It is because of these uncertainties in measurement of the development of the productive forces in price terms that we should prefer models constructed in terms of value, the only certain standard.
The major defect of analysis in price terms compared with analysis in terms of value is not due to the “open” character of Sraffa’s model (meaning that the dynamic equilibrium of supply and demand for each product—equipment goods and consumer goods—is not formulated as an internal condition of the model but simply assumed to be related externally), in contrast to the “closed” (full circle) character of Marx’s model (in which the equilibrium in question is formalized in the model itself). This defect is due to the substitution of prices, which depend on distribution, for values, which do not so depend. This means that the concept of improvement in the productivity of labor (as the measure of the development of the productive forces), which is perfectly objective in Marx’s practice (it does not depend on the rate of surplus-value), is no longer objective in Sraffa’s model or in any other model constructed in price terms.
Furthermore, the Sraffian framework does not lend itself to analysis of the conditions for dynamic equilibrium, since, unlike Marx’s framework, it is not concerned with the equilibrium of supply and demand for each type of product. It is therefore impossible to deduce from it the propositions set out above concerning expanded reproduction. What it offers is a meager empirical model, which serves at best to describe an evolution that has been observed, but not to infer from this any laws of evolution.
A system defined directly in price terms is also perfectly determined—in the sense that relative prices and the rate of profit are determined—once the rate of real wages is given.
But then there arises the question of a standard, which Sraffa, in the Ricardian tradition, defines like this: is there a standard that would leave the net product unchanged while distribution (w or r) changed independently? The answer to this question is no. Let us see why this is so.
Sraffa does not analyze the system as Marx does. He excludes labor-power from the productive process, in order to consider wages not as the value of labor-power but as a distribution category. This is why he describes the system in the following form:
He further proposes, as we know, that we select as our standard the price of the net product:
With this standard, r and w are in a linear relationship that is independent of p1 and p2:
With this standard, r and w are in a linear relationship, whereas any arbitrarily chosen standard gives a relationship between r and w that is neither linear nor monotonic, and is described by a curve (see graph on opposite page).
But is this standard any better than others? Not so at all: (a) because this standard presupposes Sraffa’s treatment of wages: if the wage is integrated in the productive process as variable capital, the standard varies when w varies: it is no longer independent of prices; (b) because, even in Sraffa’s formulation, since the net product changes with the passage of time (the result of growth), the standard is not independent of prices but is elastic.
If then we reintegrate w in the productive process, as we should, whatever the standard being used, we get three equations and four unknowns (p1, p2, r, and w). It is still possible to express r as a function of w, but the relation is no longer linear, nor even of necessity a monotonic decreasing one.
The fundamental question underlying the dispute over whether to choose value as the standard, or something else, is that of how to measure, precisely and objectively, the progress of the productive forces.
The value standard, on the other hand, enables us to measure the progress of the productive forces from one phase to another; that was why Marx chose it.
It is not fair to Marx to reduce his proposition that value should be chosen as the standard of prices to the argument that this standard “works”—that is, that with it transformation is possible. The debate on transformation remains secondary, and however much ink it has caused to flow, it is in no sense primordial.
Marx was actually seeking an instrument by which the development of the productive forces could be measured. This instrument is value. In fact, the quantity of socially necessary labor is, in the last analysis, society’s only “wealth”—and value is independent of distribution.
This value standard means comparing the progress from one system (0) to another—(1), (2), etc.—along the Y-axis w. Along this axis r = 0, and wages w absorb the entire net product. The system that maximizes w for r = 0 maximizes income, or else minimizes the socially necessary labor time needed to produce a given amount of use-values. It corresponds, therefore, to more efficient, more highly developed productive forces.
Sraffa’s standard, on the other hand, means comparing the systems along the X-axis r. For w = 0, r = R, and profit absorbs the entire product. Assuming w different from zero does not affect the conclusion since Sraffa cancels the wage by replacing it with the goods consumed by the wage earner. Sraffa therefore compares systems along a horizontal line parallel to the z axis, starting somewhere on the vertical w axis. The system that maximizes the rate of profit R will be considered the best. Isn’t that the same thing? Not necessarily. The result of the two methods of comparison would be the same only if the two curves (0) and (1) did not intersect. If they do intersect, then it is possible that the system that maximizes w does not maximize r.
Why is this? Because, along the Y-axis (r = 0) comparison between the systems takes into consideration simultaneously (for a system with two products) the four coefficients a11, a12, a21, and a22, corresponding to the commodity inputs, and the two coefficients a01 and a02, defining the inputs of direct labor. The productive systems become (for r = 0):
and the prices p are then similar to the values.
If, however, we compare the productive systems along the X-axis for which w = 0, this means taking into consideration only the first four coefficients (production of commodities by means of commodities, and not by means of commodities plus direct labor) and leaving out the two coefficients of input of direct labor. The systems then become (for w = 0):
The value standard is superior because this standard alone considers production as the resultant of all the technical coefficients that describe it.
The conclusion of this analysis is fundamental: that social system that maximizes the rate of profit (for a given level of wages) does not necessarily maximize the development of the productive forces (the reduction of social labor time).
There is no way of doing without the theory of value. This theory alone enables us to link all the economic magnitudes (prices and incomes) to a common denominator—value, that is, the quantity of socially necessary labor, which is independent of the rules of distribution (exploitation, competition, and so on), and to do this both for characterizing a given phase (static synchronic analysis) and for measuring change from one phase to another (dynamic diachronic analysis) of the progress of the productive forces.
If a single standard is chosen to describe two systems, successive or simultaneous, there is a relation between w and r that is illustrated (see page