Sraffa Versus Keynes on the Method of Economics: Measurement, Homogeneity and Independence

Abstract

The chapter investigates the methodology Sraffa and Keynes apply in their critique of economic theories and the development of their own theory. The main focus is on their views on the measurement of economic magnitudes and the assumptions required. There are striking points of similarity in the methodology of their critique, but also contrasting approaches and divergences when they set about positively constructing their own new theory.

This paper was first published in Annals of the Fondazione Luigi Einaudi. An Interdisciplinary Journal of Economics, History and Political Science, Fondazione Luigi Einaudi, Torino (Italy), vol. 52(2), pp. 137–168, December, 2018. Here it is reprinted with the permission of the publisher.

Comment

1.1 A Comment on Carabelli on Sraffa and Keynes on Measurement

• John B. Davis 

Anna Carabelli’s (2018) excellent paper comparing Sraffa and Keynes on the measurement of economic magnitudes addresses both how the issue of measurement figured into their critiques of other economic theories and how it provided grounds on which they developed their own theories. She sees strong similarities in their critiques of existing theories, but she also sees significant differences between them in regard to how they each went on to construct their own theories. What they shared was the belief that “economics is characterised by heterogeneity and variety … [and] that the material of economics is complex and interdependent and require[s] a theory and concepts with similar characteristics” (166). Where they differed concerned the philosophical foundations of the theories they themselves developed: “Sraffa moved towards an objectivisation of economic theory based on physicalism and empiricism, while Keynes always remained faithful to a money vision of the economy which opposes physicalism and empiricism” (ibid.). To make this all clearer, let us begin by distinguishing Sraffa and Keynes in terms of both methodology and method.

Consider first their common ground regarding the critique of existing theories. Here the issue is what kinds of concepts are used in economics. A problem they agreed existing theories often encountered was their recourse to vague concepts in the sense of concepts not susceptible to quantitative measurement, whether in value terms or physical terms. Vagueness in concepts, then, not only precludes theories’ application to the world but also their empirical evaluation. We can characterise this as a specifically methodological problem because, as Carabelli notes, methodology concerns “epistemology and the philosophy of science basis on which explanations are made” (ftn. 1). Since this issue concerns the adequacy of existing theories, the problem of vagueness provides a basis for their critique of existing theories.

Consider second where they differed and in particular the issue of the unit or standard of measure, or that by which measurement is carried out. Clearly a unit or standard of measure needs to be independent of the quantities it is used to measure since otherwise it fails as a standard of measure. On this they agreed. In this case, we can characterise the issue here as a once concerning method because, as Carabelli notes, method concerns “technique and how concepts are used, applied, and developed” (ibid.). Then, since which particular standard of measure one adopts affects the foundations on which new theories may be reliably constructed, the issue bears on the question of where Sraffa and Keynes went in different directions theoretically.

Thus, in regard to the measurement of economic magnitudes, Sraffa and Keynes largely agreed with one another on methodological grounds and in their critiques of much existing theory, but when it came to the theories they each developed they disagreed with one another on grounds of method regarding how economic magnitudes should be identified. An implication of this is that Sraffa’s “objectivisation of economic theory based on physicalism and empiricism” and Keynes’s “faithful[ness] to a money vision of the economy” derived not from methodological differences between them regarding what theories needed to do in order to be adequate but from differences over the technique and practice in how theories could be developed. That is, once the problem of vagueness is addressed, what techniques one employed depended on what types of theory one intended to develop.

Carabelli, then, nicely captures the dilemmas Sraffa and Keynes each encountered—indeed the economist encounters—regarding determining a method for measurement of economic magnitudes:

not all economic quantities can be expressed (measured) in physical terms … [and often] quantities can be expressed or measured in physical terms only under restricted conditions. In fact, when difficulties in their physical measurement arise, economists usually make recourse to value measures, either in price or in money terms. And yet, problems may also arise with value measures. (142)

One might prefer physical measures, then, but they may not be feasible given the goals of one’s type of analysis. At the same time, recourse to value measures encounters its own difficulties. Thus, there is no simple path forward when it comes to constructing theory, and one has no alternative but to exercise judgment regarding how one proceeds.

In Sraffa’s case, this came out in his remarks at the Corfu conference on capital theory, where he distinguished between measurement as statisticians understand it and measurement in theory. The latter is the basis for the former and in itself requires absolute precision. This depended upon having an adequate definition of capital, which Sraffa believed Böhm-Bawerk and J.B. Clark had failed to achieve. For him, this was thus a methodological failing that helped him ultimately develop his understanding of the Standard Commodity , which allowed him to advance his view that there exists a linear relationship between the share of wages and the rate of profits in the surplus. Carabelli reviews, then, Sraffa’s account and evaluation of three types of quantities possible in economics, and his rejection of two of the three and recommendation that economic theory be based on the third, specifically, quantities that have an objective, independent existence at points in time or throughout the processes of production and distribution.

In Keynes’s case, how he proceeded stemmed from his view of economics as a moral science. This in turn derived from his early thinking about the logical and theoretical difficulties involved in thinking about the nature of probability, which carried over to his economic thinking about the macroeconomy where expectations play an important role in a probabilistic world. For Keynes, economic magnitudes are not like those employed in the physical sciences, because economic magnitudes exhibit an intrinsic indeterminacy. Thus for him there is no unique unit of measurement. Indeed, as Carabelli puts it: “The impossibility of a numerical measurement of probability is not a product of mental incapacity or lack of knowledge, but arises from the nature of the case itself” (148). Needless to say, this implies a quite different approach to economic theory than the one Sraffa adopted.

However, despite their differences regarding method and theory construction, Sraffa and Keynes agreed methodologically about the centrality of the assumptions of homogeneity and independence—or rather their heterogeneity and variety—when it came to how we think about the phenomena that economics investigates. At root in their thinking was their shared conviction “that the material of economics is complex and interdependent (166). Here I only briefly summarise their thinking about homogeneity and heterogeneity, as reviewed by Carabelli, since the degree to which we can ascribe independence to economic phenomena depends first on whether and when we can judge them as homogeneous.

As she points out, then: “Homogeneity means that there is no difference (or variety) in the kind (or substance) of quantities” (150). Thus for Sraffa, the heterogeneity of economic phenomena was the starting point, though importantly under particular conditions homogeneity might be assumed. That is, the issue for him are the “relevant differences or variety” (ibid.). As Carabelli explains, the

[I]ntroduction of the assumption of homogeneity depended on judgements of relevance. For Sraffa relevance implied that although there exist differences and variety in the quantities they can be judged negligible: relevance was not empirical but logical, i.e. relevance to the problem under investigation. (ibid.)

Similarly, for Keynes, economics quantities are generally not homogeneous except under quite specific conditions. For him, economic magnitudes are a

non homogeneous complex which cannot be measured […] except in certain special cases … [they are an] incommensurable collections of miscellaneous objects. (CW 7: 38, 39)

For both, consequently, the goals of theory construction consequently determined when homogeneity in economic magnitudes might be introduced or assumed. This was in contrast to the theories they criticised, in which homogeneity of economic magnitudes was mistakenly assumed to be the case by default, thus mis-representing the nature of economic phenomena from the start.

Carabelli’s paper thus goes on to review in careful detail how on these grounds Sraffa and Keynes reasoned about the concepts of labour, capital, utility, and money. This discussion is much recommended for careful study. Indeed, it bears directly on not only how we understand the relationship between Sraffa and Keynes, but also why they moved in “opposite directions” to construct quite different types of theories—Sraffa with his search for “an absolute physical standard of value and measure” and “Keynes stuck to money measures and to the unit of labour as the sole physical unit” (160).

Of course there is much debate about the relationship between Sraffa and Keynes’s ideas and about the differences between the theories they constructed. Yet generally these debates have ignored how their methodological views and understanding of method were foundational to their critiques of existing theories and the theories they developed. However, as Carabelli nicely shows, their common ground effectively created a division of labour between them regarding their respective positive contributions. In this way, she opens the door to deeper study of the commonalities and distinctions within the Cambridge School.