Sraffian economics — a critique

by
Lars Pålsson Syll
My articles My siteAbout meMy booksMy videos
Follow on:Twitter

Sraffian economics — a critique

After being tasked with editing David Ricardo’s Collected Works in 1930, Piero Sraffa, with the assistance of Maurice Dobb, published them between 1951 and 1973. This work earned him the 1961 Söderström Gold Medal from The Royal Swedish Academy of Sciences.

For the edition, Sraffa wrote an interesting and thought-provoking introduction. Its purpose was to demonstrate that classical economists based their theory on the concept of surplus, defined as the remainder of the product after deducting the necessary production costs. Ricardo’s contribution to the classical tradition was primarily to specify in more detail the relationship between the shares of the various social classes in this surplus—the social net product—and its development and changes over time. Sraffa shows how Ricardo wrestled with the problem of how to define the surplus in a way that allows for an unambiguous determination throughout his adult life.

What has been fiercely debated among historians of economic thought is Sraffa’s thesis that Ricardo, up until 1815 and Essay on Profits, solved this problem by constructing a ‘corn model.’ According to Sraffa, Ricardo reasoned as follows:

In agriculture, only labour and corn (seed) is used to produce more corn. The real wage is assumed to be given at a certain level and defined as a specific quantity of corn. This means that inputs and outputs are homogeneous and can therefore be measured in physical terms. The rate of profit can then be defined as the quantity of output (expressed in corn) minus the quantity of input (expressed in corn) divided by the quantity of input (expressed in corn). This can be reconciled without taking into account the other sectors of society. Since, according to the “equality assumption,” there can only exist one rate of profit, the individual rates of profit in the other sectors must adjust to that of agriculture, which thereby determines the general rate of profit in the economy.

How Ricardo, according to Sraffa, determines the rate of profit in Essay on Profits implies that there is no need for a theory of value to determine the distribution of the net product between profit, rent, and wages. However, as soon as Ricardo departs from the assumption that the economy is fundamentally single-sectorial, it becomes necessary to determine the profit on capital as it is determined in the market through commodity prices. When attempting to extend a corn model to include other production sectors that are mutually interdependent, problems arise. These force one onto partially new paths: one reaches a crossroads where the choice lies between divisible and indivisible production systems. It involves a possible conception of production as natural on the one hand and a necessary concept of the social character of production on the other.

In a single-sector model, distribution and production can be reduced from social entities to natural ones, but the idea of production confined to one sector contradicts the very concept of capitalist production, which is the production of value without physical boundaries. Therefore, it becomes necessary to go beyond the corn model to achieve a more accurate understanding of distribution and value within the framework of the capitalist society. In doing so, one will also partially break with the views of some earlier classical economists on the nature of the economy and instead see production as socially determined. The importance of a theory of value becomes clear to Ricardo through the contradictions that this creates within the theory of distribution. According to Sraffa, what Ricardo is attempting to demonstrate in Principles is that the conclusions he has reached regarding distribution within the single-sector model are also generally valid in an interdependent multisectional economy.

In the corn model, labour functions merely as a mediating link in the natural production process where corn is transformed into more corn. There is no conception of labour as socially equivalent. By having all production occur in the corn sector, all labour is transformed into corn labour. This is, to use Marx’s terminology, to reduce abstract labour to concrete labour and therefore to reduce the social character of labour to labour as nature. The root of the confusion between labour and labour power, which will complicate much of the subsequent work on the theory of value, can be found here.

The perception of production as a natural process essentially means that production loses its economic character and instead takes the form of general metabolism. It becomes a logical category with the meaning of transforming inputs into outputs. However, production in this general form cannot serve as the basis for economic relations because it is indifferent to the specificity of these relations.

In the distribution theory attributed to Ricardo by Sraffa, the starting point is the corn model because only through it can Ricardo separate the determination of the value of goods from their distribution. Capital in this model takes the form of circulating capital, specifically a wage fund of labour advanced, which is also the share of the previous production period that accrues to labour. In an economy of this kind, the simplicity of the distribution problem is a function of its character as the distribution of material entities. To measure the net product, we only need access to the natural unit of corn. It is worth noting further that the distribution of the good has nothing to do with the exchange of the product, and the secret of the corn model actually lies in this assumption of the absence of commodity exchange. Since there is only one product, there is, of course, no reason for the exchange of this universal product among the different actors.

Ricardo’s solution to the distribution problem in Essay on Profits is to dissolve the distributive parts into their natural class components. To present these shares as natural, the distribution must be separated from the value form of the product. Even in his later works, Ricardo, according to Sraffa, would attempt to reduce his theory of value to the only thing he assumes it is based on, namely the distribution of material products. Wages and rent acquire the character of natural distribution variables because they are not produced. Profit, which is the return on capital, on the other hand, expresses that production in a capitalist society is the production of capital. However, Ricardo tries to reduce profit to a natural factor by considering capital as an accumulation of material objects, as a capital stock.

Once capital and its self-increase in the form of profit have been reduced to natural categories, the only task remaining is to determine the proportions in which wages and profit should stand to each other. Establishing this proportion, and above all its development, is Ricardo’s main concern in Principles. The task of economics becomes to determine the distribution of a given value production. The inability of the corn model to treat distribution as the distribution of value is related to an unwillingness to view distribution as fundamentally a social process. The social form must be subordinated to natural reality.

However, Sraffa’s interpretation of the ‘early Ricardo’ is based on rather vague grounds, and he admits (in the Introduction to Ricardo’s Collected Works, vol. I, p. xxxi) that his rational reconstruction was never “explicitly expressed by Ricardo.” Upon closer examination of Sraffa’s arguments, it proves to be untenable. When Sraffa claims that Ricardo provided an exact formula for determining the profit rate in the agricultural sector, it is more accurate to say that he indicated the dependence of the profit rate on the size of the surplus. In Essay on Profits, capital is calculated in terms of corn — capital is estimated in quarters of corn — but nowhere is it assumed that it would consist only of corn, let alone that corn would be the sole input or the only commodity included in the wage basket.

Ricardo expressing the profit rate in terms of corn quantities does not imply any specific assumption about the composition of the wage basket. Both Sraffa and his later associate Piero Garegnani claimed that Ricardo assumed wages to consist solely of corn and that Malthus criticized him for this in their correspondence after Essay on Profits. However, this is not the case. Malthus’ criticism pertains to Ricardo’s choice of corn as the numéraire and not the composition of the wage basket.

What Ricardo actually does is assume that absolute prices, and consequently relative prices, of corn and other input goods remain constant, and that due to diminishing returns in agriculture, there is a falling profit rate through “the general rise of wages.”

While Ricardo’s analysis does not address price changes, this is not because, as Sraffa believes, he assumes homogeneity in product and capital (i.e., that capital is identical to the produced goods), but rather due to the specific price assumption mentioned. In Essay on Profits, the reduced profitability in the industrial sectors is a result of the diminishing returns in agriculture raising wages without correspondingly increasing industrial product prices.

Before Essay on Profits, Ricardo subscribed to Adam Smith’s ‘adding-up’ theory, according to which the price of a good was obtained by adding wages, profit, and rent. It is even likely that in the years preceding Essay on Profits, he believed that increases in production costs in agriculture would be reflected in corresponding price increases in corn. As Malthus pointed out in a letter to Ricardo on March 12, 1815, otherwise, it would be difficult for Ricardo to maintain his assumption of a falling profit rate in the agricultural sector. The problem of demonstrating the fall in the profit rate under different assumptions about how corn prices affected industrial product prices led Ricardo into a “labyrinth of problems.”

As his strongest ‘evidence’ that Ricardo had a corn model, Sraffa cites a letter from Ricardo (in Collected Works, vol vi, p. 117) where he states that the rate of profit and rent have to depend on the proportion between the production and the consumption necessary for that production. However, this is irrelevant as the definition only pertains to the ‘rate of product’ from which the rate of profit can be determined. Central to Ricardo’s explanation of the falling tendency of the profit rate is the difficulty or ease of procuring food, which depends on the necessity of cultivating poorer lands on which, with the same quantity of labour, the same quantity of products cannot be obtained. The only permanent cause of a fall in the rate of profit is thus increasing — not constant — production costs.

Thus, it is hardly possible to address Ricardo’s problems of value, distribution, and growth without considering diminishing returns in agriculture. However, Sraffa’s interpretation of Ricardo is fundamentally based on the exclusion of diminishing returns. The first signs of this can be found in his 1925 article Sulle relazioni fra costo e quantità prodotta:

It can be said that all classical writers accept implicitly, as an obvious fact, that cost is independent of quantity, and they do not bother to discuss the contrary hypothesis.

And a year later, In a letter to Keynes in June 1926, he again claim that Ricardo’s theory presupposes constant returns.

For Sraffa, diminishing returns in agriculture are something that only has relevance for distribution theory, and changes in the distribution of the net product due to diminishing returns are assumed not to affect relative prices, as this would affect all commodity costs to the same extent. However, Sraffa’s own analysis from 1960 shows that the relationship between output and production costs affects distribution and relative prices (through its impact on the rate of profit, which influences relative prices depending on the technical conditions of production concerning the numéraire commodity).

Where did Sraffa get the inspiration for his interesting but fundamentally flawed interpretation of Ricardo’s focus on corn? In Production of Commodities by Means of Commodities (1960:99), we find the explanation:

It was only when the Standard system and the distinction between basics and non-basics had emerged in the course of the present investigation that the above interpretation of Ricardo’s theory suggested itself as a natural consequence.

This interpretation, therefore, turns out to be a projection of Sraffa’s own research struggles onto Ricardo. As a description of Ricardo’s thinking, it must be considered overly speculative and poorly supported textually. However, the fact that it led to a renewed interest in classical economics is proof of Sraffa’s intellectual acuity and his ability to highlight the relevance of the ‘classics.’ For this, we owe him recognition.

Just as post-war mainstream economics has been criticized for being founded on an axiomatic-deductive mathematical methodology, I think it is fair to also question Ricardo’s most well-known modern follower, Sraffa, since the core of his neo-Ricardianism is based on the same kind of methodology.

To solve the problem of finding a measure of value that is not affected by changes in income distribution, Sraffa — in Production of Commodities by Means of Commodities — constructs his ‘standard commodity.’ Expressed in terms of this specific ‘commodity,’ the relationship between the profit rate and wages becomes linear. In the standard system, the profit rate is determined as a ratio of commodity quantities independent of their prices. Sraffa’s standard commodity plays the same role as corn in the ‘corn model.’ However, if we compare his solution to Ricardo’s problem in Principles, Sraffa only solves a part of it. Ricardo sought a measure that would be invariant to both changes in production and changes in distribution. Due to its static nature, the standard commodity is only invariant to the latter.

Now, some of Sraffa’s neo-Ricardian followers have claimed that by using the standard commodity, we could treat income distribution independently of prices and that whatever shortcomings there may be in the labour theory of value, it leaves — as Pasinetti puts it — “the possibility of treating income distribution independently of prices completely intact.” This claim must be strongly questioned. From an informational standpoint, the standard system adds nothing. In this system, the wage and profit shares are not clearly and simply related to the ‘real’ shares, as the ‘real’ income is not ‘invariant’ to changes in the profit rate. Our conclusion must be that the analysis of income distribution, contrary to what has been claimed, is worsened by using the standard commodity as the numeraire.

In doctrinal-historical contexts, it is sometimes claimed that the standard commodity would be a modern variant of Ricardo’s ‘corn’. However, this is hardly accurate, as ‘corn’ in Ricardo’s theory is a consumption good, whereas there is no reason to assume that the standard commodity represents the workers’ wage basket in Sraffa’s analysis.

Furthermore, the standard commodity is a poor approximation. As soon as we have technology choices and joint production, it is not possible to treat income distribution independently of prices. If we follow Ricardo and Marx and calculate profits on the entire advanced capital, price equations must be written in a form that causes the relationship between distribution variables to cease being linear and instead assume a hyperbolic form.

Sraffa perceives his standard commodity not only as a means to treat distribution independently of prices but also as a means to determine the source of relative price changes when distribution changes. Just like Ricardo, Sraffa fails in the second objective, as price changes depend on both technology and distribution. In Sraffa’s standard system, it is also impossible to separate the direct price effect of a change in the production conditions of the wage basket from the indirect price effects caused by changes in the profit rate accompanying altered production conditions. The redundancy of the standard system is also evident in the fact that the inverse relationship between wages and the profit rate, which the standard commodity is supposed to demonstrate, can be shown independently of it.

Regarding the issue of economies of scale, it turns out that even for the standard commodity, Sraffa must actually assume constant returns to scale (unless he wants to make the unrealistic assumption that workers and capitalists have the same composition of their demand). Otherwise, the standard commodity can only be constructed for a given combination of wages and profits.

This would mean that the cherished idea of a linear relationship between wages and the profit rate advocated by Sraffa would become meaningless. For prices, this would imply that they only hold at a specific point in time without possessing the ‘gravitational force’ that the ‘natural’ prices of classical economists possessed.

Sraffa’s work certainly undermined a significant portion of neoclassical economics claims, especially in its Marshallian formulation. However, contrary to Sraffa’s own belief, his work does not show that income distribution is independent of supply and demand. These forces can only operate in a world where consumption and production can vary, which is excluded in Sraffa’s model. While the model is consistent, the question is how much it can tell us about a changing world where dynamics, money, and expectations play a crucial role.

Sraffa’s critique of the neoclassical productivity theory based on aggregated production factors is a severe blow to this theoretical framework. However, it doesn’t have much impact on general equilibrium theory because it does not rely on the ability to measure capital independently of the profit rate.

Sraffa’s doctrinal-historical work demonstrates great intellectual acuity and interpretive power. However, by projecting his own research efforts onto his predecessors, the sustainability of his interpretations becomes dubious.

On the political front, one could argue that since Sraffa does not address production relations — on the contrary, he overlooks them to focus on distributional relations — his message implies a call for a more equitable income distribution rather than demanding the abolition of wage labour. It could be argued that his analysis, therefore, ultimately results in an apology for capitalism itself.