A critical reorientation of Keynes's economic theory
In the system of GT we find two sides that are logically inconsistent with each other.
On the demand side, the central aim of Keynes's analysis is, it would appear, to outline a theory of the determination of output (or employment) by effective demand. The core of that theory is a physical quantity-adjustment mechanism for the whole economic system - in a word, the 'multiplier' process. It is generally recognized that this is the central feature of the Keynesian Revolution. As argued especially in sections 10.4-10.7 of this chapter, we cannot say that Keynes consistently adopted the atomic hypothesis, in particular for the demand side of his system. Indeed, when we consider Keynes's philosophical thoughts, attention must be paid to his suggestion that there is deterministic uniformity for the 'whole' system versus the uncertain and varied behaviour of individuals, or its 'parts'. By contrast, the cost side of GT explicitly depends on the atomic hypothesis, as shown by Keynes's intention to extend by analogy to a whole industry conclusions arrived at in respect of the cost behaviour of a single industry.
The other important property of GT is, as it were, its two-layered logical structure. Its fundamental layer is the 'principle' of effective demand. It is entirely independent of any behavioural relations and even of any particular institutional set-up. As indeed mentioned in section 10.7, though Keynes seems to grasp it intuitively, he can never formulate it explicitly. Nevertheless, the principle of effective demand plays a basic role in what concerns a 'production economy', particularly a technologically advanced and developed economy. In such industrialized economies, effective demand generates production in a simple but fundamental way. The other layer of the logical structure, as formulated explicitly in GT and in the article in The Quarterly Journal of Economics, is the 'theory' of effective demand. This layer is based on a particular institution, that is, a capitalist market economy, and on specific behavioural relations of economic agents (such as entrepreneurs, workers and consumers, and so on). Its kernel is, of course, the notion of an 'investment multiplier'. In short, Keynes's theory, as composed of this two-layer structure, is undoubtedly a genuine 'aggregate' model in respect of the economic system as a whole.
Compared with Keynes's approach, Sraffa's production system also forms, in a sense, a kind of two-layered structure. Its fundamental layer is the system of production before the emergence of capitalists and workers. It is independent of any behavioural relations and even of a particular institutional set-up. And its primary feature, which is ultimately found in any kind of 'production economy', is the process of value generation by the methods of production, that is, by technological inter-industry relations. Thus, we may say that this layer is situated on nearly the same logical level as Keynes's 'principle' of effective demand. The other layer is the system of production after the emergence of capitalists and workers. It explicitly assumes a particular institutional set-up, that is, a technologically advanced capitalist economy. However, Sraffa's 'distribution variables' - either the unit wage or the rate of profits - are incompatible with the case assumed by neoclassical economists, since one of them is determined from outside the system of production. In other words, they are not determined only by the price mechanism. All this implies that the values of commodities are independent of the behavioural relations of demand and supply, as they are determined 'objectively' by technological inter-industry relations - in a word, by the technology of the system. In this sense, Sraffa's theory of prices is a thoroughly non-behavioural and objective system.
A fundamental line of critical reorientation of Keynes's economic theory can be outlined as follows.
1. At first, the most serious difficulty of Keynes's theory lies in the logical inconsistency between its demand side and its cost side.
In that connection, we should, above all, leave out the micro foundation provided by the Marshallian increasing (marginal) cost based on the first classical postulate, as that may be considered as the most fragile building block in Keynes's construction. And we should replace it with the more solid foundation proposed by Sraffa, a critical successor of the classical economists (Smith, Ricardo, Marx).
- 2. In view of the logical consistency and formal elegance of Sraffa's system, which is a disaggregated multi-sector model, it would seem natural to transform Keynes's system, too, into a kind of disaggregate model, although the latter was originally aimed at the formulation of aggregate theory. Or it may be more desirable to generalize the field of application of Keynesian theory by using the original aggregate model for some purposes and the disaggregate model for other purposes. But what should not be overlooked is that there is need for a sort of disaggregation of Keynes's original model in order to integrate both its demand side and its supply side into a consistent theoretical scheme.
- 3. In order to transform the Keynesian aggregate model into a disaggregate one, the demand for output as a whole (that is, effective demand) must be subdivided into as many components as there are commodities. In this case, as Pasinetti (1993; 1997) suggests, we need to introduce the so-called macroeconomic condition23 as the overall condition of full employment, instead of the principle of effective demand.
Sraffa's subsystems (Sraffa, 1960, p. 89) represent a type of vertical integration model. This point of view provides a connection between Sraffa's system and that of Keynes (see below):
1. Let us partition Sraffa's original system into as many components as there are commodities. Any such component (that is, any 'subsystem') forms a smaller self-replacing system producing a net product consisting of only one kind of commodity. In order to produce the net product, each subsystem uses directly or indirectly not only labour but also means of production.
However, in any one subsystem, the amount of means of production existing at the starting point of the annual cycle is fully replaced during the production cycle, so that it would be possible to leave aside such means of production altogether. The primary aim would be, first, to single out merely two quantities, that is, the 'net product' and the 'amount of "labour"' required (directly or indirectly) to produce it, and, second, to relate directly the former to the latter.
In his discussion of the model of 'reduction to dated quantities of labour' (Sraffa, 1960, ch. 6), Sraffa states:
[I]n the sub-systems we see at a glance, as an aggregate, the same quantity of labour that we obtain as the sum of a series of terms when we trace back the successive stages of the production of the commodity.
- (Sraffa, 1960, p. 89)
- 2. If we suppose a general case including 'joint production', it would appear that Sraffa devised the notion of 'subsystems' in order to evaluate unambiguously the extent of validity of the classical labour theory of value, whereas Pasinetti (1990) in his 'multi-sector vertical integration' analysis proves stringently that Sraffa's idea of 'subsystems' serves as a fruitful analytical tool from the point of view of effective demand as well. Thus, the method of singling out the direct relationship between 'net product' or 'final product' and 'the amount of labour' represents, in a sense, a device to get away from the whole set of means of production used up directly or indirectly to produce the net product.
In outline, the attempt to integrate Sraffa's system with that of Keynes elucidates the fact that the theoretical scheme of a 'production economy' must be completely separated into a quantity system independent of each other. Thus, in a certain period of time, given the technology and the productive organization of the economy, the price system is determined by technological relations while the quantity system is determined by sectorally differentiated effective demand, that is to say: Technology ^ ; Demand ^ This simple but clear analytical scheme is, I think, a fundamental point of departure for disclosing the fundamental features of the structural dynamics of a 'production economy' in which the primum movens is, without doubt, continuous technological progress.
and a price system
independent of each other.
Thus, in a certain period of time, given the technology and the productive organization of the economy, the price system is determined by technological relations while the quantity system is determined by sectorally differentiated effective demand, that is to say:
; Demand ^
This simple but clear analytical scheme is, I think, a fundamental point of departure for disclosing the fundamental features of the structural dynamics of a 'production economy' in which the primum movens is, without doubt, continuous technological progress.