson,*55 who refers especially to Wicksteed. The contention is that specific or marginal productivity cannot afford a theoretically adequate method of distribution, for the reason that the sum of the products of the separate agencies, as defined by the theory, will be not equal to the total joint product, but considerably larger. The amount subtracted from the total product when "one unit" is withdrawn will, it is argued, be much greater than can be imputed to that agent alone, since the loss of any agent will more or less dislocate the organization. It, therefore, becomes impossible by this method to divide the total accurately into parts ascribable to the separate "factors" individually as the specific contribution of each. Wieser proposes an alternative method, which is identical with Professor F. M. Taylor's exposition of the productivity theory itself.*56 Hobson dogmatically declares the problem impossible.


The error in this line of reasoning lies in fixing the attention upon a comparatively small organization and comparatively large blocks or units of productive service. When account is taken of the actual size of industrial society and of the ordinary unit of most agencies, it will be seen that the "dislocation" is negligible; theoretically, to be sure, the units would have to be of infinitesimal size, separately owned and effectively competing; i.e., the proportions must be continuously variable, in the mathematical sense. But in the typical case the error resulting from this assumption is not large in comparison with other inaccuracies in the competitive adjustment. It is true that there are exceptional cases where agencies are not highly divisible, or even not divisible at all, and competition gives place to a greater or less degree of monopoly. These exceptions are relatively infrequent in the mass of industry as a whole, but are of considerable absolute importance, and we shall have something to say later on in regard to unique and indivisible agencies.*57


Padan, in the article referred to, further attacks Professor Clark's exposition of the productivity theory on the express ground that the amount received by any factor would depend on the arbitrary size assigned to the marginal unit. This point also is hypothetically sound, but irrelevant. The size of the unit is not an arbitrary matter of methodology, but a question of fact, and Professor Clark may be open to criticism only for seeming to imply the contrary. The soundness of the theory, the possibility of competitive distribution at all, in fact, depends on the actual division of productive agencies into bargaining units of small size.*58 We should hold that it is an error to say that "labor" or any "factor" gets or tends to get its product. This holds good only for the actual individual men or other agencies.


A third, somewhat philosophical, criticism is also advanced by Davenport and Adriance. It is contended that the "marginal" product of labor, for example, is as much a joint product as that of any other than the marginal unit. The laborer who uses no-rent land still has to use it, can produce nothing without it, and hence the product cannot be ascribed to the labor alone. Professor Taussig also, though like Davenport somewhat guardedly, asserts that all product is joint product and cannot be divided into parcels attributable to separate agencies, though at the same time he inclines to regard all income as the "product" of labor.*59 An examination of this reasoning would carry us into the question of the meaning of production and causality, which will be taken up presently. For the present it must suffice to point out that it involves a confusion between mechanical and economic productivity. The land used by marginal labor may be necessary to the operations in the former sense, but is not in the latter, since by hypothesis if it is withheld from use it can at once be replaced by other land equally good; otherwise it would not be free land. The fallacy is parallel to the confusion between "utility" (as usually defined) and economic value. Free goods, like air, may be necessary to life, but no particular portion being necessary, the good cannot have economic value (nor, as we have argued above, should it be said to have utility if this term is to be used to connote any sort of economic significance).


We must notice, finally, another objection raised by Hobson to the general doctrine of "marginalism."*60 With Hobson's fundamental position, that marginalism is the necessary form of a rational treatment of choice, and that the rational view of life is subject to drastic limitations, the writer is in hearty accord. It is not clear that Hobson intends his strictures to apply specifically to the productivity theory of distribution, but it may not be out of place to remark that such an application would be an error. In general we submit that there is much more deliberate, quantitative balancing of alternatives in economic conduct than the discussion under notice would have us believe, but this is a large issue which cannot be threshed out here. It does not seem to us that the composition of life is closely analogous to Hobson's painting or cake in which the proportion of the ingredients is rigidly determined by a recipe or a preconceived ideal of the whole. In any case, the production of goods by industry is very emphatically a rational process, an adjustment worked out by the producer in terms of these very separable effects of separate agencies. Nor is it true, as Hobson does argue elsewhere,*61 that technical conditions prescribe the proportions in which agencies are to be used. The proportions of labor to land and of capital to either, and to a large extent of various sorts of each among themselves, are open to variation through a range almost without technical limit, in the fundamental industries at least. Again, the final appeal is to fact. It is the value to the producer as an addition to his organization as a whole which determines the amount which he will bid in the market for the use of any unit of labor, land, or capital, or the amount of any one which he will purchase at an established price. Hence it is this "specific product" which rules the apportionment of income at large among productive agencies at large.


As remarked above, most of the objections to the productivity theory relate to the meaning of production and of product, and come down in fine to the propriety of using the word, rather than to any fundamental disagreement as to how the distributive mechanism actually works. We wish now to point out that in calling the addition made by any agency to the total output of a large organization its specific or separate product, we are using the word "product" in the same meaning and the only meaning which the words "cause" and "effect" or equivalent terms ever have. It is never true in an absolute sense that one event is the cause of another. The whole state of the universe at one moment may perhaps be said to cause its whole state at the next moment, but when we say that "A" is the "cause" of "B" we always assume that other things are equal; we never mean that if the rest of the universe were removed "A" alone would produce "B." And the imputation of any single event to another as cause or effect is always largely arbitrary. Every event has an infinite number of causes, and it depends upon circumstances, the point of view, the problem in hand, which of these we single out for designation as "The" cause. "The" cause of a phenomenon is merely that one of its necessary conditions which is for some practical reason crucial, generally from the standpoint of control. It is the one about which we must concern ourselves, the circumstances enabling us to take the others for granted. It may be quite correct to name a dozen different antecedents as "the" cause of a particular occurrence, according to the point of view. The fact that other agencies, even the whole social system, may be concerned in the production of a certain good does not therefore argue against its being the (specific) product of the particular agency upon whose activity its creation actually hinges under the actual circumstances of the case.*62


A general analytic statement of the principles of static organization, in price terms and on the basis of supply and demand, will consist of two main parts. We have to consider two valuation problems relating respectively to consumption goods and productive services. The problems are usually designated as "value" and "distribution." It will be convenient to take up the second of these problems first. We have already seen that the effective form of the law of variation of proportions of factors is the law of diminishing value productivity. It is obvious that all readjustments involve transfers of productive resources and that every such transfer implies a price change, raising the prices of goods produced by the organization from which resources are taken and lowering the prices of goods to whose production resources are diverted. And the effect of this price change coincides in direction with the effect of diminishing physical returns. We may content ourselves for the present with this superficial view of the price reactions on the side of consumption goods and proceed to work out the price conditions of equilibrium of the system in terms of the distributive shares. After which the viewpoint will be shifted to regard these shares, not as the remunerations of agencies, but as costs of the goods into which their services enter. When the adjustment and its equilibrium have been studied as a relation between prices and costs of consumption goods, we can bring the two analyses together and see the relations of the three sets of price facts—values of goods, costs of goods, and values of productive services. It is obvious that as aggregates the three concepts are identical, all being in fact the social income looked at from different points of view.


From the standpoint of the present problem of the "static state" the supplies of all productive agencies are rigidly fixed, and the theory of the valuation of their services is closely parallel to the market price theory as given in the last chapter for consumption goods. The facts of demand and supply for any particular kind of agency can be presented in the form of schedules or graphs showing the respective amounts that will be forthcoming and that can be sold at each price, and the equilibrium point would be manifest in such a presentation. The facts on both the supply and demand sides of the relation are more complicated than in the case of consumption goods. On the supply side we cannot take the amount in existence even at a moment as a given physical datum. For we are dealing with the services of a particular kind of agency, not the agency as such. The amount of the agency is fixed, but the amount of marketable service forthcoming from it may well vary with the price offered. Two courses are open. We may define and classify services on the basis of the physical characteristics of the agencies which render them or in terms of the physical result produced.*63 Let us take first agencies as physically defined. In this case the effect of the substitution of more or less similar agencies is to be taken into account in plotting the demand curve; supply means the supply of the services of a particular kind of physical agent, things which are perfectly homogeneous and universally interchangeable alone being grouped together.


It is usual, because superficially "natural" to assume that a man will work more—i.e., work harder or more hours per day—for a higher wage than for a lower one. But a little examination will show that this assumption is for rational behavior incorrect. In so far as men act rationally—i.e., from fixed motives subject to the law of diminishing utility—they will at a higher rate divide their time between wage-earning and non-industrial uses in such a way as to earn more money, indeed, but to work fewer hours. Just where the balance will be struck depends upon the shape of the curve of comparison between money (representing the group of things purchasable with money) and leisure (representing all non-pecuniary, alternative uses of time). We therefore draw our momentary supply line in terms of price with some downward slope.*64


The second alternative is to define agencies or factors in terms of the physical results which they produce. When this is done the shape of the supply curve at a moment will depend simply on the degree of specialization of the service under discussion. At one extreme we would have an unspecialized service, such as unskilled labor in a certain employment. For such a service there would be no supply at all below the established competitive price in all uses, and a virtually unlimited supply above that price. That is, the supply curve as a function of price would be a vertical line. At the other extreme would be absolutely specialized services, such as diamond cutters or aviators. For these there would be no supply below a certain minimum price, what such men can earn in other lines of work, and as the price rose the supply would rapidly increase until the men trained for the service were all employed in it, beyond which the curve would merge into the supply curve previously discussed of services from given agencies. (See accompanying graphs, which show supply as a function of price.) Graph IV.2.   Click to enlarge in new window.


In regard to demand, also, the case of productive services is less simple than that of consumptive goods; demand is (a) always indirect or derived, a reflection of the demand for the products of the agency, and (b) always joint in character. In connection with the first fact, the demand is also highly composite; identical productive agencies minister alternately to a vast range of wants and widely different agencies to the same wants. These complexities in the use of productive services make a really logical classification of them a difficult if not impossible problem. The fact of joint demand, as we have seen, differentiates producer's goods from consumer's goods in degree only, and to a relatively limited degree.


The shape of the demand curve showing possible sales of the services of any physically defined type of agency as a function of price is similar to that of the consumption goods demand curve. It is the curve of diminishing value productivity already described, descending in consequence both of decreasing physical productivity and decreasing price. That is, if the supply of any productive agency be increased the proportion of that agency in combinations in which it is employed will be raised all along the line, and at the same time there will be a relative increase in the production of those commodities in which its use is relatively important with a consequent decline in their relative price. The equilibrium price point under static conditions is practically the specific productivity of the given supply of the agency (though we must remember that there is some variation in supply of service as price varies even at a moment). In the equilibrium condition, that is to say, the value of each service is equal to the value of its contribution to the total product, and the contributions of physically similar agencies are of equal value throughout the system. It is evident that this adjustment fixes the prices of consumption goods at the same time with those of productive services, and we may apply the supply and demand analysis to consumption goods also, giving the theory of normal price in contrast with the theory of market price studied in the last chapter.


At a moment, the theoretical price of any good is the ("marginal") demand price of the existing supply, the highest uniform price that will take the supply out of the market. The supply is a given physical fact, not an economic variable, but a constant in the equation. The equilibrium price of a good over a long period is a different problem. Here it is not the amount of the good that is constant (together with the facts of demand), but (under "static" conditions) the conditions of production of goods in general (and of demand). The supply of any particular good may change freely and will do so as its price varies, other things being equal. The price must be adjusted not to dispose of a fixed supply, but to equate a rate*65 of production with a rate of consumption, both variable with or "functions of " the price.


No particular reinterpretation of the demand curve is called for, however, the only new problem being on the supply side. Assuming for the moment that the rate of supply as well as the rate of demand is in fact a function of price, it is evident that the price must move toward an equilibrium point equating the two rates; for goods cannot be consumed more rapidly than they are produced and will not be produced more rapidly than they are consumed. Any difference either way will at once react on the price and the price will react on the production and consumption rates in accordance with the assumed functional relations, and so on until the demand and supply both correspond to the existing price.


To investigate the basis and character of the relation between supply and price, we must consider the motives which control production. The productive group or establishment, however organized, must pay its members (the owners of productive services) enough to retain them; i.e., it must meet competition. When any group can hire a new member at a profit it will do so, and clearly it can get any new member by raising ever so little the remuneration he is receiving elsewhere. Clearly, also, it will dispense with any member who must be employed at a loss; i.e., any to whom competing groups can afford to pay more than it can afford to pay. The amount of any commodity that will be produced at any price, therefore, tends quickly toward the amount that will yield neither profit nor loss, for when production yields ever so little profit it will increase, and vice versa. For the study of this adjustment it is convenient to interchange the axes of our previous graph and view cost and selling price as functions of the size of supply.


It is usually assumed that cost may either increase, remain constant or decrease as supply is increased.*66 (Selling price, of course, practically always decreases.) The question is really one of the most difficult and perhaps one of the worst muddled in economic theory and cannot be adequately treated here. But examination seems to show that under the conditions necessary to perfect competition, costs must always increase as supply increases. If there is to be competition, conditions must be such that an establishment of relatively small size in comparison with the industry as a whole is more efficient than a large one; otherwise monopoly will result. New supply will then come through an increase in the number of similar establishments, not through an increase in the size of any of them, and no economies of large-scale production will be realized.


On the contrary, the increased supply must mean a diversion of productive resources from other uses, which will raise their price in those uses through the decreased output and consequent rise in price of the competing product. Of course, if competition exists the price will go up uniformly to all producers, and it goes without saying that the cost of all units of the supply is the same.*67


The precise form of the cost function will depend on the importance of the particular good in the demand for the productive services which enter into it. If its production constitutes a negligible fraction of the demand for all these services, we shall have practically constant cost; if a considerable fraction, a more rapidly rising cost. It will also vary with the character of the function representing the law of decreasing returns in the given technological situation; for as production is increased the proportions of more abundant agencies will be increased relatively to those more limited in supply. The graph on p. 91 shows the character of the functions and the meaning of equilibrium, and is applicable also to conditions of joint production.


The equilibrium condition or long-run tendency for the static state has now been formulated in three ways from as many different standpoints. From the standpoint of distribution, every agency must be in the situation where it can make the greatest possible value contribution to the social income and be valued by the contribution which it makes. From the standpoint of consumption goods, prices must be such that rates of production and consumption are equal or that costs and selling prices per unit are everywhere the same. It is important to see clearly that these statements are logically equivalent, presenting different aspects of the same phenomena. It is self-evident that costs of goods are identical in the aggregate with distributive shares, and both with prices of goods; all three are in fact different names for the total income of the society. A formulation including all these statements would be that consumption goods and productive services must be so priced that equal price amounts of the second make equal price contributions of the first which have equal utilities to all persons in the system. It is really self-evident that this condition alone can be stable, that any other sets forces to work to bring it about.


Hitherto we have dealt only with different sorts of human services as giving rise to the phenomena of competitive imputation. The meaning and rôle of property in the problem of economic organization next call for notice. We have seen that material productive goods do not modify the principles of organization so long as they are not subject to increase or decrease and not separable from the persons of their owners, to whose personal capacities the same restrictions must apply.


The conventional classification of productive agencies under the three categories of land, labor, and capital has several times in the foregoing pages been referred to adversely, and it is appropriate at this point to take up for somewhat more detailed notice the difficult problem of correct definition and classification. It is evident that all these classes are anything but homogeneous, that different human beings, different machines, and different natural agents show the greatest diversity in characteristics and in the services which they perform. Cairnes's attempt to reduce labor to more approximately homogeneous bodies gave us the famous "non-competing groups." Still more obtrusive are the dissimilarities of different natural agents—wheat land vs. pineapple land, arable vs. grazing or timber, and all contrasted with mineral-bearing and the multitudinous kinds of the latter. Capital is somewhat peculiar in this respect, its "fluidity" depending on the length of time taken into view.


On the other hand, it is if possible a more important fact that agencies from different classes and of the most divergent physical properties may be equivalent and interchangeable with respect to the results which they achieve. As Carver has observed, a (human) ditch-digger is economically as closely akin to a steam shovel as he is to a bookkeeper.*68 Indeed, the possibility of a competitive organization of society depends on the fact of varying proportions, that no particular agency is indispensable, but that within limits they may be substituted for each other and therefore each must compete with others of different kinds for its place. It is evident that otherwise producers would not be in the market for the agencies separately and they could not be separately evaluated through competitive bidding. The existence of a problem of distribution depends on the coöperation of different kinds of agencies performing physically different operations in the creation of product, and the possibility of solving the problem depends on the equivalence of determinate amounts of the several services in contributing to the value result. It follows at once that, as already observed, no classification or measurement of productive services on the basis of their contributions has any meaning for the distribution problem. According to such a standard they all form one vast homogeneous fund.*69


The problem is really a difficult one, and cannot be passed over, since we cannot discuss the valuation of things without knowing what it is that is being evaluated. Much the same difficulty, however, was met with, as will be recalled, in the sphere of consumption goods, and the answer must come from the same source in the two cases—an appeal to the unsophisticated facts of the market. Things quoted under the same name and identically priced may be taken as identical, and vice versa. Some special features of the present case may be mentioned, however. In the first place, interchangeability of productive agents depends on the use; two things may be equivalent for one purpose, entirely dissimilar for another. This is not nearly so true of consumption goods, which, indeed, are not generally open to such a complex variety of uses. Interchangeability is also a matter of time. The problem of changing the form of productive agencies and adapting them to new uses carries us into long-time considerations, and especially the meaning of capital, which will come up in the next chapter. It will be seen that examination tends to widen the capital category greatly; most productive services ultimately represent a previous investment of resources of some sort.


The variation in interchangeability in different uses introduces a special complication which has caused confusion. The consideration which finally determines is not interchangeability in creating any particular physical product, but a certain amount of value. The former variety of interchangeability is not in fact a necessary condition for the operation of competitive distribution. If agencies are combined in different uses, effective substitution is secured through relative growth or decay of the different industries. We have previously remarked that Wieser, who repudiates the productivity theory of distribution as based on variation in proportions, puts forth the really equivalent theory, based on different proportions in different combinations. Taylor, however, takes the latter method for his explanation of the productivity theory, but points out that the two are equivalent. Both sorts of variations in proportion are, of course, concerned in the actual working of the market for productive services, and systematically occur together, as explained in our exposition of distribution theory just given.*70


To conclude this brief discussion of the productive services, we may merely notice the invalidity of four commonly assumed grounds of distinction between labor and property services: (1) Activity vs. passivity. It is characteristic of the enterprise organization that labor is directed by its employer, not its owner, in a way analogous to material equipment. Certainly there is in this respect no sharp difference between a free laborer and a horse, not to mention a slave, who would, of course, be property. Closely related is (2) the question of preference in the agency itself as to (a) the kind and (b) the amount of service to be performed. But here also there is at most a vague difference in degree; the owner of property quite commonly does have moral or sentimental reasons for restricting the field of its employment. We must not confuse the agency actually performing work with the personality of its owner, and it appears that a tool or a building or a piece of land is in this regard similar to a man's hand or brain. Similarly as to (b) the amount of work done. It may be urged that material agents do not care whether they work or not. But the ground for restricting hours of labor or taking a vacation is a possible alternative use for one's personal resources or the desire to conserve them unimpaired, and the same considerations apply to property resources.*71


(3) Another superficial difference which similarly dissolves under scrutiny relates to "sub-marginal" agencies—too poor in quality to be employed. It may be urged that there is no wageless labor analogous to free land. As a matter of fact, however, marginal and sub-marginal human beings are nearly as common and significant a phenomenon as in the case of land, and far surpass capital in this respect. Every man is a sub-marginal laborer for a considerable fraction of his life at each end of it, and institutions are full of sub-marginal men. And there are thousands and millions of other idle man-hours in a year which would be devoted to anything that brought in the least return above the competitive pay which would have to be given to the equipment necessary to employ them. On the other hand, the same fallacious reasoning noted in connection with overwork undoubtedly leads to the employment of large numbers who use equipment which would yield more product if employed in the "more intensive exploitation" of more competent workers.*72


(4) The most important alleged difference between property and personal powers, the moral aspect, is not strictly within the scope of a purely descriptive discussion such as the present, but it may be in place to observe that it also is largely unreal. The contrast between personal-service income as "earned" and property income as "unearned," of which much is made by "reformers," is distinctly misleading; it is difficult if not impossible to find grounds for a moral distinction of any general validity between the two. "Some are born great, some achieve greatness, and some have greatness thrust upon them"; and the same applies quite as well to wealth. And the task of separating the portion of product or capacity to produce which is due to conscientious effort from that which goes back to inherited advantage or pure luck is about as impossible—and the evil results of making a false separation perhaps about as great—in one case as in the other. There is a difference of some significance in the practical possibility of effecting a redistribution in the two cases, which brings us back to the one specification which we found it necessary to lay down in regard to property in order to exclude it as a complicating fact; it is separable from the person of its owner, and labor generally is not, or is so to nothing like the same degree. The only conclusion as to social policy which we shall insert here is the insistence that "society" must get rid of the idea that because income is "earned" it is "deserved" and not otherwise. We are already far from this view in practice, as is shown by the indiscriminate taxation of large "service" incomes and assistance of the unfortunate and incapable. If we are to have organized society and maintain human standards of life, we must either radically eliminate weakness or impose upon strength the burdens which weakness cannot bear. (And even then there are limits to the possible toleration of weakness, and the luck element would still remain!)


Turning again now to consider the causal relations to economic organization of the one causally significant distinguishing attribute of property, let us first suppose that in our society some property is separable by lease, though not by sale, from the person of its owner. The only difference will be that the owner of such property may belong to more than one productive group and contribute more than one kind of service at the same time. The principles of organization of the system as a whole are in no wise affected by this change in the conditions of competitive arrangements.


The possibility of the permanent transfer of property by exchange, even though not subject to increase or decrease, does introduce some new factors into our problem. These results are closely related to the bearings of another abstraction hitherto made, the continuity and timelessness of the production-consumption process. Consequently, we must first get rid of this simplification and consider the effect of the abstracted element. What then will happen in a society such as we have studied when conditions are so modified in the direction of reality that, while perfect knowledge and static conditions in other respects are maintained, the production process is protracted over a considerable period of time and split up into complicated stages and subdivisions, and when, moreover, goods need no longer be consumed at once when finished, but may be stored for future use, or exchanged?


The division of the productive process into stages carried on in different groups or plants is a detail connected with the time length of the process, but which we can pass over with brief notice. It is in fact a relatively accidental matter of organization, and under the "frictionless" conditions here assumed it would make no practical difference whether successive processes in the making of an article were integrated through the internal organization of a single group or through the external mechanism of market dealings between groups. Under these conditions there will be in existence at any time a complex aggregate of partial products, goods in process, which of course will have value. We must separate that element in the value of the partial products which is due merely to the stored-up productive energy which they contain from any modification of this value due to the direct psychical influence of the time which must elapse before they are ready for consumption.


The relation of time to the production and consumption of goods is a complicated and controversial question; while only a very brief discussion can be attempted here, it is necessary to make a superficial survey. The assumption of a general preference in human nature for present over future goods is so commonly and confidently made that some courage is required to call in question the foundations of the entire body of doctrine on the subject; yet it must be done. Most discussion of the subject is, in the writer's view, vitiated by a false conception of the nature of the problem. The fact of the existence of interest in society is wrongly taken as proving that men discount the future. The relation between interest and time preference is, in fact, inverted in this view. In a free market where interest can be obtained it is natural that men should esteem a present dollar equally with its amount at the current interest rate at a future date, since one can be freely exchanged for the other. Nor does the fact that men do not postpone all consumption of goods indefinitely into the future argue an ingrained abstract preference of present to future consumption. Neither do they wish to compress all the satisfactions of a lifetime into the present moment and fast forever after,*73 which act by the same reasoning would prove a disposition to discount the present in favor of the future.


The error in the current reasoning is a wrong choice of a zero point from which to measure time preference. The correct basis is not everything to-day and nothing in the future; a more sensible form of question would be this: If one had to choose between enjoyment to-day with abstinence to-morrow on the one hand, and abstinence to-day with enjoyment to-morrow, on the other, which would be more desirable, all other things being equal? Or better still, if a man were given his entire income for a year in a lump-sum payment on January first, how would he distribute its expenditure through the year? There would clearly be no question either of eating it all up the first day or saving it all till the last day; a zero time preference obviously means a uniform distribution in time. Any piling-up of consumption at an earlier date to be compensated by reduced consumption later on would be a real discount of the future, while to skimp now for the sake of plenty or luxury in the future would be to discount the present. Of course, we abstract from the element of uncertainty as to the future. We seem justified in pronouncing either tendency irrational if other things are really reduced to equality in the alternatives.*74


As to the facts of human nature it is safe to assume that different individuals would give the most varied forms of distribution. Doubtless few, if any, of these would conform to straight lines or smooth curves of any sort, ascending, descending, or level. Most would go in waves of greater or less period and amplitude, intervals of moderation or even abstemiousness alternating with "blow-outs" of various sorts and degrees. Irregularity seems in fact to be a virtue on its own account, at least to the spirited individual.*75 Whether there would be an upward or downward trend would depend also upon the individual. To many, a bird in the hand is worth two or more in the bush, while others take much thought for the morrow. Some children, as Marshall remarks, pick the plums out of the pudding to eat first, while others save them until the last, and many do not pick them out at all; and adults differ in the same way. The improvidence of savages is proverbial. Of course, the physical conditions of life set limits to the discounting process in both directions; we cannot enjoy to-morrow unless we live to-day, and many have learned at a cost that too high a rate of living in the present may have a similar effect upon the capacity for future enjoyment. No generalization in regard to the human race at large seems to be worth making, especially in view of the unreality of any simple assumptions as to the conditions surrounding the choice. The facts of mere prodigality on the one hand and mere miserliness on the other are indisputable and may be studied without attempting to strike any precise balance.


It is perhaps even more important at this point to insist that the mere question of time preference in consumption is relatively unimportant at best as an explanation of the phenomenon of saving. The disposition to spend or to save, to consume income in the present or to store up wealth, is much more influenced, in fact, by other motives.*76 Like human conduct in other respects it is mostly a matter of social standards, of what is "good form," "the thing" or not the thing to do. The fact of possessing an accumulation of goods confers social prestige and in addition vast power over one's fellows. Even where, as we are now assuming, productive employment is not open to wealth, the rich man will be in a position to make his favor solicited, his ill-will feared, and may, of course, turn his situation to material profit if so disposed. Accumulations are necessary to lavish displays or magnificence of any kind. On the other hand, we must suppose that where accumulation is limited to consumption goods, it will be subject to considerable costs, for storage, preservation, protection, and doubtless inevitable deterioration.*77


It will be evident that differences among the individual members of society in economic position and taste with reference to the time of use of goods create a situation in which exchange will be mutually advantageous. To one, a present or early allotment of goods in advance of his own production and against an obligation to repay later will be or seem a benefit, while to another, with an accumulated and growing idle stock, a dependable obligation*78 for the future delivery of a certain amount of value, may be highly preferable to the possession of the goods themselves.


If the balance of the time preference in the population as a whole is in favor of the present, no appreciable net accumulation of goods will take place. Those disposed to accumulate will transfer their surplus production as fast as made to others disposed to draw on the future. The conditions of supply and demand will establish a market ratio of exchange between present and future goods which in this case will show a premium on the present, the magnitude of the premium depending on the strength of the excess desire to anticipate the future. Obviously the premium on the present goods will constitute an additional motive for surplus production and a deterrent to surplus present consumption. The rate established will be that at which the amount of surplus present production will equal the amount of surplus present consumption. The repayment of loans does not affect the principles involved, as it is a repetition of the original transaction with the rôles of the parties interchanged. In the aggregate an excess of present consumption over current production is, of course, impossible.


If, on the other hand, the balance of time preference is on the side of a disposition to postpone, the result will be an excess for the time being of production over consumption with net accumulation in the society as a whole. The exchanges between present and future goods will establish a premium on the latter. The ratio at which exchanges take place must constantly be such as to equate the amounts of each sort of service offered in the market to the amount that will be taken at the price. With a premium on future goods, accumulation will continue at a rate depending in part on the amount of the premium, until the premium disappears or becomes equal to the cost of keeping the accumulated stocks. Any greater premium on the future is impossible as a permanent thing. But the conditions of accumulation might well be such that an indefinitely long time would be required to reach the equilibrium result. In that case the actual condition at any time is a premium on the future with progressive accumulation taking place.


The "premium" or time preference rate under the conditions described, though similar to (positive or negative) interest, must be distinguished from that phenomenon as it is met with in modern industrial life; it is, indeed, an element, but a relatively insignificant one, affecting the interest rate on loans of productive capital.*79


Time value, presentness or futureness, is perhaps best regarded as a special sort of utility in a good, like nutritive value or beauty or any other quality conferring or enhancing desirability. The rate of payment for it, where separated from other considerations, is evidently determined by "psychological" considerations on both the demand and supply sides, and the current interest theory of the psychological school is based on a confusion of this phenomenon with interest proper as a distributive share. The subject of interest proper will claim attention at a later stage of the discussion. We shall find that interest in the correct sense may not be met with at all in a society where uncertainty is absent, even if accumulated wealth is productively used and even if the society is progressive with respect to the accumulation of capital, if knowledge and foreknowledge are complete.


We may now return, and in view of the knowledge obtained of the rôle of time in economic conduct take up the relations of property in the simple sense of productive agencies separable from the persons of their owners and subject to lease and sale. It must be borne in mind that for the present we exclude any possibility of either increase or decrease in the property or any physical change of such a character as to modify its functioning. Such changes and their effects belong to our third division of economics, which deals with changes in the conditions of the production and consumption of wealth. To realize static conditions they must be abstracted. It will be convenient to refer to property of the sort we have in view as "land,"*80 since land has been conventionally treated as if qualitatively and quantitatively given once for all by nature. This is not at all the view of land which will be presented in this study when the time comes to discuss the subject. But it is a convenient name at this point for a productive agency of a certain described character. We assume, as a matter of course, that such property is limited in amount (i.e., subject to "diminishing returns") and that there is no other sort of property present in the society. On the production side, then, the side of demand, and in relation to functional distribution it will be exactly like other agencies (human services), but its presence may affect the personal distribution of income very considerably.


Supposing the final adjustment to have been reached in the organization of production, any piece of property such as described may be regarded as a right or title to a commodity or money income in perpetuity. As such, its bearings on conduct are closely related to the time distribution of consumption. A piece of land represents future goods in the very special form of a value income distributed uniformly throughout all future time. We may assume without argument that such a piece of property will be desirable and that under conditions of free contract a definite market rate of exchange between land and consumption goods will be established. More accurately this price will be a ratio between the income from the land (of which there is no significant measure other than its income) and a quantity of present goods also measured in value terms. The price could, therefore, be stated as a certain number of years' purchase or a rate per cent per annum, and represents the familiar phenomenon of capitalization. Our present problem is to formulate the conditions determining this capitalization rate.


Land will be in demand especially by persons disposed to store up wealth for future use; i.e., to discount the present. It is in effect future goods, but the manner of their distribution in the future imposes a new special limitation on the conditions of their demand. We have seen that it is reasonable and common for human beings to prefer future goods to present, within limits, as compared with a uniform distribution in time. Most civilized persons, in fact, plan for a rising standard of living through life rather than a constant, much less a falling one. But when infinite time comes under consideration the case is different.


Any finite amount of consumption or enjoyment distributed uniformly through infinite time becomes a zero rate of real income. Hence there must be an apparent discount on the future in the demand for perpetual income goods. Indeed, it is self-evident that future incomes must be discounted at some rate greater than zero or they would have infinite present worth. The discount of the present in favor of the future can hold good only for finite periods of time in a society where present goods are limited at all; i.e., under economic conditions. We must note also, however, that when a capitalization rate and a market price for land have been established, the land will be convertible at will into a fund of present consumption goods The existence of a free market for permanent income goods makes the apparent rate of time preference uniform for all real (finite) intervals. The individual who may not wish to keep on postponing to the end of a long period knows that he does not need to do so unless he wishes; for at any time he can realize upon his accumulation in present consumption form as rapidly as he may wish. There must be a premium on present over future goods in the market for perpetual income property; but such a premium, even if high, is not incompatible with a premium on the future over the present for any finite interval, and might perfectly well exist in a society where every individual and the group as a whole distributed its consumption in time in a curve ascending at any finite slope.


Under these conditions a person could arrange, by the purchase and sale of income property, for any desired distribution of consumption over any specified period, or, through an appropriate life insurance organization, over the uncertain period of his life. Those wishing to postpone consumption, to secure a rising distribution of real income, would buy such property in the earlier years and gradually sell it off in the later ones. Those wishing to anticipate future production and secure a descending curve of consumption would progressively sell off their land. (Persons possessing no land could make the anticipation arrangement only in the manner described above in discussing a situation where such goods were absent.) The society as a whole cannot anticipate future production unless there is some other society from which it can borrow. It can postpone in the aggregate only as in the situation above described, through an actual accumulation of consumption goods. The process of net accumulation would again tend toward an equilibrium with current production and consumption equal, though the goal might be an indefinite distance in the future. There must at any time be an equilibration of the two sorts of motives through the discount rate established, together with, in the case just mentioned, a certain rate of net accumulation.


The rate at which perpetual income goods are capitalized in the market is not yet a rate of interest in the sense of a distributive share. Nor would there be any necessity under the conditions we have described for lending money in connection with the transfer or use of income-bearing property (though consumption loans might be effected in much the familiar form). The capital loan for productive purposes is, as we shall presently see, a device for separating the ownership of value equities in production goods from the direct ownership of the goods themselves. It is mainly the presence of the risk or uncertainty factor which makes such a separation desirable. In a progressive society some motives for specializing to individuals other than the savers the function of making the investment might exist even in the absence of uncertainty. In the society which we have described with both uncertainty and progress absent, there would be no motive for lending or borrowing value funds for the purchase of productive agencies.

Notes for this chapter

See above, chapter III, pp. 76-80 [II.III.39-50—Ed.], for the assumptions under which we are working.
See note above, p. 86 n. [nn31—Ed.], on indifference as to the presence and use of money.
Competitive relations between similar establishments are much complicated in real life by the fact that practically every business enjoys a certain degree of partial monopoly. It does not turn out exactly the same product (bundle of utilities) as its competitors. An extreme example is the case of railroads where a part of the output, the through traffic, is competitive while the other part, the local traffic, is monopolistic. This whole question of the relation between the size of an industry and the size of an establishment seems to the writer badly mixed up in the literature. Professor Bullock has distinguished between the three principles of diminishing returns with varying proportions between the factors, diminishing costs in an industry as a whole and decreasing costs in the single establishment, or economy of large-scale production. (Cf. Quarterly Journal of Economics, vol. XVI, pp. 473 ff.) But no one, so far as I know, has worked out these cost laws adequately. (Cf. also Davenport, Economics of Enterprise, chap. XXIV). Davenport does not go as far as Bullock in the analysis of the problem.
See F. M. Taylor, Principles of Economics, chap. IV, for a very thorough and sound non-mathematical discussion of the whole question of variable proportions and diminishing returns. I must remark, however, that Taylor's treatment of the economy of large-scale production seems to me to be based on fallacy.
The second statement of the law is deducible from the first. All that is involved in the law of diminishing returns is properly to be regarded as a deduction from the following self-evident premises:
    1. The proportions of agencies in a combination may be varied without destroying its productivity.
    2. If to a certain amount of one agency (say, labor) another agency (say, land) is added in amounts varying continuously from zero to infinity, a definite amount or range of amounts of this second agency (neither zero nor infinity) will yield a larger total product than will larger or smaller amounts. In other words, if the proportion of one agency to another is increased without limit, the product per unit of the decreasing agency will first increase and then decrease; i.e., there is a maximum point, or range, beyond which in either direction the product (per unit of the increasing agency) will decrease.
    3. It is demonstrably true, and is necessary to the theory of distribution that extreme variation (short of infinity) in either direction will yield a zero product.

It is most essential in regard to this law that it relate to any variation in proportions irrespective of the absolute amount of any factor present and of the direction of the change. But the conventional case of the application of labor to land, or rather of land to labor, is easy to visualize and suitable for illustration. Let us imagine a group of new settlers on a virgin continent faced with the problem of how much of the unlimited supply of land to use with their limited supply of labor. It is surely evident: (1) that they can use different amounts and still get some product (Ax. 1); (2) that they can use too little or too much to get the largest amount of product (Ax. 2); (3) that they might conceivably try to use so little or so much land that no product at all would be secured (Ax. 3).

It is to be noted that we must assume the size of individual establishments to be nearly a matter of indifference.

The above reasoning proves also that the curve itself cuts the X axis positively as drawn in our figure, and does not pass through the origin. It follows further from the symmetry of the relation between factors that the curve will cut the X axis again beyond the maximum point and not become asymptotic, as it should do if it passed through the origin. Professor Taylor's curve was incorrectly drawn in this detail as it should either become asymptotic or else not pass through the origin.

Really on the other agencies applied to the land, but we follow the usual formulation. The assumption must be borne in mind that men know what they are doing and are motivated by the desire to maximize production. In fact, the results are much distorted by ignorance, the effect of tradition carried over from a place where land is scarce to new countries where it is abundant, ingrained land hunger, etc., and in the United States by the conditions of land settlement and preëmption.
Cf. below, p. 108 and note [II.IV.19-21 and nn48—Ed.].
The fall in specific or marginal contribution is an easy inference from the law of the variation of product per unit. For a detailed demonstration see Taylor, loc cit., especially pp. 101, 102. The "added product" of a unit in the text above is what Taylor and most writers call "the marginal product" of the "factor." For reasons which will presently appear I prefer to avoid the misleading terminology of factors and margins altogether.
This terminology is more or less arbitrary, but is one way of straightening out the current confusion and giving different names to different things. Taylor (loc. cit.) uses both expressions "diminishing returns" and "diminishing productivity," in connection with the instrumental law; in fact in virtually the same sense, and does not bring out the contrast between the variation of physical product and that of value product. Strange to say, he does not use the principle of diminishing returns which he so well formulates in his discussion of distribution, but adopts a different line of reasoning through different proportions of factors in different industries without variability of proportions in single industries. That this same principle is involved is recognized by Taylor, who thus shows a considerable advance over Wieser. This author, it will be recalled, uses the same theory of imputation which Taylor uses, but advances it in place of the specific productivity theory, applied to industries independently, which he repudiates. (See below, p. 110 [II.IV.23—Ed.].)
Cf. above, chapter III.
As Davenport has remarked. (Cf. Economics of Enterprise, chap. XXII.) But Davenport's position will come up for criticism later on. (Below, p. 124. [II.IV.44—Ed.])
The mode of internal organization of the groups need not trouble us here. It might take any form which would produce effective common action and responsibility. In life, it is, of course, generally worked out through a responsible entrepreneur as intermediary, but it is necessary to exclude such a functionary at this point in the argument, and in fact his services would be superfluous, except, perhaps, temporarily while the adjustment was being worked out. Greater violence is done to reality by the specification of perfect competition among organizations for members. This assumption involves, in the first place, perfect knowledge and intercommunication throughout the society. In addition it calls for a large number of groups exploiting every sort of service, and entire absence of collusive action among them. The number of establishments in any line of production depends upon the size of each, which in turn depends on the divisibility of the factors being combined. Hence the principle laid down above (p. 98 [II.IV.7-8—Ed.]) that competition depends on a degree of divisibility in productive factors. That division of labor is limited by the scope of the market is true; but commodities sold in different markets do not represent the same aggregations of utilities, and are different commodities.
There is a difficulty in regard to the meaning of the value contribution to a social total. Exchange values being essentially ratios, an aggregate of exchange value has very little meaning. We cannot be sure that the value income of society as measured by the market, in terms, say, of a particular commodity, would be larger when the final adjustment was reached than under any other arrangement, and, of course, it will not do to say that the individual gets the physical commodities which he enables the society to produce. The answer is that he will get the value of the physical contribution which he makes, enough value income to buy it. The actual physical contribution should theoretically consist of infinitesimal increments of practically all the commodities produced in the society, perhaps including an increment of "leisure."
For a full discussion and demonstration of the theoretical exhaustiveness of the distributive process as described above (though in a somewhat different setting), see Wicksteed, Common Sense of Political Economy, book II, chap. VI, and The Coördination of the Laws of Distribution, passim. The reader will notice that the lines along which the adjustment is supposed to be worked out above are very different from the "dosing method" familiar in American economic literature. (Cf. especially J. B. Clark, The Distribution of Wealth, chap. XII.) This latter procedure seems to the writer unnecessarily abstract and unreal and more difficult to follow than the realistic method of tracing out the effect of competition among establishments.
See especially pp. 8, 9.
Quarterly Journal of Economics, August, 1901.
Political Science Quarterly, June, 1915.
The Economics of Enterprise, chap. X.
"Specific Productivity," Quarterly Journal of Economics, vol. XXIX, pp. 149 ff., esp. pp. 159 and 160.
Der Natürliche Werth, 3. Abschnitt, "Die Natürliche Zurenchnung des Productiven Ertrages," § 22.
The Industrial System, chap. V, appendix, pp. 112-20. A somewhat different (quasi-mathematical) line of argument to the same end is put forth by R. S. Padan, Journal of Political Economy, March, 1901 (vol. IX, pp. 161 ff.).
Cf. above, p. 104, note [nn43—Ed.]. Taylor is right in the contention that specific productivity can be imputed through differences in the proportions of agencies in different industries alone without variability of proportions in the industries individually. In fact, both elements come into play. We have mentioned and shall presently discuss further the fallacy involved in the concept of the "factor" of production.
We may notice here another point raised by Padan, the bearing of increasing returns upon the theory. It is generally recognized that in the earlier stages of a hypothetical dosing process, increasing returns will be secured, up to a certain point. By "supposing" this stage of increasing returns to last throughout the process, Padan easily makes the application of the method appear absurd. This line of reasoning is still more arbitrary than his earlier point, however, and need not detain us. We have shown at sufficient length that increasing returns is an absurdity; that an agency worked under such conditions is negatively productive and had better not be used at all. Professor A. Landry, in criticizing Professor Carver, has also overworked this supposition. (See Quarterly Journal of Economics, vol. XXIII, pp. 557 ff.)
Proceedings, Twenty-Second Annual Meeting of the American Economic Association, p. 143. Taussig's statement that labor produces all wealth, but is not entitled to all of it, would better, it seems to me, be reversed. Labor cannot claim to be the only causal source of goods, but may put forth a superficial claim to a right to consume them all.
Work and Wealth, chap. XXII.
The Industrial System, cited above.
In the writer's opinion, the hostility to the productivity theory is due mainly to the notion that the productivity of labor and capital represents their moral desert in distribution, joined to the conviction that the existing order is not morally ideal. The theorists who treat a productivity remuneration as synonymous with ideal justice are merely uncritically voicing the popular view. It is this popular dogma which is the seat of the difficulty, and which represents a confusion of the most egregious sort and leads to equally muddled reasoning on the question of causality in order to avoid a repugnant conclusion as to the justice of things as they are. The question cannot be gone into here, but a little consideration will show that there is almost no case at all for an identification or close assimilation of causal contribution to production with moral desert in distribution. The inequalities in inherited property and opportunity in several senses are obvious, but it must also be recognized that natural differences in personal capacity are equally powerless to create a valid moral claim to favored treatment.
It seems to me a manifest absurdity to define them in price terms as does Professor J. B. Clark. (The Distribution of Wealth, chap. VI.) There would be only one factor if measured in price terms, and the theory of distribution would be a pure petitio principii.
If this conclusion is not evident after a little reflection it may be demonstrated by reasoning as follows. Suppose that at a higher rate per hour or per piece, a man previously at the perfect equilibrium adjustment works as before and earns a proportionally larger income. When, now, he goes to spend the extra money, he will naturally want to increase his expenditures for many commodities consumed and to take on some new ones. To divide his resources in such a way as to preserve equal importance of equal expenditures in all fields he must evidently lay out part of his new funds for increased leisure; i.e., buy back some of his working time or spend some of his money by the process of not earning it. The conclusion is enforced by the important practical consideration that the expenditure of money also requires time and energy which must be saved from the work period if the best results are to be secured.

The facts as to the shape of the supply curve of labor from given laborers are well known to employers of native workmen in backward countries, especially the tropics. White men in the advanced industrial nations have not always behaved so rationally; their traditions give them a higher preference for the kinds of satisfactions purchasable with money in comparison with the more inward and spiritual enjoyments. But the effect which was to be anticipated was very conspicuous after the outbreak of the World War, when the wages for certain kinds of work rose to unprecedented heights and produced increased loafing and dissipation instead of increased production. (It is important to bear in mind that we are speaking of a permanent change; it would be in keeping with rationality to work harder at a temporarily higher rate in order to purchase more leisure later on.)

While on the subject we may observe that it is also an error to assume that in this respect land or other property services will be different from labor. These agencies also have alternative non-pecuniary uses, and if, say, the rent on land were to rise, landowners could afford to use more of it for lawns, flower gardens, athletic grounds, game preserves, pleasure parks, etc., and less for cultivation and marketable crops; and if they calculated closely they would do so.

Marshall correctly treats long-time demand and supply as time rates, but does not sharply contrast this form of the variable with the absolute amounts dealt with in market price.
Cf. Taussig, Principles of Economics, chaps. 12, 13, 14.
Economic literature is full of the contrary assumption, but it is a definite error, in dealing with long-time normal price. The existence of differences in costs in different establishments in an industry is proof, when not due to differences in accounting practice, that the competitive adjustment is imperfect. The current conception of marginal cost necessarily falls away through the same reasoning. The producer's calculations are made in terms of cost per unit and selling price per unit.
The Distribution of Wealth, p. 85; cf. also Davenport, Economics of Enterprise, chaps. XI and XXII.
Reference has been made to the absurdity of the two-factor analysis, as exemplified particularly in the work of Professor J. B. Clark. The same author falls into the closely related fallacy of measuring separate agencies by their productive contributions. He recognizes and clearly states the difficulty (The Distribution of Wealth, p. 374, note) and ostensibly gets around it by setting up an absolute subjective standard of measurement. It is very difficult for the present writer to criticize this reasoning, and out of the question in the space available; I can see nothing in it but a complete failure to make connections, a palpable non sequitur. It is to be observed that the fallacy is equally involved in all other distribution theory which makes use of "factors" at all—the number is immaterial—and this includes most of the literature of the subject.

A conspicuous exception is Davenport's discussion (Economics of Enterprise, chaps. XI and XXII) already mentioned, which is excellent for this phase of the question. Where it falls short is in failing adequately to separate the long and short period problems of distribution. It is this failure which in the writer's view explains most of the controversial differences between economists in so far as they relate to the scientific explanation of distribution, and not to questions of propriety or policy. It is essential to take account of the fact that from the long-time point of view the question of classification takes on a different aspect, becoming a question of the conditions of supply of different types of agents. The case for the conventional tripartite division (or more especially the separation of land and capital) is argued at length in A. S. Johnson's Rent in Modern Economic Theory. (See especially pp. 35 ff.) This phase of the problem will presently come up for discussion, and it will be pointed out that there is danger of over-simplification here also. (See below, chapter V.)

It may strike the attention of the reader that while the tripartite classification is emphatically repudiated, the factors are still commonly referred to in the present essay as "land, labor, and capital." If explanation is called for, it is to be found in the necessity, for mere expository purposes, of some expression which explicitly covers the whole group. The significance is the opposite of classificatory; "animal, vegetable, and mineral," or "solid, liquid, and gaseous agencies" could have been used but for their unfamiliarity in this connection. Also the familiar terms have social and ethical significance if none of a strictly economic sort.

See above, p. 119 [II.IV.33-34—Ed.].
The notion of sacrifice has been overworked in economics. Economists as well as employers have been too prone to assume that subjective willingness is the principal limitation on the amount of labor obtained from given persons or for a given outlay. And employers as well as economists are waking up to the efficiency of well-paid labor. There is no doubt that employers as a class have lost much money (not to mention the higher considerations involved) through working their employees beyond, and feeding, clothing, and amusing them below, the point of maximum physical efficiency. This would not be done with a dumb animal! Of course it may be profitable to the individual employer to pay a wage below what is necessary to maintain maximum efficiency and an adequate supply of labor from generation to generation (if the working class maintains the labor supply partly at its own cost); what is meant is that they have paid uneconomically low wages even from the standpoint of the short periods for which they have to deal with the same individual laborer. The presence of idle equipment is a great temptation to an employer, and the debit side of overworked help is less conspicuous to view. Of course the ignorance and imprudence of the workers are as much in point as those of the employer. It is of interest that Lord Leverhulme has recently put forth the contention that a six-hour day, without decreased pay, would be profitable to British employers in many industries, if the men would consent to two shifts during each twenty-four hours.
This is being recognized in the case of child labor by many employers who refuse to employ children simply on the ground that it does not pay in the business sense. This whole problem becomes more important as the amount of capital per worker increases. It is also true that the increasing use of machinery provides tasks which a lower and lower grade of human capacities are required to perform. The net result is difficult to estimate. The social problem of the "unemployable"—how to identify him and what to do with him—is surely forbidding enough. Like most of our new troubles, it is partly a product of the disintegration of the family as well as of industrial changes directly.
The point may be illustrated by the anecdote of a tramp who, finding a hundred-dollar bill, made a bee-line to the nearest quick lunch and excitedly ordered a hundred dollars' worth of ham and eggs. That men do not behave after this fashion does not prove that, other things equal they prefer a future satisfaction to a present one of the same magnitude.
H. Sidgwick similarly takes the view that a preference on the ground of time alone is irrational, criticizing Bentham for including "propinquity" as a basis of preference between otherwise similar enjoyments. See History of Ethics, p. 241, note. Cf. also Jevons's discussion, Theory of Political Economy, pp. 72 ff., where the same position is taken. Jevons's illustrative problem of the consumption of provisions on a vessel at sea is very effective in bringing out the issue.

It will be noted that the effect of the uncertainty of the future is very complex. Against the chance of loss of future enjoyment through death or incapacitation must be set the danger of future privation due to other contingencies. We are more likely to suffer loss of earning power than of power to enjoy, and the consequences of need without ability to gratify need are very unpleasant. Perhaps the perfectly rational homo œconomicus would discount the present up to the point of making provision for the more urgent necessities as far ahead as he was at all likely to live and discount the future beyond this point in increasing degree. The point is significant chiefly as showing the absurdity of hedonistic rationalism as a theory of actual behavior.

Cf. Spencer, First Principles, chap. X, "The Rhythm of Motion."
It is fundamental to the actual phenomenon of capital accumulation, that the principal, once saved, never is consumed; if it is consumed later, there is no net addition to the capital supply of society. Men save in large measure with no thought of ever consuming the capital, or even the income which it yields. For this reason the older term "abstinence" seems to me far more descriptive than its modern substitute "waiting." To be sure, an income of five dollars a year in perpetuity represents more consumption than one hundred dollars now; but no one consumes an income in perpetuity or expects to do so. Even if the saver consumes the entire income from his investment as long as he lives, he may or may not consume a total amount equal to the principal saved. Capital formation is the result of abstinence rather than waiting.

In fact, the term "saving" itself is misleading. Men do not generally produce wealth to consume it and then decide to invest it instead. Most of that which is invested is destined to that purpose in the first place and would otherwise never be produced at all.

We pass over here the effects of divergence in suitability for accumulation of different classes of goods, due to differences in bulk, perishability, universality of appeal, elasticity of demand, etc.
We must here assume it to be made absolutely dependable by insurance or otherwise.
Wicksteed has an excellent discussion of this point. (See Common Sense of Political Economy, chap. VII.) It is noteworthy that the "usury" against which moralists have universally thundered in pre-industrial society corresponds to the phenomenon just described rather than to modern interest. The productive investment of accumulated wealth was nearly unknown in earlier times and even the purchase of existing productive property was rare. Practically the only productive agencies known were land and slaves. Land was not private property in the modern sense and was hardly ever bought and sold commercially, while slaves were used almost exclusively in connection with land and by its owner even when not legally attached to the land itself. If there had been a free market for consumption loans the correspondence with the phenomenon we have described would have been complete except for the element of risk. The absence of a competitive market was the source of much of the evil of usury, and the payments made doubtless did represent extortion largely. Be it observed, also, that historically speaking modern interest developed out of the consumption loan through the intermediary of passive partnerships in trade ventures and not out of dealings in canoes, fish nets, etc., in which the fancies of a certain school of interest theorists are prone to revel.
With the actual history of property we are, of course, not concerned. Doubtless, the first approximation to private productive property was in human beings, slaves, or, perhaps, women or children, while the last thing to become really privately owned was land. But the proper order for our purpose is not chronological, but rather that of increasing complexity.

Part II, Chapter V.

End of Notes

Return to top