Misunderstanding Classical Economics? A Reply to Blaug. Pierangelo Garegnani

Transcription

1 Misunderstanding Classical Economics? A Reply to Blaug Pierangelo Garegnani From Professor Blaug's 1999 article (unspecified page references are to it) there emerges the paradox that an interpretation pointing to the basic role of institutional and historical factors in the classical theory of distribution and capital accumulation is criticized as "amazingly narrow" (215), one that looks at those economists through the lenses of the linear production model (234). Further, the contrast between the seeming narrowness of the interpretation and the evident broad character of classical works induces Blaug to view that interpretation as one that reads backwards into Smith's and Ricardo's texts Sraffa's theoretical alternative to neoclassical theory (215) as if the contrary was not rather the case, with those texts being the sources to which one had to resort in order to understand Sraffa's terse Production of Commodities (1960). A need for clarification is therefore present. Blaug is critical of the interpretation advanced by Sraffian authors (I may note in passing how such collective references may easily if unwittingly result in blurring the issues) because, in his view, the notion of Correspondence may be addressed to Pierangelo Garegnani, Centro Sraffa, Dipartimento di Economia, Università degli Studi di Roma Tre, Via Ostiense 139, Rome, Italy, 00154; sraffa@uniroma3.it. The original reply submitted in July 2001 has been appreciably shortened as editorially requested for the present discussion and is forthcoming as a Quaderno di Ricerca of the Centro di Studi e Documentazione Piero Sraffa. The version originally submitted is available from the author on the Internet. History of Political Economy 34: by Duke University Press.

2 242 History of Political Economy 34:1 (2002) social surplus they refer to, while adequate for the French physiocrats (Blaug 1987, 439), does not provide a valid interpretation of Ricardo and the English classical economists. A first reaction to this is one of some surprise. In his concrete work on the classical economists, Blaug has often acknowledged elements that are at the basis of Sraffa s interpretation of Smith and Ricardo. Recall his references to Ricardo's corn model and fundamental theorem of distribution. Or remember his recognition of permanent labor unemployment in Ricardo (Blaug [1958] 1973, 179), a feature basic enough to raise the question of a theoretical approach alternative to the modern one and along the lines of Sraffa 's interpretation. The contradiction in Blaug comes, we shall suggest, from his seeing those classical elements in isolation, and not in the connections that make them aspects of an approach to distribution and prices basically different from the modern one: an expression, one may argue, of the difficulty of conceiving the possibility of an approach to distribution and prices other than some form of the modem one, based on demand and supply functions for factors of production. One result is that Blaug has to repeatedly fall back on ascribing the undeniable and striking differences between Smith's or Ricardo's analyses and the modem ones, to confusions and primitivism (see below). A preliminary observation is, however, in order here. Sraffa 's interpretation was explicitly advanced for the old classical economists from Adam Smith to Ricardo (Sraffa 1960, v; emphasis added); Blaug argues instead as if the interpretation he criticizes referred to a wider British tradition of classical political economy, starting more or less with Smith in the third quarter of the eighteenth century and ending more or less with John Stuart Mill and Karl Marx in the third quarter of the nineteenth century" (215). This is surprising from the author of a book (Ricardian Economics), a chapter of which was titled The Half Way House of J. Stuart Mill, and who is therefore aware of the reasons for which Sraffa may have wished to exclude Mill (for one) from his purview. However, the inconsistency may have been favored by Blaug's tendency to locate in subject matter (220) the domain of Sraffa's interpretation, which is instead focused on the logical structure of Smith's and Ricardo's theory. 1 1 This should of course not be confused with Sraffa providing a rational rather than historical reconstruction of the old classical economists (213). The case here is similar to that of a contemporary grammar of classical Latin, in which a Latin writer of the period should

3 A Central Misreading Garegnani / Blaug on the Sraffian Interpretation 243 Blaug describes the Sraffian interpretation as one for which the classical economists addressed the question of how an economic surplus is generated, spent, and augmented from period to period, where the economic surplus is the disposable part of total output that is left over after capital consumption has been made good and the workers fed (216; see also 214). However, what was originally meant in that interpretation was more specific: it was the tracing in Adam Smith and Ricardo of an approach to distribution and relative prices alternative to any however primitive form of modem demand and supply of factors. 2 Aware perhaps of the vagueness of his definition, Blaug focuses then on what he describes as the hard version of the interpretation, which he attributes in particular to my work. And it is here that a basic misreading occurs. He characterizes that version by the role of data, which, he claims, I ascribe to real wages, outputs, and technical conditions of production in the analyses of Smith and Ricardo. But what I contended (e.g., Garegnani 1987, 560) was different: it was that those three sets of circumstances were data when determining the nonwage distributive variables and the relative prices, and not data for the theory as a whole, as Blaug imputes. The question, we shall see, is one and the same as the one above of the alternative theory of distribution in Smith and Ricardo, who determined the nonwage distributive variables on the basis of the difference between the (net) product and the wages of the workers producing it, governed by subsistence. Within that determination, the physical product and the real wage appeared therefore as givens. But since obviously no economist, least of all Smith and Ricardo, can avoid being concerned with determining product and wages, that meant that, unlike in modem theory, those magnitudes (and the technical conditions with them) were determined separately from the nonwage distributive variables and the prices. recognize the rules he consciously or unconsciously followed in expressing himself; as Blaug would agree, the grammar would constitute a historical reconstruction of that Latin. 2. This was what I argued when advancing that interpretation on the basis of Sraffa's introduction to Ricardo's Principles ( ) in a Cambridge Ph.D. dissertation (1958), comparing classical and neoclassical theories on the problem of capital. For the applicability of that interpretation in particular to Smith, see Garegnani 1958, 13-24; also Garegnani 1987, , and the bibliography therein. See also Blaug himself who, in Blaug 1987, 439, writes that the interpretation captures much of the drift of the Wealth of Nations.

4 244 History of Political Economy 34:1 (2002) They only appeared as data in a purely quantitative core of the theory in which the relationships between prices and distributive variables enforced by free competition were seen to determine profits of capital and/or rents of land as the mentioned residual or surplus. They were so to speak intermediate data and not ultimate, as Blaug interprets. It may be immediately noted how this misreading about data is directly responsible for the drastic reversal in the thrust of the Sraffian interpretation we emphasized above in the opening section: the institutional and historical richness of classical works, rightly stressed by Blaug, emerges in the treatment of wages, outputs, and technical conditions precisely what Blaug's misreading neatly excises from the interpretation. However, the substance of the question of intermediate data, as distinct from its definition, also needs clarification. It consists, essentially, of a distinction between two fields of economic analysis that the classical authors had drawn instinctively (Garegnani 1987, 562). 1. The first was the determination of the nonwage variables and prices by means of the competitive relations of the core that, whatever the language (algebraic, arithmetical, or literary) expressing them, entailed quantitative properties general and definite enough to warrant a mainly deductive analysis. 2. The second covered the rest of the theory, and in particular the determination of wages, outputs, and technical conditions. Given the multiplicity and variability of the factors felt to be involved, a determination of those circumstances by functional relations similar to those of the core such as the demand and supply functions of later theory would have corresponded to what F. Y. Edgeworth ([1881] 1967, 4) was aptly to describe as using arbitrary functions representing not merely not numerical knowledge but ignorance Blaug briefly mentions the above criterion for my distinction of two parts in classical theory (217), but seems to forget it when in the article as a whole he objects to my interpretation by simply arguing that the classical economists were much concerned with explaining wages, outputs, and technical conditions, as if that had ever been denied. Indeed, had I denied that, something would have been seriously amiss in Blaug's argument too. He insists on my distinction between what is in, and what is out, of the core (e.g., ) but what could ever be out of it, if its data happened to be the ultimate ones of the theory? Perhaps in order to temper this contradiction latent in his argument, Blaug slips, at times, into taking the different tack that, by excluding, wages, outputs, and techniques from the core, the interpretation takes

5 Garegnani / Blaug on the Sraffian Interpretation 245 The distinction between fields 1 and 2 thus left no space, as a matter of general theory, for a single system of quantitative relations determining prices, distribution, and outputs simultaneously. It rather meant that wages, outputs, or technical conditions were studied separately from the prices and nonwage distributive variables determined in the core of the theory. 4 We have referred to the distinction between the two fields of analysis as drawn instinctively by Smith and Ricardo. We can therefore hardly expect to find the distinction explicitly stated by them (see, however, Marx on wages, below). What we find are unambiguous expressions of it, like Ricardo's table in his Essay on Profits ( , 4:17), where corn wages and corn outputs are independent variables in determining profit rate and rents or like that determination of prices separate from outputs, which forced Alfred Marshall ([1920] 1962, 814, ) to attribute to Ricardo the two assumptions of constant return in manufacturing and an absolutely inelastic demand for the products of agriculture where constant return would obviously not have done. Blaug, we said, objects to my interpretation by arguing that the classical economists, far from taking technical conditions, outputs, and wages as (ultimate) data, were much concerned with explaining them. With respect to technical conditions, limits of space confine us here to considering how Blaug points to the [Sraffians ] historical misrepresentation of technology as an exogenous variable in classical economics (221), exemplified by their utter indifference...to the opening three chapters of the Wealth of Nations on the division of labor, a subject they never discuss or even mention (220). However, in a work of mine that Blaug cites elsewhere in his article, I had stressed that, unlike modem authors, classical economists like Adam Smith or Karl Marx had thought that the technical conditions of production... were largely determined by phenomena with which economic theory had to deal, just as with wages and outputs (think of the analysis of the division of labour in Adam Smith) (Garegnani 1990, 127). them to have been of secondary importance for classical authors, and at the periphery of their interests (219, 222} a criterion that, however, has nothing to do with that of method, on which my distinction rests. 4. This may incidentally explain what Joseph Schumpeter (1954, ) famously perceived as the Ricardian vice of [bundling up] as large parts [of the general system] as possible... so that [they] should be frozen as given.

6 246 History of Political Economy 34: 1 (2002) Blaug's similar charge about the classical high concern for outputs is more interesting since it brings out the need to clarify an important point in Sraffa. In the preface to his 1960 book, Sraffa famously writes, No changes in output and... no changes in the proportions in which different means of production are used by an industry are considered in this book, and he states that this was the standpoint of the old classical economists (v; emphasis added). Outputs, it may then seem, are ultimate and not intermediate data for the classical economists, since their standpoint excludes changes in such data, and therefore, in Blaug's interpretation, it excludes a determination of them within the theory. However, as the oddity of the conclusion might have led one to suspect, Sraffa 's point is a quite different one: it regards that core of the theory with which Sraffa's 1960 book is almost exclusively concerned and not the theory as a whole. Sraffa states that changes are there absent from the logical relations by which he determines the prices of each given position of the economy. 5 This is quite unlike what a neoclassical economist does when he considers demand functions for factors and supply functions of products, the specification of whose points by means of marginal products or marginal costs or any other means entails the consideration of changes in outputs and in the proportion of factors: therefore he has to include changes in the very logical relations (equations) that define each particular position of the economy. What Sraffa is saying is in effect that, like the old classical economists, he is determining prices independently of any demand and supply functions. Nothing grotesque, then, in all that (223); in particular, no denial of the classical concern for capital accumulation or output changes of any sort. 5. It may seem that this is contradicted by Ricardo's treatment of rent (224). But his extensive cent results from a difference between coexisting outputs, and not a change in them, and the same is true for the intensive rent resulting from two productive methods coexisting on the same kind of land. In this connection Blaug also lays stress on what he describes as Ricardo's moving equilibrium and the latter's concern for intertemporal rather than intratemporal comparisons of value (223-25). No doubt Ricardo was much concerned with a declining rate of profit and he studied it by the method of comparative statics, as it was later called. However, for those very intertemporal comparisons, each compared position had to be determined first in both its profit rate and intratemporal relative prices; therefore the intertemporal comparison has its logical foundation in the intratemporal ones.

7 Garegnani / Blaug on the Sraffian Interpretation 247 Wages and Outputs: The Analysis of the Old Classical Economists Blaug opens his section on wages by writing, We come... to the most disputable of all the three variables that are said to be given in classical economics (225). That given wage might, however, have seemed less disputable had Blaug recalled the following passage: The foundation of modern political economy [is] the conception of the value of labour power as something fixed, as a given magnitude (Marx [1905] 1969, 45), written by an author whom Blaug (1988, 29) describes as thoroughly steeped in the concepts and modes of thought of Ricardo. But, of course, Marx intended that wages in the modern political economy of Smith and Ricardo were given in precisely the sense of my intermediate data: he surely did not mean to exclude from economic theory the determination of wages on which he had of course much to say. Blaug proceeds, then, to contrast the alleged given wage of Sraffians with its determination by classical authors. 6 His description of a subsistence-based wage trend resulting frorn the population principle is essentially uncontroversial, and its determination, being prior to that of profits and prices, is in fact consistent with a wage treated as an intermediate datum as Blaug himself in effect explains (227). Where controversy may arise is on the key question of the mechanism by which Smith and Ricardo supposed divergences between the relative speeds of capital accumulation and population growth were corrected so as to achieve that wage trend. A difficulty, which does not seem to have been sufficiently attended to in the literature, emerges here in the attempt by Blaug and other authors to interpret that mechanism in terms of demand and supply along neoclassical lines. 6. Blaug, however, ascribes to Sraffians the view that the classical economists must have taken the real wages as a datum because the logical consistency of their theory demanded it (229). No references are given, but clearly the wage is an (intermediate) datum in the classical economists for the reasons that Blaug himself recognizes (227) logical consistency then taking natural care of itself. Blaug's point appears, however, to relate to the even more surprising opinion that the degree of freedom we find in Sraffa's equations of distribution and prices is due to his assumption of fixed coefficients of production (Blaug 1987, 441; 1988, 22). However, the entire part 3 of Sraffa 1960, and the critical literature on reswitching, have been concerned with the choice of technique, without in the least affecting that degree of freedom; surely a demand-and-supply determination of the wage is no direct result of the assumption of alternative techniques.

8 248 History of Political Economy 34: 1 (2002) Unlike the impression often given, the inverse relation between the wage and the rate of growth of capital on the one hand, and the direct relation between wage and population growth on the other, cannot replace or imply the neoclassical demand and supply. An inverse relation between the wage and the labor employment possible with a given amount of capital, that is, an elastic, proper neoclassical demand function is required for the purpose. Without it, a demand and supply mechanism along neoclassical lines, far from ensuring that wages adjust the growth of population with that of capital, would rather lead to absurd zero wages when the former growth overtakes the latter, zero quasi-rents in the opposite case, and indeterminacy when the two happened to balance. Now, no such elastic demand function can be traced in either Smith or Ricardo. 7 The nonexistence of it and of the corresponding neoclassical mechanism is made clear, for example, by Ricardo's admission of permanent labor unemployment. 8 And no confusion should here be caused by the wage-fund theory in the form it took after Ricardo, with, for instance, John Stuart Mill. The mechanism by which Smith and Ricardo envisaged wages to signal balance or unbalance in the relative growths of labor and capital appears to have been an entirely different one. It can be traced in passages on wages by those authors that, significantly enough, have been endless puzzles to modern historians. Think of Smith's observations on the existence of a minimum wage, or about combinations of masters or laborers, tacit or explicit, imparting advantages to the respective parties in the determination of what, notwithstanding all that, appear to be competitive wages. And texts of Ricardo have constituted similar puzzles (see, for example, Garegnani 2000). Those passages may be seen to point in the direction of a real wage determined by what Marshall ([1920] 1962, 82, app. J) once described (in order to reject it) as the relative strength of the competing parties. Especially in Smith 's and Ricardo's times that strength could be seen to be broadly measured by the proportion of the supply to the demand of labour (Ricardo , 7. The issue is the one that Paul Samuelson (1978, 1423) describes as..showing that the Canonical system is... globally stable. An even clearer recognition of the distinct roles of the two kinds of relations comes from Frank Knight ([1935] 1956, 81). For a fuller discussion of the issue, see Garegnani Blaug ([1958] 1973,179) notes how Ricardo assumed the existence of Marxian unemployment. Paul Samuelson's repeated attempts (e.g., 1978, 1428; 1988) to deny that Ricardo's admitted permanent labor unemployment for his conclusions in the chapter On Machinery in the Principles are discussed in Garegnani 2000.

9 Garegnani / Blaug on the Sraffian Interpretation 249 1:94) the two single quantities indicated by the word proportion used for their relation which thus specified rather than contradicted what is often indicated as a bargaining theory of wages standing alongside a demand-and-supply theory in Smith (O'Brien 1975, 111). 9 The relative speed of the growth of population and capital, by a1tering the proportion, and hence the pressure, of labor underemployment, affected the relative strength and explained changes in wages capable of adjusting population to capital accumulation, without any need to appeal to labor demand and supply functions along neoclassical lines. 10 The absence of demand and supply functions for labor is on the other hand quite naturally associated with an absence of demand and supply functions for products-and, therefore, with a treatment of outputs fundamentally different from the modern one. Demand for products influences normal or natural prices by affecting supply prices. And the neoclassical demand functions mainly do so through the relative demands for factors and therefore the levels of the wage and profit rates. 11 The classical theory of wages would therefore be sufficient to deprive neoclassical demand functions of their influence on prices via direct effects on wages. On the other hand, the effects that quantities demanded could have on supply prices through increasing physical returns from the division of labor, or decreasing physical returns from scarce natural resources, were seen by the classical economists to fall largely outside the sphere of the determination of prices, and in that of capital accumulation. Moreover, the classical theory of distribution further undercut the potential relevance of demand functions by doing without the equalities between demand and supply of productive factors determining the individual outcomes from which the demand functions for products should come. It is only natural, then, that the notion of predefined functional relations between prices and quantities demanded of products should have 9. Indeed, the way out of these puzzles adopted by modem interpreters seems mainly to consist of tracing several conflicting theories of wages in Smith, in whose work the position of the old classical economists is more fully articulated. 10. A fuller discussion of the question is in Garegnani 2000; see also Garegnani 1990, Contemporary economists have of course been reminded of this by the nonsubstitution theorem.

10 250 History of Political Economy 34: 1 (2002) remained foreign to Smith and Ricardo, where they would have had no clear general determining role on prices or, even, a definite basis in individual incomes. The several circumstances on which outputs depend (general level of activity, technical conditions, distribution of income, individual choices, etc.) then faced the classical economists with the complexity and dependence on circumstances characteristic of the second of the two fields of analysis we distinguished earlier and thus, essentially, led them to taking normal outputs as intermediate data in the determination of prices in the core. That, of course, also did without any supply functions of products: as we said returns to scale only became relevant when output changes were considered outside the core where they could be dealt with according to the circumstances of the case. In particular, there was no need to assume the constant return in manufactures or the rigid demand for corn that Marshall read into Ricardo's work in his attempt to reconcile it with his own demand and supply apparatus. With this we have come full circle to see how far from taking an amazingly narrow view of the classical economists (215, 233) the Sraffian interpretation uncovers in their works the roots of their sociological richness. Blaug rightly stresses that richness, but he does not explain its presence there and its absence in modem works. The intermediate data and the reasoning by stages that classical works entail respond, on the other hand, in a striking fashion to the need for short chains of deductive reasoning and for specific experience, which Marshall ([1920] 1962, 773) saw to be required by the subject matter of economics the same need he tried to satisfy by his method of partial equilibria. We are now also in a position to appreciate how the absence of demand and supply functions for products in Smith and Ricardo, far from being a sign of primitivism or confusion, should rather be seen as part of an alternative, consistent view of how the economy works. High among the advantages of this view apart from freedom from the basic difficulties to be presently recalled is its independence from the assumption of reversibility of economic changes that underlies demand and supply functions. And it is well known how that assumption, which a perceived lack of alternatives seems to have induced the majority of the profession

11 Garegnani / Blaug on the Sraffian Interpretation 251 to forget, or forgive, was worrying to Marshall. He knew well that in economics every event causes permanent alterations in the conditions under which future events occur ([1879] 1930, pt. 2, 12): Smith, Ricardo, and their predecessors had instinctively taken good care of that. Conclusion Before concluding, we may briefly follow Blaug from history into theory. It is not difficult to see now that Blaug cannot be correct in commenting on how thoroughly Sraffa is steeped in Walrasian general equilibrium theory (229). Both Sraffa and Walras, it is true, deal simultaneously with prices in all sectors of the economy, but the equations they advance have little else in common. Sraffa's equations are simply the algebraic expression of what Blaug himself says classical theory focuses on, namely, long-period... prices characterized by a uniform rate of profit on capital,... in short, what Smith called natural prices (233) quite a different matter from Walras's equations, further characterized by the equalities between demand and supply of productive factors. Failure to perceive those differences becomes evident in the section on competition of Blaug s article. I strongly sympathize with his stress on the decisive importance of showing that prices and outputs converge on an end-state (229). That question assumes, however, a very different aspect in the two approaches, as a result, essentially, of their different explanations of the wage. In classical theory, given the stability that the wage derives from its socio-institutional determinants, the competitive tendency to uniform rates of wages, profits, and rents will generally suf - fice to ensure a tendency to the normal prices of the theory. 12 The problems lie on the neoclassical side, where uniqueness and stability come to depend on the shape of labor demand and supply functions. It comes 12. See Garegnani Differing positions on the matter are also discussed there (140, , ). However, Blaug writes that in Sraffa not a word is wasted on telling us how we got [to the competitive prices he supposes] (231); Sraffa's reference to the distinction between market price and natural price (1960, 9) should, however, have been quite unambiguous in referring readers used to classical terminology to Smith s analysis of competition. Moreover, Blaug contrasts the perfect competition he believes I uphold with the classical notion of free competition (231). However, the perfectly competitive horizontal demand function facing the firm is a (Marshallian) descendant of Marshallian demand and supply function for the industry, and is therefore unlikely to find place in any contemporary revival of classical economics.

12 252 History of Political Economy 34: 1 (2002) therefore to depend on the difficulties besetting the conception of capital as a factor of production, difficulties that are inexistent in a classical context (as, for that matter, are the analogous difficulties caused by income effects). Our remarks on theory have brought out, I believe, that same difficulty of recognizing the existence of an alternative theoretical approach in the old classical economists, which, we suggested at the beginning, may explain the contradictions between Blaug's criticism of Sraffa's interpretation and his acknowledgment of basic elements of it. Thus we have noticed Blaug's recognition of labor unemployment in Ricardo and his references to Ricardo's com model. But perhaps the most striking contradiction is found when Blaug has to admit in Ricardo (and Smith) the very absence of demand and supply functions for products, which he describes as a grotesque distortion when noted by Sraffians. The distortion would consist in holding that any appeal to the forces of demand and supply in determining prices is necessarily alien to classical economics and that classical natural prices have nothing whatsoever in common with Marshall's long run normal prices (Blaug 1987,442; emphasis added), where, as the reference to Marshall makes clear, the forces of demand and supply denied by Sraffians are the neoclassical ones and not the quite different classical concepts that we shall presently mention. A few lines after that passage Blaug has, however, to notice that Ricardo (and Marx after him) propagated the misleading idea that demand-and-supply explanations only pertain to market prices [and, moreover, there was] Ricardo's tendency to think of demand and supply as quantities actually bought and sold not as schedules of demand and supply prices. (1987, 442; emphases added) A confusion, Blaug concludes. 13 The misleading idea is there mentioned rather incidentally and is confined to Ricardo and Marx. But, as Blaug surely knows, the notion that demand-and-supply explanations only pertain to market prices that is, to deviations of the actual prices from natural or normal prices 13. Blaug's quantities bought and sold in this passage is clearly a slip, since they would be one and the same, and not Smith's and Ricardo's effectual demand and quantity brought to market.

13 Garegnani / Blaug on the Sraffian Interpretation 253 and not to the prices themselves-is no peculiarity at all of Ricardo and Marx: it was Adam Smith who had first treated the idea quite systematically in the whole of chapter 7, book 1, of the Wealth of Nations, titled Of the Natural and Market Price of Commodities. And it is precisely from there that Ricardo also got his confusion concerning demand and supply as single quantities, demand being the [effectual] demand of those who are willing to pay the natural price of the commodity (Smith [l776] 1950, 1:49), supply being the quantity brought to market. It is then difficult to see where lies the grotesque distortion of maintaining that Marshall's forces of demand and supply were alien to Smith, Ricardo, and Marx, and in particular that they were alien to their determination of natural prices. Do not the texts of those authors argue just that? Blaug also writes: Ricardo had no theory of... why output is made up of one set of goods rather than another. But then neither did any other premodem economist (223), which means, I believe, that Ricardo and other premodern economists did not determine the composition of outputs by demand and supply functions of products. How did they treat outputs then, if not as intermediate data? References Blaug, M. [1958] Ricardian Economics. Westport, Conn.: Greenwood Press Classical Economics. In The New Palgrave Dictionary of Political Economy, edited by J. Eatwell, M. Milgate, and P. Newman. London: Macmillan Economics through the Looking Glass. London: Institute of Econornic Affairs Misunderstanding Classical Econornics: The Sraffian Interpretation of the Surplus Approach. HOPE 31.2: Edgeworth, F. [1881] Mathematical Psychics. New York: Kelley. Garegnani, P A Problem in the Theory of Distribution from Ricardo to WickseIl. Ph.D. diss., Cambridge University Surplus Approach to Va1ue and Distribution. In The New Palgrave Dictionary of Political Economy, edited by J. Eatwell, M. Milgate, and P. Newman. London: Macrnillan Sraffa: Classical versus Marginalist Analysis. In Essays on Sraffa, edited by K. Bharadwaj and B. Schefold. London: Routledge On Some Supposed Obstacles to the Tendency of Market Price Towards Natural Price. In Equilibrium and Economic Theory, edited by G. Caravale. London: Routledge Samuelson on Sraffa's Hits and Misses. Unpublished manuscript.

14 254 History of Political Economy 34: 1 (2002) Knight F. H. [1935] The Ricardian Theory of Production and Distribution. In On the History and Method of Economics. Chicago: University of Chicago Press. Marshall, A. [1879] The Pure Theory of Foreign Trade, the Pure Theory of Domestic Values. Clifton, N.J.: Kelley.. [1920] Principles of Economics. London: Macmillan. Marx, K. [1905] Theories of Surplus Value. Vol. I. London: Lawrence and Wishart. O Brien, D. P The Classical Economists. Oxford: Clarendon Press. Ricardo, D The Works and Correspondence of David Ricardo. Edited by P. Sraffa, with the collaboration of M. H. Dobb. Cambridge: Cambridge University Press. Samuelson, P. A The Canonical Classical Model of Political Economy. Journal of Economic Literature 16 (December): Mathematica1 Vindication of Ricardo on Machinery. Journal of Political Economy April: Schumpeter, J. A History of Economic Analysis. Oxford: Oxford University Press. Smith, A. [1776] The Wealth of Nations. 2 vols. London: Dent and Sons. Sraffa, P Production of Commodities by Means of Commodities. Cambridge: Cambridge University Press.