Blog

In a series of texts published during the 1960s, the most important of which is undoubtedly the essay Equilibrio generale ed equilibrio macroeconomico [General Equilibrium and Macroeconomic Equilibrium] (1965), Augusto Graziani presented a critical appraisal of general economic equilibrium theory and its historical development. Starting from the inconsistencies in Walras’s theory of capitalisation –already discussed by Gragnani (1960), Tosato (1963) and Trezza (1963) – Graziani emphasised the need to reformulate general equilibrium in aggregate terms in order to address the issue of dynamics. The invariance of prices over time – and thus the persistence of equilibrium – could be ensured only by assuming a uniform rate of expansion for all capital goods figuring in the system. Such a condition amounted to treating the various capital goods as dimensions of a single commodity. This led Graziani to argue that, even without relying on the highly unrealistic assumptions of perfect foresight and the existence of forward markets for every future point in time, a disaggregated general economic equilibrium framework was inherently unable to provide more than a static description of the economy at a given point in time.

A transition towards dynamics was feasible only on the basis of an aggregate analysis. From this perspective, the analogies that Graziani highlighted between von Neumann’s model – which sought to translate Walras’s perspective into dynamic terms – and the macroeconomic accumulation models of Harrod and Kaldor were far from coincidental. While this move made it possible to extend equilibrium analysis to dynamics from a formal standpoint, the economic significance of the result remained unsatisfactory. First, the opposition between the classical and the marginal theories of distribution remained unresolved, since von Neumann’s model – despite of Champernowne’s (1946) interpretation in terms of a rehabilitation of the classical perspective – continued to exhibit a functional relationship between wages and labour supply. Moreover, this approach could not dispense with the assumption of a given technological configuration, thereby abstracting from the interrelation between technical progress, disequilibrium and capital accumulation. By assimilating the economic system to the case of a single commodity, it introduced an undue simplification that made it impossible to account for differences in both the rates of profit and productivity growth across the various sectors of a developing economy.

In what follows, I shall first present Graziani’s discussion in General Equilibrium and Macroeconomic Equilibrium of Walras’s theory of capitalisation. I shall then analyse the attempts to extend Walras’s analysis to dynamics, first from a microeconomic perspective, centred on the problem of the compatibility of individual agents’ plans, and then from a macroeconomic perspective, where the aim is to define the conditions ensuring, under constant prices, the maintenance of full employment of resources over time. This in turn will lead to Graziani’s account of the various post-Walrasian versions of equilibrium: the distinction between disaggregate and aggregate equilibria is intertwined with that – based on different interpretations of consumption – between classical and neoclassical equilibria. Finally, I shall examine the reasons why, in Graziani’s view, the general economic equilibrium paradigm was inadequate for the analysis of economic dynamics. In this respect, his approach, both in its references and in the theoretical requirements it sets out, appears to be in line with the project that Pasinetti would later carry out to develop a structural economic dynamics.

 

General economic equilibrium in Walras

In his analysis of Walras’s general economic equilibrium model, Graziani discusses the main – and opposing – criticisms levelled against the Elements of pure economics: namely that Walras formulated either an indeterminate (Tosato and Trezza) or an overdetermined (Garegnani) system.

According to the first criticism, Walras’s model is compatible with different values of the interest rate. In order for prices to be determined, the interest rate must therefore be treated as exogenous. This indeterminacy arises because, if one accepts Walras’s assumption of fixed production coefficients, the demand for each capital good will not be influenced by the equations of capitalisation. These equations will merely determine the proportion between the different capital goods which are compatible with a uniform rate of net return, without fixing the absolute level of demand for each of them. The level of demand is instead determined by the aggregate demand for investment. However, determining the demand for investment requires to know a demand curve relating, for given endowments and utility functions, the interest rate to the quantity of capital invested. Yet, as Joan Robinson (1953) had already observed, in the presence of heterogeneous capital goods it is not possible to construct a measure of the quantity of capital other than in value terms. Such a measure, however, presupposes that the interest rate is known. The interest rate must therefore be taken as a given if the system is to be determinate.

Furthermore, according to Graziani, the assumption of fixed coefficients led Walras to a decidedly non-neoclassical result: the structure of production was sufficient to determine prices, while the quantities produced were determined by utility functions. Since the indeterminacy of the system under fixed coefficients stemmed from the inability of utility functions to influence the production methods employed, the adoption of variable coefficients was sufficient to ensure the existence of a unique solution. Under the particularly restrictive assumption of perfect substitutability among both consumption goods and factors of production, it also became possible to rule out the possibility of the solution containing negative prices – a problem to which Walras’s system would otherwise inevitably be exposed. Alternatively, in line with Abraham Wald’s (1935, 1936) reformulation of general equilibrium, this difficulty could be overcome by replacing the equilibrium equations with inequalities. It was thus acknowledged that supply might exceed demand – in which case the good was regarded as non-scarce, and therefore free – and that the price of a good might fall below its cost of production – in which case the production process would not take place.

While accepting the indeterminacy of Walras’ original formulation of general equilibrium, Graziani rejects Garegnani’s charge of overdetermination as entirely unfounded. According to Garegnani, the interest rate is determined in Walras’s model by the equations equating the cost of production of each capital good with its price. Let r denote the interest rate,  the prices of the m productive services employed for producing the new capital goods, and b₁₁…b_1s,… b_m1…b_ms the production coefficients employed in the s productive processes:

b₁₁v₁ + b₂₁v₂ + ··· + b_m1v_m = v₁ r

b_1sv₁ + b_2sv₂ + ··· + b_msv_m = v_s r

Garegnani assumes that the m prices of productive services are independent of the prices of newly produced capital goods, being determined, together with the prices of consumption goods, exclusively by the equations establishing equilibrium between the supply and demand for each productive service and by those establishing the equilibrium between the supply and demand for consumption goods. They must therefore be taken as given within this new set of equations.

To support this claim, Garegnani argues that the differences between the production coefficients required to produce different capital goods can be regarded as negligible. Consequently, changes in the relative demand for capital goods would have no effect on the relative prices of productive services and, therefore, on the relative prices of consumption goods. Given not only the production coefficients but also the prices of productive services, each equation ultimately determines a different interest rate, thereby rendering the system overdetermined.

For Graziani, however, the alleged independence of the prices of consumption goods and productive services from the prices of capital goods was entirely unjustified. Not only did Walras make no mention whatsoever of such an assumption, but it also ran counter to the logic of general economic equilibrium, in which all endogenous variables are interdependent and, consequently, determined simultaneously.

Furthermore, the equations in question were not used to determine the interest rate, but rather the proportions in which new capital goods had to be produced to ensure a uniform interest rate. The interest rate, on the other hand, was determined by the equation equating gross savings with the value of the capital goods produced:

S = v₁ r K₁ + v₂ r K₂ + ··· + v_s r K_s

Although Garegnani’s objection was unfounded, different interest rates could still arise. They arose, however, not from the determination of the rate of interest on newly produced capital goods, but from the coexistence of these newly produced capital goods and durable capital goods produced in previous periods and, therefore, under conditions of production distinct from the current ones. As Pareto (1896–1987) had already observed, in equilibrium previously produced capital goods will inevitably benefit from negative or positive quasi-rents arising from the gap between their cost of production – determined in earlier periods – and their current value, which reflects the conditions of the present equilibrium. The divergence of their rate of interest from the one on newly produced capital goods does not constitute a problem; rather, it is the inevitable consequence of the existence of past periods of production. Only when the structure of production has become stationary is the system be compatible with the absence of quasi-rents and, consequently, with a uniform rate of interest, as Walras himself appears to have recognised:

If we suppose that old fixed capital goods proper of the types (K), (K’), (K’’)… are already found in the economy in quantities Q_k, Q_k’ Q_k’’… respectively and that their gross and net incomes are paid for at prices determined by the system of production equations and by the rates of depreciation and insurance, it is not at all certain that the amount of savings E will be adequate for the manufacture of new fixed capital goods proper in just such quantities as will satisfy the last equations of the above system. In an economy like the one we have imagined, which establishes its economic equilibrium ab ovo, it is probable that there would be no equality of rates of net income. Nor would such an equality be likely to exist in an economy which had just been disrupted by a war, a revolution or a business crisis. (Elements of Pure Economics, p. 308)

While the difficulties in Walras’s theory could thus be overcome from a formal point of view, the economic object described by the equations ultimately took on a rather unexpected form. Since prices were determined only for the current period, equilibrium was reduced to a point-in-time concept, entirely indifferent to both past and future states of the economic system. The temporal structure of economic processes was thus sacrificed – a cost Graziani found unacceptable:

With the introduction of durable goods, time ought to become part of the general equilibrium system; but the general equilibrium system, by its very structure, is such that it describes only instantaneous situations devoid of duration. (Equilibrio generale ed equilibrio macroeconomico, pp. 66–67)

Each period described by the equations “thus represents a self-contained episode; there is no mechanism leading from one equilibrium to the next, because the equilibrium solution for each period does not depend on the solutions of previous periods” (p. 67). Walras’s model thus describes a world “in which history starts afresh every day” (p. 67). As Graziani goes to argue, this perspective implies not only that there is no guarantee that the productive structure rules out the existence of quasi-rents, but also that there is no guarantee that the existing stock of capital is sufficient to ensure full employment.

Full employment could only be met only if capital goods were perfect substitutes or if the available stock of capital were fully compatible with the composition of demand. The first possibility, equivalent to the existence of a single capital good, undermines the distinctive nature of general equilibrium as an analytical framework capable of accommodating the heterogeneity of both capital goods and consumption goods. The second cannot be guaranteed by any economic mechanism and can only occur by chance: since the capital goods inherited from previous periods were produced in response to past demand conditions, which may differ substantially from the present composition of demand.

Only under conditions external to the price-determination mechanism does Walras’s equilibrium, “unbalanced in substance and evanescent in duration” (p. 70), admit solutions that are not only stable over time, but also guarantee, in addition to formal equality between supply and demand, the full employment of available resources. It follows that, if the analytical significance of the general equilibrium framework is to be preserved, the model must be reformulated so as to overcome the independence of the different periods. Such a reformulation may be pursued in two ways: either by ensuring that the decisions taken by economic agents remain mutually compatible for as long as their effects persist, thereby affecting not only the current period but also the future ones; or by imposing conditions that ensure the price system remains unchanged over time.

 

Neoclassical equilibrium after Walras

According to Graziani, there are three possible approaches to safeguarding the significance of general economic equilibrium while addressing the problem of the compatibility of individual decisions over a multi-period horizon.

The first, which he traces back to Pigou (1935), von Wieser (1889) and Eucken (1954), consists in assuming that all decisions are made by a single agent. In this way – an approach that would later find favour in the concept of the representative agent – “any potential incompatibility between the plans of autonomous agents is automatically eliminated” (p. 71) by definition. By sidestepping the problem of compatibility between the plans of heterogeneous agents, this approach is deemed, despite its popularity, untenable.

A second solution consists in assuming perfect foresight. Knowledge of all future states of the economy enable agents to formulate plans that will always be realised. Moreover, since all agents formulate their plans based on an identical knowledge of the future, these plans will be mutually compatible. This position, associated with Knight (1921) and Hayek (1941), cannot be accepted in a world where “in the course of events, forecasting error is the rule; correct forecasting is the exception” (p. 72).

Finally, it possible, as proposed by Hicks in Value and Capital (1939), to assume the existence, alongside spot markets, of forward markets for every commodity. Bargaining in these markets enables agents to acquire knowledge of prices that will prevail in future periods. While this proposal represents an advance on Hayek’s insofar as it provides a justification for agents’ knowledge of future prices – since they themselves negotiate these prices – it ultimately sidesteps the problem of the temporal nature of the economic process by confining price negotiation to the initial period alone and interpreting subsequent periods merely as the execution of previously established agreements.

The limits of Hayek’s and Hicks’s approaches stem not merely from their lack of realism, nor from their implicit return to a static perspective, but also from the fact that, while ensuring the absence of quasi-rents, they tell us nothing about the conditions an equilibrium must satisfy to ensure the full employment of available resources. The formulation of a substantive general economic equilibrium therefore appears to require a departure from the neoclassical perspective.

 

Macroeconomic equilibrium

While the introduction of expectations within a framework of dynamic general equilibrium inevitably leads either to the unrealistic assumption of perfect foresight or to the static conception that all prices are negotiated in the very first period of the economic process, a return to the idea that agents make their decisions on the basis of current prices may suggest a different approach to the problem, provided that these prices remain unchanged over time. This perspective can apply both to a stationary economy, in which no new capital is formed, and to a dynamic economy, provided that capital accumulation does not disrupt the relative costs of production and that the additional income created does not alter the composition of demand. In the dynamic case, the economy must grow at a constant rate while maintaining the proportions in which the various capital goods enter the production process. In such a dynamic equilibrium, the total value of capital goods will represent a constant share of total income, while the rate of return on investment – which must remain constant – is given by the ratio of the growth in monetary income to the system’s propensity to save.

In an economy in which the composition of both income and product remains constant, every good can be viewed as a fraction of a single commodity that is either reproduced on the same scale or expanded at a constant rate of growth. If general economic equilibrium is to acquire substantive significance, expressing more than a static picture where “the past and the future appear as mere exogenous data” (p. 78), its disaggregated character must inevitably be sacrificed. It is therefore no coincidence that “modern dynamic theory has been developed almost exclusively through the use of aggregate models” (p. 78).

The more prominent examples are the models of Harrod (1948) and Domar (1957). Graziani highlights the indeterminacy of investment rate in these models: once the level of investment that ensures equilibrium with saving (and with it the full employment of resources) has been determined – as the model indeed allows – “there is no guarantee that actual investment will conform to this level” (p. 81). Unlike Walras’s approach, which includes a mechanism for convergence towards equilibrium based on agents’ propensity to buy dear and sell cheap, the macroeconomic perspective merely establishes the conditions required for a substantial equilibrium over time, i.e. the maximum growth compatible with full employment of resources, without providing any guarantee that these conditions will arise spontaneously.

Although it appears as a disaggregated equilibrium, von Neumann’s model too must, according to Graziani, be classified as an aggregated model. Indeed, in formulating the conditions for balanced growth, it defines a single rate of expansion for all productive sectors. Since the goods figuring in the system are required to maintain the proportions in which they appear as components of output, von Neumann’s framework ultimately proves to be, for the reasons already discussed, consistent with a macroeconomic approach. Moreover, his model can be interpreted as a special case of Harrod’s model, namely the case where the marginal propensity to save is equal to one and the whole national income is therefore reinvested. Under these conditions, the growth rate and the rate of profit then coincide.

 

Classical and neoclassical equilibria

According to Graziani, the shift to the macroeconomic perspective gives rise to two distinct approaches to consumption. Consumption may be regarded either, as in the neoclassical tradition, as constrained by the quantity of labour supplied or, as in the classical tradition, as depending exclusively on the quantity of available resources. In the neoclassical interpretation, there is necessarily a functional relation between the labour supply and the real wage. Since wage goods account for the largest share of the demand for consumption goods, higher per capita consumption is possible only through the provision of additional labour – and hence through the creation of an additional quantity of output. At the same time, an increase in production is possible only through an increase in the resources allocated to consumption. Distribution and production, the level of consumption and the level of production thus appear to be closely interdependent.

The situation is different, however, in the classical tradition, where the equilibrium wage was, at least in the short run, “unbound from the labour supply” (p. 87). While the level of consumption did depend on the level of production, the latter was in no way influenced by the level of consumption, owing to the inelasticity of the labour supply with respect to changes in the real wage. In the long run, this inelasticity was certainly called into question by the link, according to Malthus’s theory, between changes in wages and in the population: such a relationship ensured that wages remained constantly at the subsistence level. On the other hand, the Malthusian mechanism was more than offset by the positive effect that capital accumulation and technical progress had on the equilibrium wage, constantly raising the threshold of subsistence.

The classical approach thus made it possible to “conceive of the economic process as aimed at the production of consumption goods, attributing to the act of consumption the role of the final stage of all economic activity”, whilst the neoclassical approach treated consumption – “functionally linked to the provision of labour services” as “one link among many in the overall production process” (p. 88). This position stands in stark contrast to Garegnani’s interpretation of the opposition between classical and neoclassical economists. According to Garegnani (echoed on this point by Napoleoni), it is classical thought that calls into question the autonomy of consumption by treating wage goods as just one input among many in a circular economic process aimed exclusively at its own reproduction. Conversely, the neoclassical economists introduce a linear representation of production that inevitably makes consumption the ultimate end of economic activity. Although decidedly less influential than Garegnani’s interpretation, this view nevertheless seems to resonate with Pasinetti’s interpretation of the economic process. In Pasinetti’s framework, which relies on his own reformulation of the labour theory of value, workers’ consumption is the sole true income of the production process, whereas profit, by contrast, is seen as the necessary cost for the expanded reproduction of the system.

Another original feature of Graziani’s analysis is the way he employs this distinction between the classical and neoclassical schools to emphasise the continuity between Walras and von Neumann, in contrast with the interpretations advanced by Champernowne (who had first proposed an anti-marginalist reading of von Neumann) and Napoleoni (who had attributed to Walras the desire to safeguard the autonomy of consumption).

In Walras’s system, as a consequence of the general-equilibrium framework, the demand for consumption goods depends on equilibrium prices of the productive services, and therefore also on wages. In turn, wages were inevitably a function of the supply of labour.

In Graziani’s view, the same functional relationship between the level of consumption and the labour supply could also be found, albeit implicitly, in von Neumann. The absence of demand functions for consumption goods was, for Graziani, “more apparent than real” (p. 83): if labour was a factor of production reproducible in the same way as capital goods, and a given vector of consumption goods simply represent the bundle of inputs required to produce one unit of labour, then the conditions governing the production of labour expressed nothing other than “the demand for consumption goods by workers as a function of the labour they performed” (p. 83). Here too, access to consumption was mediated by the service rendered to the production process.

The situation was different in Kaldor’s model (1957). This model assumed that the economic system was in a state of full employment of all available resources “in the strictly Keynesian sense – a state of affairs in which the short-run aggregate supply of goods and services is inelastic and unresponsive to further increases in monetary demand” (A Model of Economic Growth, p. 593). Under these conditions, the possibility of increasing consumption by supplying more labour was ruled out and, consequently, any functional relationship between the level of output and the real wage disappeared. Given full employment – and hence a fixed level of output in the short run – the distribution of income will be determined, for given marginal propensities to save of workers and capitalists, by the level of investment. An increase in the level of investment will in fact result in the formation of additional savings and, given the higher marginal propensity to save of capitalists, will raise the share of profits in national income. The same level of output is therefore compatible with a variety of distributional configurations.

This principle was subsequently radicalised by the final general equilibrium model analysed by Graziani, namely Sraffa’s scheme. By demonstrating the compatibility of a single net product with the most diverse distributional configurations, Sraffa provides, for the first time, “a rigorous formulation” of the “possibility of altering the distribution of the social product without altering the production process, a notion cherished by Stuart Mill, and which drew so much harsh criticism upon him” (p. 95). This result is achieved through a model which, unlike von Neumann’s, remains “rigorously disaggregated” (p. 94), albeit at the cost of abandoning any claim to develop a dynamic perspective, leaving the employment of the surplus entirely undetermined.

While Graziani rightly points out the incompatibility of Sraffa’s model – where any abstraction from change, not only in the proportions between commodities but also in the volume of production, is deliberately made on methodological grounds – with a dynamic perspective, he nevertheless seems to overlook the fact that Sraffa’s demonstration of the compatibility of a given ratio between net product and inputs with multiple distributional configurations ultimately rest on reducing the disaggregated case to the single-commodity case – as in the corn model – where this distributional phenomenon is immediately visible. This reduction, made explicit through the construction of the standard commodity, was implicit in the equivalence between the rate of profit when the real wage is set at zero and the rate of profit that would result if all commodities had the same ratio between the quantity produced net of reintegration and the quantity used as input within the system. This measure thus expressed not only a value ratio valid in a single distributional configuration, but the average ratio in physical terms between net product and the means of production in the real system. It was therefore invariant to changes in the distribution, allowing to show as linear the relation between real wage and the rate of profit.

The opposition between classical economists and marginalists regarding the autonomy of consumption is probably one of the weakest points of Graziani’s position. First, the relationship between the level of wages and population growth – and, in Marx’s view, the formation of the reserve army of labour – seems in any case to be an essential part of the process of accumulation. But it is above all the analysis of accumulation that shows how, from the classical perspective, wages are inevitably an item of cost, and the opposition between consumption and the productive use of resources remains central. The reproduction of the working class through consumption appears, in this perspective, more as a consequence than as the goal of the process of production and accumulation. These considerations in no way detract from two fundamental aspects that Graziani’s argument helps to highlight. Firstly, as Claudio Napoleoni had already observed in La posizione del consumo nella teoria economica [The Position of Consumption in Economic Theory] (1962), the tension between preferences and the structure of production remains an essential aspect of neoclassical attempts to build a general equilibrium model. Secondly, it remains doubtful whether the analogy between the neoclassical labour market – where the labour supply relies on workers’ preferences – and the classical economists’ and von Neumann’s framework, where labour is reproducible provided that the means of subsistence are produced, can be sustained. Nevertheless, the thesis of continuity between the various general equilibrium models remains fruitful, as does the view that a rigorous formalisation in terms of general equilibrium does not unequivocally allow us to move beyond the neoclassical perspective. This argument has been plausibly developed, among others, by Christian Bidard (2004), who demonstrated the need to reintroduce utility functions in cases of joint production where the number of production methods does not coincide with the number of goods produced. The contrast between the classical and neoclassical approaches must therefore be resolved outside the narrow constraints of simultaneous equations – a conclusion that Graziani himself had already adopted by rejecting, in his review (1962) of Claudio Napoleoni’s Il pensiero economico del 900, the idea that general economic equilibrium represented both the ultimate criterion of epistemic truth in economic science and the approach followed by that science in its most significant and interesting developments. Graziani seems to pursue the same line of argument – a line that would later prove significant in his rehabilitation of the Marxian theory of value as independent of the question of price determination (1986) – in the conclusion to his 1965 essay.

 

Dynamics beyond equilibrium

In the light of Kaldor and Sraffa’s analysis, Graziani concluded that the rehabilitation of the classical perspective did not provide a way out of the dilemma between extending the model to dynamics and maintaining a disaggregated framework. Both alternatives appeared unsatisfactory, leaving the impression that economic theory, “in its attempt to achieve a coherent representation of development processes, has set itself on a dead-end path” (p. 97). Moreover, the very hypothesis of macroeconomic equilibrium appeared to be a highly restrictive approach, since the assumption of constant prices could be maintained only by abstracting entirely from technological innovation and intersectoral competition. Changes in production methods brought about by technical progress would inevitably disrupt the conditions required for equilibrium. Referring to the work of Pasinetti and Spaventa (1960) and Pedone (1962), Graziani therefore emphasised the need to abandon the fixed proportionality among the various sectors of production.

This approach, which requires abandoning the assumption of uniform growth rates and rates of return across sectors, represents the only way forward through the development of a formal analysis capable of incorporating the fact that “every process of development takes place, and can take place only, outside of equilibrium” (p. 97). Graziani observes that this implies nothing less than the abandonment of the very idea of competitive equilibrium. Such a move entails not only renouncing “a specific form of market” (p. 97) but also abandoning the very idea that economic evolution takes place in accordance with conditions of efficiency, along a path of full employment and optimal growth. Yet this is the only approach compatible with the existence of innovation and technical progress: in a world where the rate of profit were uniform across sectors, “there would be no incentive to transfer resources from one sector to another; the structure of the economy would remain unchanged; and ultimately the process of development – which, for the most part, is precisely identified with a process of gradual structural evolution – would come to a standstill’” (p. 98). Rather than as a “unified block of perfectly interconnected markets perpetually in equilibrium” (p. 98), the economy should be interpreted as a “set of multiple economic systems operating simultaneously and, much like in the international economy, exchanging factors of production and characterised by different rates of development and different rates of return” (p. 98).

While the emphasis on the nexus between dynamics and technological innovation undoubtedly reflect Schumpeter’s influence (see, for example, Costabile 2015 and Bellofiore 2019), the direction Graziani appears to take here does not seem to be entirely consistent with that of the Viennese economist, whose theory of money and credit would in any case prove fundamental to the subsequent development of the theory of the monetary circuit. For Schumpeter, equilibrium and disequilibrium represented complementary analytical dimensions: the latter always originated in a disruption of an initial equilibrium and was ultimately exhausted in the transition to a new one. Equilibrium therefore retained considerable analytical value, both in the study of interdependent relationships within a stationary context and in the comparison between successive stages of economic evolution. The disruption of equilibrium – which Graziani regards here as lacking any rationale – had in Schumpeter a justification in his theory of the relationship between credit and innovation.

On the contrary, Equilibrio generale ed equilibrio macroeconomico reveals a profound dissatisfaction with the distinction between a rigorously tractable but static moment and an evolutionary process that seems to resist formalisation. Graziani argues that disequilibrium and evolution should be regarded as the normal state of the economic system, and as such deserving of analysis in formal terms. The reference to Pasinetti – who would later propose analysing the economic system by breaking it down into vertically integrated subsystems, each characterised by its own rate of growth and rate of profit – therefore appears far from coincidental, and not merely because of the dissatisfaction they shared with aggregate models. At a time when the limitations of general economic equilibrium are widely recognised, the aspiration to place the dynamics of economic evolution as the centre of economic theory remains a path that has yet to be fully explored.