“Quelli che ti spiegano le tue idee senza fartele capire …”
1 – Introduction
1“Modern classical political economy” is a field of research that has flowed directly from the analytical framework proposed by Sraffa (1960) and from the works of certain scholars who gravitated around Cambridge (UK) between the sixties and the eighties. It is well known that Cambridge (UK) has been a melting pot of innovatory views on fundamental issues in political economy for over 40 years. The first major strike against the predominant position came from the principle of effective demand and the general theory of employment, interest and money, formulated by Keynes in (1936). A second strike was the “capital critique” originated by Sraffa’s (1960) framework of production of commodities by means of commodities. Together with this critique, Sraffa’s book made it possible to reappraise classical political economy with a particular focus on issues concerning production and income distribution. Sraffa’s book, together with the edition of Ricardo’s opera omnia, fascinated many of his Cantabrigian colleagues, but only the younger scholars were able to grasp the real significance of his contribution. Pierangelo Garegnani and Luigi Pasinetti were two of these scholars, perhaps the first two who engaged in a deep investigation of the analytical framework proposed by Sraffa and its theoretical implications. Garegnani and Pasinetti have always worked independently of one another, and both are at the head of two rich lines of scholars who have dedicated their research to this new field.  The present paper aims to observe, in a comparative way, how Sraffa’s research programme has been carried out by these two leading figures among Sraffa scholars. It is undeniable that Garegnani and Pasinetti are clearly rooted in classical political economy and in Sraffa’s approach. Yet their divarication is clear: Garegnani focused on the delimitation of the logical structure of classical political economy, while Pasinetti’s focus is the analysis of structural change. Garegnani tried to connect the bare analytical framework proposed by Sraffa to the actual phenomena it can explain and the conditions implicitly required to make this effort fruitful. He described how classical analysis can be separated into two “stages” of analysis, with a particular focus on the determination of the prices of commodities and their relations with the distributive variables. To achieve these goals, Sraffa and Garegnani emphasize the necessity to consider the quantities produced as given in this stage of analysis. Moreover, they refer this analysis to a situation characterized by a uniform rate of profit. Garegnani justifies this uniformity as a result of a “gravitation” process induced by the phenomenon of capital mobility in searching for the highest return; hence he explicitly refers the entire analysis to a capitalist system. Pasinetti, on the other hand, managed to extend Sraffa’s framework to analyse the processes of structural changes of actual economic systems. To this end he developed a multisectoral system where population, technology and final consumption evolve over time in a sectorally differentiated way. Consequently, the quantities produce change as time goes by; moreover, he refers his analysis to a pre-institutional stage, where an entire structure of differentiated sectoral rates of profits is determined to meet the sectoral necessities of accumulation of productive capacity in relation to the evolution of the final demand of each commodity.
2It will be argued in this work that these research projects are not incompatible with one another; they are in fact complementary. We will see how the structural change model is built upon, and may be regarded as, an extension of the “core” equations of the surplus approach: in particular, in order to determine a structure of rates of profits connected to the evolution of final demand of commodities, Pasinetti needs to avoid any biunique (mechanical) co-determination—typical of neoclassical analysis—between prices and distributive variables on the one hand, and quantities on the other. This will be possible simply because the evolution of quantities is the object of a separate determination with respect to the relations between prices and distributive variables. In fact, when Pasinetti formulates the laws of evolution of the parameters of the system, he follows a route which is parallel (though not equivalent) to that followed by Sraffa and Garegnani.
3In this work I do not intend to provide the correct interpretation of the two authors considered, nor of their approaches. Rather, I will present, in a constructive way, what I have learned from their approaches, and what I consider a common and solid basis for grounding an economic investigation of our contemporary systems. Hence, I will be more inclined towards uncovering connections and indicating their possible integration than emphasizing the differences—which I do not deny—in their approaches.
2 – The logical framework of classical surplus theories
4For expository reasons, it is convenient to start from the reconstruction of the logical framework of classical surplus theories provided by Pierangelo Garegnani (see, in particular, Garegnani, 1984 and 2007).
5The main focus of Classical Political Economy was the determination of the size of the social surplus and of its distribution. The peculiar feature of these theories is the view that the shares of the product other than wages are determined residually. This means that once the replacements of the means of production employed and wages (pre-determined on the basis of an institutional mechanism) are deducted from the social product, what remains, i.e. the residuum, goes to profits and rents. In other terms, profits and rents arise because wages do not absorb the entire net product. This view clearly delineates a society where capitalists and land-owners have a prominent position in the distributive process. This principle can be expressed by the following equation:
7where Π are profits, R are rents, X is the social product, A is the replacement of the means of production and W are wages. There is a fundamental logical requirement in order to give relation (1) a theoretical meaning, i.e. to interpret it as an equation and not just as an accounting identity: all the magnitudes on the r.h.s. of (1) must be considered as given in the stage of analysis where we study the determination of the magnitude of the shares other than wages. In other terms, X, A and W must be “intermediate data,” to use a term quite recently introduced by Garegnani (2007).
8It is well known that the satisfaction of this logical requirement represented an Achilles’ heel for Smith’s surplus theory of profits, which Ricardo perceived very clearly but was unable to solve satisfactorily. The difficulties arise from the fact that X, A and W are values of aggregates with different compositions. The determination of the prices of the commodities entering them should thus be prior to the determination of the shares other than wages, but the determination of “natural” prices for Smith requires the knowledge of wages, profits and rents. Sraffa’s (1960) framework provides a solution to these difficulties, and Garegnani (1984) carefully defined the logical requirements of this determination. It is useful to briefly recall the main steps. Consider a system where commodities are produced by themselves and by labour, all capital is circulating and there is no joint production. Free competition ensures a tendency of prices to cover wages and gross profits (we are ignoring rents for the sake of simplicity). The prices of commodities must thus satisfy the following equations
10where p is the price vector, x̂ is the diagonal matrix of gross output of the various industries, A = [aci] is a square matrix where aci is the quantity of commodity c annually employed by industry i, where c, i = 1, 2, …, C = I (processes are represented on the rows), q is the vector of the gross rental prices (they include depreciation) of the various commodities used as capital goods, w is the wage rate and ℓ is the vector of the annual quantities of labour employed in each industry. This formulation, which recalls Smith’s notion of natural prices,  is still incomplete, as it overlooks the elementary fact that it is impossible to fix all distributive variables, q and w, independently of one another.  This inconsistency, typical of an adding-up theory of prices for which the claim of each class can be accommodated by a suitable variation of prices, is eliminated once it is recognized that in normal conditions, gross rental prices of capital goods, q, are linked to the prices of production of these goods, p, by the relation 
12where π is the uniform rate of return or of profit. After replacing equation (3) into equation (2) we obtain the usual formulation of the price system:
14It should be recognized that price equations (2) plus equations (3) or, their equivalent, system (4), is common to both classical and long-period neoclassical approaches.  What really differentiates them are the forces which regulate income distribution: the relative scarcity of factors in the neoclassical approach—expressed by the supply and demand curves which co-determine the prices of commodities and the distributive variables—or social and historical (i.e. institutional) factors in the classical approach.
15It is well known that the neoclassical determination of income distribution is affected by the logical difficulties connected with the notion of “quantity of capital,” both on the demand and the supply side.  In order not to undermine the validity of Sraffa’s logical construction overall, following the classical tradition, one must consider the quantities produced, x, the quantities of commodities employed as means of production, A, and the quantities of labour ℓ as given when writing price equations (4). In the Preface of his book Sraffa warns:
[a]nyone accustomed to think in terms of the equilibrium of demand and supply may be inclined, on reading these pages, to suppose that the argument rests on a tacit assumption of constant returns in all industries. If such a supposition is found helpful, there is no harm in the reader’s adopting it as a temporary working hypothesis. In fact, however, no such assumption is made. No changes in output and […] no changes in the proportions in which different means of production are used by an industry are considered, so that no question arises as to the variation or the constancy of returns. The investigation is concerned exclusively with such properties of an economic system as do not depend on changes in the scale of production or in the proportions of “factors.”
17In this way, once a numéraire has been chosen, it is possible to deduce from the price equations (4) the relation between the rate of profit and the wage rate. As known, when the Standard commodity is chosen as numéraire this relation takes the simple form
19where R = (1 – λ*)/λ* is the maximum rate of profit and λ* is the dominant eigenvalue of Ax̂-1. Formula (5) depicts clearly the trade-off between profits and wages, which should be impossible to see by looking only at the price equations (2); in other terms, it displays very clearly the residual character of one distributive variable with respect to another. With a different numéraire, the relation between π and w takes a more complicated form, but it also shows the trade-off between profits and wages. Relation (5) (or the analogous relation between π and w entailed by the chosen numéraire) shows that income distribution must be determined outside the price equations, that is, outside the sphere of production.
20While the attitude of marginalists has been one of searching for this determination in the “factors market,” the attitude of classical economists (both “old” and modern) has been that of searching for it in the “institutional sphere.” This point has not always been clear in the literature about Sraffa’s work. One recurrent point was the idea that Sraffa’s price equations represented just one side of the economic relation, the supply side, and that they needed a set of demand equations to close the model: for example, Samuelson writes:
[m]y fundamental point, let it now be clear, was that Piero Sraffa sought to have but one leg to stand on. Competitive prices, everyone now knows, must stand squarely on the two legs of (1) tastes, desires, needs and distribution of endowments (in short, on consumer-demand factors), and (2) technology and production costs. At one time or another, Adam Smith (very briefly), David Ricardo, and Frank Knight (briefly), have tried to concentrate on subcases of reality where competitive prices (price ratios, and goods prices relative to factor prices) can be determined autonomously in terms of technology and costs alone: the one-leg case. What is consistent throughout the lifeline of Piero Sraffa—in 1925, 1926, between 1926 and 1930, in 1951 and 1960—is the attempt to emphasise the singular cases in which the theory of value happens to be dependent only on technology and costs independently of the composition of demand.
22While Joan Robinson writes:
We are concerned with equilibrium prices and a rate of profit uniform throughout the economy, but we are given only half of an equilibrium system to stand on. We need a fence to prevent us plunging off into the abyss.
24In order to set up Sraffa’s theoretical framework within the realm of classical political economy, and to outline the main characteristics of this approach, it is useful to adopt a device proposed by Pierangelo Garegnani, who enucleated the “core” of this theory, which consists in a subset of relations (equations) that, given the value of some economic magnitudes which are provisionally considered as independent variables (also called “intermediate data”), determine the remaining variables as dependent or endogenous variables.
25For “old” classical economists (Smith, Ricardo and Marx) the intermediate data are:
- the social product, x,
- the real wage rate, wT = [w1, …, wC], i.e. a bundle of commodities,
- the technology of the system, A;
26the relations belonging to the “core” determine the following dependent variables:
- labour employment, E* = ℓTx,
- the shares other than wages, that is, profits—here rents are not considered— whose rate is given by  Π = (1 – λM)/λM,
- the price system, pT = p*T,
27where λM and p*T are the dominant eigenvalue and the corresponding eigenvector of the socio-technical matrix A + wℓT.
28The attitude of “modern” classical economists, since Sraffa (1960, § 44), is that of considering the rate of profit as the independent distributive variable. Hence, in this case the intermediate data are
- the social product, x,
- the rate of profit, π,
- the technology of the system, A;
29the relations belonging to the “core” determine the following dependent variables:
- labour employment, E* = ℓTx,
- the wage rate, w* = 1 – π/R,
- the price system, p*T = w*[I – (1 + π)A]-1.
30This distinction makes clear that the relations of the “core” of the system can be adequately expressed by “necessary quantitative relations” (Garegnani, 2007, p. 186), i.e. by equations and formal relations. On the contrary, the magnitudes which are taken as given in the core, i.e. the intermediate data, are determined by forces that are less susceptible to be represented by means of formal relations.
31Obviously, no one in the surplus approach denies that there are influences and feedback relations among the various intermediate data, and between the dependent variables on the one hand and the intermediate data on the other. But these influences do not have the same level of generality and unambiguousness: they may change significantly according to the institutional circumstances and may go in both directions, partially compensating one another. These relations are thus better studied outside the “core” of the system, and in a partially different way: other disciplines, like political or social sciences, economic history, etc., can here usefully support the economic investigation.  For example, it is obvious that relative prices affect the composition of final output by affecting the demand of the various commodities.  But these effects may change according to the historical or social circumstances; moreover, they are not univocal, and can partially compensate each other. For all these reasons, it is preferable to consider the intermediate data (i), (ii) and (iii) of old classical economy or (I), (II) and (III) of modern classical political economy as determined separately from the endogenous variables of the core. Formally this is realized by considering the intermediate data as given when studying the forces that determine the endogenous variables of the “core”: (a), (b) and (c) or (A), (B) and (C).
32There are cases where some specific institutional problems could be analysed in a formal way: for example the study of gravitation of market prices around production prices, or the study of the accumulation process in Ricardian frameworks. But, again, the formalization concerns the specific problem at hand, not the working of the entire economic system. Hence, in these stages of analysis, other variables are kept frozen at some given levels: for example, the wage rate is taken as given when one studies how market prices gravitate around natural prices in consequence of capital mobility (see, for example, Boggio, 1990, p. 48; Duménil and Lévy, 1993, chapter 5; Garegnani, 1990, p. 333); again, the level of the wage rate is taken as given when one studies how profits are accumulated and new plots of land are cultivated (see Bellino, 2014, in particular § 3, and Kurz and Salvadori, 2006, pp. 110-11).
33This attitude—of analyzing the working of an economic system in different stages, each for a specific problem or situation, sometimes in formal terms, in other cases by using the instruments of social, historical and institutional analysis—is in sharp contrast with the attitude followed by neoclassical economists, which normally refer to a general model. 
3 – Structural economic dynamics
34More than any other, Luigi Pasinetti is probably the scholar who succeeded in building a bridge between Keynesian themes or subjects and the modern Classical reappraisal led by Sraffa. Pasinetti’s model of structural change is in fact grounded on Keynes’s principle of effective demand and on the classical theory of distribution and value. He originally presented this model in his PhD dissertation in 1963, and then published it in various versions (1965, 1981 and 1993).
35His analysis has a twofold objective:
- to study the consequences of structural change of technology and of final demand on output, value and employment;
- to study the conditions that have to be satisfied in order to accomplish the potential of the system concerning growth, employment and the satisfaction of final wants.
36In brief, we could say that objective i) is descriptive, while objective ii) is normative. To handle objective i) Pasinetti starts from Sraffa’s price equations and extends them to the case of an economic system undergoing a process of structural change, i.e. a change in the proportions of the various industries. Hence, some of the magnitudes that have been kept as given in the price equations here must be left free to change. But, as we will now see, this extension is done in line with the methodological requirements imposed by the logical structure of surplus theories.
37The price equations considered by Pasinetti are
39We are considering an economic system with I single product industries. Thus, 1 unit of commodity i requires ℓi units of labour and 1 unit of a capital good, specific to the commodity, i = 1, …, I (for this reason, we will call “capital good i” the capital good employed in the production of commodity i);  in each production period a constant proportion, δi, of capital good i wears out; 1 unit of capital good i is produced by λi units of labour. Let ci be the units of final good ci required by each individual as final consumption; let ji be the individual demand for capital good i by the final sector (net investment). Let pi and vi be the prices of commodity c and of its specific capital good. System (6) contains 2I + 1 equations in 2I + 2 unknowns: p1, …, pI, v1, …, vI, w and π. The first 2I equations can be written as:
41After having chosen the numéraire, there remains one degree of freedom which expresses, as usual, the fact that income distribution is determined outside the price equations. For example, if we chose commodity ‘1’ as numéraire, i.e. if we set p1 = 1, by substituting the first equation of (p) we obtain 1 = p1 = (δ1 + π)wλ1 + wℓ1, which engenders the following inverse relation between the rate of profit and the wage rate expressed in terms of commodity ‘1’:
43There remains, however, a further equation in system (6), the last one,
45which expresses the condition that all incomes (wages + profits) must be entirely spent.
46In addition to the price system, we have a quantity system. Let xi and ki be the quantities produced of final good i, and of its productive capacity; let xN be the quantity of labour employed in all production activities. The quantity equations are
48System (7) contains 2I + 1 equations in 2I + 1 unknowns: x1, …, xI, k1, …, kI, xN. It is a homogeneous system; once the condition to ensure non-trivial solutions is satisfied (see Pasinetti, 1981, ch. II, § 3), we have one degree of freedom; the quantity which is more apt to be fixed from outside is the employment level, xN; if we want to guarantee full employment we fix
50where N is the amount of the labour force.
51The last equation of system (7) is
53Equations (6N) and (7N) together express the Keynesian principle of effective demand: wages and profits must be entirely spent in order to ensure that labour requirements employ the available amount of labour force. As we will see in brief, this condition is never automatically satisfied.
54Pasinetti assumes that all magnitudes which enter as “parameters” in the description of the industrial system of the previous section are now allowed to change. In particular, Pasinetti supposes that technical coefficients, final demand coefficients and population vary according to the following exponential functions:
56There remains a set of coefficients in the quantity equations (7) whose dynamics has not yet been specified: they are the coefficients of net investment in capital good c, kcN. Consistent with his normative attitude, Pasinetti fixes them in such a way that the productive capacity of each commodity increases in line with the evolution of final demand for that commodity. This amounts to imposing the following:
58(see Pasinetti, 1981, ch. V, § 4).
59The dynamics of parameters envisaged by equations (9) and entails a structural change for the endogenous variables of the system considered: for prices, for sectoral output, for sectoral employment, as well as a macro-dynamic for the aggregate level of employment. This is a relevant result for a growth model (for details see Pasinetti, 1981, ch. V).
60The second objective pursued by Pasinetti in his investigation derives from interpreting the equilibrium conditions simply as relations describing an ideal (efficient) situation, where the “potential” of the economic system concerning growth, employment and satisfaction of final wants is best obtained independently of the study of the forces needed to implement these conditions in an actual system. This allows him to “separate” the level of analysis where the various conditions are described and which must be satisfied in the ideal or “natural” configuration of the economic system, from the level of analysis where the institutional mechanisms needed to achieve these conditions are described and compared.  We could summarize the relations that must be satisfied in the “natural” system in the five following points:
- (N1) a price system which guarantees the reproducibility of the various commodities, identified by the first 2I equations of system (6);
- (N2) a set of output levels which satisfy the final demand of each commodity, identified by the first 2I equations of system (7);
- (N3) the “macro–economic condition,” i.e., equations (6N), (7N) and (8), which together guarantee full employment of the labour force;
- (N4) an income distribution configuration which guarantees the growth of the productive capacity in each sector in line with the growth of the final demand of the respective commodity. Pasinetti envisages the fulfilment of this goal by a set of differentiated rates of profit, called “natural rates of profit,” determined by the total rate of growth of the final demand of the correspondent commodity (πi = n + ri, i = 1, …, I). It is possible to prove that in this case the price of each commodity becomes proportional to the quantity of labour which is necessary i) to reproduce the commodity, ii) to reproduce its means of production and iii) to expand these means according to the growth rate of final demand of that commodity. A renewed form for the theory of labour value thus takes shape in this case. Consequently, each individual receives a fraction of the net product equal to the proportion of the quantity of labour he contributes with respect to the total labour of the system (“labour principle” of income distribution);
- (N5) a natural rate of interest, which guarantees that the debt and credit relationships among individuals do not distort the income distribution process from its “labour principle.”
61The connection of the “natural” relations with actual systems is seen by Pasinetti as an “institutional” problem. In particular, as regards the achievement of goals listed above as (N1)-(N5), we can observe that goals (N1) and (N2) are normally fulfilled in capitalist systems (for details, see Bellino, 2011): free competition ensures that prices tend towards their normal levels while the Keynesian principle of effective demand aligns the output of each commodity to its demand.  On the contrary, goal (N3) is not automatic in modern capitalist economies; moreover, the phenomenon of structural change is for Pasinetti a force that contrasts with the fulfilment of this objective; it can be pursued by means of several policies, according to the specific situation of the system (introduction of new goods, search for new markets, public investments, monetary policies, etc). Goals (N4) and (N5) are probably never feasible, or only partially feasible, in capitalist systems; nevertheless, they may represent some sort of benchmark levels, to evaluate how far an actual system is from its “natural” configuration. 
4 – Two directions, one framework
62To carry out our comparison between the surplus approach and the structural dynamics research programmes, it is convenient to start with two (quite long) quotations from our two authors. Garegnani writes:
[W]hy take as given some magnitudes that the theory has also to determine, and are therefore ultimately in the nature of unknowns? We shall see […] below how this method of “intermediate data” has its basis in the distinction, implicit in the application of the notion of surplus to a market economy, between two fields of inquiry and the corresponding different methods of analysis. On the one hand, we have the necessary quantitative relations, which competition entails between commodity prices and distributive variables and, which, in their comparative simplicity, are of a nature allowing for a mainly deductive treatment. On the other, we have the circumstances determining what we have described as the “intermediate data”: the subsistence or, more generally, the wage, the outputs, the technical conditions of production. These circumstances were seen to be closely related to institutional and historical factors, which, because of their complexity and variability according to circumstances, prevented deducing the corresponding variables from a few basic principles as was possible for prices and profits in the “core.” Those intermediate data rather required, for their study, methods of a more inductive kind. This distinction, concerning both contents and methods, which underlies the notion of surplus, appears to be what has entailed the separation between the two fields of analysis and the corresponding logical construct of the “intermediate data.”
In the previous chapters, the analysis has been concentrated on the theoretical scheme of what has been called the “natural” economic system of a simple […] production economy. The natural economic system represents so to speak the framework skeleton of the present theoretical construction. It is a set of relations that possess characteristics of analytical relevance and logical consistency, with strong normative properties.
But the natural economic system does not come into existence automatically. For any actual economic system, the problem arises of inventing and setting up those organizational devices – in other words those “institutions” – which put into motion processes actually able to bring the natural economic system into existence. […] The institutional problem does not need to have a unique solution, nor does it emerge once for all. By being a problem of construction of organizational devices (the institutions) in order to achieve certain results (the natural economic system), it is obviously susceptible of being faced in different ways, from place to place, from time to time, and at the variation of many external circumstances, without mentioning that the organizational field is itself subject to continuous evolution and innovation.
Moreover, an economic system does not come about in a vacuum. It presupposes a complex network of political, juridical, and legal institutions. These institutions may have been shaped through different historical processes or according to different traditions in different countries, sometimes with even stronger requirements than those behind economic institutions. With this wider institutional framework, the economic institutions must merge and intermingle, while carrying out the task entrusted to them.
66From the above quotations, we can observe in both authors a “separation” of the analysis into a more restricted theoretical level of relations (the “core” and the “natural system”), and a second level which includes in both cases the institutional level. This “separation” does not overlap for the two authors. Pasinetti’s natural system is a theoretical construction that includes the essential relations among sectoral outputs, relative prices, employment and wages, profits and interest, that must be satisfied in an economic system in order to best exploit its potential, independently of the institutional set-up to enforce these relations. Garegnani’s core is a subset of relations among the same set of magnitudes in the case of a specific institutional set-up: capitalism.
67The purpose of each logical construction is different; they are not alternative but complementary to one another. The core is mainly concerned with providing a description of how capitalist economies work: it thus has a positive purpose. The natural system is essentially a normative tool. Yet there are two fundamental characteristics that are common to the approaches:
- the study of the relations among the magnitudes belonging to the “core” as well as those belonging to the natural system are carried out, in both cases, by “necessary quantitative relations” (i.e. equations), while the relations outside of the core as well as those concerning the institutional system are better studied in a separate way, in connection with social, historical, political disciplines;
- all the parameters considered as given by Sraffa and Garegnani (see the quotation from Sraffa, 1960, reproduced here at p. 9) change in Pasinetti’s model. But there is a device that avoids these changes allowing what was left out the front door to re-enter through the back: the co-determination of prices and quantities and, consequently, the return to a “mechanical” determination of income distribution based on supply and demand forces. This is prevented by the choice to assume that the rates of change of labour inputs, of final demand and of population (n, ρi, ρji and ri) are independent of any other endogenous variable of the model. This formal independence is just a way to preserve the theoretical choice to provide a separate determination of some variables, which appear more subject to the influence of institutional factors. The methodology of “given quantities” and “given technology” is thus now extended to the case of an evolving economy: final consumption and technical coefficients are assumed to evolve according to a given pattern of evolution. This is a methodological aspect induced by a specific theoretical need. On this point Pasinetti highlights thatThe […] innovative methodological line of research consists in separating sharply the distinction between variables and constants from the distinction between unknowns and data. In traditional economic analysis these two distinctions tend to coincide because of the essentially static approach which is adopted: those magnitudes which are considered as unknowns are also considered as variables, and those magnitudes that are considered as data are also considered as constants. But in a dynamic context, to insist upon this coincidence makes no sense. Or rather, to insist on this coincidence is equivalent to frustrating the purpose of any investigation into dynamics.
68In solving a particular problem there is thus no contradiction in considering some magnitudes as given, even though they may change as time goes by—as in Pasinetti’s structural change model—or may be affected by changes of the endogenous variables they contribute to determine—as in the core of the surplus approach. This kind of feedback can be disregarded when there are reasons to believe that they are non-univocal, non-systematic or non-persistent.
69A common methodology is thus adopted for analysing an economic system from different perspectives: the description of the behaviour of a capitalist economy and the identification of an “ideal” configuration that constitutes a norm to orientate the actual working of system.
5 – Some concluding remarks
70Modern classical analysis is a school which has always appeared compact united in its critical attitude towards neoclassical economics. On the positive side, while at first glance the positions may appear quite un-homogeneous and discordant, this is primarily a direct consequence of the authors having developed their own research independently of one another. Nevertheless, by comparing their contributions, one sees substantial affinities regarding the methodology followed in order to keep the analytical framework coherent with the common classical matrix (the separation of an analytical subset of purely economic relations from the broader set of “institutional” relations is what prevents classical theory from getting lost within the difficulties connected with the notion of capital affecting the supply and demand approach). At this point, it is instructive to reproduce the list of the nine “quite clear characteristic features” that Pasinetti attributes to that group of scholars who became known as the “Cambridge Keynesian School” (see Pasinetti, 2007, 217-37).
- Reality (and not simply abstract rationality) as the starting point of economic theory.
- Economic logic with internal consistency (and not only formal rigour).
- Malthus and the Classics (not Walras and the Marginalists) as the major inspiring source in the history of economic thought.
- Non-ergodic (in place of stationary, timeless) economic systems.
- Causality vs. interdependence.
- Macroeconomics before microeconomics.
- Disequilibrium and instability (not equilibrium) as the normal state of the industrial economies.
- Necessity of finding an appropriate analytical framework for dealing with technical change and economic growth.
71A strong, deeply felt social concern.
72It will now be shown how the overwhelming majority of the characteristics described above are also deeply shared by the group of scholars led by Pierangelo Garegnani, who developed and systematized the modern version of the surplus approach.
731. Reality: the pure logic which supports Sraffa’s framework is never conceived as an exercise for its own sake; Sraffa’s idiosyncratic opposition to mathematical formalism was known to all of his scholars.  Additionally, the interpretation of Sraffa prices as the “normal positions” of the economy is a clear attempt to establish a correspondence between theoretical and observable variables (on this, see Garegnani, 2007).
742. Internal consistency. Self-evident!
753. Classics, not the Marginalists. Self-evident!
764. Non-ergodicity. The interpretation of Sraffa’s framework as a “steady state” has been rejected on several occasions. For example, normal positions around which actual (marker) prices “gravitate” are characterized by the prevalence of a uniform rate of profit: this does not exclude the possibility of conceiving a dynamics of normal positions (on this see, for example, Cesaratto, 1995).
775. Causality. It has been shown throughout this paper how specific features of the surplus approach have been obtained by a suitable choice of asymmetrical links among the relevant magnitudes. For further details on this point see Bellino-Nerozzi (2014).
786. Macroeconomics before microeconomics. The distinction between macro- and micro-economics is extraneous to the classical tradition. Pasinetti’s intention, however, is that of recalling that
[t]he Cambridge economists caught very clearly the principle that the behaviour of the economic system as a whole is not reducible to, in the sense that it does not emerge as the exclusive result of, the sum of its single individual parts […] [t]here are many examples of fallacy of composition that the Cambridge School have highlighted, as against the attempts to extend what is true for the single individual to the behaviour of the economic system as a whole.
80The works of Garegnani and of other surplus theorists evidently fulfil this characteristic.
817-8. Disequilibrium and instability + technical change and economic growth. These are, probably, two points where the methodology of modern surplus theorists diverges, but—I argue—only partially, from that of the “Cambridge Keynesian School.” The preference of surplus theorists for using the Marshallian “short chains of reasoning,” to investigate “economic change step by step,” is undeniable (Cesaratto, 1995, p. 274). Pasinetti, on the contrary, on the basis of Frisch’s (1935-36) notion of “moving equilibrium,”  has studied how an equilibrium configuration changes as a consequence of a change in one or more parameters. As we saw in Section, the relevant parameters (population, technology and final demand) are supposed to change as time goes on. The possibility of slipping into a model of full interdependence (like a Walrasian model) where all variables are mechanically determined is avoided by a specification of the dynamics of these parameters on the basis of a given pattern, which replicates, in a reasonable way, long-term historical tendencies—like the supposition of reducing labour input coefficients due to technical change, or an evolution of final demand coefficients like that described by Engel—rather than by a “microfoundation” of the variations of the parameters.
829. Social concern. Self evident!
83From this perspective it can be recognized that the elements at the basis of the surplus approach as described by Garegnani are not contradictory with the analysis of structural change of industrial systems developed by Pasinetti. Clearly, Pasinetti adopts a perspective which in some way is pre-institutional. He is interested in describing the conditions that an industrial system, undergoing a process of structural change, must satisfy in order to accomplish its potential concerning growth, employment and the satisfaction of final wants. Furthermore, the logical separation, emphasized by Garegnani, of the analysis of prices and distribution from that of the quantities produced is preserved in Pasinetti’s analysis in the choice to formulate the evolution of final demand (and of technology) independently of the price system.
84We thus do not have two (or more) ways to develop Sraffa’s framework, but a unified “core” to describe the economic workings of capitalist societies and to orientate the system towards its ideal “natural” configuration.
85Recently, Pasinetti has expressed his worry about a certain negative attitude developed by the members of Cambridge school—which he called the “Cambridge prima donna syndrome”—that sometimes leads to disregarding or ignoring the works of the other members of the group in favour of emphasizing the peculiarities of their own contributions (see Pasinetti, 2007, p. 46, fn. 18).
86Without wanting to disregard the differences and specificities of the two approaches here presented, I hope the present work may assist in the convergence of the research undertaken in this school towards a common modern classical-Keynesian school.
The relations with Keynes’s analysis
87The present paper represents a good occasion to hint at a further issue, which has often seemed to be an obstacle to integrating the approaches here discussed: the rôle given to Keynes’s analysis. On the one hand the Keynesian principle of effective demand is recognized by all members of the modern classical approach. Pasinetti devoted an entire essay to explaining it in its “pure” version, i.e. not contaminated by the Walrasian interpretation contained in the neoclassical synthesis (see Pasinetti, 1974, Essay II), while Garegnani used it as a base for his theoretical and empirical research concerning the Italian post-war system (see Garegnani, 1962 ). There are however two aspects that see the surplus approach scholars on one side of the debate and the post-Keynesian growth theorists on the other.
88The first aspect concerns the theory behind Keynes’s marginal efficiency of capital curve. Garegnani underlines how this curve is deeply rooted in the theory of marginal productivity:
[h]owever, the price which Keynes has to pay for the traditional strand in his thought becomes clear with respect to the schedule of the marginal efficiency of capital.
90This link, he says, has significantly reduced the importance of Keynes’s critique to the traditional theory of employment:
[t]he critique of the traditional theory of interest becomes then the key to an acceptance of Keynes’s arguments—and the concept of the marginal efficiency of capital proves to be the Achilles’ heel of that very critique.
92Most likely, the critique would have been more effective if it had been paired with one of the fundamental results of the capital debates of the sixties, i.e. the non-existence, in general, of a monotonic and inverse relation between the rate of profit and the capital-labour ratio. This result would have contributed to disproving the misleading idea that Keynes’s results were essentially due to rigidities and that a suitable flexibility of price factors (wages and profits) is sufficient to restore full employment.
93Pasinetti, on the other hand, is more of a possibilist, and maintains that:
the marginal-efficiency-of-capital schedule, which might, at a first superficial look, appear as belonging to the marginal economic analysis, when examined more deeply turns out to have a rather different origin. Keynes’ ranking of all investment projects in a decreasing order of profitability is more akin to Ricardo’s ranking of all lands in a decreasing order of fertility than to any marginal economic elaboration. And in any case, there is absolutely no need to consider Keynes’ marginal-efficiency-of-capital schedule as an expression of the marginal productivity theory of capital.
95More recently, Pasinetti returned to the issue and specified his position:
Keynes was not able, or was not in time, to take advantage of Sraffa’s ongoing critical elaborations. But we are in a position now to state the results of the critique of the neoclassical production function, which would have been needed to debunk the demand-for-investment side of the orthodox theory. […] The conclusions [of the reswitching of techniques controversy] are strictly logical and devastating. The downward-sloping investment-demand function, to the extent that it relies on a continuous process of substitution of capital for labour, as the rate of interest falls, is theoretically unsound; it has no logical foundations. […] The “reswitching” result only means that, if such a downward-sloping relation exists, it cannot be explained by a process of substitution of capital for labour (i.e. by a neoclassical production function); it cannot be explained by more and more capital-intensive techniques as the rate of interest falls. Such a relation, if it exists, must be explained by something else – by some other theory or circumstance.
It is to this effect that we must logically search for a meaning (non-orthodox meaning) of Keynes’s notion of the “marginal efficiency of capital.”
97As we can see, the interpretative disagreement on this point does not undermine the acceptance of the principle of effective demand.
98The second disagreement concerns the result entailed by the Cambridge equation; the critical position was expressed in particular by Pierangelo Garegnani, Ferdinando Vianello and others, who argued against the compulsory negative relationship that the Cambridge equation establishes between the growth rate and the real wage rate. They aim to break the idea that a higher growth rate entails a lower real wage. An explicit statement of this idea can be found, for example, in Kaldor:
[t]he theory thus serves to explain the long-observed fact […] that distributive shares are constant over long periods whilst they fluctuate over shorter periods […] as well as the fact that in fast-growing economies the share of profits is generally appreciably greater than in economies which grow at a relatively slow rate.
100Marglin is still more explicit:
[i]n the short run, fluctuations in investment demand are reflected in fluctuations in output; the rate of capacity utilization changes in accordance with aggregate demand. The distributional conflict between capitalists and workers is, as it were, a non-zero-sum game. […] But in the long run, the period with which neo-Keynesian analysis concerns itself, there is no excess capacity to accommodate investment demand. Distribution must bear the brunt of adjusting aggregate demand to supply. In contrast with the short period, the long-run conflict is a zero-sum game—at least in the absence of technological substitution or technological change.
102The surplus approach school envisages the additional resources for accumulation in the variations of the rate of capacity utilization: this is possible because the productive capacity of firms is never fully utilized, even in the long run (entrepreneurs prefer to leave a margin of available productive capacity, to face unexpected peaks in demand). Hence, a higher rate of accumulation does not require higher profits (and lower wages) as long as the degree of utilization of productive capacity can be increased: the emphasis is thus placed on the forces of demand for their activating power in creating new and permanent increments of income along purely Keynesian lines (see Vianello, 1985, 1996 and Garegnani, 1992).
103Undoubtedly this point of view is in opposition to the one expressed by Kaldor and, mostly, by Pasinetti. Nevertheless, there is no apparent barrier to integrating the possibility of varying the degree of utilization of productive capacity in the post-Keynesian theories of income distribution. Moreover, the normative meaning given by Pasinetti to the Cambridge equation allows one to read it along two perspectives: i) it sets a minimal level under which the profit rate cannot fall, i.e. a maximum level that the wage rate cannot exceed if the economic system has to grow at a given rate; but, at the same time, ii) it identifies a reference level for the rate of profit: it identifies a threshold level to evaluate when a rate of profit is no longer justifiable on the basis of the accumulation needs of the system.
Università Cattolica del Sacro Cuore, Milano; email@example.com
A detailed analysis of these classifications is beyond the purpose of this essay. Recall, though, that the term ‘post-Keynesian’ is normally used to refer to the group of scholars directly connected with Keynes, including Richard Kahn, Nicholas Kaldor, Michael Kalecki and Joan Robinson. Luigi Pasinetti belongs, in some ways, both to the post-Keynesians and to the Sraffians: he represents a ‘bridge’ between the two groups. These groups of scholars are identified by various labels, such as Sraffians, modern Classicals, neo-Ricardians, post-Keynesians (although this latter adjective is less appropriate), etc. There are many surveys of the various aspects of these research programmes: for example, Harcourt (1972) reconstructs the first steps of the capital critique, Pasinetti (2007) gives an overview of the main topics debated in light of the personal relations among the original members of the Cambridge school, while Baranzini and Mirante (2015) give an up-to-date review of all the research fields explored by the scholars in this school.
Smith (1776, ch. VII) writes: “[w]hen the price of any commodity is neither more nor less than what is sufficient to pay the rent of the land, the wages of the labour, and the profits of the stock employed in raising, preparing, and bringing it to market, according to their natural rates, the commodity is then sold for what may be called its natural price.”
Smith (1776, ch. VII) writes: “The natural price itself varies with the natural rate of each of its component parts, of wages, profits, and rents.” On this, see Garegnani (1984, § 9) and Sraffa (1951, p. xxxv).
Equations (3) hold for the case where all commodities are circulating capital goods. In the case of fixed capital (with a constant depreciation rate) they should be replaced by , where is the diagonal matrix of the depreciation rates of the various commodities used as capital goods.
Equation system (2) corresponds to the price equations of capital goods of Walras’s system, while conditions (3) are the conditions of uniformity of the rates of return on the supply prices of capital goods; the same conditions, written in (4), are the Sraffa price system.
We are referring to the impossibility of ensuring the inverse monotonicity of the demand of capital with respect to the rate of profit highlighted by the reswitching debate and the impossibility of obtaining a scalar expression of the quantity of capital independent of the rate of profit.
In this case, the price system becomes pT = (1 + π)pT(A + wℓT) (in compliance with old classical economists, wages are supposed to be paid in advance). The ensuing rate of profit is the maximum rate of profit, Π.
As we will see later, Luigi Pasinetti has proposed an analogous, although not coinciding, ‘separation’ between a theoretical stage of analysis (to be developed by the deductive methods of pure economic theory) and an institutional stage (to be developed with the support of other disciplines).
Interestingly, a letter on this issue was sent by Sraffa to Arun Bose (SP, C32/3). I reproduce it here in its entirety.Cambridge,
9th December, 1964
I am sorry to have kept your MS so long – and with so little result.
The fact is that your opening sentence is for me an obstacle which I am unable to get over. You write: “It is a basic proposition of the Sraffa theory that prices are determined exclusively by the physical requirements of production and the social wage-profit division, with consumers demand playing a purely passive role.”
Never have I said this: certainly not in the two places to which you refer in your note 2. Nothing, in my view, could be more suicidal than to make such a statement. You are asking me to put my head on the block so that the first fool who comes along can cut it off neatly.
Whatever you do, please do not represent me as saying such a thing.
This initial and to me quite maddening obstacle has prevented me, in spite of many attempts, from reading understandingly your article. You must find a more detached reader to advise you about it. I am very sorry to seem so unhelpful, but I have spent quite a lot of time upon your work, to no purpose. I do not think that it would be any good keeping it longer, so I now return it to you.
This is probably one of the reasons for the communication difficulties between the two approaches. In principle, however, neoclassical theory also adopts the same methodological choice to consider some (other) magnitudes as given when studying how endogenous variables are determined: in general equilibrium analysis, for example, preferences, endowments, technology and property rights are taken as given when prices and allocations are determined. The obvious links between these groups of variables and data are intentionally not analysed.
Following Pasinetti’s reasoning the capital good used to produce 1 unit of commodity i can be considered a composite commodity, what he calls ‘productive capacity of final good i’. Thus, we can denote, by the same single magnitude, a set of heterogeneous means of production. The advantage of this procedure is that a change of the physical form of productive capacity of a final commodity, induced for example by technical change, can be ultimately reduced to a decrease in quantity of the vertically integrated labour needed to produce the commodity (for further details, see Pasinetti, 1973, § 15).
Moreover, we are here considering the case where capital goods are produced only by labour. The general case, where capital goods are produced by labour and other capital goods, is presented in (Pasinetti, 1981, ch. II, §. 7).
It is interesting to observe that a similar perspective is suggested by Sraffa himself in a note written in 1942:This paper deals with an extremely elementary problem; so elementary indeed that its solution is generally taken for granted. The problem is that of ascertaining the conditions of equilibrium of a system of prices & the rate of profits, independently of the study of the forces which may bring about such a state of equilibrium. Since a solution of the second problem carries with it a solution of the first, that is the course usually adopted in modern theory. The first problem however is susceptible of a more general treatment, independent of the particular forces assumed for the second; & in view of the unsatisfactory character of the latter, there is advantage in maintaining its independence.
It could also be argued that goals (N1) and (N2) are both fulfilled by competition only (without the need to invoke the Keynesian principle of effective demand): the gravitation process envisaged by classical economists presupposes a joint movement of prices and quantities towards their normal position.
An interesting perspective through which to penetrate the meaning of the natural configuration comes from the fact that in it the same commodity normally receives a different price according to the vertically hyper-integrated subsystem which uses it as a means of production. This result, which concerns the general case—not considered here—where capital goods are produced by capital goods and labour, emerges clearly in Pasinetti (1988, §3). The apparent oddness of this result disappears if one recognizes that, while the uniformity of the rate of profit as well as the law of one price are typical of competitive systems, in the natural configuration prices measure the difficulty of reproducing the various commodities. Within the vertically hyper-integrated subsystem, this difficulty will be suitably modified according to the diversified structure of the growth rates of the final demand of each specific commodity. This diversification is at the basis of different prices for the same commodity.
A simple but meaningful example is the way Sraffa handled the case of self-reproducing non-basics having a physical rate of surplus lower than the average rate of the other (basic) commodities. Newman (1962) was inclined to consider this case as symmetrical to the case of a self-reproducing non-basic commodity with a physical rate of surplus higher than average. Sraffa, in his reply, argued that the former situation can be regarded as exceptional on the basis of ‘reality’ arguments (for the entire exchange between Sraffa and Newman see Sraffa, 1970).
I owe this reference to Ariel Wirkierman: Pasinetti quotes Frisch’s paper only in his mathematical formulation of Ricardo’s system (Pasinetti, 1960, p. 84, fn. 3).
The theoretical part of this work has been published in Italian in Garegnani (1964-65).