The direct factor substitution mechanism: from Thunen to Marshall
Awareness of these problems is initially nearly absent. First, this is shown for the economists who base their theory principally on the direct factor substitution mechanism. In such an approach, consideration of marginal utility is not indispensable; income distribution can be determined by marginal productivity even with a rigid composition of consumption demand and rigid factor supplies, and in fact the approach is born before the “marginal utility” revolution.
The starting point is the realization that, in the classical approach to intensive differential rent in the production of corn with corn, the role of land and the role of corncapital (inclusive of advanced wages at a given wage rate and therefore also representing an input of labour) can be reversed because land rent, determined as a residue after the determination of profits as the “marginal product” of the last unit of corn-capital on a given quantity of homogeneous land, can also be seen as the “marginal product” of the last unit of land on the given quantity of corn-capital; both rent and profits can be seen as “marginal products” determined by a symmetrical mechanism; it is then concluded that the same intensive-differential-rent principle, which classical economists had used to determine only the division of the net product between land rent on one side and the sum of profits and wages on the other side, can be generalized to determine the division of this second part between wages and profits, a division that in the surplus approach was determined totally differently.
The first author to reach this conclusion is Heinrich von Thunen, whose influence upon Marshall, J.B. Clark, and indirectly upon the Austrian School through Roscher, is enormous. In Die Isolierte Staat (1850) von Thunen depicts an economy similar to Ricardo’s corn sector in that rye is produced by rye (subsistence consists of rye); but land is assumed free, and Thunen uses intensive differential rent to determine the division of the product between interest and wages.
The notion of marginal productivity is clearly formulated, and in some examples he also anticipates the idea of isoquants, as well as the idea that the optimal combination of capital and labour requires the equality between ratio of rentals and ratio of marginal products. In the same book he formulates detailed examples of a society where initially there is only production by unassisted labour, until someone by saving subsistence goods makes it possible to dedicate a year to produce tools first, which raise labour’s subsequent productivity of consumption goods, so workers are ready to pay interest for the savings; when all workers producing consumption goods have been supplied with the most useful tools producible by one year of labour, further savings allow dedicating another year to the production of further and different tools, which add less to productivity since the most useful tools have been produced first, so the interest rate decreases and wages rise (see Leigh 1946, for a summary); thus Thunen also introduces the idea that savings allow a delay between application of labour and production of consumption goods that permits the productivity of labour to increase, the idea to be then found in Jevons and in the Austrian school. Heterogeneous capital is a value magnitude, equal to the wages paid to the labour that produces capital goods.Clark studies Thunen, grasps the foundation of the new approach in a generalization of rent theory, and explicitly states that all incomes are determined by “a law of rent”. But he does not notice the essential role of the physical homogeneity of corn-capital in permitting the symmetry between land and capital in classical intensive rent theory; he treats heterogeneous capital unproblematically as a quantity of value independent of income distribution, both in determining its total endowment, and in order to determine the factor proportions chosen by entrepreneurs. He does not realize that production functions or isoquants in which one factor is value capital can only be derived on the basis of given prices of the capital goods, and therefore cannot determine how optimal factor proportions change with changes in factor rentals, because the prices of capital goods will change, causing shifts of isoquants and of marginal product curves (as Wicksell 1934: 149 points out).
The same mistake can be found in many other authors of the period.Marshall, too, has read Thunen, but prefers a short-period approach where durable capital goods are given, and avoids an explicit reference to a value endowment of capital; but the value conception of capital as a factor of production emerges when he speaks of “free” or “floating” capital as what must be absorbed by new investments in a proportion to labour that depends negatively on the rate of interest. This “free” capital is the flow supply of savings, treated as measurable independently of income distribution, and entrepreneurs will demand it up to the point where the marginal value product of this investment becomes equal to the rate of interest. Thus although only at the margin of new investments, capital in Marshall is still the single factor produced by savings, an amount of value but analogous to physical factors in the marginalist factor substitution mechanisms. In determining the demand for the flow of savings, the dependence of prices on income distribution is neglected as much as in Clark. The short-period framework obscures but does not alter the basis of the theory, and the subsequent writings of his pupils Robertson or Pigou confirm this fact.