1.1 – Capital in the theory of international trade

6The first post-Sraffian analyses of international trade were undoubtedly a joint product of the controversies on capital theory. In the preface to Fundamental Issues in Trade Theory, an edited volume which brought together the main contributions of the neo-Ricardian authors, one reads that

7

the capital controversies of the 1960s reminded economists that […] there are intrinsic logical difficulties in the notion of “capital” as a “factor of production”. Since that notion has played a significant role in the dominant theory of international trade in recent decades, it is not surprising that in the late 1960s and throughout the 1970s a number of writers have sought to trace through the implications of the capital debates for orthodox trade theory and, where necessary, to provide some new trade theory.

8In the theory of international trade the main target was obviously the Heckscher-Ohlin-Samuelson model, the demonstration of which makes use of production functions with aggregate capital. In his introductory essay to Fundamental Issues in Trade Theory, Steedman wrote concerning this model: “it is more important here to consider further the treatment of production within that theory, for not only is that treatment open to criticism but such criticism also suggests the lines along which one can start to create a trade theory with an alternative ‘vision’, a theory which is concerned with growth and accumulation” (Steedman, 1979b, p. 3). The aim was twofold: on the one hand, questioning the foundations of the Heckscher-Ohlin-Samuelson theory; and on the other, providing the basis for a different approach. Steedman quickly provided the first elements of this new approach in another book, Trade Amongst Growing Economies (Steedman, 1979c). However, in this book he made use of formalizations explicitly referring to von Neumann, and not to Sraffa. Likewise, it must be noted that among the sixteen articles collected in Steedman (1979a), most of which had previously been published, only Mainwaring (1976 and 1978), Metcalfe and Steedman (1973) Parrinello (1973) and Steedman and Metcalfe (1973a) explicitly referred to Sraffa (1960).

9Though contributions by neo-Ricardian authors were discussed at the time within the context of criticism of the standard theory of international trade, they were never really considered later because their attempts to rebuild the theory of international trade had failed to convince supporters of the neoclassical theory. For example, while assessing Steedman’s two books, Avinash Dixit noted that:

10

Steedman provides a clear statement of their [the volume’s contributors’] aims. He begins by criticising the usual static two-by-two model for treating capital as a scalar input made exogenously available in fixed quantity. According to Steedman, a proper treatment must recognize three features: the heterogeneity of capital goods, the fact of their being produced means of production, and the time needed for such production. It is clear that on all three counts the Heckscher-Ohlin-Samuelson model of elementary textbooks must plead guilty.

11Dixit also noted that though Steedman’s statement (1979c) is “simple, clear, excellently organized, and integrates geometric and algebraic methods really well. […] The model is very special. While it serves to illustrate some general features of the approach, some of the techniques of analysis, as well as some of the detailed results, will not survive generalization” (ibid., pp. 280-81). Furthermore, Dixit rejected Steedman’s approach, stressing that “models which confine themselves to steady states are […] simply inappropriate for answering real-world policy questions” (ibid., p. 182). His views on Steedman (1979a) were even more critical, remarking that “some of the general models […] are well-expounded, others are not,” and adding: “some others papers, however, spend an inordinate amount of time deriving and discussing the wage-interest frontier, first without input substitution, then with a choice of activities, etc. Surely most of this is standard stuff by now. The collection would have been much more readable if the papers had been revised to give a streamlined exposition, and that could have been done much more simply using unit cost functions” (ibid., p. 281). Dixit concludes that “Steedman and associates have performed a most valuable service to the community of international trade theorists by alerting them to serious defects of the static two-by-two model that need to be remedied if it is to be useful in important dynamic contexts. The questions they have posed, and their analysis, have opened up useful avenues for further research. But I do not think success lies in the neo-Ricardian direction” (ibid., pp. 293-94).

12In his survey of “capital theory and trade theory,” Alasdair Smith expressed the same feeling: “the ‘neo-Ricardians’’ objective seems to have been to make a radical revision of, rather than an extension to, the theory of international trade” (Smith, 1984, p. 318). After discussing contributions published in Steedman (1979a), he asked: “what then of the neo-Ricardian contribution? There is no doubt that it is substantial and useful, and at least some of the ‘neoclassical’ theoretical developments surveyed above would have been different and poorer without the stimulus provided by the neo-Ricardians. On the grand question of whether we really are given an alternative ‘vision’ of international trade, only a personal view can be expressed and my answer is negative” (ibid., pp. 319-20). Smith added: “on the positive-economic side, all the neo-Ricardian analysis can be interpreted from a ‘mainstream’ viewpoint as variants of or extensions to the ‘mainstream’ theory, but the converse statement cannot be made” (ibid., p. 320).

13Although these comments recognize the relevance of the neo-Ricardian critique, they minimize its impact on the theory of international trade: “the limitations of the static two-by-two textbook model do not arise solely from its neglect of capital. There are other purposes for which it has been augmented or altered, while preserving its character. The specific factor model is perhaps the most eminent example, but there are several others. The criticisms of Steedman et al. do not destroy the usefulness of such models in the appropriate contexts” (Dixit, 1981, p. 294). Moreover, since 1979, international trade theory has taken a new direction by studying the implications of increasing returns and imperfect competition (Helpman, 1984). In making a connection with industrial organization (Krugman, 1989), this new field dismissed the questions raised by neo-Ricardian models. It followed that the debate was restricted to post-Sraffian authors who discussed issues raised by the extension of Sraffa’s price system to the analysis of international economic relations.

1.2 – The extension of Sraffa’s price system to international economics

14The main difficulty in applying Sraffa’s model to international trade is the determination of international prices. Indeed, if one considers the simplest case of two economies of the same size, producing the same two basic commodities without fixed capital, one only has to suppose that they differ in their productive structure and (or) in the structure of their income distribution for their relative prices to be different. This constitutes sufficient ground for the establishment of an international exchange; however, nothing ensures that it is beneficial for both countries (Mainwaring, 1974). Then, with an assumption of constant returns to scale, it becomes possible to replace the two autarkic systems by a new “international” system describing the complete specialization of each economy in one of the two commodities. However, without any additional hypothesis, such a system does not determine international prices since, unlike situations of autarky, there is no reason why distributive variables should be identical in the two countries after the specialization (Montet, 1979).

15To remove this indeterminacy, the first solution was to introduce an assumption of steady growth at a constant rate for each economy (Parrinello, 1973; Mainwaring, 1974; Steedman, 1979c). This procedure needs to add to the price system of each economy a system of equations defining the quantities, assuming that they depend on the rate of growth and on consumption levels of both commodities. For the system defining the activity levels to be consistent with the corresponding price system, neo-Ricardians have generally retained a traditional savings function, specific to each country, in which the rate of growth depends only on the rate of profit through the capitalists’ propensity to save. To avoid one of the economies becoming too large, this approach requires the assumption that the two economies are growing at the same rate, thus preventing a complete specialization of each economy; and it also requires that the saving functions do not change with trade openness (Parrinello, 1973; Steedman, 1979c). [2] Although this solution solves the indeterminacy of international prices, it has nevertheless two important ambiguities with respect to the Sraffa analysis. The first is that, contrary to the Sraffian logic, the prices thus obtained depend mostly on consumer preferences, since they are based on the propensities to save and consumption patterns characterizing each economy. The second ambiguity is related to the assumption of equal growth rates that results from introducing a relationship between the two countries. This relationship comes prior to international trade, since the given propensities to save necessarily imply that the two economies are characterized by a stable differentiation in their rate of profits (Ravix, 1990).

16Subsequently, a second solution was proposed—first introduced by Ghislain Deleplace (1976)—which excluded the hypothesis of steady growth in favour of the notion of the “branche fictive.” This solution is connected with Sraffa’s approach (1960), according to which, when prices and wages are measured in terms of the Standard commodity, the wage–profit relationship becomes linear. Then, the maximum rate of profit or “Standard ratio” can be a parameter, defined as a solution of the “Standard system,” because it only depends on the technical conditions of production (Sraffa, 1960, Ch. 4; Deleplace, 1976). However, the application of this method in international trade is problematic since, because of specialization, the Standard ratio in each country is different before and after free trade takes place. Moreover, in the case of an open economy, the calculation of the Standard ratio after specialization may not be the same as in a closed economy, because the production structure of the country is reduced to a single industry. To resolve this problem, Deleplace introduced into the Standard system a “fictitious industry” in each country, after complete specialization. This “fictitious industry” simply reflects the fact that the quantity of goods that the country no longer produces, but still uses as a means of production, must be obtained in exchange for the amount of the other product which is required by the second country. This procedure is then used to show that, after specialization, “the rate of profit and prices are determined solely by the production methods, depending on the distribution of the value of the net product regardless of the consumer preferences” (Deleplace, 1976, p. 53). However, such a result is not satisfactory since, in an open economy, the means of production are evaluated on the basis of prices expressed in different and incommensurable standards (Montet, 1979). [3]

17This notion of “fictitious industry” was taken over by Gilbert Abraham-Frois (1981), who gave it a different content. Indeed, the idea was to complete the price system of each country, after specialization, by a particular industry which ensured the supply of the foreign commodity, assuming that the latter requires a certain amount of labour and that there is “no reason for the capital invested in the international exchange to receive different remuneration than in other industries” (Abraham-Frois, 1981, p. 457). Then, the particularity of this fictitious industry lies in the fact that the production coefficient characterizes itself as the ratio of the prices of the two commodities that balance the international market. Applied to the international economy, this procedure leads to two systems of national prices. Both of them include “two kinds of prices: domestic production prices in an open economy […] and international relative price” (Maurisson, 1986, p. 51). But this feature is purely formal, since domestic production prices depend largely on the international equilibrium price which directly affects the conditions of production of fictitious industries in each country. It follows that, on this approach, domestic prices follow the same logic as the international equilibrium price.

18In order to correct this last difficulty, and to resolve the logical problem raised by Deleplace’s model, Patrick Maurisson (1986) proposed to consider this fictitious industry as a branch of foreign trade, without making any particular hypothesis about the quantities, nor fixing prices in terms of the Standard net product. This solution allows us to determine price systems at the international level which, because they are autonomous, ensure the determination of real domestic production prices in open economies. However, nothing ensures that these systems really describe specialized economies that are engaged in trade, and therefore the very idea of international trade vanishes.

19Thus, it appears that these attempts either leave us incapable of explaining international trade, or force us to abandon, explicitly or implicitly, Sraffa’s theory of prices in favour of prices determined by supply and demand. The fundamental reason for this paradox comes from the fact that the non-equalization of distributive variables prevents the determination of international production prices. It is indeed easy to check that the indeterminacy disappears as soon as this assumption is abandoned, but in this case it becomes impossible to distinguish between a closed economy and a fully integrated global economy. Of course, it is always possible to admit, as Antoine Delarue (1975) does, that such an economy exists: it is also the only solution that really guarantees that international prices are actually production prices. However, in this case, it becomes impossible to explain the formation of such a global economy, since the specializations are given in advance (Ravix, 1990). To overcome this last difficulty, one should be able to demonstrate that in closed economies, characterized by differentiated distributive variables, the transition to international trade through specialization leads to a situation in which these variables are equalized. Unfortunately, such a proposal is unverifiable in a post-Sraffian approach (Steedman and Metcalfe, 1973a; Mainwaring, 1976 and 1978).

1.3 – The post-Sraffian approach of the Ricardian theory of international trade

20Following up on their analysis of the Heckscher-Ohlin-Samuelson theory, post-Sraffian authors also criticized the neoclassical interpretation of the Ricardian theory of international trade. So, Ian Steedman disputes “what is commonly, if unfortunately, called ‘Ricardian’ trade theory in the textbooks, for that theory generally ignores the role of produced means of production, proceeds as if the rate of profit were zero, ignores the choice between alternative techniques of production and says nothing concerning growth” (Steedman, 1979c, p. 14). More generally, Takashi Negishi reported that “the wage rate in the neoclassical interpretation of Ricardian theory is the quasirent imputed to the given supply of labour, which is determined exclusively by the demand of labour, hence by the productivity of labour. This is quite un-Ricardian” (Negishi, 1985, p. 132). He can then assert that

21

the neoclassical interpretation cannot explain the implication of Ricardo’s numerical example in which labour productivity in the cloth industry is lower in England, which seems to be the more advanced country, than in Portugal, which seems to be the less-developed country, since it is strange to consider that the wage is lower in the former than in the latter. It is, however, not the wage but the rate of profit which should be considered to lower in England, while the natural wage may not be lower in England.

22Yet, in their attempts to reformulate the foreign trade analysis developed by Ricardo, post-Sraffian authors provide no real reinterpretation of Classical thought in this area. In particular, Steedman and Metcalfe recall that,

23

basing his analysis on the assumption that no-trade price ratios will equal the corresponding ratios of quantities of embodied labour, Ricardo obtained, amongst others, the following three important conclusions regarding foreign trade. First, trade will increase the rate of profits in a country, if and only if the imported commodities enter the real wage in that country. Secondly, the pattern of trade will be determined entirely by the methods of production available in the several trading countries. Thirdly, trade will ‘increase the mass of commodities, and therefore the sum of enjoyments’ in each of the trading countries”.

24In adopting a Sraffian reformulation, they show that “only the first of the three conclusions mentioned above is independent of the relation between no-trade price ratios and embodied labour ratios; in general, the pattern of trade could not be predicted by Ricardo from a knowledge of production methods alone and nor could Ricardo have shown that trade must increase the ‘mass of commodities’ available in each trading country” (ibid., p. 100).

25According to Negishi (1982 and 1985), another interpretation is possible: namely, to consider that capital is not absent from the Ricardian analysis, because in an open economy “the reason given by Ricardo for this unequal exchange of labour is the difficulty with which not the labour but the capital moves from one country to another” (Negishi, 1985, p. 135). In this new interpretation, Ricardo’s example is considered as a special case in which the capital advanced to produce each commodity would consist of labour whose wages may be defined in terms of amounts of the two commodities. On this basis, Negishi developed the idea that, although Ricardo assumes that capital does not move from one country to another, it can nevertheless be admitted that some international capital movements would be possible, if only to organize import and export operations. Indeed, Ricardo wrote: “when merchants engage their capitals in foreign trade, or in the carrying trade, it is always from choice, and never from necessity: it is because in that trade their profits will be some-what greater than in the home trade” (Ricardo, 1817, p. 293). According to Negishi, the whole problem is then to define the minimum rate of profit necessary to prevent capitalists transfering their capital abroad, or to define a “rate of conversion” between profit rates in each country. With this additional relationship, Negishi can conclude: “By specifying the rate of conversion by which home capitalists discount a higher rate of profit in the foreign country, then, Ricardian theory can determine the terms of trade on the basis of cost-price relations without having recourse to reciprocal demands” (Negishi, 1982, p. 208).

26Though it is more consistent with Ricardo’s text than the so-called “Ricardian” neoclassical model, Negishi’s interpretation does not differ very much from the neo-Ricardian approach developed by Mainwaring, Steedman and Metcalfe, or by Parrinello. Indeed, in both cases, the determination of international prices takes place through a simple linear relationship in which an element is set exogenously: either a rate of direct conversion between the rates of profit; or the propensity of the capitalists to save, thus guaranteeing the equality of the steady state rate of growth in each country. However, this kind of solution is based on an approach that is not fundamentally different from that which is at work in the neoclassical theory. Both of them define the proper procedure to circumvent the lack of international competition, stated at the beginning of the analysis, so as to describe a situation similar to that which would prevail if there were international capital mobility.

2.1 – The Classical debates on the gains from international trade

28The origin of the Classical analysis of international trade lies in Adam Smith’s Wealth of Nations. Two passages are particularly important. The first is in Chapter 9 of Book 1, Of the Profit of Stock: according to it, for a stock of capital which remains unchanged, international trade will reduce the competition amongst capitalists by offering them new markets. This will lead to an increase in the general rate of profit:

29

The acquisition of new territory, or of new branches of trade, may sometimes raise the profits of stock […]. The stock of the country not being sufficient for the whole accession of business, which such acquisitions present to the different people among whom it is divided, is applied to those particular branches only which afford the greatest profit. […] In all those old trades, therefore, the competition comes to be less than before. The market comes to be less fully supplied with many different sorts of goods. Their price necessarily rises more or less, and yields a greater profit to those who deal in them.

30The second passage where Smith deals with the advantages derived from international trade is in Chapter 1 of Book 4, Of the Principle of the Commercial, or Mercantile System. There Smith explains his “vent for surplus” theory (Myint, 1958), according to which international trade enables countries to sell the surplus of the goods they produced in excess for the home market. In addition to this “vent for surplus,” international trade leads to an increased international division of labour. The effects of this increased division of labour are the same as those experienced in a single country and described in the famous chapter Of the Division of Labour: an increase in the productivity of the means of production, in the amount of commodities produced and in the revenue of the country. In other words, in addition to his “vent for surplus” theory, Smith also developed a “productivity doctrine” (Myint, 1977) according to which international trade leads to a better division of labour and therefore increases wealth:

31

Between whatever places foreign trade is carried on, they all of them derive two distinct benefits from it. It carries out that surplus part of the produce of their land and labour for which there is no demand among them, and brings back in return for it something else for which there is a demand. It gives a value to their superfluities, by exchanging them for something else which may satisfy a part of their wants, and increase their enjoyments. By means of it, the narrowness of the home market does not hinder the division of labour in any particular branch of art or manufacture from being carried to the highest perfection. By opening a more extensive market for whatever part of produce of their labour may exceed the home consumption, it encourages them to improve its productive powers, and to augment its annual produce to the utmost, and thereby to increase the real revenue and wealth of the society.

32Smith’s first argument, according to which international trade may increase the general rate of profit by reducing competition, seems to have been overlooked by most economists and scholars, with the exception of Ricardo. [6] This may be explained by the fact that Smith’s argument dealt with the acquisition and the exploitation of new territories—such as North America and the West Indies—more than with a trade between independent and equally powerful nations. By contrast, Smith’s “vent for surplus” theory was much discussed; [7] J. S. Mill considered it a “surviving relic of the Mercantile Theory” (J. S. Mill, 1848, vol. 2, p. 592). James Mill opposed this doctrine in his 1808 pamphlet Commerce Defended. In this pamphlet, Mill developed for the first time his law of market according to which “the production of commodities creates, and is the one universal cause which creates a market for the commodities produced” (Mill, 1808, p. 81), and argued that “foreign commerce, therefore, is in all cases a matter of expediency rather than of necessity. The intention of it is not to furnish a vent for the produce of the industry of the country, because that industry always furnishes a vent for itself” (ibid., p. 86).

33Ricardo was more moderate than the “Mills”; but he nevertheless observed that this thesis was in contradiction with Smith’s own analysis of the gravitation of market prices around natural prices, and with the mobility of capital within the country which Smith had so ably described:

34

One would be led to think by the above passage [Smith vol. 1, p. 372], that Adam Smith concluded we were under some necessity of producing a surplus of corn, woollen goods, and hardware, and that the capital which produced them could not be otherwise employed. It is, however, always a matter of choice in what way a capital shall be employed, and therefore there can never, for any length of time, be a surplus of any commodity; for if there were, it would fall below its natural price, and capital would be removed to some more profitable employment. No writer has more satisfactorily and ably shewn than Dr. Smith, the tendency of capital to move from employments in which the goods produced do not repay by their price the whole expenses, including the ordinary profits, of producing and bringing them to market.

35While the “vent for surplus” doctrine was largely rejected by the Classical economists, Smith’s “productivity doctrine” was generally adopted. This was the case in 1808 when both James Mill and Torrens published pamphlets designed to refute Spence and Cobbett’s protectionist position on international trade.

36In Commerce Defended, Mill developed the Smithian idea that the “use and advantage of [foreign commerce] is to promote a better distribution, division and application of the labour of the country than would otherwise take place, and by consequence to render it more productive” (Mill, 1808, p. 86). This “productivity” argument was illustrated by an example showing how much Great Britain could increase her wealth by producing corn and hardware and exchanging them against a greater quantity of flax than she would be able to produce with the same labour and land:

37

Now commerce tends to increase this annual produce by occasioning a more productive application and distribution both of the land and of the labour of the country. Instead of raising flax, for example, or hemp, on our land, we raise corn; with that corn we feed a number of hardware manufacturers, and with this hardware we buy a greater quantity of flax than the land which raised our corn, and fabricated our hardware, would have produced.

38Torrens developed an original view in The Economists Refuted, differing from Mill on several important analytical points. [9] However, his defence of commerce runs along the same lines as Mill’s pamphlet: “commerce, by establishing divisions of labour between the individuals of different nations, is a mean of augmenting wealth” (Torrens, 1808, p. 34). This he exemplified by illustrating the gain that England would derive from a trade with France:

39

The only way, therefore of ascertaining the amount of the benefit derived from commerce, is to ascertain the degree in which the foreign division of labour augment the productiveness of human industry. Thus, if I wish to know the extent of the advantage, which arises to England, from her giving France a hundred pounds worth of broad cloth, in exchange for a hundred pounds worth of lace, I take the quantity of lace which she has acquired by this transaction, and compare it with the quantity which she might, at the same expense of labour and capital, have acquired by manufacturing it at home. The lace that remains, beyond what the labour and capital employed on the cloth might have fabricated at home, is the amount of the advantage which England derives from the exchange.

40Both James Mill and Torrens thus admitted Smith’s “productivity doctrine.” Both also agreed that the benefit of this trade be estimated for any country by the additional quantity of commodities that international trade provided to this country in comparison to the commodities this country would have been able to produce in autarky. However, neither Mill nor Torrens established a theory of the determination of international prices and therefore, though the gain from international trade was said to be “always reciprocal” (Torrens, 1808, p. 34), “no exact estimate can be made of what any nation gains by commerce” (Mill, 1808, p. 106).

2.2 – The discovery of the comparative advantage theory

41In 1821 James Mill published his Elements of Political Economy, which ran through three editions (1821, 1824 and 1826). Chapter III, entitled “Interchange,” included a section devoted to foreign trade where Mill dealt with “occasions on which it is the interest of Nations to exchange commodities with one another” (Mill, 1821, p. 83). After having reviewed “obvious cases” such as “commodities [that] can only be produced in particular places” (ibid.), Mill moved on to “another cause which requires rather more explanation” and broached the issue of comparative advantages.

42Mill compared two situations where Poland and England both produce cloth and corn. In the first case, Poland has an absolute advantage in the production of cloth and of corn but no comparative advantage: “the cloth and the corn, each of which required 100 days’ labour in Poland, required each 150 days’ labour in England” (ibid., p. 85). In the second situation, Poland has an absolute advantage in the production of cloth and corn but England has a comparative advantage in the production of cloth: “while the same quantity of cloth which, in Poland, is produced with 100 days’ labour, can be produced in England with 150 days’ labour, the facts are such, that the corn which is produced in Poland with 100 days’ labour, requires 200 days’ labour in England” (ibid., p. 85).

Case 1: Days’ labour required for producing a given quantity of Corn Cloth Poland 100 100 England 150 150 Case 2: Days’ labour required for producing a given quantity of Corn Cloth Poland 100 100 England 200 150

43In the first case, Mill concluded that “no exchange would take place” because

44

The cloth of 150 days’ labour in England, if sent to Poland, would be equal to the cloth of 100 days’ labour in Poland: if exchanged for corn, therefore, it would exchange for the corn of only 100 days’ labour. But the corn of 100 days’ labour in Poland was supposed to be the same quantity with that of 150 days’ labour in England. With 150 days’ labour in cloth, therefore, England would only get as much corn in Poland as she could raise with 150 days’ labour at home; and she would, on importing it, have the cost of carriage besides.

45In the second case, when England has no absolute advantage but a comparative one in the production of cloth, “an adequate motive to exchange would immediately arise.” Indeed:

46

With a quantity of cloth which England produced with 150 days’ labour, she would be able to purchase as much corn in Poland as was there produced with 100 days’ labour; but the quantity, which was there produced with 100 days’ labour, would be as great as the quantity produced in England with 200 days’ labour. England, therefore, would obtain her corn with less labour, through the medium of her cloth.
Poland would profit in the same manner. A quantity of corn which cost her 100 days’ labour, being equal to the quantity produced in England by 200 days’ labour, would purchase, in England, the produce of 200 days’ labour in any other commodity; for example, in cloth. But the produce of 150 days’ labour in England in the article of cloth, is equal to the produce of 100 days’ labour in Poland. If, with the produce of 100 days’ labour, she can purchase, not the produce of 150, but the produce of 200, she gains to the amount of 50 days’ labour; in other words, a third.

47As several scholars (Pennington, 1840, pp. 35-36; Einaudi, 1930, p. 166; Sraffa, 1930, p. 540) stressed, James Mill’s reasoning contains an error. Indeed, he attributed to both countries the whole gain of the exchange. While focusing on England, Mill supposed that the international exchange ratio of corn for cloth was one for one (that is, the Polish internal rate of exchange in autarky). Therefore, the whole gain derived from international specialization benefited England. However, in the second paragraph, where he focused on Poland, Mill supposed that the international rate of exchange of corn to cloth was one to one-and-a-third (that is, the English rate of exchange in autarky). In this case, the whole gain benefited Poland. [10] Of course, such a “double” international price was impossible, and Mill’s calculation of the gain derived from international specialization was faulty.

48Mill corrected this error in 1826. In the third edition of his book, he indicated that if the exchange of corn for cloth was made at the Polish autarkic price (one for one), “the whole of the advantage would be on the part of England; and Poland would gain nothing, paying as much for the cloth she received from England, as the cost of producing it for herself” (Mill, 1826, p. 121). On the other hand, if the English autarkic relative price was adopted as the international rate of exchange, it is Poland which “would obtain the whole of the advantage” (ibid., p. 122). Mill finally adopted a middle way between these two extremes and concluded that “the result of competition would be to divide the advantage equally between them” (ibid.).

49Scholars have formulated different hypotheses on how this correction happened. Einaudi (1930) conjectured that it may have originated from James Pennington, who might have made the correction “on the occasion of some scientific debate” at the Political Economy Club in 1824 or 1826 (Einaudi, 1930, p. 170). Einaudi’s thesis followed the authority of Robert Torrens who had affirmed, in the preface to the second edition of his Principles and Practical Operation of Sir Robert Peel’s Act of 1844 (1857), that the “error was corrected by Mr. Pennington; and the correction, if I rightly remember, was adopted by Mr. James Mill in the third edition of his Principles of Political Economy” (Torrens, 1857, pp. xv-xvi).

50However, the “Pennington hypothesis” was unfounded. Indeed, as Sraffa demonstrated, Torrens’s allusion to Pennington did not refer to any debate at the Political Economy Club but to a pamphlet Pennington published in 1840—that is, 14 years after the third edition of Mill’s Elements. This pamphlet, A Letter to Kirk-man Finlay, Esq., on the Importation of foreign Corn, and the Value of the Precious Metals in Different Countries, contained

51

A very able tho belated criticism […] of James Mill’s fallacy; and […] also the conclusion that the ratio of interchange will depend upon the strength of the demand of each country for the other country’s product (pages 40-41). The long quotations from the Elements given by Pennington make it clear that he was using the second edition [1824], in which the argument of the first is repeated with merely verbal alterations; and that he was quite un-aware that the error had already been corrected in a third edition.

52To explain Torrens’s (and after him Einaudi’s) mistake, Sraffa recalled that he (they) may have ignored the 1826 edition of Mill’s Elements, the sales of which were so poor that the remaining copies were used in a 1844 edition with a new title page. [11] It is then possible that “Torrens was misled by one of these copies, which bear the imprint of 1844, when he suggested that in the third edition the correction had been adopted from Pennington’s pamphlet of 1840” (Sraffa, 1930, p. 543). However, according to Sraffa himself, a more likely explanation is that Torrens just ignored the date on which the third edition of Mill’s Elements was published, and that on this point “his only source was J. S. Mill’s essay of 1844” (Sraffa, 1930, p. 543). Indeed, in the first of his Essays on Some Unsettled Questions in Political Economy (1844) J. S. Mill had asserted that in the third edition of the Elements his father had corrected the mistake he had made in the two former editions. However, J. S. Mill did not specify the publication date of this third edition.

53However, Sraffa went further than just demonstrating that Pennington cannot be considered as the one who first pointed out the error contained in the first two editions of James Mill’s Elements. He also demonstrated that the correction did not come from James Mill either. The proof emerged from J. S. Mill’s Autobiography. In a passage alluding to the meetings of his study group which started in 1825, J. S. Mill wrote:

54

Our first subject was Political Economy. We chose some systematic treatise as our text-book; my father’s ‘Elements’ being our first choice. One of us read aloud a chapter, or some smaller portion of the book. The discussion was then opened, and any one who had an objection, or other remark to make made it […] The theory of International Values which I afterwards published, emanated from these conversations, as did also the modified form of Ricardo’s theory of Profits, as laid down in my Essay on Profits and Interest.
The theories of International Values and of Profits were ex-cogitated and worked out in about equal proportion by myself and Graham […] I may mention that among the alterations which my father made in revising his Elements for the third edition, several were grounded on criticisms elicited by these Conversations; and in particular, he modified his opinions (though not to the extent of our new speculations) on both the points to which I have adverted.

55Thus, as Sraffa concluded, it was not James Mill who first identified the error regarding the gain of the exchange in the first two editions of his Elements, but his son John Stuart along with George Graham.

56Finally, Sraffa’s intervention in this debate on “who first corrected the error” contributed to clarifying another question where J. S. Mill, Torrens and Einaudi all appear to have been misled. This second question was “who first committed the error.” According to J. S. Mill, the error first appeared in Ricardo’s Principles (1817) and was then repeated by James Mill. In his essay Of the Laws of Interchange between Nations; and the Distribution of the Gain of Commerce among the Countries of the Commercial World, J. S. Mill wrote:

57

Mr. Ricardo […] unguardedly expressed himself as if each of the two countries making the exchange separately gained the whole of the difference between the comparative costs of the two commodities in one country and in the other. But, the whole gain of both countries together, consisting in the saving of labour; and the saving of labour being exactly equal to the difference between the costs, in the two countries, of the one commodity as compared with the other; the two countries taken together gain no more than this difference: and if either country gains the whole of it, the other country derives no advantage from the trade. […] This, […] was first corrected in the third edition of Mr. Mill’s Elements of Political Economy (pp. 122-24).

58The same thesis attributing the responsibility of the error to Ricardo was repeated by Torrens in 1857: “In the view of the question presented by Mr. Ricardo, the advantages derived from foreign trade were confined to only one of these countries” (Torrens, 1857, p. xv). [12] Finally, Einaudi, who took Torrens’s statement at face value, propagated the same story on the origin of this error. [13]

59Once again, it was Sraffa who first corrected this error regarding Ricardo’s theory. Indeed, Ricardo gave the famous example of English cloth exchanged for Portuguese wine in his Principles, but did not commit the error. He supposed that the wine produced in Portugal by the labour of 80 men for one year would require the labour of 120 men in England while cloth produced in England by the labour of 100 men for one year would need the labour of 90 Portuguese men:

Number of men required for one year in order to produce a given quantity of Cloth Wine England 100 120 Portugal 90 80

60Ricardo concluded that it was advantageous for both countries if England produces cloth and Portugal wine, England exporting cloth for Portuguese wine while Portugal exports wine for English cloth: “England would give the produce of the labour of 100 men, for the produce of the labour of 80” (Ricardo, 1817, p. 135). As Sraffa stressed: “We find no trace of the error, or even of a mere oversight” (Sraffa, 1930, p. 541).

2.3 – Ricardo on the gain from trade

61Ricardo’s contribution to the theory of foreign trade is characterized by his opposition to Smith’s doctrine of foreign trade reducing competition and thus increasing the rate of profit. This he clearly stated at the beginning of Chapter 7 On Foreign Trade of the Principles:

62

It has indeed been contended, that the great profits which are sometimes made by particular merchants in foreign trade, will elevate the general rate of profits in the country, and that the abstraction of capital from other employments, to partake of the new and beneficial foreign commerce, will raise prices generally, and thereby increase profits. […] They who hold this argument agree with me, that the profits of different employments have a tendency to conform to one another; to advance and recede together. Our variance consists in this: They contend, that the equality of profits will be brought about by the general rise of profits; and I am of opinion, that the profits of the favoured trade will speedily subside to the general level.

63This position may appear surprising if one considers Ricardo’s position in the Corn Laws debate. Indeed, for almost ten years Ricardo advocated the repeal of the Corn Laws on the grounds that this would increase the general rate of profit in Britain. So how can Ricardo argue at the same time that “any change from one foreign trade to another, or from home to foreign trade, cannot, in my opinion, affect the rate of profit” (ibid., p. 345).

64Actually, according to Ricardo, two different cases have to be clearly distinguished. First, the case of an extension of foreign trade leading to the importation of corn (or more generally of basic goods): this will decrease the value of corn as well as the natural wages, and thus will increase the general rate of profit. Second, the case of an importation of luxury goods which has no consequence either on the natural wage or on the natural rate of profit, but only increases the general wealth of the country:

65

It has been my endeavour to shew throughout this work, that the rate of profits can never be increased but by a fall in wages, and that there can be no permanent fall of wages but in consequence of a fall of the necessaries on which wages are expended. If, therefore, by the extension of foreign trade […] the food and necessaries of the labourer can be brought to market at a reduced price, profits will rise. If, instead of growing our own corn, or manufacturing the clothing and other necessaries of the labourer, we discover a new market from which we can supply ourselves with these commodities at a cheaper price, wages will fall and profits rise; but if the commodities obtained at a cheaper rate, by the extension of foreign commerce […] be exclusively the commodities consumed by the rich, no alteration will take place in the rate of profits. The rate of wages would not be affected, although wine, velvets, silks, and other expensive commodities should fall 50 per cent., and consequently profits would continue unaltered.

66Now, while he considered the gains of the extension of trade to foreign countries, Ricardo always considered the “productivity doctrine” as the only one and overlooked the possible effect of foreign trade on the general rate of profit:

67

Foreign trade, then, though highly beneficial to a country, as it increases the amount and variety of the objects on which revenue may be expended, and affords, by the abundance and cheapness of commodities, incentives to saving, and to the accumulation of capital, has no tendency to raise the profits of stock.

68Ricardo’s point may be explained by distinguishing between the consequences deriving from opening the market to foreign trade that are “necessary effects” and those that are only “possible effects.” Indeed, according to Ricardo, the extension of foreign markets will necessarily lead to an increase in the division of labour and then to a general increase of wealth. [16] Foreign trade may also have an effect on the rate of profit, but only in the particular case that the commodities exchanged are wages goods. Furthermore, even in this case, only the importing country will see its general rate of profit decrease. The exporting country will have to increase its production of corn and then to have recourse to inferior soils, thus increasing the natural price of its corn:

69

No very great quantity [of corn] could be obtained from abroad, without causing a considerable increase in the remunerating price of corn in foreign countries. […] To raise a larger supply, too, those countries would be obliged to have recourse to an inferior quality of land, and as it is the cost of raising corn on the worst soils in cultivation requiring the heaviest charges, which regulates the price of all the corn of a country, there could not be a great additional quantity produced, without a rise in the price necessary to remunerate the foreign grower.

70The consequences of this increased price of corn in Ricardo’s theory are well known: an increase in the natural wage and a fall in the general rate of profit.

71Ricardo considered that the repeal of the Corn Laws would have a positive effect on the general rate of profit in England, which was an importing country. [18] However, from a theoretical point of view, it was impossible for him to draw any necessary consequence about the effect on the rate of profit of opening the market to foreign trade. The effect could happen (in the case of basic commodities) or not (in the case of luxury goods), and even when it happens, it could increase (in the case of importing countries) or lessen (in the case of exporting countries) the general rate of profit. Therefore, the only effect that Ricardo ascribed to opening the market to foreign trade was the “productivity doctrine” described by Smith and adopted by James Mill and Torrens.

72There is however an important difference between Ricardo’s exposition of the comparative advantages and that of J. Mill and Torrens. This does not concern the figures but the kind of commodities considered. Indeed, in J. Mill’s case, Poland produces corn while England produces clothes. Torrens developed the same example of corn from Poland and clothes from England when he explained his own version of the comparative advantage theory in his Essay on the External Corn Trade (Torrens, 1815, pp. 263-65). In his example Ricardo substituted wine for corn—that is, he substituted what he considered a luxury good (see Ricardo, 1817, p. 132), for a basic good (corn).

73This point has rightly been stressed by Gehrke (2013, pp. 20-21) while discussing “the standard view (according to which) Ricardo had failed to provide a rule for the determination of international values, had supposed an arbitrary distribution of the gain from trade, and had omitted to specify the limits for the terms of trade.” Indeed, in the case considered by Ricardo, the commodities are luxury goods not entering into the wages basket. Then, the distributive variables in the two exchanging countries (the rates of wage and the rates of profit) are not affected by the exchange. The international prices of these commodities are determined only by their conditions of production in the exporting country at the average rates of wage and profit.

74The same principle, according to which it is the natural price of commodities in the exporting country which ultimately regulates the international prices, also applied in the case of basic goods. As stated by Ricardo:

75

Corn, like every other commodity, has in every country its natural price, viz. that price which is necessary to its production, and without which it could not be cultivated: it is this price which governs its market price, and which determines the expediency of exporting it to foreign countries. If the importation of corn were prohibited in England, its natural price might rise to 6l. per quarter in England, whilst it was only at half that price in France. If at this time, the prohibition of importation were removed, corn would fall in the English market, not to a price between 6l. and 3l., but ultimately and permanently to the natural price of France, the price at which it could be furnished to the English market, and afford the usual and ordinary profits of stock in France; and it would remain at this price, whether England consumed a hundred thousand, or a million of quarters. If the demand of England were for the latter quantity, it is probable that, owing to the necessity under which France would be, of having recourse to land of a worse quality, to furnish this large supply, the natural price would rise in France; and this would of course affect also the price of corn in England. All that I contend for is, that it is the natural price of commodities in the exporting country, which ultimately regulates the prices at which they shall be sold, if they are not the objects of monopoly, in the importing country.

76Then, in the case of basic goods, the problem becomes more complex. Indeed, as explained in the first part of this paper, in case of basic goods the system becomes undetermined and it is impossible to determine the natural prices of commodities as well as the rates of profits in the different countries.

77However, whether the commodities considered are basic or luxury goods, their natural price in the exporting country always regulates their international prices for Ricardo. In this respect, his theory differed essentially from that of John Stuart Mill, who considered that in foreign exchanges, “the principle, that value is proportional to cost of production, being inapplicable, we must revert to a principle anterior to that of cost of production […] namely, the principle of demand and supply” (J. S. Mill, 1844, p. 8). Another problem appeared in J. S. Mill’s interpretation of Ricardo with regard to the gain of trade. Mill considered the question “in what proportion the increase of produce, arising from the saving of labour, is divided between the two countries” (ibid., p. 5), and as “this question was not entered into by Mr. Ricardo,” this led him to the error of attributing to each country the whole gain of trade—an error which, as we have seen, James Mill committed but not Ricardo. Once again, J. S. Mill’s interpretation is erroneous. For Ricardo, the gain of trade did not consist in a division of an increased produce as in J. S. Mill, but consisted in the saving of labour that each country experienced while obtaining the commodity for which it was comparatively less productive by trade rather than by producing it. In his example, England imported wine that would have been produced by 120 men and exported cloth produced by 100 men. The gain was then 120 – 100 = the labour of 20 men. As for Portugal, the gain was the labour of 10 men (90 – 80).