How To Understand The Core Of A Profit Rate

Transcription

1 This article was downloaded by:[universita' degli Studi Roma Tre] [Universita' degli Studi Roma Tre] On: 14 May 2007 Access Details: [subscription number ] Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: Registered office: Mortimer House, Mortimer Street, London W1T 3JH, UK Review of Political Economy Publication details, including instructions for authors and subscription information: Professor Foley and Classical Policy Analysis To cite this Article:, 'Professor Foley and Classical Policy Analysis', Review of Political Economy, 19:2, To link to this article: DOI: / URL: PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: This article maybe used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. Taylor and Francis 2007

2 Review of Political Economy, Volume 19, Number 2, , April 2007 Professor Foley and Classical Policy Analysis PIERANGELO GAREGNANI Università di Roma Tre, Italy 1. I must first of all say that I welcome Professor Foley s well-informed views and stimulating comments on what I have elsewhere called the theoretical world of the old classical economists. I will here be concerned with three issues among those he raises in his review article, namely: (a) the structure of classical theories and the existence of an analytical core in them; (b) whether in the classical approach to distribution and prices the conception of capital is subject to the same critique raised against the notion of a quantity of capital in neoclassical theories: this will require some clarification of the critique itself; (c) the implications of the classical approach for policy analysis. I. Surplus Principle and Core in the Classical Theories 2. Foley writes as follows on the question of a core in the theories of the old classical economists: The process of competition among capitals that tends to bring about profit rate equalization, can be modeled at a level of mathematical explicitness comparable to the neoclassical general equilibrium analysis; but the fundamental determinants of distribution, either the wage or the profit rate, and of the composition of social output, lie in a different methodological realm, and must be explained on different principles and with different methods. There are things to be said both in favor of and against this methodological position, but it is not at all clear that it would be recognizable to Smith or Ricardo (or Marx). (2001, pp ) My answer to the question raised at the end of the passage, is that the position would be as recognizable to Smith, Ricardo or Marx as would be, for them, the notion of a surplus determination of the non-wage distributive variables i.e. their ultimate determination as the difference between the social product and what is physically Correspondence Address: Pierangelo Garegnani, Centro studi Piero Sraffa, Dipartimento di Economia, Facoltà di Economia, Università di Roma Tre, Via Silvio d Amico III, Roma 00145, Italia. sraffa@uniroma3.it ISSN print/issn online/07/ # 2007 Taylor & Francis DOI: /

3 222 P. Garegnani necessary for its production, in particular real wages, broadly identified with workers subsistence. I think it has not always been made sufficiently clear that the notion of a core in classical theories is but an aspect of that surplus interpretation of their theories (see Garegnani, 2002, pp ) which, undisputed in the case of Quesnay, is also widely if not universally acknowledged for Ricardo (often in the garb of the so called corn model ) as well as for Smith, not to mention Marx who attempted to again give to the surplus determination the direct form it had in Quesnay. Indeed, to determine the non-wage shares of the social product as the difference between the social product and the wages is non-circular only to the extent that product and wages can be taken as given in that determination. And since no economist, least of all Smith and Ricardo, can avoid being concerned with explaining product and wages, their role as givens in determining the surplus could only mean that those magnitudes were conceived as determinable separately from (though not necessarily in a way independently of ) prices and the non-wage distributive variables. This implies that the theory put forth by those economists in effect contained a purely quantitative core in which product and wages appeared as givens and where profits (and, in the appropriate way, land rents) were determined as a residuum together with relative prices. Taking as data in that determination both the product and the wages implied, on the other hand, also taking as data the technical conditions, on which the labour required for that production (and, of course, the replacement of the means of production) depend. We may then say that those three sets of circumstances were intermediate data, separating out from the rest of the theory that core in which the non-wage distributive variables and the entailed relative prices could be determined. 3. But why these intermediate data? Why not go from the circumstances determining them, directly on to the non-wage distributive variables and to relative prices, as happens in neoclassical theory, where tastes and factor endowments determine outputs and wages simultaneously with prices and profits? The answer lies essentially in the kind of circumstances that were seen to determine the division of the product between wages and profits, and the resulting distinction between two fields of analysis requiring different methods of inquiry. In one of them, as Foley recalls in his passage, free competition and the corresponding tendency to a uniform rate of profits, and uniform wages and rents, do establish quantitative relations between those distributive variables and between them and prices which are of a nature to allow for deductive methods in their study. But with respect to outputs and wages as is made clear for the latter by the explanation of them on the basis of a historically determined subsistence the deductive methods of the core could not be applied at the level of universality appropriate to theory. Those deductive methods would only be usable there at the level of what we would today call models, dealing, that is, with special problems or with special assumptions (as, for example, in Marx s treatment of outputs in his reproduction schemes ). The attempt to fully quantify relations as complex and variable as those affecting wages and outputs would entail introducing, in the apt words of Edgeworth (1881, p. 4), arbitrary functions representing not merely not numerical knowledge but ignorance. And the irreversibility of most

4 Professor Foley and Classical Policy Analysis 223 of these relations makes it difficult even to conceive of their representation as functions. This distinction between two fields in the theory implied therefore a view of the nature of economic phenomena basically different from that of later theory. The deductive, purely quantitative relations of the core which later theory attempted to extend to the whole of economics by means of demand and supply functions resulting from substitution between factors of production were instead woven by the old classical economists into an analysis conducted by the more inductive methods appropriate to those intermediate data, for whose determination institutional and historical factors played a central part. A reading of the Wealth of Nations, or of Marx s Capital makes clear what we mean. It is in this sense that the notion of a core is of a piece with the surplus determination of non-wage shares. It indeed provides the answer to Jevons s criticism of Ricardo, according to which the doctrine that if wages rise profits must fall, [drawn from the equation] Production ¼ Profits þ Wages [is] radically fallacious: it involves the attempt to determine two unknowns from one equation. (Jevons, 1871, p. 269; see a similar criticism in Walras, 1954, p. 425) Clearly the doctrine is not fallacious if, as was the case for Ricardo, Production and Wages can both be taken as independent variables, i.e. can be determined separately from the core of the theory which the equation represents. 4. Foley s doubts about the attribution of that core to the classical economists, are however specified with regard to the role of both wages and outputs in the analysis. On wages he writes: the modern classical position seems to underemphasize the explicit theorization of the wage rate as arising from a subsistence standard of living determined by the fertility behavior of workers in Malthus and Ricardo. This theory [...] depended on much the same type of reasoning in terms of tendencies and feedbacks that are at the heart of the classical account of capital mobility and profitrate equalization. (Foley, 2001, p. 375) Foley seems here to think here that (a) the demand and supply of commodities to which the classical economists appealed to explain profit rate equalization in their theory of market prices were also broadly applied to labour and, therefore, (b) wages could hardly be supposed to have been data in a core where prices were instead unknowns. I would not disagree with part (a) of the statement, if Foley were referring to classical demand and supply: these however only explain the tendency of actual or market prices towards natural or normal prices and not those prices themselves which are determined independently of such demands and supplies (as made clear by Smith in Book I, Chapter VII of the Wealth of Nations). The classical application of those concepts to labour 1 and here we turn to part (b) of 1 On the classical treatment of demand and supply of labour as single quantities see Garegnani (2002, pp ).

5 224 P. Garegnani Foley s statement does not therefore mean that we find there a determination of wages by demand and supply functions any more than we find a determination of commodity prices by such means in the classical account of capital mobility and profit rate equalization. Foley s tendencies and feedbacks would contradict the treatment of wages as an intermediate datum only if Smith s and Ricardo s demand and supply were interpreted along lines like those of Samuelson s (1978) canonical classical model, that is, along the lines of a demographic specification of the neoclassical functions that determine wages simultaneously with prices. The admission of labour unemployment think of Chapter XXXI of Ricardo s Principles (1821) should suffice to rule out any such demand and supply mechanism for Smith s or Ricardo s wages. The attribution to those authors of some version of that neoclassical mechanism on the basis of their population principle is however so widespread that it deserves some closer consideration to make clear by contrast the sense of Smith s or Ricardo s treatment of the wage as an intermediate datum. 5. In fact recognizing the role of the population principle in Smith and Ricardo in no way entails the above mentioned canonical interpretation of their determination of wages. 2 Indeed if we were to reason in terms of neoclassical demand and supply functions, the classical population principle taken by itself, could only contribute a rigid demand for labour to be coupled with an equally rigid supply. It would thus absurdly entail, as Samuelson admits (1978, p. 1423), zero wages when population growth has overtaken capital accumulation; or zero quasi-rents on capital goods (i.e. a negative net return on the capital goods supply prices) in the opposite case; or, finally, indeterminacy of distribution between wages and interest when the two rates of growth happened to balance. That would indeed be so unless we quite arbitrarily attribute to Smith and Ricardo a properly elastic labour demand on neoclassical lines operating at each given level of accumulated capital. Without that elastic labour demand, the interpretations put forth by Samuelson (1978), Hicks & Hollander (1977) and other authors with the then implied zero wages or zero quasirent or indeterminacy would not only contradict anything to be found in the classical texts: it would also fail to explain Smith s and Ricardo s argument about wages bringing population growth and capital accumulation into line with each other i.e. the very basis on which that interpretation is supposed to rest. A sequence of positions of the economy with zero wages or zero quasirents, or indeterminate division of the product is in fact hardly a basis for arguing the tendency to a wage ensuring balance in the growth of population and capital. As mentioned, the admission of permanent labour unemployment at natural or normal wages by Ricardo, Smith and Marx makes it quite evident on the other hand, that no such elastic labour demand function and corresponding neoclassical-like demand and supply mechanism can be found in those authors. And no confusion should be caused in this respect by the wages-fund theory in the 2 The questions sketched in this and the following paragraph are considered more fully in Garegnani (2002).

6 Professor Foley and Classical Policy Analysis 225 form it took only after Ricardo. Even less should doubt be cast by Smith s and Ricardo s already recalled different notions of demand and supply of labour. As suggested by those authors references to a proportion of demand to supply, single quantities and not functions are indeed meant there for labour (see for example Ricardo, 1821, p. 165), no less than they are meant for commodities. Demand for labour is there the broadly defined given employment possible at the stage reached by the accumulation of capital, while supply is, equally broadly, the amount of population the two providing by their difference some measurement of labour unemployment or underemployment, but certainly no schedules determining the natural or normal wage by their intersection. Thus, in conclusion, current interpretations of Smith s and Ricardo s theories of wages along neoclassical lines cannot in fact rest on the classical principle of population: they ultimately rest, it appears, only on tautologically presupposing in the classical authors that very concept of elastic labour demand function and of the resulting neoclassical demand-supply mechanism which is the object in dispute. 6. However a question arises then: if it is not the interaction between supply and an elastic demand function for labour, what else can trigger the wage increases that classical authors suppose when accumulation overtakes population growth, or the wage decreases they suppose for the opposite case? The answer, I believe, lies quite simply in that comparative strength of the competing parties to which Marshall would later refer in order to reject it (Marshall, 1920, Appendix J). As I have contended elsewhere, this is what emerges from those passages of Smith s and Ricardo s theories of wages which, significantly enough, have long puzzled modern interpreters, starting from Smith s notion of a minimum wage depending on common humanity (Garegnani, 2002, pp ). The relative growth rates of population and capital, with the resulting changes in Smith s and Ricardo s proportion of the demand to the supply of labour i.e. the varying pressure on wages of labour underemployment can indeed be seen to affect the comparative strength of the competing parties, and to trigger the variations in wages and the population adjustments envisaged by the classical economists, without passing at all through neoclassical-like labour demand and supply functions. We may now better understand in conclusion how the classical tendencies and feedbacks mentioned by Foley were in fact compatible with the treatment of the wage as an intermediate datum : it was determined separately from profits and prices by the circumstances affecting the level of subsistence, on the one hand, and, on the other, by those causing sufficiently persistent divergences from it like the state of the balance between population and capital accumulation and the corresponding levels of labour underemployment. It was in that sense that Marx, for example, could write that the foundation of modern political economy [is] the conception of the value of labour power as something fixed, as a given magnitude ( , Vol. I, p. 45). I have discussed in some details Foley s passage on classical wages because contrasting classical and neoclassical concepts of labour demand and supply illustrates well, I believe, the institutional and historical, character of much of classical

7 226 P. Garegnani theory and the corresponding inductive method, as opposed to the generalized deductive nature of neoclassical theory But we have noted that Foley s doubts about a core in the classical theories also apply to outputs treated as intermediate data. He writes: I doubt that Ricardo had any strong allegiance to the methodological proposition that the composition of social outputs was, in fact, determined at a different level from the equalization of the profit rate. On the whole, he seems to have thought that the kernel of his results simply did not depend much on the composition of social output, and therefore did not focus his attention on it. (Foley, 2001, p. 375) What Foley writes here is in fact an argument for what I describe as the assumption of given outputs in Ricardo. Saying that Ricardo s essential results on distribution and prices did not depend much on the composition of outputs which I take to imply that they did not so depend systematically as is the case in neoclassical theory is the same as saying that prices and the residual distributive variable were determined separately ( at a different level, as Foley puts it) from outputs, and outputs were therefore treated as data or independent variables in that determination with respect to the possibility of variable returns to scale. This treatment of normal outputs is on the other hand made clear by the equality of normal outputs with the respective effectual demands, unambiguously taken as given by Smith (1776, Book I, Chapter VII) and by Ricardo. It is also made broadly evident by a perusal of Ricardo s Essay on Profits (1815), or of his Principles (1821), or of Marx s writings. It is indeed the treatment of prices separately from outputs that forced Marshall to attribute to Ricardo the assumption of constant supply prices, and then, in agriculture, where that attribution would have been obviously incorrect, the further assumption of an absolutely inelastic demand for corn, the garb that a given output takes in neoclassical eyes (Marshall, 1920, p. 671 and Appendix L). 8. A word of clarification may be useful at this point on a further question concerning the classical core raised by Foley and, in different forms, also by other authors. Foley (2001, p. 375) writes that Smith or Ricardo (or Marx) [...] seem rather to have thought they were generalizing or abstracting from real economic phenomena, not proposing models in the contemporary methodological sense at all. Yes, I would entirely agree, if Foley means that those authors were attempting the universal explanations that is proper of a theory, and not exploring the logical implications of admittedly special assumptions as is done in today s models (cf. par. 3 above). But, in attempting such universal explanations they would certainly have instinctively strived for consistency, and, therefore operating as they did, within a science which had begun generating systematic treatments of the 3 This may seem to conflict with the abstract deductive reasoning whose introduction in economics is often ascribed to Ricardo. In fact Ricardo s specific contributions pertained to the deductive core of classical theory, but he mainly deferred to Smith for the rest of the theory where induction dominated.

8 Professor Foley and Classical Policy Analysis 227 economy since at least Quesnay s Tableau Economique ( ) their work would quite naturally result into something like a system. The question of a core in classical theory is not, that is, the question of a model in the contemporary sense which is consciously proposed : it is a question of a theory, in which the authors adapt instinctively to a reality in which the forces of competition, acting to bring about the uniformity of prices and rates of remuneration reflected in the core, are seen to operate within a framework where broader institutional and historical forces operate and quite naturally suggest a different method of analysis (Garegnani, 2002, p. 397). Of course, classical authors needed not have been fully aware of the terms of their system, as no alternative analytical structure existed to impose reflection on their own. And it is to that lack of awareness that Foley (2001, p. 374) appeals for his doubts about the claim that the classical political economists had a complete and well worked out method of inquiry into economic phenomena different from, but on the same level as, 20th-century neoclassical economics. Of course a practiced system is hardly ever complete and well worked out, but this does not prevent us from detecting its basic logical structure; just as, for example, the fact that good Latin writers may have been only partly conscious of the rules they followed in their language does not prevent modern grammarians from tracing a well worked out structure in their texts. II. The Classical Approach and the Quantity of Capital 9. Let me now come to our second subject: capital. Professor Foley (2001, p. 372) writes: it seems to me that it is impossible for the classical tradition to eliminate the value of the capital stock and of investment as central theoretical concepts. I agree: it clearly is impossible to explain distribution, prices or growth without referring to produced means of production. However, the analytical structure of the classical theories imposes conditions for the measurement of capital which are different from those holding under neoclassical theories, and can be more easily fulfilled. 4 In order to deal, however briefly, with the issue we need first to specify more precisely the question of capital to which we are referring. Just because it provides parameters for the determination of prices (by, e.g., entering the representation of the technical conditions of production) capital has to enter that determination as a magnitude, or a set of magnitudes, which are independent of prices, and this is true both for classical and neoclassical theories. And, it should be stressed, when that independent measurement is achieved, the value of the capital of Foley s passage is no problem, since it can be obtained in the same way as the value of any other set of products. The central point here is that, after the attempts of Jevons, Böhm-Bawerk, Wicksell and others at defining an average period of production independent 4 The question was in fact at the centre of my 1958 Cambridge PhD thesis (cf. Garegnani, 1960).

9 228 P. Garegnani of distribution and prices, it has had to be recognized that an independent measurement of capitol cannot be effected by means of a single magnitude: it can only be effected in terms of a set of magnitudes, be they the physical quantities of the several capital goods of Walras or Sraffa, or the dated quantities of labour of Wicksell or Dmitrieff. Now, the measurement of capital in terms of a set of magnitudes raises no logical difficulties in the classical theories. The independent measurement of capital is only needed there for expressing the technical conditions of production and a set of quantities can do. This is shown by the equations of Sraffa or, by those of Dmitrieff and others. But a measurement in terms of a set of magnitudes will not do in neoclassical theory which, I have contended, ultimately needs the single magnitude for the reasons we shall now briefly recall. 10. As Foley well knows, the question of capital in neoclassical theory is both complex and controversial. It must also be first recalled that it is of basic, if not always transparent, importance. Thus, to Foley s pragmatic mainstream scholar asking what difference does the capital critique make in the research on central banks and economic growth for which he is using the Solow model (Foley, 2001, p. 378), my answer would be that it may ultimately make the whole difference as to whether growth is controlled, as Solow (1956) supposes, by the saving decisions of the individuals taken jointly with population growth and technical progress or, on the contrary, some or all of the latter circumstances are broadly controlled by growth itself, in turn governed largely by aggregate demand. Briefly, the controversy on capital turns on nothing less than whether neoclassical demand and supply analysis provides us with a valid theory of the economy and, therefore with valid applications. Coming now to why neoclassicism needs the single magnitude, the point from which we must start is that, for savers, heterogeneous capital goods are perfect substitutes for each other in proportion to their values, as providers of the single commodity future income. Unlike the demand for consumption goods, savers demand for capital goods results from individual preferences that are non-specific to the particular good, and are only specific with respect to the values of aggregates of them, i.e. to the quantity of that single commodity (see Garegnani, 2000, pp recalling Walras s treatment of savings as a demand for that commodity). Hunger can be satisfied by corn and not by coal, but the demand for future income by savers can be satisfied by looms as well as tractors or any of the other thousands of capital goods, whichever provides the highest return on its value (i.e. the lowest price of a numeraire unit of future income). Now, it was this perfect substitutability of the capital goods for wealth holders, and the resulting quantity of capital that, at the end of the 19th century, provided the basis of the attempt by the main marginalist stream, to extend to the distribution between wages and profits, the classical principle of rent, (cf. Garegnani, 1970, p. 407). Indeed, had it been question there of taking as factors the individual capital goods, the possibility of varying their proportions to labour (and to each other) in analogy with the classical proportions of land to labour would have been almost nil and almost nil would have been the substitutability between factors, on which that neoclassical attempt relied. Alternative production processes differ in fact by the kind of capital goods employed rather than

10 Professor Foley and Classical Policy Analysis 229 by their proportions to labour and among themselves. The idea of a variability of factor proportions, rested therefore on taking the different kinds of capital goods as quantities of the same single factor capital, i.e. of the single commodity future income demanded by savers. Most importantly from a logical point of view the same fluidity of the capital endowment taken as the single magnitude needed to validate factor substitution, was required to allow its physical composition adjusting also in order to fulfil the condition of a uniform effective rate of return on the capital goods supply prices as necessary for equilibrium as are uniform wages or rents for each quality of labour and land. 5 Indeed these are the conditions that (since Adam Smith s natural prices ) have been held necessary in order to determine (under competition) normal prices whose persistence would allow them to be effective centres of gravitation for actual prices and therefore warrant a sufficient correspondence with some average of the observable prices. 6 The single magnitude was accordingly needed in neoclassical theory for something as basic as the possibility of correspondence between theoretical and observable variables. Thanks to the notion of capital as a single factor therefore, and only thanks to it, the way appeared to be open for replacing the historical and social factors by which the old classical economists had explained the distribution between wages and profits, with the desired generalization of classical rent. In particular an inverse relation between rate of interest and quantity of capital employed with given quantities of other factors would have ensured (from the side of the forces originating from the production system) a uniqueness and stability of distribution and of the whole system analogous to the uniqueness and stability of the classical distribution between rents on the one hand, and wages plus profits on the other. Nothing of the above holds on the other hand for the classical surplus theories of Smith, Ricardo or Marx, where factors substitution and the resulting demand functions play no role. The independent measurement is there essential only for representing the methods of production, where a set of magnitudes can also do. 11. This dependence of neoclassical, but not of classical, analysis on capital as a single magnitude explains why its abandonment, comparatively innocuous within 5 The qualification of the rate of return as effective is intended to take care of the case in which changes in relative prices over time are considered in the definition of the equilibrium and, accordingly, what is effectively a uniform rate of interest is expressed by different own commodity rates. On the confusion which has frequently marred the capital controversy between, on the one hand, the divergence of the effective rates of return on the capital goods supply prices mentioned in the text, due to the arbitrary initial physical composition of the Walrasian capital endowment, and, on the other hand, the divergence of own commodity rates due merely to changes of relative prices over time, see Garegnani (2003, Appendix II), which examines the confusion with reference to Hahn (1982). 6 Once the tendency of actual prices to their normal level (i.e. the stability of the latter) is preliminarily established, the persistence of normal prices would allow for a repetition of transactions sufficient to let the normal price emerge as some average of the actual prices (see Garegnani, 1976, pp ).

11 230 P. Garegnani the classical approach, was likely to entail a deep change indeed in neoclassical theory and in its nature and significance. The generalized abandonment of the concept in mainstream neoclassicism occurred essentially as a result of the first phase of the capital controversies in the 1960s and early 1970s, though the way out of that concept had in fact been already attempted within mainstream theory by Hicks, in his Value and Capital (1939). Clearly he had by then come to realize the impossibility of proceeding with the conception of capital as the single quantity on which he had based his Theory of Wages (1932a). The only conceivable alternative to that conception, attempted within neoclassical theory, had been that used by Walras for his general equilibrium system, where each capital good constituted a separate factor. And that alternative was the one adopted by Hicks. As we noted, that conception allowed however little scope for the substitutability between factors on which the theory rests: a circumstance which may well suffice to explain why the conception, despite its greater definiteness, and the fame of its author had failed to enter the neoclassical mainstream in the near three quarters of a century since Walras had advanced it. On that lack of substitutability Hicks had in fact to turn a blind eye in his revival of Walras s theory of capital. 7 Above all, Hicks had then to cope with the second difficulty also seen in par. 10, one that had caused a logical inconsistency in Walras s original system, where the treatment of the capital endowment as a vector had been in contradiction with the condition of the uniform effective rate of return on the capital goods supply prices, which Walras, like all his predecessors and contemporaries, had introduced in his system. 8 Hicks coped with this basic inconsistency by silently dropping the condition of the uniformity of returns and, with that, the notion of a persistent normal position of the economy that would allow for a correspondence between theoretical and observable variables (par. 10 above). That notion, central to all previous mainstream neoclassical authors had then to be replaced with that of intertemporal and temporary equilibria, dependent on future prices. These left a choice 7 It is interesting to note how Hicks, whose Value and Capital (1939) was to introduce the Walrasian conception of capital in the neoclassical mainstream, would write in (1932b) that [Walras s and Pareto s] theories of capital [...] are the last part of their work which one can consider as final, or accept without the most careful consideration [...]. For it is surely significant that Walras s elaborate theory of capitalisation does not reappear in the later work of Pareto, while there is nothing very substantial to take its place (Hicks, 1932b, p. 297). Further, and more specifically, in his Theory of Wages Hicks (1932a, pp ) had contrasted the marginal product of labour he was using then obtained by keeping constant the quantity of co-operating capital [but not its] form with the corresponding short period marginal product obtainable on a Walrasian basis, where, capital could not change its physical form. He then curtly dismissed the latter because it is very doubtful that this marginal product [...] can be given any precise meaning that is capable of useful application. Although little noticed in the literature, Hicks s about turn on a matter as basic as the conception of capital and equilibrium and the adoption of equilibria whose deficiencies had made for his explicit dismissal not many years before, is I believe, a symptom of the depth of the crisis which neoclassical theory has undergone because of the inconsistency of its original notion of capital. 8 On Walras s inconsistency and the need to conceive the capital endowment as a single magnitude in order to satisfy that condition of uniformity of returns, see Garegnani (1976, p. 34) and the literature mentioned there.

12 Professor Foley and Classical Policy Analysis 231 only between the Scylla of assuming nonexistent complete futures markets and the Charybdis of indeterminate subjective expectations. From this and from the indefinite number of Walrasian factors replacing the traditional trinity, there also came, incidentally, the formalism Foley rightly laments (2001, p. 367). 9 It is not surprising therefore that replicating Walras s earlier fortunes, this overturning of the then accepted notions of equilibrium exerted at first little influence at least in England and, in particular, in the Cambridge of Pigou, Keynes, Robertson and Shove, then still at the centre of economic theorizing. It was only when the early phase of the capital controversy had really left no other choice, that the Hicksian reformulation could become dominant in the forms it had assumed in the meantime at the hands of Samuelson and other mathematical economists. 12. It can perhaps now be better realized why the specific difficulties pointed out in the capital controversy pertain only to neoclassical theory and also how the abandonment of the notion of a single quantity of capital was a momentous change in neoclassicism a change which left the theory deeply different from what it originally was intended to be, thus marking what we may call a Hicksian divide in its evolution. The toll which the potential explanatory capacity of the theory has had to pay for the change was heavy. Foley himself recalls elsewhere its drastically reduced capacity of explanation (see Foley, 2003, p. 237). And, what is more it begins to emerge that that toll has been paid in vain. As a correct understanding of the neoclassical problem of capital, with its roots in the homogeneity of capital goods for savers, might well have suggested, the Hicksian reformulation has not eliminated the ultimate dependence of neoclassical theory on the notion of capital as a single magnitude. As we said, savers decisions are taken in terms of that quantity, and the latter must therefore be present in a neoclassical demand and supply system, like that of any other good demanded by the individuals. Removed from the immediately visible level of the capital endowment, it remains at the less transparent but more fundamental level of the savings investment process. It can then be argued that the non-existence of a measurement of capital independent of prices entails the same consequences which the controversy had pointed out in the traditional formulations of normal prices, and as we saw, had enforced its abandonment (Garegnani, 2000). III. The Classical Approach and Economic Policy 13. We may pass now to the third of the issues listed at the beginning: the classical approach and policy analysis. Professor Foley writes that while the criticism of neoclassical theory identifies fundamental weaknesses in neoclassical theory, yet a neoclassical mainstream continues to dominate teaching and research in economics in America and increasingly in the rest of the world. After what we saw in par. 12, the continuance and increase in neoclassical domination Foley 9 It is indeed to the period following Hicks s Value and Capital that some authors referred for a Formalist Revolution, whose origin they see essentially in that book, but whose causes they seem to have difficulty in tracing (see for example Hutchison, 2000; Blaug, 1999).

13 232 P. Garegnani refers to might need some qualification as to its true nature, but we are here interested only in Foley s own explanation of the situation. He writes: the market for analytical economics is fundamentally a market for policy analysis. [...]In committing methodological sin by putting the theories of price and distribution on the same level of abstraction, neoclassical practice gains the tremendous advantage of being able to make routine predictions about the composition of output and distribution. (Foley, 2001, pp ) And he continues by arguing that, despite the excellence and persuasiveness of its doctrinal critique and researches in the history of economic thought, little work has been done in the ambit of the classical resumption in the direction of an applied classical political economy as an alternative to neoclassical practice. It is undoubtedly true that much work connected with the resumption of classical analysis has been on pure theory and, partly, on history of thought and there are, I think, good reasons for that, which we shall see below. We should however not miss first the work that has already been done concerning policy. That work may perhaps have attracted less attention than it deserves, because of some expectation that it should have taken shapes (e.g. formal links between prices, outputs and distribution) not unlike those of the contemporary neoclassical applied work making for routine predictions, rather than the more institutional shapes pertaining to classical theory beyond its core and exemplified, at the time, by the work of the classical economists themselves. Indeed, even apart from detailed critical considerations on contemporary applied work, the complexity of the economic world might justify some a priori scepticism, shared by most neoclassical economists of the past, and also by some of the present, about the real importance of applications of theory making for such routine predictions. I believe that in fact the contributions on various aspects of policy analysis from the classical side have not been inconsiderable especially with respect to the relevance of aggregate demand for the long-period, and not only for short-period fluctuations in activity But before coming to that, we may consider more closely why it seems natural that in the present theoretical situation, work on the classical revival should have been focused on pure theory and its history, even to the temporary detriment of applications. I believe two elements stand out for explaining that. The first is that in economics, when a theory is as dramatically new to the profession, as what we now understand to have been the theory of the old classical 10 On aggregate demand, see note 14 below. Work has on the other hand been done on wages and more generally on income distribution; on capital accumulation and growth with particular attention to the relation between consumption and investment; on technical change and its effects on employment; on issues in monetary and fiscal policy, including some specifically related to the European Monetary Union. Instances of works in these areas are Levrero & Stirati (2004); Stirati (1994, 2001); Cavalieri et al. (2004); Piccioni & Ravagnani (2002); Garegnani (1990); Palumbo & Trezzini (2003); Trezzini (1995); Cesaratto et al. (2003); Garegnani & Trezzini (2005); Ginzburg & Simonazzi (2004); Pivetti (1999); Ciccone (2003); Cesaratto (2006); Garegnani & Palumbo (1998); Palumbo (1994, 1996); Pivetti (1998); Barba (2006); De Vivo & Pivetti (1980); Vianello & Simonazzi (2005).

14 Professor Foley and Classical Policy Analysis 233 economists and Marx, then, its impact on applied analysis and policy can only follow, rather than precede, some initial understanding and acceptance of the theory. The obvious example is Keynes s analysis, whose tremendous impact on policy followed from an initial understanding and acceptance of the theory. His policy proposals had in fact been broadly anticipated by the writers Keynes referred to as cranks to no effect whatsoever, because they were not backed by a theory capable of challenging orthodoxy. The second element in explaining those priorities in the classical revival is that the novelty of the theory inevitably raises a number of problems also among those who are inclined to a broad acceptance of it. The clarification and solution of these problems is evidently essential for a sufficiently solid grasp of the theory and, therefore, for fruitful applications of it. It was therefore natural that, even when the choice was not predetermined by the need to answer criticism or dissipate misunderstandings, scholars working for the resumption of the classical approach should primarily concentrate their efforts in directions susceptible of widening and deepening the understanding and acceptance of the theory in the profession. 11 And these directions were essentially the three of the critique of dominant theory, the investigation of the analytical structure of the classical theories and its comparison with that of dominant demand and supply analysis, and finally, the establishment of the continuity with the work done by classical authors, from Smith to Ricardo and Marx (not a minor matter when we realize that theories destined to last have rarely been built in just a decade or two by a few scholars however brilliant). 12 Not to give some priority to these directions of research would mean, in effect, accepting a subordinate role for a theoretical approach which is felt by its practitioners to have the potential of replacing the deeply flawed one which is still dominant. 15. As we return now to the relevance of the classical revival for policy analysis, I will confine myself here to mentioning what I believe are some basic implications of the classical approach regarding content and also methods of economic policy analysis. 11 Foley (2001, p. 378) refers to the dramatic tactic of debating an empty chair used in American politics, and to how the revival of classical political economy faces a definite scarcity of scholarly opponents willing to show up for debate. That chair might perhaps appear less empty when we remember all the agitation which followed the publication of the New Palgrave Dictionary (Eatwell, Milgate & Newman, 1987), or recall Arrow (1991), Samuelson (1987a, 1987b, 1990, 1998, 1999, 2000), or Hahn & Petri (2003), and recall also the debate on the interpretation of the classical economists launched in the 1970s by Hicks, see Hicks & Hollander (1977) and Samuelson (1978), and which, shows no sign of subsiding. 12 All three directions of research centre on the core of classical theory, but would not seem to imply, because of that, any more aversion to risk (Vianello, 2007) than, say, looking for a lion in its den, rather than in the forest where it sometimes strolls. We may here recall also Ginzburg s criticism (2000, p. 140) that I confine Marx s argument on the transitory nature of capitalism to the inverse relation between wage and rate of profits. However, as Ginzburg well knows, that inverse relation is present also in neoclassical theory, where it has generally coexisted with a harmonious view of capitalism. The antagonistic character of the relation I referred to was therefore meant to come largely from the out-of-core considerations specified in the very volume in which the interview Ginzburg refers to appeared (see Garegnani, 1981, p. 64).

15 234 P. Garegnani With respect to content, the basic change with respect to neoclassical work can be summarized in the different meaning to be given to competitive prices. These can no longer be seen as indexes of the scarcity of the commodities, reflecting, that is, the relative scarcity of the factors of production entering their production. The change in the explanation of the distribution of the product between wages and profit undercuts that conception, and competitive prices emerge a priori as merely the result of a particular way of distributing the surplus, the one pertaining to the institutional set up of competitive markets. In that sense competitive prices are just one of the many ways of distributing the surplus which are conceivable. We have emphasized the phrase a priori. Nothing of course in the classical theory denies that the mode of distribution of the surplus may have important and even decisive effects on the working of an economy and society. What is denied is that such effects might be summarized in terms of a Pareto optimality for competitive prices. The denial of a determination of the distributive variables by an equilibrium between demand and supply of factors, i.e. by the full employment of them, breaks open the magic circle within which policy analysis tends to be confined by the neoclassical preoccupation with distortions of the Pareto-optimal allocation of resources allegedly brought about by competition. There vanishes in particular the notion that the full employment of productive resources, and therefore the maximization in some sense of the social product, requires that factors prices should be those realized by competitive market forces. The circle is thus broken open, for example, for the use of prices to influence the distribution of income. In effect the Pareto optimality of competitive markets then gives way to what I would provocatively call the principle of the underutilization of resources in a market economy. At its basis we find a second main implication of classical theory for applied and policy analysis, besides the distributive rather than allocative role of prices. Once the neoclassical idea of distribution determined by an equilibrium between factor demands and supplies is left behind, the full utilization of productive resources emerges as something likely only under special circumstances and limited periods of time in markets left to their spontaneity. Indeed these are reasons, which sheer observation imposes in various forms even upon orthodox analysis, for arguing that some underutilization of resources, in particular some unemployment of labour, is systematically required for the stability of a market economy. (The turning away from full employment policies in the advanced capitalist countries in the 1970s is revealing in that respect). 13 However, the principle of the underutilization of productive resources draws its full importance from what I see as a further, third, main implication of classical theory for applied and policy analyses. It is the conclusion as to the relevance of aggregate demand for capital accumulation and growth which has been the result of both the critical and constructive work carried out so far within the classical revival, and has been at the centre of applied and policy analyses there. (It should be noted here how in classical theory this third implication concerning aggregate demand is independent of the above one on possible inequalities 13 This turning away has been analyzed in Cavalieri, Lucii & Garegnani (2004).

16 Professor Foley and Classical Policy Analysis 235 between factor demands and supplies: Ricardo, while sharing Say s law admitted the labour unemployment of Chapter XXXI of the Principles.) 14 In marginal theory Say s law is in fact ultimately an implication of the theory: the forces which should ensure the tendency to the full employment of labour and the other factors if an equality between aggregate expenditure and aggregate output at normal prices could be achieved, are the same that should ultimately warrant that equality. The alleged tendency of investment to adjust to full employment savings is arguably one thing with the alleged tendency to an equality between the demand and the supply of capital postulated in the theory. The critique of the neoclassical theory of distribution entails therefore the critique of the idea of a long-period irrelevance of aggregate demand for the level and growth of the social product. With respect to the constructive side, on the other hand, we do not find in the classical theory of distribution and relative prices any argument for a tendency of investment to adjust to saving decisions out of the income corresponding to a full utilization of existing productive capacity, so that the theory is in fact open to a role of aggregate demand in determining outputs. 15 Indeed we can go further and argue that such a role is a natural result of classical theory: aggregate demand takes its place with other factors in the separate determination of outputs which characterizes those particular intermediate data. This long-period relevance of aggregate demand constitutes then the mainstay of our principle of the underutilization of resources in a market economy. That principle in fact derives its strength not only from the inexistence of the neoclassical distribution with its equality between factor demands and supplies, but also and even more importantly from the compound-interest-rate-like effects which pertain to the missed potential increases of future productive capacity entailed in the missed income and corresponding savings associated with that missed equality. The observable unused capacity thus grossly understates the real underutilization of resources in the economy, which should include the missed opportunities of growth due to lack of aggregate demand in the past just as it understates the opportunities for expanding labour employment and outputs should aggregate demand be sufficiently high. 16 The contradiction of Pareto optimality under free competition is patent: outputs and labour employment could be increased, and economic welfare conceivably expanded for the many without decreasing it for anybody. 14 On the relevance of aggregate demand and other implications of classical analysis for labour unemployment, cf. Garegnani (1978). 15 Ricardo s adherence to Say s law was essentially the result of a failure to distinguish between decisions to save and to invest. As noted in the text, that law did not entail for him any tendency to the full employment of labour; it only meant no demand limit to what the stage reached by capital accumulation allows the economy to produce. On the sense of Ricardo s adherence to Say s law see Garegnani (1978, pp ). On the way in which competitive wages were seen as altogether compatible with labour unemployment by classical and pre-classical economists, see Stirati (1994) and Levrero (2003). 16 See for example Garegnani (1992) and Garegnani & Palumbo (1998). See also Palumbo (1994) for further consideration of how that fits with the view of several economic historians and the dissatisfaction of many of them with dominant economic theory.

17 236 P. Garegnani We have thus found three strictly connected features of classical theory, of importance for applied and policy analysis: (1) the role of competitive prices as just one of the conceivable criteria open to society for the distribution of its surplus; (2) the possibility of labour unemployment and capacity underutilization resulting directly from the inexistence of the neoclassical factor demand and supply equalities; (3) the relevance of aggregate demand for the level and growth of the social product. Points (2) and (3) then converge into what we have called the principle of underutilization of resources in a market economy. These elements should suffice to show the extent of the implications of classical theory for policy analysis both in advanced economies and even more importantly in developing economies where the social obstacles to demand expansion, joined also by some obstacles to attaining long-period supply expansion (like illiteracy, lack of technical education, etc), are patent. 16. As for method in policy analysis, I believe it can be said that there is bound to be less space for routine predictions about the effects of policy, whether from theoretical analysis, or econometric models: the fact that as we said the analysis would largely have to be carried out outside the core of classical theory will see to that. There will in fact be few grounds left for the idea that policy may be left to technicians who will steer an exact course in some objectively specifiable collective interest. This idea would seem to have to give way to recognition that policy decisions will generally favour some groups and damage others, and accordingly have primarily to deal with relations of power. For example, even a policy of full labour employment, which might seem to be in the obvious interest of the whole community, meets and has in fact met obstacles which the naked eye unambiguously reveal to lie in how that policy affects the relative power of the classes or groups concerned. Postscript A. After the preceding remarks were written, Professor Foley published a further review essay (Foley, 2003) concerning a second collective volume devoted to classical matters, namely Heinz D. Kurz s Critical Essays on Piero Sraffa s Legacy in Economics (2000) and I permit myself this postscript because what I have said in my intervention above may be of use for clarifying some points Professor Foley raises in his later essay, with respect also to my chapter on intertemporal general equilibrium (2000) in the volume reviewed. We may leave aside here Foley s interpretation of Sraffa s work which, I believe, underplays somewhat the rediscovery and revival of the surplus theory of the old classical economists and Marx. We may focus instead on the dividing line Foley traces between what he calls the neoclassical parable which treats capital as a single productive factor, and the contemporary neo-walrasian theory which treats it as a vector of capital goods. Foley s distinction coincides in effect with the Hicksian divide in the evolution of neoclassical theory to which I refer in par. 11 above. Of course I share Foley s critical stance towards both those positions, but it appears to me that he raises the possibility of misunderstandings when he proceeds to write:

18 Professor Foley and Classical Policy Analysis 237 Neo-Walrasian general equilibrium theory is similar to Sraffa s construction in that neither suggests a systematic relationship between the equilibrium rate of interest and the value of the capital stock per worker [and the two theories] are thus incompatible with the neoclassical parable of a uniform capital substance. (Foley, 2003, p. 229) I would rather stress, on the one hand, the strict continuity between neo- Walrasian general equilibrium and the parable, since the former is as founded on the neoclassical principle of substitution between factors of production and on the resulting demand and supply forces, as is the parable which constituted in fact the dominant formulation of neoclassical pure theory (inclusive of its explicit general equilibria versions) 17 until, as we saw, the neo-walrasian version replaced it in that role when the first stages of the capital controversy had made that formulation absolutely untenable. This continuity clearly sets neo-walrasian theory apart from Sraffa s revival of the classical surplus theory and from his critical aims. On the other hand, and more specifically, it has been recently emerging that post-hicksian, neo-walrasian theory may depend on the capital substance and the corresponding systematic relationship with the interest rates, no less than does the pre-hicksian neoclassical parable. As recalled in my intervention above (par. 12) I have argued so in my (2000) chapter on intertemporal general equilibrium of the volume reviewed by Foley, the dependence has merely become less evident of the neo-walrasian formulation because the latter avoids its most obvious form, i.e. the parable s open reliance on that substance for the given capital endowment, logically necessary for its normal positions. B. I have just referred to my (2000) chapter. Foley understands my purpose there to be reconstructing virtual (out-of-equilibrium) supply and demand functions for saving and investment and showing that these schedules [...] imply implausible out-of-equilibrium dynamics. To which he objects: But there are many ways to reconstruct out of equilibrium behaviour schedules in general equilibrium models, [...]. Garegnani s approach is ingenious but no more compelling than any others. My main purpose in that paper was not, however, to advance a specific method of analysis of the stability of intertemporal general equilibria. That was incidental to the basic purpose which was to show that, as just said, the neoclassical difficulties with capital persist in those reformulations, and that once capital goods are properly introduced in the system the resulting overall properties of the equilibria are 17 Cf Wicksell (1934) and note 18 below for the location of Walras own original general equilibrium system. Contrary to what seems to be often believed the change entailed by the neo-walrasian versions has not been the adoption of general equilibria. It has been a change in the kind of general equilibria (Marshallian partial equilibria themselves rest on general equilibrium): i.e. away from normal positions and towards intertemporal or temporary equilibria.

19 238 P. Garegnani correspondingly in such a sharp a contrast with observation as to throw into doubt the theory determining those equilibria (Garegnani, 2000, p. 392). And the properties in question relate to the possibility of multiple equilibria and zero prices for factors and commodities which experience indicates as undoubtedly scarce, no less than to the instability of those equilibria. As for the latter specific result (largely implied already independently of any dynamic assumptions by the possible multiplicity of isolated equilibria argued there) I of course did not overlook that there are many ways to reconstruct outof-equilibrium schedules besides those I introduced. On the contrary, the multiplicity of possible dynamic hypotheses on the matter was explicitly stressed by me: my point was rather that, given the implications of reverse capital deepening which my schedules were intended to highlight, only acrobatic hypotheses about out-of-equilibrium behaviour could ensure that the system was stable. As I wrote: our critical aim strengthens the legitimacy of assuming that the dominant outof-equilibrium movement will be along the [savings supply and investment demand] schedules [since] if instability were to result under that assumption, it would be all the more plausible when obstacles to the adjustments to equilibrium are also considered in the connected markets which the schedules assume to be kept in equilibrium. (2000, p. 427) C. On the other hand Professor Foley (2003, p. 237) points out that the internal development of [neo-walrasian] general equilibrium theory by its own adherents has shown its sharp limitations. The Debreu Sonnenschein Mantel theorems, which show that any excess demand functions that satisfy continuity and Walras s law can be supported by some appropriately defined Arrow Debreu economy underline the fact that the Arrow Debreu theory is extremely weak in explanation power. It certainly is interesting and revealing that proponents of the theory should themselves conclude that the latter is extremely weak in explanation power. However the Debreu Sonnenschein Mantel results apply under conditions of pure exchange or when production is so represented as to in effect leave aside the equilibrium problems raised at each date by, on the one hand, the outputs flows of capital goods entailed in each period by the decisions of firms about the techniques by which to produce and, on the other hand, the saving decisions of individuals. At the roots of those theorems there appear therefore to lie no more than the income effects considered long ago in less general, ways by Marshall, Walras, Wicksell or Hicks. These effects were then argued to be compatible with a more precise definition of the substitution principle in consumption, and were generally taken to be ultimately no more worrisome for the theory than a backward rising supply of labour or of savings which may well explain why today, despite the allegedly radical nature of those theorems, demand and supply theory remains dominant. Now, the difficulties pointed out in my 2000 paper differ in that they originate from a quite distinct source, i.e. capital and, as I have argued, they appear to throw into question the very principle of substitution between productive factors, on which the theory was originally founded and still ultimately rests.

20 Professor Foley and Classical Policy Analysis 239 D. We noted under point A above how Foley s confinement of the problem of capital to what he describes as the neoclassical parable appears to underestimate the implications of the problem for neoclassical theory as such. The same underestimation emerges when Foley wonders along with Burmeister (2000, p. 394) why the victory of the Cambridge UK side in the capital controversy has left no mark on the reality of neoclassical economic research method. This seems to paradoxically overlook the change, which Foley himself stresses, from the neoclassical parable and towards contemporary neo-walrasian theory where the parable as we recalled, is in effect, pre- Hicksian-divide neoclassicism. 18 This momentous change was conceived, it is true, before the capital controversy erupted, but it could achieve dominance, only after the controversy had swept away the traditional versions of the theory (par. 11 above). And the key point is that the versions thus swept away were those which had originally allowed people to think that neoclassical demand-and-supply constituted an acceptable theory, i.e. that it had the explanation power which, as Foley admits together with many of its present proponents, neo-walrasian theory lacks. That seems a very substantial mark indeed which the Cambridge UK victory has already left on present day neoclassical theory. E. Foley also reports of two tests proposed by Burmeister for whether a serious consideration of the issues of heterogeneous capital [...] can be given by modern economists. In common with Burmeister, he appears to believe that the question of heterogeneous capital fails them. The two tests are (i) the possibility of identifying interesting economic questions whose correct answers require a model with heterogeneous capital goods, and (ii) to find ways to give such models empirical and policy content (Burmeister, 2000, p. 313). On the first of these tests, I may refer readers to my observations on Foley s pragmatic mainstream scholar in par. 9 of the intervention above. But a more direct reflection on it seems to be also in order. Reswitching of techniques and reverse capital deepening clearly require heterogeneous capital goods: do, then, Burmeister and Foley seriously think that such phenomena do not raise interesting economic questions? On the second test, and the policy implications of the question of capital, I can however only refer readers to Section III of my intervention above and to the policy implications of the deficiencies of neoclassicism brought to light by the heterogeneity of capital. 18 Even Walras can be seen to have shared in the parable, in that he originally and inconsistently aimed at determining the normal position of the economy, requiring the capital endowment as a single magnitude (par. 11 and n. 8 above). In effect, in his analysis of economic progress he treats capital as a single magnitude as if we could suppose all kinds of capital goods to change in the same proportion; cf. for example his summary of par. 331 in the Elements (Walras, 1954, p. 389) about economic progress: Assume land fixed, population doubled, capital more than doubled.