Sraffa, Keynes, and the Post-Keynesian Revival of Classical Political Economy
The Marshallian approach soon received criticisms from within the Cambridge tradition. The most famous critiques of the Marshallian framework were made by Sraffa, who came to Cambridge after publishing his critique, and Keynes. Sraffa’s criticism was much more devastating than Keynes’s because it undermined not only Marshall’s theory, as Keynes did, but also Marshall’s method, something Keynes did not do, at least not as consistently as Sraffa did. It was Sraffa who noted that Marshall’s method was developed with the problem of internal relations in mind, as we can see from Sraffa’s (D3/12/11/11) unpublished manuscripts at the Wren Library, Trinity College, Cambridge University, where Sraffa discusses the passages from Industry and Trade mentioned above, in which Marshall explains his method. Sraffa finds Marshall’s use of the method of Newton and Leibniz an intelligent, but fallacious effort.
Sraffa (D3/12/42/10, sixth proposition) argues that Newton’s method is not applicable to economics, and that Marshall must have been short of arguments to use such an unfair trick. The method Marshall adopted was used by Newton and Leibniz for the study of infinitesimal changes. But Sraffa notes that in economics we do not deal with infinitesimal changes. Thus, we cannot rely upon the method of Newton and Leibniz, which presupposes them. In economics, we must consider not only direct effects but also indirect effects, which are not negligible if changes are not infinitesimal. [1]
In two famous articles, Sraffa (1925, 1926) shows that we cannot use supply and demand curves while assuming that everything else remains constant. Movements in the demand curve will lead to changes in the quantities of factors of production used. These are used also in other markets, leading to changes in the prices of the factors of production across several markets, and to changes in the supply curve of the market under analysis. We cannot assume that everything else remains constant, or that these effects are negligible, since we are not dealing with infinitesimal changes. Thus, supply and demand curves cannot be taken to be independent determinants of prices and quantities. Also, we cannot assume that one curve moves while the other curve remains constant. Lastly, changes in prices lead also to changes in the number of firms operating in the market, contributing to further changes in the supply curve, which cannot be known using only the information provided by the supply and demand curves.
What Sraffas critique implies is that Marshall’s method does not really solve the philosophical problem of internal relations, which had been perceived early on by Marshall, and by Cambridge philosophers like Russell, Moore, and the young Wittgenstein. The problem of internal relations had been addressed in Cambridge in the early twentieth century by assuming that we could look at components of reality while assuming those components to be isolated from everything else. We can see this both in Marshall’s economics (and in the Cambridge economic tradition he founded) and in Russell’s and Moore’s philosophy (and in the Cambridge school of analytical philosophy that they founded).
Sraffa rejected Marshall’s solution to this problem, and also influenced Wittgenstein’s later works, where Wittgenstein abandoned his early approach to the problem. Sraffas rejection of Marshall’s partial equilibrium analysis together with Wittgenstein’s abandonment of his earlier logical atomism (which was, in turn, influenced by Friedrich Ludwig Gottlob Frege and Russell) are connected to the problem of internal relations.
The solution Sraffa found for the problem of internal relations was to look at the conditions for the reproduction of the economy and society as a whole, at a given moment in time. By focusing on the whole economy, we can take into account that reality is deeply interconnected. Moreover, by focusing on a given moment in time, we can abstract from changes that lead to complex processes through time which we cannot fully grasp analytically.
However, the question arises as to which moment oftime we should look at in order to understand the economy. Reality is permanently changing, and different configurations appear at different moments. Pierangelo Garegnani (1984) suggests that we look at the normal position of the economy, which is simply an average of various positions that take place over time. Over a sufficiently long period of time, we can then have a firmer grasp of the more persistent elements that constitute the underlying conditions of reproduction of the economy and society (see, in addition, Harcourt (1981) on how this idea also appears in Marshall, Keynes, and Sraffa).
Sraffa (1960) stressed that the method he adopted was simply the old standpoint of the classical political economists like Smith and Ricardo, which had been developed by Marx. Ronald Meek (1961) quickly interpreted Sraffas 1960 book as a revival of classical political economy. Sraffa provided a system where prices are determined by the cost of production, for a given distribution of income. Marshall’s (1890) [1920] theory of value was a development of the supply and demand framework that had been advanced by Malthus in his critique of Ricardo. Sraffas theory of value, in contrast, was a revival of Ricardo’s theory of value, where prices depend upon the cost of production, and supply and demand merely explain deviations from this cost, and cannot provide a systematic account of the underlying causes of value, as Sraffa (1925, 1926) demonstrated in his critique of Marshallian supply and demand curves.
In the first footnote of The General Theory, Keynes (1936 [1973]: 3, fn. 1) writes that although Marx invented the term ‘classical economists’ to denote Ricardo and his predecessors, Keynes himself had become accustomed to using the term ‘classical’ to designate those he believed were developing Ricardian economics, namely Mill, Francis Ysidro Edgeworth, Marshall, and Pigou. For this reason, Keynes (ibid.) presents The General Theory as a critique of classical theory, where classical theory is understood as a project that was to be found in its most advanced form in the writings of Marshall and Pigou. We can then see that while Sraffa presented his approach as a revival of classical political economy, Keynes presents his theory as a critique of classical political economy.
This happens because Sraffa and Keynes had very different, or indeed opposing, conceptions of classical political economy in mind. Keynes, like Marshall, interprets classical political economy as being no different from what Marx called vulgar political economy. Like Marshall, Keynes also sees Ricardo as part of this project. Sraffa (D3/12/4/10), by contrast, follows Marx’s distinction between classical political economy and vulgar political economy, and sees Marshall’s approach as a further development of vulgar political economy, with a more sophisticated technique. Thus, Sraffa saw Malthus, rather than Ricardo, as the true precursor of Marshall’s supply and demand analysis. The confusion is even greater because Keynes also believed that Malthus was the most important precursor of his own [Keynes’s] theory.
For Keynes, the defining feature of classical political economy was the adoption of Say’s Law, that is, the idea that supply creates its own demand. Keynes (ibid.: 32) notes that authors like Malthus and Marx had rejected this idea. According to Keynesian theory, we cannot presuppose that the economy will tend automatically to full employment of labour. Rather, the economy may fail to achieve full employment for an indefinite period of time.
Keynes’s own contribution was much influenced by the economists with whom he met regularly to discuss his ideas, such as Richard Kahn, James Meade, Joan Robinson, Austin Robinson, and Piero Sraffa, the so-called Cambridge Circus (before the 1930s, Keynes worked mainly with Dennis Robertson). Ralph Hawtrey and Roy Harrod are also important in this connection. Michai Kalecki had developed the same ideas as Keynes, which appeared in articles written before The General Theory was published. Coming from a Marxist background, Kalecki developed these ideas in a way which shows more evidently how they are clearly compatible with the classical surplus theory as developed by Marx.
However, the fact that Keynes presented his theory as a critique of what he saw as the classical theory, and the fact that he followed the vulgar interpretation of classical theory, obscured important connections between his theory of employment and the classical theory of value. In fact, if we interpret the classical theory in line with Sraffas perspective (which is in line with Marx’s original distinction between classical and vulgar theory), we find that Keynes’s theory is much more compatible with Ricardo’s perspective than with Malthus’s, contrary to what Keynes himself thought. Malthus’s approach is a forerunner of the theories that Keynes is really criticising, developed by Marshall and Pigou.
This issue can be seen more clearly if we note that Ricardo’s theory actually presupposes that there is no inherent tendency towards full employment, as is the case with Keynesian theory, while the marginalist supply and demand analysis that follows from the vulgar approach that stems from Malthus presupposes a case of full employment. Ricardo assumes Say’s identity between supply and demand, but there is nothing in Ricardo’s theory to sustain such an assumption, and the theory can do equally well without it, or in fact it becomes much more consistent without it, since Ricardo’s theory of value presupposes that there is no full employment of labour.
Walras saw this point, but used it to criticise Ricardo’s theory. Walras notes that, for Ricardo, the price tends to the cost of production because if demand drives the price above the cost of production, more commodities can be produced so that the price is brought back to the cost of production. The reason why more commodities can be produced, Ricardo argues, is because he presupposes that labour is available to do so. Walras criticised this Ricardian assumption, that more commodities can be produced to respond to demand and bring the price back to the cost of production. While Ricardo presupposed an economy where typically there is not full employment of labour, that is, labour is available for further production, Walras presupposed an economy where typically there is full employment of all factors of production, including labour, and so labour is not available for further production. There will be only a redistribution of fully employed supplies of factor services between the production of different commodities, according to supply and demand. Hence, prices will be determined by supply and demand, since variations in demand cannot be offset by variations in quantity produced. Demand is thus an independent determinant of prices, explained in turn in terms of marginal utility.
Marshall and Pigou further developed this supply and demand analysis, where prices and quantities are determined mutually. Keynes’s critique is targeted at the belief that supply and demand brings automatically full employment (while integrating theories of money and finance and prices with the determination of income and employment), not at what Ricardo actually wrote. Ricardo, like classical analysis in general, did not even address the determination of quantities that Keynes is focusing on. Rather, the classical authors focused on the determination of value while taking quantities as given (and, in the case of Ricardo, clearly presupposing that full employment cannot be taken for granted, as Keynes also did). Thus, the classical theory can be more readily combined with a theory that explains quantities, such as the Keynesian theory.
The contributions centred around the work of Keynes, Kalecki, and Sraffa led to the emergence of a new branch within the Cambridge economic tradition, the Cambridge Keynesian tradition (see Harcourt 2006 and Pasinetti 2007). Of course, depending on whether the emphasis is on Keynes, Kalecki, or Sraffa, we find differences within various sub-branches of the Cambridge Keynesian tradition. But the writings of the various authors of the Cambridge Keynesian tradition consistently reject the neoclassical approach of Marshall and Pigou. The Cambridge controversies in the theory of capital show us a moment where Keynesians, Kaleckians, and Sraffians stand united in the rejection of neoclassical economics, while advancing a perspective in line with Marx’s interpretation of classical political economy (see Harcourt (1972, 1976) and Cohen and Harcourt (2003) on the Cambridge controversies in the theory of capital).
A central theme within the Cambridge Keynesian tradition was a study of cumulative causation processes, undertaken by Kalecki, Joan Robinson, and
Nicholas Kaldor who, like the classical economists, saw how the distribution of the surplus is an essential aspect of the circular process of reproduction and accumulation of capital.
- [1] See Add Msc 96/2, Henry Sidgwick Papers, Wren Library, Trinity College, Cambridge University.
