Sraffa: The Origins of ‘Marginal’ Analysis

Since the advent of the ‘marginal’ method, the doctrines of the old classical economists have been submerged and forgotten. It is this standpoint that Sraffa revives in his 1960 book Production of Commodities by Means of Commodities. Being third in the series of posts [Post 1; Post 2] on Sraffa, this post examines the origin of the ‘marginal’ method and its subsequent (mis)use by the neoclassical economists. The posts concludes with a brief mention of how history of economic thought is important so as to place theories in a proper context.

In the preface of his book, Sraffa points out that in a system of production where the scale of an industry or proportions of factors of production remained unchanged, one would not be able to locate marginal product and marginal cost. To put it differently, marginal analysis is done by considering ‘potential change’. That is, we try to find out variations in equilibrium quantities and prices with respect to infinitesimal changes in the neighbourhood. [Bharadwaj 1986, p 39]

What we do not pay adequate attention to, is that the most familiar case of ‘marginal analysis’ is that of the product of marginal land (also known as no rent land) in agriculture, when lands of different qualities are cultivated side by side. This refers to the well known differential rent theory of David Ricardo. In fact, it is the case of diminishing marginal returns on land which is at the junction of the ‘fundamental methodological shift from classical to equilibrium theory’. [Bharadwaj 1986, p 40] This can be understood only through a discussion of ‘extensive’ and ‘intensive’ margins.

Cultivation on lands of different qualities is visualised as the outcome of a process of ‘extensive’ diminishing returns. On the other hand, successive use of more output producing techniques refers to the process of ‘intensive’ diminishing returns. [Sraffa 1960, p 76] In the case of ‘extensive’ margins in cultivation, ‘the rents can directly worked out on the basis of the single observed situation.’ [Bharadwaj 1986, p 41] Whereas, in the case of ‘intensive’ margins, the calculation of rent requires a quantitative change in the situation. That is, successive doses of labour and ‘capital’ need to be added to the land. And, a further assumption is made on the nature of these ‘doses’. These ‘doses’ are considered to be homogeneous. As Krishna Bharwadwaj explains: ‘At any moment of observation, no dose is distinguishable from each other. No ‘marginal product’ can, therefore, exist in this case without introducing potential change.’ [Bharadwaj 1986, p 42]

Thus, it is the Ricardian theory of rent which provided the basis for the neoclassical theory of distribution by providing an inverse relationship between successive doses of labour and ‘capital’ and their remuneration. This theory of Ricardo was intended to explain the origin of rents. In the hands of later authors, this was generalised to labour and ‘capital’. Hence, we see the inverse relation between ‘capital intensity’ and rate of profit in microeconomics textbooks of today.

From this excursion into the Ricardian theory of rent, two aspects are very clear. First, the concept of ‘marginal’ or ‘margins’ was used exclusively in the domain of cultivation. In ‘intensive’ cultivation, it is obvious that the output would increase only until a certain point, owing to the quality of that piece of land. Whereas, in the case of ‘extensive’ cultivation, the output would increase till all the acres of land are cultivated- notice the scarcity element here. What is not clear is the rationale of extending such an analysis into the area of manufacturing! Also, it is well accepted that land is scarce; but, is ‘capital’ or produced commodities scarce in a similar way’

No book of microeconomics mentions the origins of the famous ‘marginal’ analysis. And this method is so entrenched in the profession, that it is almost impossible to throw it away. It is in this context that other conceptual frameworks, that pay more attention to the changing historical conditions, assume importance. Probably, we need to revisit earlier theories and theorists not just for their own sake but for our sake as well in throwing light on contemporary issues. Sraffa’s work has inspired a lot of work on the history of economic thought, which will be summarised in a later post.


Bharadwaj, Krishna (1986), ‘Classical Political Economy and Rise to Dominance of Supply and Demand Theories‘, Universities Press: Calcutta.

Sraffa, Piero (1960), ‘Production of Commodities by Means of Commodities: Prelude to a Critique of Economic Theory‘, Cambridge University Press: Cambridge.

Sraffa: Production as a Circular Process

This is the second in the series of posts on Sraffa. The objective of this post is to clarify the assumption of scarce resources made frequently in neoclassical economics. This is then contrasted with the notion of mass-production in capitalist economies. This facet of capitalism is understood by concepts such as ‘circular production’ and ‘production of commodities by means of commodities’ in Classical Economics.

A brief look at the history of micro and macroeconomics becomes essential. Elements of Marshall and Walras are found in modern microeconomics. Specifically, partial equilibrium analysis comes from Marshall; whereas, Walras contributed ‘general equilibrium analysis’ to economics. Usually, emergence of macroeconomics is considered to have originated with the work of Keynes. This has been contested and it has been shown with considerable evidence that William Petty (1623-1687) was the first macroeconomist. [See Murphy 2009] And that economists like Adam Smith, David Ricardo and Karl Marx were talking about macroeconomics when they discussed production, distribution and accumulation. Neoclassical macroeconomics can be loosely said to comprise New Classical Economics, Neo-Keynesian Economics, variants of Computable General Equilibrium Models (CGE), etc. One of the unifying features of the above mentioned neoclassical schools/models is the assumption of ‘scarce factors’. It is owing to the assumption that factors are scare, that optimization is carried out.

Lionel Robbins defined economics as ‘the science which studies human behaviour as the relationship between ends and scarce means which have alternative uses.’ Here, scarce means refers to scarce factors of production ‘ land, labour and capital. Yes, land can be considered scarce in an economy where the pressure of population is high (or for environmental reasons). But, wouldn’t labour be scarce in some countries and abundant in others’ Now for the tricky ‘capital’. Capital is understood as produced means of production. That is, tools, machinery, plants, conveyor belts, electrical appliances, tractors, etc are ‘capital goods’. Are they scarce’ They would be scarce if nobody produced them. Usually, in a capitalist or quasi-capitalist economy, capital goods are produced by the private sector, the government and often, imported from abroad. Therefore, a priori, we have no reason to maintain that capital is a scarce factor. Or for that matter, even labour.

Marshall provided a theoretical partition through which one could say that factors are scarce. He introduced the concept of ‘short period analysis’. Till Marshall, the early classicals and neoclassicals analysed economies using the ‘long period method’. Through the short period, Marshall introduced an imaginary period wherein one factor is fixed (usually, capital) and the other factor (labour) is variable. In this period, it is as if one factor is scarce. In a later post, it will be shown that this sort of analysis is an improper generalisation of Ricardo’s theory of rent.

Sraffa’s Production of Commodities by Means of Commodities deals with ‘value and distribution’. That is, he focuses on the relationship between relative prices, wages/profits and technique of production. Throughout the whole analysis, output/quantity is treated as given. This is in tune with the ‘sequential analysis’ of classical economics. Value & distribution is one level of analysis or the ‘core’, as was popularised by Garegnani. Once, the foundation is well-established, the next level is growth & accumulation. In the first level, quantities are treated as given and in the second level, prices are assumed to be given. This is done keeping in view the complexity of the economic processes. Whereas, as we know, in neoclassical theory of general equilibrium, there is a simultaneous determination of quantity, price, wage rate, employment, rate of interest and quantity of capital. That is, all kinds of prices and quantities are simultaneously determined.

A set of equations from classical and neoclassical production theory is given below. This is so as to bring out the differences in a clear way.

Production function: Ya = f(La, Ka)

Sraffa’s equations:
(AaPa + BaPb + … + KaPk) (1 + r) + LaW = APa
(AbPa + BbPb + … + KbPk) (1 + r) + LbW = BPb
. . . . . .
(AkPa + BkPb + … + KkPk) (1 + r) + LkW = KPk

where A, B …. K are the output produced in various industries, L is the labour employed in each industry, Pa refers to price/value of output A and Pk refers to value of output K, r is the rate of profit. [As these are for purposes of illustration alone, some conditions have not been mentioned]

In the first case, it represents the transformation of inputs ‘ labour and capital into an output Y. Let me reproduce what Sraffa writes about his particular conception of production: ‘It is of course in Quesnay’s Tableau Economique that is found the original picture of the system of production and consumption as a circular process, and it stands in striking contrast to the view presented by modern theory, of a one-way avenue that leads for ‘Factors of production’ to ‘Consumption goods’.’ [Sraffa 1960, 93]

The concept of circular production also brings to the fore the web of connections between different production structures. Both classical and neoclassical economics attempts at reducing the complexity of economic phenomena. Neoclassical economics, at the outset abstracts away from interrelated production structures through the concept of ‘representative firm’ in microeconomics. In a similar way, in the area of consumption, man as a social being is reduced to man as an individual whose utility does not depend on that of others. Classical economics carries out its analysis by taking prices as given so as to analyse interrelated production structures.


Sraffa, P (1960), Production of Commodities by Means of Commodities, Cambridge: Cambridge University Press.

Murphy, A (2009), The Genesis of Macroeconomics: New Ideas from Sir William Petty to Henry Thornton, New York: Oxford University Press.

Piero Sraffa: An Introduction

In the first half of 20th century, Marshallian economics dominated economic theory and policy. Keynes in his 1936 book (The General Theory) tries to break away from the orthodoxy by challenging its concepts such as full employment and equality of savings and investment. In 1960, Sraffa mounted a strong critique through his book Production of Commodities by Means of Commodities.

Concepts such as utility, capital, prices, marginal product, marginal cost, etc were questioned by Sraffa. In fact, not only were they challenged, but they were also shown to be problematic. The following paragraphs will show how Sraffa pointed out the inadequacies with the above mentioned neoclassical conceptions of the economy.

Traditional theory lists land, labour and capital as ‘factors of production’. This means that production is carried out using some combination of the above factors. However, the process of production is seen as a one-way avenue from factors of production to production of final goods for consumption. Sraffa, in the tradition of classical economists argued that production is a circular process, or to use Myrdal’s phrase- production is a circular and cumulative process. This was in the tradition of the Classical economists who visualised the economy as an interdependent entity- Francois Quesnay was the first to provide a systematic account of interdependence in his Tableau Economique.

As we know, neoclassical theory of equilibrium prices (value) is built on the twin pillars of production and consumption. It is the production side that provides the supply function; and the consumption side provides the demand function. Through the interaction of demand and supply, equilibrium prices or value is created. This seems plausible and true, at the very outset. Why’ When we think of the production of, say, a car, we presume that both the buyer and the seller has some role in price fixing. Yes, this would be the case if we viewed production (of cars) as an independent activity. Whereas, in reality, production is a social activity and so is consumption. Marshall’s partial equilibrium analysis sought to understand equilibrium price and quantity formation in isolated firms/industries. Sraffa’s 1960 book provides a framework for understanding values (or in neoclassical terms, equilibrium prices) in an economy where production is a social activity and where the nature of production is circular. Hence the title of his book: production of commodities by means of commodities. This highlights the interdependent production structure, which is the case in today’s economies.

In the preface of his 1960 book, Sraffa points out the obvious problems of the ‘marginal method’. To have marginal cost, one needs to pay attention on change. To illustrate, suppose Hero Cycles produce 100 cycles a day using 10 machines and 10 labourers. And production is carried out in this way. How does one calculate marginal cost’ Do we add a machine and see how much extra output is produced’ Or do we ask one worker to work in Hero Cycles for a day’

The above paragraphs are only meant to introduce a reader to Sraffa. To me, Sraffa’s work has brought about a change in how I visualise the economy. Earlier, the linkages between various macroeconomic variables and their micro counterparts were vague. In the following posts, the above mentioned concepts will be explained in more detail. And, concepts such as increasing/diminishing returns, scarcity and prices, joint production, surplus approach, capital theoretic problems, etc will be tackled in the following posts.

On Prices/Values

Economics, rather Political Economy attempted at providing a coherent theory of value. Economists such as Adam Smith, David Ricardo, Karl Marx, etc are associated with a ‘theory of value’. Currently, in economics, ‘value’ is not discussed in courses of relevance. However, students are exposed to value theories such as labour commanded, labour embodied and so on.

This post is the second in the series of posts ‘On Prices’. This posts attempts at clarifying concepts such as values, prices and costs of production. Note that all prices which are mentioned in economics textbooks (microeconomics, introductory economics, principles of economics, etc) pertain to relative prices or long-run prices. That is, they do not talk about market prices. The reason for this is because it is assumed/believed that market prices tend to fluctuate or hover around these relative prices. In other words, given a particular technology, these relative prices, in some sense, reflect the interrelationships in the economy. Hence, these natural/normal proces are studied in order to understand the workings of a capitalist economy.

Let me start with what is usually taught in various economics and management institutes across the world and even in higher secondary schools. Prices are determined by the interaction of supply and demand. This implies that an excess dmand leads to a price hike. Let us look at an example: Suppose I go to a toy shop and ask for a Meccano set and immediately, another customer asks for the same set. But, the shop has only one Meccano set. Will the price of the Meccano set increase’ Is such an explanation intuitive or common sensical’ This example talks of an isolated case.

Economics is interested in the formation of a spectrum of prices at the level of the economy. Interestingly, macroeconomics has nothing to offer on price formation. Often, or rather everywhere in the world, economics is taught as microeconomics and macroeconomics. The interdependence and interrelationship present in any economy is inadequately addressed. The closest one comes is probably through the ‘circular flow’ diagram which highlights the role of the firm as well as the households. In this diagram, the complex and strutural interdependence is oversimplified to that of a 2-way interaction between the firm and household via labour market, capital market, etc and the state is shown to play the role of a facilitator. The interrelated production structures goes unnoticed or is seldom mentioned. Why is this important’

How can prices be determined’ (The dominant factor will be mentioned.)

1) Demand & Supply – The prices which are determined in this way are the prices of vegetables and fish, prices of shares in the stock markets, price of real estate, etc. In some sense, these prices can be said to be supply determined. For, these commodities are more often subject to variations in supply than in demand.

2)Costs of Production – An alternative view which is present in the literature is that the prices of commodities are prices according to the prices paid to the means of production as well as adding a certain percentage as profit. According to Kalecki, the percentage depends on the monopoly power of the firm. What if the firm’s final product is an input for another firm’ Will this affect the price of the product’ This aspect is often forgotten in economics.

This forgetfulness is strongly associated to the lack of importance mainstream and even some heterodox economic theories gives to interrelationships in the production structures in an economy. To have a glimpse into this, one needs only to look at an Input-Output table.

If we assume (correctly) that production structures in a capitalist economy are interrelated then we can conceptually distinguish goods/services into – Basics and Non-Basics. [Sraffa 1960] Basic goods are those goods which directly or indirectly enter into the production of every commodity in the economy including its own. An obvious example would be foodgrains because they are needed for labourers and labour is required in all activities. And a tax on a basic good will have cascading effects on the prices of all the goods in the economy.

I shall quote Sraffa to point out the significance of accepting and studying interdependence.

The exchange-ratio (or relative prices) of non-basics is “merely a reflection of what must be paid for means of production, labour and profits in order to produce them – there is no mutual dependence.” [p 8, Sraffa 1960]

“But for a basic product there is another aspect to be considered. Its exchange-ratio depends as much on the use that is made of it ….” [pp 8-9, Sraffa 1960]

It is because of these issues that Sraffa uses values/prices than costs. Also, Sraffa knew that in an economy, “costs of production cannot be measured independently of, and prior to, the determination of the prices of products.” [p 9, Sraffa 1960] To conclude, dan we therefore think that Sraffa’s analysis is similar to the neoclassical analysis of price using demand and supply’

In brackets, Sraffa writes “one might be tempted, but it would be misleading, to say that ‘it depends as much on the Deamnd side as on the Supply side.'” [p 9, Sraffa 1960]


1) Kalecki, Michal (1971), Costs and Prices, in Selected Essays on the Dynamics of the Capitalist Economy 1933-1970, Cambridge University Press, pp. 43-61.
2) Sraffa, Piero (1960), Production of Commodities by Means of Commodities, Cambridge University Press.