I.

HOW FAR does the conception of marginal units of production assist toward a theory of distribution? It is easily shown that if an unlimited amount of labor is procurable for a business where the other factors of production are given quantities, the last unit brought into employment will receive no aid – or a minimum aid – from the other factors and will take in wages virtually the whole addition to the productivity of the business which follows its employment. This is clearly set forth by Professor Marshall in the illustration of the marginal shepherd. A farmer with a given farm and farming capital calculates that it is just worth while to employ a tenth shepherd, whose addition to his staff enables twenty more sheep to be marketed in a year than would be the case were nine shepherds employed. These twenty sheep must be accredited to the marginal shepherd as the specific product of his labor, and he will receive them, or their value, for his wages. For though he receives the same assistance from the land and capital as the other shepherds do, it is necessary to assume that no more productivity is got out of these factors when ten shepherds are employed than previously when nine were employed; or, put in another way, any assistance the marginal shepherd receives from capital and land is attended by a corresponding shrinkage in the assistance rendered to the other nine. For unless the full economy of the fixed factors were exhausted by the employment of nine shepherds, there would be a surplus gain to the farmer on the employment of the tenth shepherd after paying his wages; in that event it would pay to employ an eleventh shepherd, and the tenth would not be marginal. It is necessary to admit that the marginal shepherd – or, to be more precise, the marginal part of his labor – adds a product which is entirely attributable to his labor and (nearly) the whole of which is returned to him in wages.

Similarly, it is argued, the last unit of capital borrowed for a business, the other factors of which are given quantities, will just produce what must be paid for interest, leaving no margin of profit. If our farmer working a farm with his ten sons finds it is just worth his while to borrow another £100 for working capital, the additional product created must be attributed entirely to this last or marginal unit of capital and leaving only a nominal profit after paying the interest. For though the labor of the farming group co-operates with this last £100 worth of capital, we are obliged to assume that the enlarged capital is less completely utilized than the smaller capital: otherwise there will be some profit from this last £100 of capital which in that case is not really marginal.

If we were to suppose that agricultural land were in the same condition as capital, an unlimited amount being available for renting on the same terms, it is clear that our farming family entering such a country with a given capital would take on a marginal unit of land which would pay as rent the whole of the increased product of the farm due to its use.

Proceeding further, let us suppose a farmer entering agriculture with a given stock of personal ability and enterprise, a pure entrepreneur, able to buy all sorts of productive energy in free and virtually illimitable markets. His increments of productive energy will be composites of land, various forms of capital and labor, varying in composition as he expands the area of his farming operations, whether extensively or intensively; he will buy and apply the units of the various factors according to his estimates of the technical economy of the farming business. The limit to the quantity of each sort of productive energy he buys, and to the aggregate which he will employ, is determined by the economy of the utilization of his given amount of personal power; the last increment of land/labor/capital he takes on will only just pay the rent-wages-interest and will leave a merely nominal surplus-gain for him. But this last or marginal unit of land/labor/capital must be considered just as productive as any oi: the others, and, ex hypothesi, receives the same payment in rent-wages-interest. This payment will exactly cover the product attributable to the specific productivity of each composite unit, and, on our hypothesis of an indefinitely large supply of each factor, we must conclude that each unit of land, capital, and labor receives in payment just what it produces. Our farmer-entrepreneur cannot be exploiting them, for the marginal increment which is as productive as any other increment receives the whole of its product as payment for its use, so therefore must all the earlier increments.

How are we to conceive the profits of our farmer? If there is no surplus in the employment of the last unit, and the last unit is just as productive as any other unit, it would appear that no profit could arise. The ordinary diagrammatic representation of the “dosing” theory does indeed show a surplus derived from the employment of each increment except the marginal one. The familiar figure on the following page runs thus:

Here A B represents the entrepreneur’s personal power; to it are added ten increments of land-labor-capital, to which a diminishing amount of productivity is attached, the first unit yielding the figure A B b a, the tenth yielding a b D C. The marginal increment alone receives in payment virtually its whole product; each of the others yields a surplus receiving the same payment as the last, but affording a larger product. Here the entrepreneur appears to make a large profit on all the earlier increments. But this figure is a most fallacious one, if designed to explain how the aggregate product of the fully organized business is apportioned. For it appears to show that the ten increments are unequally productive, or that they receive unequal amounts of assistance from the energy of the entrepreneur; and neither supposition is warranted. For when there are ten units of land/labor/capital employed by the entrepreneur, if a separate productivity be attributed to each, whether or not that productivity includes the assistance of the entrepreneur, it must be the same for each of the ten.

The diagram as it stands says this: If to the farmer’s ability A B one dose of productive energy B b be applied, the product is A B b a; if a second dose be applied, the smaller product a b b a must be attributed to it, because, though in itself equally productive, it receives a smaller assistance from A B, So with each subsequent dose: it receives a diminishing amount of assistance from A B, until the tenth dose receives a minimum or negligible assistance. Now, it is evident that we have no right to represent the second dose as receiving less assistance from A B than the first dose, when we are analyzing the composition of a two-dose business; each dose must be supposed to have the same relation to A B. So with the full ten-dose business: the last dose receives the same assistance from A B as is now rendered to the first dose, though of course much smaller than was rendered to the first dose when it was the only dose, or one dose among four. It is folly to retain a diagram which suggests that the product of the first dose is A B b a when this is only true in the hypothesis that one dose only is applied – a hypothesis which is denied by the application of each subsequent dose.

It is clear that, if we are to represent a business in which ten doses yield the maximum economy to the entrepreneur, we must assign an equal productivity as well as equal payment to each dose. Beyond these payments, however, there emerges a surplus claimed by the entrepreneur, as profit or wages of management, which, on the ten-units basis, must appear to have the same relation to each unit. If ten doses be applied to A B, this equality of productivity, payment, and surplus will appear in the following simple figure:

The entire product A B F D is divided by C E into two parts, the lower, C B F E, going in ten equal payments to the units of land-labor-capital; the upper, A C E D, forming one or ten units of surplus or profit or wages of ability for A B.

But, it will be objected, in this diagram the last unit, which we call marginal, appears to carry, in addition to the product which represents its payment, C B F E, an extra product or surplus, A C E D. If the tenth unit be removed, this surplus seems to disappear with it. The existence of such a surplus, however, is excluded by the terms of our hypothesis.

Now, as we have already seen, we are obliged to assume that the aggregate productivity of our entrepreneur reaches its maximum in co-operation with nine units; if then, ten are employed, any assistance it appears to give to the productivity of the tenth implies a corresponding reduction of assistance to the other nine. In other words, on a nine-unit basis in which A B F D is eliminated, the product associated with each of the nine units is larger than it is found to be after the tenth is added, by the presence of a larger surplus, represented in the diagram by the substitution of the dotted line for the line A A. The employment of the tenth unit has simply substituted the surplus A C E D for the nominally smaller surplus a A A d which existed on the nine-units basis.

If we assume the operation of a law of diminishing returns, it seems self-evident that he will take on just so many units of land-labor-capital at their current price as will exhaust his economy of personal ability, the last unit evoking a merely nominal amount of this personal power. The surplus thus arising will be claimed as the product separately attributable to the entrepreneur’s personal energy and ability. So it appears that each unit of labor, capital, land, gets out of the general product of the business just what that unit produces. And does the entrepreneur get just what his ability produces? It seems as if this were necessarily the case; for, in addition to the land, capital, and labor, the only productive force which remains is his ability; hence, if we take away that part of the entire product due to the three former, what remains must be the product of the latter.

So there can be no exploitation of labor or of capital, and the profits of the entrepreneur, however large they seem, are the specific product of his personal productive energy.

But to this conclusion it will be objected: “Are you not attributing to the productivity of the entrepreneur all the effect of economy of division of labor and co-operation of the units of labor and capital? Are you not paying the several laborers on the basis of their separate productivity as individual workers working without co-operation of other workers and of capital, whereas their aggregate productivity when working under these conditions greatly exceeds the added productivity of their separate labor? Is not the main object of employers in perfecting individual bargains with employees the desire to obtain their labor-power on the reckoning of its productivity as a separate producing unit and then to make it co-operate with other units taking as profit the increase of productivity thus attained? Is it not, on the other hand, the object of labor organizations by ‘collective bargaining’ to obtain for the laborer the equivalent of his productivity as a collective laborer? Similarly with units, of capital: Is not the basis of the economy of a joint-stock corporation the desire to borrow units of capital at rates which are equivalent to their productivity when separately employed, and to raise the value and the yield of these units by employing them collectively? A banking firm manifestly makes its profits largely out of this difference between the productivity of separate bits of capital and that of the same bits used collectively.”

The ability of an entrepreneur is no doubt essential to this economy, but are we at liberty to assume that the whole results of it are attributable to the separate productivity of his ability? The separatist treatment which the close method employs upon the other factors of production expressly serves to exclude from these factors any share in their collective productivity as distinct from their individual productivity, and to impute it to the only agent who functions collectively, viz., the entrepreneur. The fact that the entrepreneur gets no appreciable advantage for himself from the employment of a supposed last unit of other factors is no evidence that he does not get an advantage from the co-operative economy of all the units; the phenomenon of the marginal unit is only another way of expressing the fact that there is in most industries a necessary limit to the area of exploitation of labor and capital by entrepreneur ability. From the standpoint of labor it may mean that there is always a limit to the number of laborers whom an entrepreneur can employ so as to get profit out of their co-operative labor; when this limit is examined, it necessarily appears that a single laborer added or substituted makes no difference to the aggregate of this profit.

It is now evident that in our illustration of the effect of a marginal unit, whether of labor or of capital or of land, or of all three together, we are suffering from a false conception of causation or attribution due to the arbitrary use of constants and variables.

The whole results of the co-operation and division of labor of the other factors of production have been imputed as the exclusive product of the ability of our entrepreneur by assessing separate products for each unit and then assigning to him the entire residue as his product. In the history of the theory of distribution this residual claimant game has been played in various forms. It has already been shown by various writers how the attribution of the residue depends entirely upon which of the factors in production is taken as constant and which as variable.

Of the variations of this game that which takes the entrepreneur as the constant and the other factors as variable is certainly the most conformable to the facts of modern industry. The function of the entrepreneur is to buy the use of units of labor, capital, and land, to organize them for effective production, and to sell the product, so as to get the largest margin of gain. Now, assuming that each unit of labor, capital, and land gets its separate product for its payment, does the surplus which we see remains do more than pay to the entrepreneur the “product” of his productive energy of management? The owner of each unit of labor, capital, and land appears to get just what he produces, and this payment is a minimum necessary cost. Is this the case with the entrepreneur? There is nothing in the setting of the problem, as we have hitherto set it, to place the entrepreneur on the same footing with the owners of the other factors of production, so as to insure that what is taken by him is (a) his specific product, (b) a minimum “cost.” For in this setting the entrepreneur factor has been taken as a constant quantity with which variable quantities of labor-land-capital co-operate. By this method it appears that the three latter receive in payment their product, but it does not follow that the residue, taken by the entrepreneur, corresponds in the same sense with his product. If we are to eliminate the possibility of any entrepreneur’s “unearned” gain, and to place his payment on the same footing with the others, we must remove the “constant” position of the entrepreneur. As we assume that our entrepreneur (A B) was able to purchase an unlimited number of units of labor-capital-land at the same price, so now we must assume an unlimited number of entrepreneurs in the position to buy the other factors.

Professor Clark, whose theory of distribution in a “static” society we are endeavoring to reach, clearly recognizes the necessity of this assumption:

May not all entrepreneurs be making the same rate of net profits and making them at the same time? May there not be a condition of equal and universal profit? Clearly not; for this would be a universal invitation to capitalists to become entrepreneurs, and, as such, to bid against each other for labor and capital, until the profit should everywhere vanish, by being made over to laborers and capitalists in the shape of additions to wages and interest. The pay of each of these agents, therefore, under perfectly free competition, is bound to stand at the productivity level. (p.291, note.)

It is not always quite fully recognized that there are two conditions to a working of industrial society which equalizes the economic position; of the several factors and reduces all their payments to costs. One is complete fluidity between the several grades of each factor, the other is the existence of an unlimited number of units of each factor. The former, indeed, is commonly admitted as that equality of economic opportunity requisite to secure equal rates of interest to different forms of capital and equal net payment of wages to various sorts of labor.

But if industry be regarded as built up out of increments of productive power in entrepreneurs’ ability, labor, capital, and land, and if the condition of remuneration by bare “cost” or “specific productivity “ be that each last or marginal increment receives just what it adds to the total product, this condition can exist only so long as there remains another unit hitherto unemployed. For it is surely evident that if all the units of any one factor of production available for a particular industry are absorbed, while units of the other factors are not absorbed, the marginal increment of the “short” factor will be able to take a scarcity rent, or an addition to the product representing its bare cost. However free the competition within the ranks of each factor, a short supply of any factor as a whole compared with the others enables a scarcity value to arise at the margin, and this value or surplus-pay will be taken by each unit of this factor in use. This is the source of such inequality as exists (a) in the distribution of wealth as between land, labor, capital, ability, regarded generically as factors of production and, (b) in the elevation of the value of some classes of commodities and the depression of other classes, as exhibited in processes of exchange.

In stating his theory of distribution for a “static” community, Professor Clark essays to isolate the static forces by supposing a stoppage of five dynamic forces, increase of population, increase of capital, changes of industrial method, changes of business organization, and growth and change of human wants. But in the “static” community thus reached it is not the case that we should, as Professor Clark avers, reach a condition in which “values are here ‘natural’ in the Ricardian sense, for everything sells at its ‘cost of production’ and no entrepreneur makes a profit.”

For if in a “static” community any one of the factors of production were less abundant than the others, either as regards a particular sort of production or as regards production in general, we should have “net profit” or “scarcity rent,” or some other surplus, which was no cost of production “in the Ricardian sense,” emerging at the margin and upsetting the theory of distribution by “cost.”

On the other hand, if we suppose a “static” community in which none of the factors is relatively short, how can we apply the method of marginal increments in a theory of distribution? So long as we could take a capitalist adding to a fixed capital successive units of labor, or an entrepreneur adding to a fixed quantity of business enterprise successive units of capital and labor, we were able to measure something that could be called the separate productivity of the last unit. But if we take what from the standpoint of distribution is the real condition – a number of units of all the factors of production spontaneously co-operating for the most efficient production – the last increment in such a complex will be composed of all the factors of production, and the knowledge of what it adds to the total product will be of no assistance in determining the productivity of the several factors which constitute this complex unit. All we should know would be that the last unit was as productive as any other unit, and that its owners received the same return for its use. But this is not taught us by experiments with marginal increments; it is an assumption prior to any such experiment. In a static society where a limited production is conducted by co-operation of freely competing and fluid units of labor, land, capital, ability, none of these factors being able to take a “scarcity” rent, speculation and diagrams on marginal units, would teach nothing. For every unit would be ex hypothesi as productive, and as well remunerated, as any other, and every unit would be an indissoluble complex of the several factors of production.

It is quite true that, with complete fluidity of the factors of production and no shortage of any factor, we should get equality in distribution. It might even be said that each unit of each factor would get its “product” in payment for its services. But this statement could not be proved by any “marginal” method, for every marginal unit would be composed of all the factors in such wise that the hypothetical elimination of any one of these would sterilize to an unascertainable extent the others.

Apart from this, the hypothesis of a “static” society in which no single factor was in a condition of shortage, and so able to take a scarcity rent, is a self-contradictory hypothesis. For such a society, containing ex hypothesi a number of unemployed units of each factor, could not remain “static” in any intelligible sense.

As applied to the case of any single industry, the “marginal” method implies a fixed or given supply of some one of the factors and an unrestricted supply of the others. Such an argument may be applied to show that the last increment of the unrestricted factors gets what its presence adds to a complex of productive power in which one portion is restricted, but it cannot show that no surplus accrues to the owner of the restricted factor by virtue of its restriction.

The whole elaborate argument regarding distribution in a “static” society seems to come to this, that under conditions of “free” competition there would be no “marginal” rents; i.e., no payments beyond cost. Under equality of economic opportunity there is equality of distribution; i.e., if everyone has an equal chance of making a gain, gains will be equal.

The marginal unit of labor or of any other factor does not assist us to determine, or even to measure, the part played by labor in the production of wealth due to the co-operation of capital, labor, law, and ability in a so-called static society. It gives us no information not already assumed in the terms of the problem.