I. INTRODUCTION
The economic impact of preferential trading arrangements is one of the most complex and controversial questions in economics. Intense debates rage about the effects of regional and bilateral trade agreements on the welfare of different groups in different countries, and the consequences of these agreements for the multilateral trade liberalization process.Footnote 1
Jacob Viner’s The Customs Union Issue, published in 1950, is the one undeniable classic in the field of the economics of preferential trading arrangements.Footnote 2 Appreciation of the work began with the early reviews, and it quickly became the point of departure for international economists who further developed the theory of customs unions. It remains the most-cited work on preferential trading arrangements.Footnote 3 Looking back over the history, Mark Blaug (1991, p. 298) judged that Viner’s book “supplied the starting point for every subsequent work on the economics of common markets and free trade areas.”
The first purpose of this paper is to provide an account of what shaped Viner’s thinking on customs unions; to clarify what Viner wrote and why. This aspect of the paper is historical, aimed at understanding Viner in context. Second, the paper will consider the reception of his work, examining interpretations offered by practitioners who further developed the economics of customs unions. The divergence between historical and practitioner accounts of Viner’s work raises general questions about the relationship between contextual history and practitioner history. Why have so many practitioner misreadings of Viner been scientifically fruitful? The third purpose of the paper is to explore how Viner’s work functions as a classic for contemporary international economists. There appears to be a reading cycle moving from use of a text by practitioners to the text’s becoming a classic, then different uses by practitioners and historians of economics.
II. WHAT SHAPED VINER’S THINKING ON CUSTOMS UNIONS
Classical Liberalism and Free Trade
Viner’s intellectual commitment to classical liberalism is clear from his published work and correspondence.Footnote 4 For example, through the 1940s, he corresponded with Hayek about the formation of the Mont Pelerin Society, although he did not join because he considered that it had become a political organization. There are many letters, especially to Alec Macfie and Lionel Robbins, where his admiration for Adam Smith is clear, and a 1967 letter to Bertil Ohlin, in which he describes himself as a disciple of Hume.
As an international economist, he was a lifelong advocate of free trade in a period when free trade was not popular, even among his economist colleagues.Footnote 5 Viner’s background commitment to free trade was the most important reason for writing The Customs Union Issue.
The Classical Economists on Customs Unions
The picture painted in the international trade literature is of Viner in 1950 announcing to a stunned world that customs unions are not always beneficial to the countries involved, let alone the world as a whole. Viner was keenly interested in the history of economic thought from his student days at McGill and Harvard, and aware of discussions of the topic by the classical economists.Footnote 6
As Denis O’Brien (Reference O’Brien1976) has documented, this supposedly novel insight that customs unions were not always beneficial was the almost universal understanding of the classical economists. For instance, Adam Smith was strongly opposed to the Methuen treaty with Portugal signed in 1703, which stipulated that there was to be no tax on Portuguese wines or English textiles. J. R. McCulloch was particularly incensed by the treaty, as it forced him to purchase inferior and more expensive Portuguese wine, rather than the French wine of which McCulloch was rather fond. Other treaties, such as the 1860s Cobden–Chevalier Treaty, which included a most-favored-nation clause, received support from economists because, according to O’Brien’s carefully documented argument, the benefits of the trade created were believed to exceed the diversion losses, and also because the treaty was thought likely to provoke reductions in general tariff levels. The writings of the classical economists on preferential trading arrangements were the first important influence on Viner’s work.
Taussig and the US Tariff Commission
Viner’s interest in preferential trading arrangements seems to have been aroused by his Harvard teacher Frank Taussig, who commented extensively on US trade treaties, especially their reciprocity and most-favored-nation clauses (for instance, Taussig Reference Taussig1892, Reference Taussig1910, 1915). Taussig was generally optimistic about reciprocity arrangements promoting free trade, perhaps because they were then the only tariff-reduction game in town.Footnote 7 Taussig’s underlying question when assessing reciprocity arrangements was whether they helped or hindered the cause of free trade, and he was sensitive to the problems of the American practice of making application of the most-favored-nation principle (MFN) conditional. Viner learned from Taussig a practical case-by-case approach and the importance of attention to the institutional details of trade agreements. Even some of his teacher’s specific examples influenced him; for instance, Viner’s wool and woolen cloth example (Viner Reference Viner1950, p. 48) seems based on Taussig’s earlier discussion (1910, p. 171).
After Viner had completed his PhD coursework, Taussig arranged for Viner to work in 1917 and 1918 with the US Tariff Commission as a special expert, contributing to its 1919 report Reciprocity and Commercial Treaties (United States Tariff Commission 1919, discussed by Meardon Reference Meardon2011). This detailed work on US trade treaties must have given Viner some insight into the effects of trade preferences and the political process surrounding them. Again, some specific examples found their way into Viner’s later work, such as the US–Hawaii–Cuba sugar trade discussed in the Tariff Commission report, which appeared in his later work (Viner Reference Viner1931b, p. 7).
Early Papers on Reciprocity and the Most-Favored-Nation Principle
After leaving the Tariff Commission and establishing himself in his new position at the University of Chicago, Viner began to write on these issues. An article on preferences in US trade agreements (Viner Reference Viner1924b) was followed by another article covering similar ground in a Swedish journal (Viner Reference Viner1931b). In these articles, Viner showed himself a fierce defender of the MFN principle, and not overly concerned about customs unions, which he saw as an unimportant and justified exception to the principle. Customs unions at that time, certainly from an American perspective, did not seem to be a major threat to the MFN principle or the wider free-trade agenda.
Aspects of the international situation in the 1920s may have influenced Viner’s approach to the economics of discriminatory trade arrangements. Following WWI and the 1916 Paris Allied Economic Conference, some European nations were reluctant to extend MFN to Germany, reversing their pre-WWI practice. The US was the strongest supporter of the MFN principle, and, in 1922, moved to unconditional acceptance of the principle (discussed in Irwin Reference Irwin, De Melo and Panagariya1993). Viner’s involvement with the United States Tariff Commission predisposed him to support MFN, and these further post-WWI European threats to the principle perhaps inclined Viner to search harder for economic arguments for non-discrimination to place his and his country’s position on a firmer footing. The Tariff Commission work and the Paris Conference are mentioned in Viner’s early articles, and, in the 1931 article, he is explicit about his motivation. He states that since the MFN principle is “definitely on the defensive,” it is time “to examine the objections which have been made” against “equality of treatment in tariff matters” (Viner Reference Viner1931, p. 96), irrespective of whether customs unions were perceived as a major current threat to free trade.
Many of the ideas, and even the language, of Viner’s skeptical account of customs unions in his 1950 book The Customs Union Issue can be found in the 1924 and 1931 articles. The concept of trade diversion and the harm done by trade-diverting agreements are clearly explained. In 1924 he wrote that for the granting country and the world as a whole,
reciprocity treaties, even on free-trade grounds, are ordinarily not an amelioration, but on the contrary are an intensification of the evils of customs tariffs. They not only do not counteract the tendency of protective import duties to divert international trade from the channels which it would follow under free trade, but they may cause an even wider departure of trade from its “natural” channels than would result from a regime of uniform protective tariffs at the levels prevalent prior to the grant of partial reductions of duties through reciprocity arrangements. (Viner Reference Viner1924b, p. 107)
The idea of trade diversion becomes more concrete in his 1931 article on the most-favored-nation principle:
A tariff that is high, but uniform in its treatment of imports regardless of their origin, may divert trade from the channels which it would follow if allowed freely to choose its own path much less than would a moderate tariff which applies different treatment to imports according to their country of origin. Suppose that under free trade country A would find it to its advantage to import a particular commodity from country B, and that even with a high duty it is still not possible to produce a commodity at home at a profit to its producers, so that it continues to be imported from B. While the tariff reduces the volume of trade, it does so only as a revenue measure, and still permits the commodity to be produced there where it can be produced most cheaply. Suppose, now, that the duty is reduced by half on imports from a third country, C, and that by virtue of this preferential treatment C can undersell B and capture A’s trade. The result of the discrimination in favour of C is that the commodity which could be most cheaply produced in B, and would be produced even if A had high tariff, provided it was nondiscriminatory, is now produced in C, where the conditions for its production are comparatively unfavourable. The reduction in duty, because it is discriminatory and not uniformly extended to all, operates as a deterrent instead of a stimulus to the optimum allocation of the world’s resources in production. (Viner Reference Viner1931b, p. 5)
Overall, he still viewed customs unions favorably: “as a rule, customs unions probably constitute a forward step towards freer trade” (Viner Reference Viner1931, p. 10), though he later noted (Viner Reference Viner1950, p. 53) that his questioning in print of customs unions began with the 1931 article.
Manitoba Consulting Assignment
Viner’s thinking on customs unions developed further through consulting work in the late 1930s for the Province of Manitoba as part of the Canadian Royal Commission on Dominion–Provincial Relations.Footnote 8 His work framed in customs union terms Manitoba’s argument about the detrimental impact on Manitoba of its being part of the Canadian federation.
For our purposes, Viner’s (Reference Viner1938) supplementary statement to the Royal Commission is of most interest. There, he explained Manitoba’s problem that the Canadian federal tariff priced lower-cost US producers out of the market, so that goods were instead imported from higher-cost Canadian producers, mostly located in Ontario. This diversion of trade affected consumer goods and inputs to Manitoba’s agricultural export industries, as well as government purchases. There was little possibility of expansion of the Manitoban export industries as a result of the Canadian tariff, but large benefits to Ontario producers. Viner noted that while terms-of-trade benefit might also be expected, as the tariff would reduce demand for imports, the fact that the foreign demand for Manitoba’s exports was almost perfectly elastic meant that the terms-of-trade benefit to Manitoba would be minimal, and any terms-of-trade benefit would accrue to Ontario and the industrial provinces. The case of Manitoba in the Canadian federation was a clear example of losses due to trade diversion that Viner was later to discuss in The Customs Union Issue.
The Manitoba work was significant not for any conceptual innovation, but because it provided Viner with a concrete example that guided his thinking about the circumstances in which a customs union would lead to losses, and where the losses appeared to be substantial.
Post-war Context
The immediate context for Viner’s book on customs unions was the difficult post-World War II situation. He was heavily involved in discussions of the post-WWII trade and financial architecture, both directly and through his extensive networks of friends and correspondents on both sides of the Atlantic, and through his writings (such as Viner Reference Viner1946, Reference Viner, Wilcox and Hall1947).Footnote 9 Advocates of free trade struggled with a system clogged by the residues of interwar protectionist measures, wartime controls, and macroeconomic imbalances. Intellectually, free traders were on the defensive. Non-discrimination, however, had been favored by the US after the 1934 Reciprocal Trade Agreement Act and the support of Roosevelt’s Secretary of State Cordell Hull. European discussions of some form of union were going on, though the 1957 Treaty of Rome, which led eventually to the European Union, was still a long way off. The beginning of the post-war multilateralism was signaled by the 1947 Havana Charter for the stillborn International Trade Organization. Customs unions were treated as an exception to the general principle of non-discrimination (as would be in subsequent GATT and WTO agreements), but the relationship between customs unions and the multilateral process was then quite unclear.
Stimulus from Other International Economists
Viner’s papers indicate extensive discussion of preferential trade issues with other international economists, and in The Customs Union Issue (Viner Reference Viner1950, pp. 53–54), he particularly notes the work of his close friend and fellow supporter of free trade Lionel Robbins (1937, especially ch. V), who criticized “tariff unions” as economically detrimental. Robbins’ book originated as lectures at the Institut des Hautes Études Internationales at Geneva in the summer of 1935, where Viner was a regular visitor in the 1930s. For Viner, it did not provide any new theoretical material but a trenchant public attack (e.g., “the only completely innocuous tariff union would be directed against the inaccessible produce of the moon” [p. 122]) by a prominent free trader on the common view, which, I suspect, was important encouragement for Viner’s taking further his earlier tentative questioning of customs unions.
Gottfried Haberler (with whom Viner had clashed in the 1930s over the question of real costs versus opportunity costs as a basis for judgments about the gains from international trade) is singled out as an example of the common position that customs unions advance the cause of free trade, and are, therefore, a good thing. Like Robbins, Haberler was among the economists who strongly supported free trade, so his public statements on the issue mattered to Viner. The relevant sections of Haberler (1936, pp. 377–391) criticize “preferential duties,” but support the formation of customs unions (“Customs unions are to be wholeheartedly welcomed” [Haberler Reference Haberler1936, p. 390]). This puts Haberler in disagreement with the position he attributes to Viner: that customs unions “cannot bring an advantage which could not be attained … by a general reduction of duties”(Haberler Reference Haberler1936, p. 390). The disjunction between Haberler’s opposition to tariff discrimination and support for customs unions, which involve political connections between nations, makes sense if there are strong political reasons for these connections—and this seems to be the case for Europe in the 1930s. Haberler explicitly discounts another economic reason for customs unions: that they are stepping stones to a universal free trade. A later essay (Haberler Reference Haberler and Harris1943) expresses strong support for regional trade blocs, for political more than economic reasons.
Viner’s combative nature and the opportunity of refuting a conventional view supported by prominent international economists perhaps contributed to his decision to write The Customs Union Issue. In my view, the overriding reason, though, was a concern for the cause of free trade, in particular the fear that the formation of customs unions would be a distraction for free traders and an opportunity for others to disguise protectionism. As he explains in the opening pages of the book, “Projects for customs unions and other special tariff arrangements between states independent of each other politically are today widespread, and many persons look to them as at least a partial solution for the major economic and political problems in the international field.” He wishes to “examine them from the point of view of the economist” (Viner Reference Viner1950, p. 3).
III. THE CUSTOMS UNION ISSUE
The Customs Union Issue resulted from an invitation in 1944 from George Finch, director of the Carnegie Endowment in Washington, to write an historical and analytical essay to serve as an introduction “to a planned collection of customs union documents” (Finch to Viner, 16 May 1944). Viner accepted the invitation and completed the manuscript by 1948 (Viner to Finch, 24 August 1948). However, the accompanying volume of documents was abandoned by Carnegie, much to Viner’s regret (Viner to Finch, 26 October 1948). A compromise was reached whereby a list of approximately 250 documents held by the Carnegie Endowment would be included as an appendix to Viner’s essay. His rather mournful note about the Carnegie collection of documents may be found on page 141 of The Customs Union Issue, preceding the list of documents. Viner’s frustration grew with delays in publishing, until he wrote, “I am simply disgusted” and “I wish to God that I had never started on this venture with the Carnegie Endowment” (Viner to Joseph Shotwell, 26 April 1950).
I turn now to what exactly Viner wrote (and didn’t write), about which there has been much misunderstanding in the literature.
Viner took a broad view of customs unions, and his analysis is relevant to most preferential trading arrangements. The introductory chapter discusses customs unions as “one of a number of arrangements for reducing tariff barriers between political units while maintaining barriers against imports from outside regions” (p. 4), while a “perfect customs union” is defined as an arrangement which “meets the following conditions:
(1) the complete elimination of tariffs as between the member territories;
(2) the establishment of a uniform tariff on imports outside the union;
(3) apportionment of customs revenue between the members in accordance with an agreed formula” (p. 5).
The economic analysis is in chapter 4, where twelve pages are devoted to the economics of customs unions. Viner begins his review of the economics of customs unions by noting the “strange phenomenon” of support for customs unions uniting free traders and protectionists (p. 41). Their views on trade issues are generally opposite, so their unity on the customs unions issue must come from misunderstanding about whether customs unions represent a move towards free trade or away from free trade. He explains his central question will be the effect of the customs union on each country separately, on the countries together, and for the world as a whole (p. 42).
Before launching into the analysis of the effects of customs unions, Viner points out that “None of these questions can be answered a priori, and the correct answers will depend on just how the customs union operates in practice. All that a priori analysis can do is to demonstrate, within limits, how the customs union must operate if it is to have specific types of consequence” (p. 43).
He then makes the crucial distinction between commodities “which one of the members of the customs union will now newly import from the other which it formerly did not import from at all because the price of the protected domestic product was lower than the price at any foreign source plus duty” (i.e., which involve “a shift from high cost to a lower cost point, a shift which the free trader can properly approve, at least a step in the right direction, even if universal free trade would divert production to a source with still lower costs” [p. 43]) and “other commodities which one of the members of the customs union will now newly import from the other whereas before the customs union it imported them from a third country, because that was the cheapest possible source of supply even after payment of duty” (i.e., which involve a shift between “a low-cost third country and the other, high-cost, member country” [p. 43]). He noted, “This is a shift of the type which the protectionist approves, but it is not one which the free trader who understands the logic of his own doctrine can properly approve” (p. 43), and commented that “simplified as this exposition is, it appears to cover most of the basic economic issues involved” (p. 44).
It is worth bearing in mind that Viner had been commissioned to write a brief essay introducing a collection of customs union documents aimed at policy makers. Viner here is not claiming to engage in a detailed exercise in welfare economics; instead, he is making a simple distinction that he believes is helpful in evaluating customs unions. This way of operating is consistent with his other writings, where he resists self-contained theoretical systems and demonstrative general propositions that purport to settle real-world policy questions.
Reading the Viner passage as some sort of general proposition also ignores the comments that precede and follow it about the impossibility of answering questions about the impact of customs unions a priori. The following passage is: “From the free trade point of view, whether a particular customs union is a move in the right or in the wrong direction depends, therefore, so far as the argument has yet been carried, on which of the two types of consequences ensue from that customs union” (p. 44).
If Viner offers welfare propositions, it is not on pages 43–44, but at the end of his brief discussion of the economics of customs unions, where he gives several rules of thumb:
A customs union is more likely to operate in the free trade direction, whether the appraisal is in terms of its consequences for the customs union area alone or for the world as a whole:
(1) the larger the economic area of the customs union and therefore the greater the potential scope internal division of labour;
(2) the lower the ‘average’ tariff level on imports from outside the customs union area as compared to what the level would be in the absence of customs union;
(3) the greater the correspondence in kind of products of the range of high cost industries as between the different parts of the customs union which were protected by tariffs in both of the member countries before the customs union was established i.e. the less the degree of complementarity—or the greater the degree of rivalry—of the member countries with respect to protected industries, prior to the customs union;
(4) the greater the differences in unit costs for protected industries of the same kind as between the different parts of the customs union, and therefore the greater the economies to be derived from free trade with respect to these industries within the customs union area;
(5) the higher the tariff levels in potential export markets outside the customs union area with respect to commodities in whose production member countries in the customs union would have a comparative advantage under free-trade, and therefore the less the injury resulting from reducing the degree of specialization in production as between the customs union area and the outside world;
(6) the greater the range of protected industries in which an enlargement of the market would result in costs lower than those at which the commodities concerned could be imported from outside the customs union area;
(7) the smaller the range of protected industries for which an enlargement of the market would not result in unit costs lower than those at which the commodity is concerned imported from outside the customs union area but which would nevertheless expand under customs union. (p. 51)
He concludes this discussion of the economics by emphasizing that “customs unions are from a free-trade point of view, not necessarily good nor necessarily bad; the circumstances discussed above are the determining factors” (p. 51).
As was Viner’s habit, he follows a simple and suggestive piece of economic theory with a list of qualifications and additional matters that need to be considered in a policy discussion. Here, the additional matters are the possibility of a customs union’s being able to exploit market power and improve its terms of trade (bringing gains at the expense of the outside world) (p. 55), better possibilities of bargaining for tariff reductions (p. 56), administrative economies flowing from customs unions (pp. 58–64), revenue from duties (pp. 65–68), and the impact of forming a customs union on the degree of monopoly within the union (pp. 75–78). This last issue was especially relevant to the French push at the time for a variant customs union that closed the domestic market to all but members of the customs union, as a way of restricting competition.
Viner showed a keen awareness of the political issues. An important question was the relationship between economic union and political union, because some economists advanced customs unions as a step towards political unions, which would help secure a lasting peace. Viner argued that, historically, most customs unions that had actually been formed were driven by political rather than economic considerations (p. 91), that political union always preceded economic union—with the German Zollverein as the only exception (p. 96), and that this is for good reasons (such as the absence of workable redistributive mechanisms). Moreover, the most economically beneficial customs unions are the hardest to form (p. 135), a point rediscovered by Grossman and Helpman (Reference Grossman and Helpman1995) and the contemporary literature on political economy of trade.
A comment on this “strange phenomenon” of customs unions’ pleasing both free traders and protectionists shows his awareness of the political economy issues. He suggests that “Businessmen, however, and governments which have had to try to simultaneously satisfy both special interests seeking increased protection and voters hostile to protection have long known ways of making increased protection look like movement in a free-trade direction” (p. 48), and then gives an example of another way protectionism can be made to look like free-trade policy. This is when reducing a duty on an input (woolen cloth) increases protection overall (on woolen goods), a situation that would later be considered by the effective protection literature, and an example of what would later be known as the ‘theory of second best.’
Viner also offers perceptive commentary on the relationship between customs unions and the multilateral trade liberalization process that was beginning to take shape as he wrote. He feared that in the post-war period, customs unions will “almost inevitably operate as a psychological barrier to the realisation of the more desirable but less desired objectives of the Havana Charter—the balanced multilateral reduction of trade barriers on a non-discriminatory basis” (p. 139), and this has sadly proved true, as discussed by Bhagwati (Reference Bhagwati2008) and others.
These sorts of political considerations are at least as important as the economic arguments about trade diversion to his conclusion that “customs union is only a partial, uncertain and otherwise imperfect means of doing what worldwide non-discriminatory reduction of trade barriers can do more fully, more certainly and more equitably” (p. 135).
IV. RECEPTION OF VINER’S WORK
Early Reviews
Sales of The Customs Union Issue were healthy, and all the reviews in major journals positive.Footnote 10 Most reviewers praised Viner for his command of the historical material, and commented on the relevance of the book for current controversies.
A few of the reviews by economists highlighted the analytical contributions for which the book would later be mostly known. For instance, Virgil Salera (1951, p. 84), in the Journal of Political Economy, considered it “the first rigorous treatment of the subject” and focused on how Viner’s trade diversion argument overturned the conventional wisdom that customs unions are always a good thing. Wolfgang Stolper described the book in his review for the American Economic Review as “the most cogent analysis of the economics of customs unions to be found in the literature. But this means perhaps only that the book was written by Jacob Viner” (Stolper Reference Stolper1951, p. 990). He correctly identified Viner’s main concern as sorting out whether customs unions contributed to the cause of free trade, and understanding why customs unions might be supported by protectionists for trade diversion or cartelization reasons. James Meade’s review for Economica (Meade Reference Meade1951) grasped the importance of Viner’s trade diversion argument. He wrote to Viner, “I have just read your book on customs unions and would like to say how much I admire it. It needed doing very badly … and the result is most successful” (Meade to Viner, 5 February 1951). His review for Economica praised it highly: “A good book on the subject of the customs union has for long been wanted; and now it is provided by Professor Viner’s study, which it is difficult to praise too highly” (Meade Reference Meade1951, p. 186); and, “Professor Viner’s study on the economic aspects of customs unions will be of central interest to economists. Indeed for many years this is likely to remain the locus classicus for the economic analysis of customs-union problems.… Professor Viner’s own comparatively short analysis of the problem marks the book so clearly with the stamp of excellence” (Meade Reference Meade1951, p. 187). There is an exposition of Viner’s argument about trade diversion and creation, but no comments on Viner’s assumptions or their limitations.
Meade
James Meade was the first economist to further develop Viner’s work in a series of invited lectures in Rotterdam (published as Meade Reference Meade1955). He repeats the admiring comments from the review, and he begins a thorough evaluation of Viner’s work: “like many basic ideas which have broken new ground and suggested the lines of work, his analysis is in my opinion in some respects incomplete; and when an attempt is made to complete it, it no longer remains as simple in its application as may at first sight appear to be the case. It is my purpose in these lectures to make a contribution to the analysis which Professor Viner has started” (p. 34). Meade’s criticism is that “Professor Viner’s analysis does not tell one how to weigh up the economic gains from some element of trade creation against the economic losses from some other elements of trade diversion” (p. 34). In order to do this, Meade sets out a model that he views as “completely in harmony with the basic analysis which Professor Viner employs” (p. 36), where “all elasticities of demand are zero, and all elasticities of supply are infinite.” In the following chapters, this model is analyzed and extended in various ways, and Meade concludes that “it is impossible to pass judgment upon customs unions in general” (p. 107). There is no denying the richness of Meade’s discussion and the acute powers of insight demonstrated, especially in the “rash generalizations” he gives at the end of the lectures (pp. 117–124). They are similar in style to Viner’s propositions discussed above, but more detailed.
Meade claims that his model is “in harmony” with Viner’s. However, later in the book, there are statements that Viner assumes “constant costs of production” and “zero elasticities of demand” (for instance, Meade Reference Meade1955, p. 77). While nothing Meade writes directly contradicts statements by Viner, Meade is not offering an adequate exegesis of Viner’s text, let alone an historical account of text. His purpose is different; he is using Viner’s work as an anchor for his own development of customs union theory.
Lipsey
Richard Lipsey was, at the time, undertaking a PhD at the London School of Economics on customs unions (his thesis was subsequently published as Lipsey Reference Lipsey1970), and marks of Meade’s reading of Viner are evident in an article Lipsey (Reference Lipsey1957) published in Economica. The article takes Viner’s comment about a free trader’s properly approving of trade creation, but not trade diversion (see Viner Reference Viner1950, p. 43; the passage quoted earlier in this paper), as a formal proposition of welfare economics, and proceeds to demolish it as such. In a competitive general equilibrium model, consumption effects can work in the opposite direction to those Lipsey sees (presumably based on a reading of Meade) as emerging from Viner’s problematic analysis of the production side. A similar position is taken in his survey (1960), although the statement about Viner’s assumptions is softened slightly to “Viner’s analysis implicitly assumed that commodities are consumed in some fixed proportion which is independent of the structure of relative prices” (p. 499).
Lipsey also suggests his analysis is more general and therefore superior, that Viner’s insights on customs unions are merely a special case of his general theorem of second best: “There can be no doubt that the theory of customs unions provides an important case study in the application of the general theory of second best (Lipsey and Lancaster Reference Lipsey and Lancaster1956, p. 14). Conceptually, Lipsey and Lancaster are right that the economics of customs unions is a special case of their theory of second best, but, historically, the theory of second best grew from customs union theory, public finance, and work on distortions in other branches of economics. The importance of work on customs unions theory is suggested by Meade’s work on customs unions, leading to his own independent statement of the theory of second best in his classic Trade and Welfare (Meade Reference Meade1956). Viner’s work, filtered through Meade, was undoubtedly a major influence on Lipsey, as he himself acknowledges in the discussion of precursors of the theory of second best (Lipsey and Lancaster Reference Lipsey and Lancaster1956, pp. 13–14).Footnote 11 There are many further statements of what looks like a generalized theory of second best in the literature before 1956, including in Viner’s work (for example, the introduction to Viner Reference Viner1951, p. 12, which contains a clear statement of the general principle of second-best welfare economics), but Viner had no interest in pressing claims.Footnote 12
Apart from the claim about having set out a general theory of second best, Lipsey’s approach is much the same as Meade’s, though Lipsey is less careful about what Viner actually wrote.
Corden
Max Corden’s (1965) very fine survey of developments in international trade theory unfortunately relied too much on Meade’s and Lipsey’s accounts of Viner’s work on customs unions. His opening statements—“The theory of customs unions is a completely new branch of the theory of tariffs. It originated with Viner’s The Customs Union Issue 1950. Since 1955 it has been rapidly elaborated by Meade, Lipsey and others, who have shown that it is one of the most important applications of the theory of second-best” (Corden 1965, p. 52)—are open to question on a number of grounds. He also claims Viner ignored certain important theoretical issues, though Corden praises Viner’s work overall.
I have chosen to consider Cordon’s survey among many published in the 1960s and 1970s that recycle Lipsey’s account because Corden himself contributed to the customs union theory, and because we have some illuminating correspondence between Corden and Viner.Footnote 13 Viner wrote, after reading Corden’s survey, “I must say that I found some difficulty in recognizing the relevance to what I published in 1950 of some of what you attribute to me with respect to customs unions.” Viner especially objected to the assertion that he assumed “that commodities were consumed in fixed proportions unaffected by the changes in relative prices brought about by the customs union” (quoting Corden 1965, p. 52). Viner added that “if this is nevertheless a necessary implication of something I did say, I would regard it as a monstrous lapse on my part, but would ask where precisely I was guilty of this lapse.”
Viner also objects to Corden’s charge that he “assumed constant costs,” believing he “had been sufficiently explicit on my pages 45 and 47,” and states that he worked with “increasing cost as the one generally prevailing (static) pattern of costs relative to output.” On this, he adds that “if there is even the slightest implication in anything I wrote in the customs union issue of an assumption of constant costs, I would certainly like to have it identified, so that I could engage in what I would regard as due penance.” The point is reinforced by a reference to increasing costs in his famous 1931 article on cost and supply curves (Viner Reference Viner1931a). He acknowledges that interpretation of his work has been affected by terminological confusion between constant costs in the short-run sense, which he never assumed, and constant returns to scale in the long run. Viner concludes the letter: “I have so much respect for the superior competence in analytical matters of the present generation of economists that I am ready humbly to accept their verdicts … provided only that they base their verdicts on what I have actually written.”
Corden wrote back quickly, although unable to access his notes: “I assume that I was in error, since you will know best what you wrote and what your intentions were. It seems to me possible that I may have been (over) influenced in my interpretation by Lipsey’s writings” (Corden to Viner, 5 April 1965).
If Corden misread Viner, then it was a fruitful misreading. Corden (1972) developed the theory of customs unions with non-constant returns to scale, remedying an omission by Viner. He acknowledges Viner’s discussion (1950, pp. 45–46) of increasing returns to scale as an argument for customs unions, and comments that some of his conclusions differ from Viner’s, “possibly because his (unspecified) assumptions differ” (Corden 1972, p. 465). In the end, Corden reaffirms Viner’s suggestion that increasing returns do not greatly affect the analysis of customs unions.
Cooper and Massell
The point of departure from Viner for Cooper and Massell’s (Reference Cooper and Massell1965) supposed new view of customs unions was that nations had no incentive to form customs unions because any benefit from a customs union could be more efficiently obtained from trade liberalization on an MFN basis. Viner had made exactly the same argument in his earliest paper on the subject:
In so far as the advantages derivable from such [reciprocity] arrangements are economic, however, they are obtainable in at least equal degree through the outright reduction of their tariffs by both countries. Moreover, if the countries are really so related to each other economically that a mutual reduction in the tariffs on each other’s products would result in a closer, instead of in a less close, conformity of trade to the channels which it would follow under universal free trade, a reciprocity treaty will not operate very differently from the outright extension to the entire world by both countries of the concessions granted to each other. (Viner Reference Viner1924, p. 108)
The same point was made in his subsequent book (Viner Reference Viner1950, p. 135).
Krauss
Melvyn Krauss’ landmark (1972) Journal of Economic Literature survey relied on Lipsey’s and Corden’s readings of Viner’s work. He identified Viner as the pioneer of customs unions theory and developer of the concepts of trade creation and diversion. The subsequent literature was a response to the limitations of Viner’s analysis—in particular, that it was “based on the number of simplifying assumptions—fixed proportions in consumption and constant costs” in production (Krauss Reference Krauss1972, p. 414).
Michaely
Michael Michaely’s (Reference Michaely1976) Journal of International Economics paper was the last detailed practitioner engagement with Viner’s work, with a title promising a resolution of the tricky question of Viner’s assumptions. He begins with the fair comment that “Viner’s discussion is expressed in a rather intuitive manner, lacking a precise specification of the analytical model and its basic assumptions. Consequently, possible misinterpretations of Viner’s position have been accepted as conventional wisdom” (Michaely Reference Michaely1976, p. 75). But then follows a common but dubious procedure of building a convenient model and trying to squeeze the writer in question into it.Footnote 14 The model Michaely builds (pp. 75–82) is a standard, small, open-economy, general equilibrium model of a country considering entering into a customs union. With given factor endowments and appropriate assumptions about technology, there is the usual convex production transformation frontier.
Michaely examines the key trade creation and diversion passage on page 43 of Viner’s book with the aid of this reference model. He observes that, in the passage, “Nothing is said here explicitly about the nature of the cost curve” (p. 83). In Michaely’s model, an equilibrium with both trade creation and trade diversion occurs if the transformation frontier is strictly convex, while an equilibrium with one or the other occurs if the transformation frontier is a straight line. Since Viner suggested in the passage that a customs union will lead to either trade creation or trade diversion, Michaely takes this to imply that Viner cannot be working with a strictly convex transformation frontier, and, therefore, must be assuming constant production coefficients. Michaely believes he has confirmed the view of Meade and Lipsey that Viner assumed constant, not increasing, costs.
The difficulty with Michaely’s logic is that he foists on Viner all the other assumptions of his reference model. Equilibria with either trade creation or trade diversion outcomes are possible with different models, especially the more complex models that Viner seems to be working with in his other publications. In his discussion of the real-cost/opportunity-cost controversy, Viner rejected the assumption of fixed labor endowments that is part of Michaely’s reference model.Footnote 15 The basic problem is Michaely’s rather forced reading of the Viner passage as a description of alternative equilibria, one where all trade is diverted trade, and another where all trade is created trade.
Michaely (p. 84) then examines a comment on page 45 of Viner’s book that a tariff can act as a partial barrier to trade, which Michaely interprets as an assertion that a tariff will reduce but not completely cut off imports of a good. In Michaely’s reference model, this happens with a strictly convex transformation frontier, but not if the transformation frontier is a straight line. So, in Michaely’s view, Viner of page 45 is assuming a strictly convex transformation frontier and thus increasing costs. Comparing with the Viner of page 43, “A basic inconsistency is thus revealed.”
Michaely then (p. 85) examines Viner’s other works to decide whether the Viner of page 43 or the Viner of page 45 is the true Viner. Either way, Viner would be guilty of a basic inconsistency. First stop is Viner’s famous 1931 article on costs and supply, where, as he pointed out to Max Corden in correspondence, there is no doubt that he is working with increasing costs. Next stop is the 1937 Studies in the Theory of International Trade. Michaely thinks that since Viner is defending the real-cost doctrine of the classical economists, and since everybody knows that the classical economists assumed constant costs, Viner must be assuming constant costs. And, we have: “Thus it appears that Viner’s concept of the nature of costs in international trade is ambiguous and inconsistent” (Michaely Reference Michaely1976, p. 86). However, just as the classical economists were not working with what became known as the Heckscher–Ohlin–Samuelson model, neither was Viner.
Next is an examination of consumption affects, and here, in Michaely’s view, Viner is at least consistently mistaken: “Viner’s position with respect to the consumption effect of a customs union, unlike his treatment of costs, is unambiguous: he consistently overlooks this element in his analysis…. Indeed, he shows no trace of awareness of the welfare effect of substitution in consumption,” and so, “the issue to be examined here is not, therefore, whether the consumption effect was neglected or not, but rather the possible sources or interpretations of such neglect” (p. 86). Michaely’s evidence includes Viner’s Reference Viner1950Customs Unions Issue and his 1937 Studies in the Theory of International Trade. In relation to the customs union book, he makes the extraordinary statement (p. 92) that “Viner’s disregard of the consumption effect of the establishment of the union was due neither to any specific assumption about the nature of this effect, nor to a decision to concentrate only on the production effect while regarding it as only part of the analysis. It was rather due, it seems, to Viner’s lack of awareness of the effect of trade and of tariffs on welfare through substitution in consumption.” Instead, Viner was simplifying as much as possible to make the conceptual point about the distinction between trade diversion and creation. A few minutes spent with any of Viner’s works that discuss the role and limits of economic theory would have been sufficient to save Michaely from this appalling misreading. In relation to Viner’s purported lack of awareness, one needs only to open Studies at page 521, where Viner draws consumption indifference curves, and describes substitution around the indifference curve in the accompanying text, to realize the inaccuracy of Michaely’s claim that Viner was unaware of the effect of trade on welfare through consumption substitution.
We would have to conclude that Michaely has been unsuccessful in resolving the question of Viner’s assumptions in the customs union book. Viner obviously had a model or models in mind when describing the effects of customs unions, but it is unclear exactly what they are. His lack of clarity about his assumptions is partly due to the nature of the book and partly to his skepticism of the sort of exercises of which Michaely’s paper is a particularly bad example.
Viner’s approach is perfectly consistent with views he expressed elsewhere on modeling. Viner (1937, p. 526) suggested it is appropriate to leave the analysis “in that state of persuasiveness associated with incomplete demonstration which seems to be a universal characteristic of propositions economic theory relating to questions involving human welfare.” Viner also suggested (1955, p. 128) that “relevance is of supreme importance for economic theory” and “leads to certain rules of guidance as to the procedure we should follow in constructing our theoretical models,” including “the practice to start with the simplest and the most rigorous model, and to leave it to a later stage, or to others, to introduce into the model additional variables or other complicating elements”. He is scathing in various writings about taking propositions deduced from models directly to policy discussion.
Viner’s Subsequent Reflections
Viner makes little reference to The Customs Union Issue in his subsequent writings on international economics, though, in the years after 1950, he was increasingly absorbed with questions of intellectual history, and published very little on any topic in international economics.
He did, however, comment in a letter to Paul Wonnacott (a Princeton PhD student working on customs unions) that, with the benefit of hindsight, he would now put more emphasis on economies of scale and transportation cost savings from relocating production in his analysis, though he admitted a “skepticism … about capacity of economic analysis with its present kit of tools” to answer the important positive and normative questions about customs unions (Viner to Wonnacott, 3 December 1967).
V. CONTEXTUAL VERSUS PRACTITIONER HISTORY
In the 1950s, Viner’s work was the point of departure for economists such as Meade, Lipsey, and Corden, who developed the modern theory of customs unions. However, their readings of Viner were poor contextual history, sometimes even difficult to reconcile with clear statements in the text of The Customs Union Issue. The reasons these interpretations have been discussed is not to embarrass these trade theorists or vindicate Viner, but to prepare the ground for the following argument about contextual and practitioner histories. This section and the following section take a broader perspective on the relationship between economics and the history of economic thought.
A much-cited manifesto for contextual history is Quentin Skinner’s (Reference Skinner1969) essay on meaning and understanding in the history of ideas. Skinner wrote: “The perpetual danger, in our attempts to enlarge our historical understanding, is that our expectations about what someone must be saying and doing will themselves determine that we understand the agent to be doing something which he would not—or even could not—himself have accepted as an account of what he was doing” (Skinner Reference Skinner1969, p. 6). He describes three common transgressions of this standard for contextual history: first, a “mythology of coherence”—trying too hard to find coherence in an author or criticizing authors who are not coherent in our sense (p. 16); second, a mythology of prolepsis—concentrating so much on significance for us that there is no place for what the author meant to say (p. 22); third, a mythology of parochialism—wrongly attributing our ideas to authors (p. 24).
The historical sins identified by Skinner can be found in abundance the work of Meade, Lipsey, Corden, Cooper and Massell, and Michaely. Skinner’s mythology of coherence is displayed particularly in the attempts of Lipsey and Michaely to fit Viner into a standard, competitive, general equilibrium trade model. Cooper and Massell seem so keen to use Viner as a point of departure for their argument for general tariff reductions that there is no place for Viner’s own discussion of the topic, and they fall victim to the mythology of prolepsis. The mythology of prolepsis is also seen in all the practitioners’ emphasis on trade creation and diversion, taken out its context in Viner’s argument for free trade. The mythology of parochialism is seen in a curiously inverted way in all the practitioners’ desire to find shortcomings in Viner’s work that provide an opening for their own analyses.
There seems to be an inverse relationship between historical adequacy of the practitioners’ readings of Viner’s Customs Union Issue and the practitioners’ success in advancing the theory of customs unions. Is this a spurious correlation, or is there some causal connection between poor history and fruitful science?
One causal connection might be found in the rhetoric of scientific introduction. Introductions to economics papers through this period were becoming more like those in scientific papers, where the contribution in relation to the previous literature had to be presented to the satisfaction of referees. Space was limited, there were no rewards for historical adequacy, and a decreasing number of practitioner referees had enough background in the history of economics to make judgments about historical accuracy or sensitivity. Introductions had to emphasize problems with previous work and identify gaps in the literature that had to be filled by the paper.Footnote 16 The practitioner history in the work of Lipsey, Corden, and especially Michaely perhaps can be seen as longer versions of the scientific paper introduction, serving the cause of scientific advance rather than historical scholarship.
Among historians of economics, there has been a lively debate in recent years between proponents of contextual and practitioner history—sometimes conducted in these terms or in terms of absolutism versus relativism, rational versus historical reconstruction, or thin versus thick history.
Perhaps the most vigorous champion of practitioner history has been Paul Samuelson, who, in a presidential address to the History of Economics Society, proposed that the “history of economic thought more purposely reorient itself toward studying the past from the standpoint of the present state of economic science,” a proposal he called “Whig history of economic analysis” (Samuelson Reference Samuelson1987, p. 52). He proudly adopted the term of “denigration” introduced by Herbert Butterfield, as well as connected with Schumpeter’s call for a history of economic analysis. Schumpeter (Reference Schumpeter1954), of course, actually gave us far more than history of economic analysis in his great work. Mark Blaug (Reference Blaug1990, Reference Blaug2001) is a contemporary champion of the importance of historians of economics’ writing for the contemporary economist practitioner audience, as is Roger Backhouse (Reference Backhouse1992) in milder and more inclusive form.
In the movement among historians of economists towards contextual history, Viner himself was a pioneer, and similar lists of sins to Skinner’s can be compiled from his often acidic book reviews; for example, his famous Brown University Convocation Address, “A Modest Proposal for Some Stress on Scholarship in Graduate Training” (published as Viner 1953); and an unpublished Harvard lecture, “Why Study the History of Economic Thought” (Viner 1962).
In the mid-twentieth century, Viner was in the paradoxical position of being both a practitioner of economics and a pioneer of contextual historical scholarship, as he watched the practice of economics and historical scholarship being driven apart. In the Brown University Convocation Address (1953), he discussed how “scholarship” (for Viner, this means contextual historical scholarship), which was increasingly absorbing his time, is something different from economic theory and policy, which he referred to as “research.” He sharply contrasts the sensibilities of scholarship and research, observing that research increasingly generates “material rewards” for the producer and “tangible benefit to the community” (1953, p. 386), while historical scholarship can now be pursued only for its own sake. Historical scholarship can “distract attention from more vital matters” and be “out of place” (1953, p. 387).
Viner’s position is amplified in the later Harvard lecture, where there is a strong distinction between the value of the history of economics for “problem solving” or “throwing light on current issues,” and its value “as part of the history of ideas” (Viner 1962, p. 1). His answer to the question of whether economists should know the history of their profession was equivocal, and he commented that only “imperfect” disciplines are interested in their history. This would probably have surprised the graduate students who had been assembled to listen to a scholar invited to Harvard because of his reputation as an historian of economics.
Perhaps most influential in the movement towards contextual history among historians of economics has been Donald Winch, through his writings (such as Winch Reference Winch1962, Reference Winch2009) and role in the pioneering program at the University of Sussex School of Social Sciences.Footnote 17 A recent article argues that history of economics and intellectual history are separate enterprises (Winch Reference Winch2009, p. 1), and identifies similar “sins in the intellectual historian’s handbook” to Skinner of “anachronism and prolepsis” (Winch Reference Winch2009, p. 8).Footnote 18 Like Viner, however, Winch does not disparage economists who use the work of past economists in ways that do not meet these standards; he just argues it is something different from history (Winch Reference Winch2009, p. 11) and should not be advertised as such.
The divergence between practitioner readings and contextual history of The Customs Union Issue discussed in this article serves as another example of the different nature of the two enterprises.
VI. THE CUSTOMS UNION ISSUE AS A CLASSIC
Viner’s Customs Union Issue is no longer read by practitioners, or stimulates theoretical development. Among practitioners, it is considered to have been fully mined and its interpretation settled. This is clear in the recent surveys of the economics of customs unions, which usually open with a mention of Viner’s book and laud his contribution to the field, without engaging with his work in any depth. For instance, Pomfret (Reference Pomfret2001) lauds Viner, who “made the greatest single contribution to the theory of discriminatory trading arrangements” (p. 182). Freund and Ornelas (Reference Freund and Ornelas2010) identify Viner as the originator of the distinction between trade diversion and trade creation, and this distinction gives their survey shape, as they assess if recent empirical and political economy studies indicate whether trade diversion or creation predominates. For Krishna (Reference Krishna, Durlaf and Blume2008), Viner is the pioneer of the static welfare analysis of customs unions. Bhagwati (Reference Bhagwati2008) emphasizes Viner’s argument that the formation of customs unions is something fundamentally different from free trade, but neglects parts of Viner’s book that connect with his own discussion of dynamic and political economy aspects of customs unions. Panagariya (Reference Panagariya2000) takes a similar approach.Footnote 19 The picture that emerges from the contemporary surveys of Viner as lauded but unread is reinforced by a scan of the contemporary journal literature on preferential trading arrangements, where there is even less serious engagement with his work.
I suggest that Viner’s work now functions as a classic. What is a classic? One answer is that it is a book you haven’t read but somehow feel you should read, or at least pretend to have read (Bayard Reference Bayard2007). Another somewhat florid account of a classic is Sainte-Beuve (1837, p. 7):
A true classic, as I should like to hear it defined, is an author who has enriched the human mind, increased its treasure, and caused it to advance a step; who has discovered some moral and not equivocal truth, or revealed some eternal passion in that heart where all seemed known and discovered; who has expressed his thought, observation, or invention, in no matter what form, only provided it be broad and great, refined and sensible, sane and beautiful in itself.
A better account that pays attention to how classic works actually function is Condren (Reference Condren1985). He shows that classic texts in political science are not texts that discuss an essentially political set of issues; neither are they part of a continuous historical discussion, nor are they better in some way than other texts. In other words, classics are not as Sainte-Beuve depicts them. What distinguishes classics is that they are exploited by practitioners—“the intellectual community needs them” (Condren Reference Condren1985, p. 284). They can be exploited either as an emblem—“a figure exploited as an emblem is a badge of the cohesive group, an abridgement of its values and a mechanism for maintaining identity” (Condren Reference Condren1985, p. 256)—or authority, which “carries weight across dispute and is part of the shared but contentious vocabulary” (Condren Reference Condren1985, p. 256).
In the customs union literature, Meade, Lipsey, and Corden all used Viner as an authority in Condren’s sense. Viner stimulated their work, and they established the importance of their contributions in relation to Viner, as criticisms or extensions of particular points in The Customs Union Issue. In the contemporary practitioner literature, Viner functions more as an emblem, marking out and justifying the existence of customs union theory as a subfield of international economics. Viner’s book is one that practitioners must connect their work to, pay due respect to, without having to seriously engage with it.
One piece of evidence for classic texts functioning this way is the continuing interest in classic texts and authors among practitioners, despite their disdain for the history of economics. The twenty-year explosion of interest in Adam Smith is a good example, reinforced by the superficiality of much of it and the astounding diversity of causes to which his name is attached. In customs union theory, there is an extended treatment by the practitioner Andre Sapir (Reference Sapir2011) of Bela Balassa’s classic Theory of Economic Integration in the Journal of Economic Literature, a journal that has not published a serious contextual historical piece for many years.Footnote 20
Strangely, there has been little discussion in the history of economics literature of the function of classic texts, though history of economics courses tend to be structured around a small number of classic economic texts, as are textbooks that support such courses.Footnote 21 In the vast literature on late eighteenth- and nineteenth-century “classical” economics, discussion of what actually makes economics “classical” tends to be restricted to distinguishing classical economists from those who preceded and followed them on the basis of shared theoretical content, rather than the way these texts and authors function for subsequent economists.Footnote 22 It is interesting that the authors’ names, as much as the texts themselves, come to be used by subsequent economists. Names are perhaps more convenient emblems than texts. In my view, this group of late eighteenth- and nineteenth-century authors and texts can be described as classical because of the way they functioned as emblem and authority for a subsequent group of economists, but a full justification of this claim is beyond the scope of the current article.
A further question is whether there is a reading cycle for economic texts, something like the industrial product life cycle where a new product sells rapidly, then sells based on its name in the mature phase, until the product is eventually retired. Viner’s Customs Union Issue, as we have seen, progressed from active use to use as authority and emblem. Does this apply to other economic texts, at least to the minority of texts that avoid eventual obscurity? Even if such a reading cycle applies to the political theory texts Condren considered, it may not apply to economics texts, because the distinction between history of economics and actual practice is stronger in economics than in political theory. What about our own classic texts in the history of economics; for instance, the works of Viner, Schumpeter, Samuelson, Blaug, and Winch? Do classics function differently for historians of economics than for economist practitioners?
VII. CONCLUSIONS
The historical account offered here of Viner’s The Customs Union Issue emphasizes how the book flows from his core intellectual commitments to classical liberalism and free trade; and how his thinking on customs unions was shaped by his teacher Taussig and others, sharpened by work for the US Tariff Commission and the Canadian Royal Commission on Dominion–Provincial Relations, was further shaped by the difficult post-war trade environment and discussions with other international economists, then eventually produced after an invitation from the Carnegie Endowment to write an introduction to a planned collection of customs union documents. With an appreciation of the context, we are in a better position to understand what Viner was trying to do, which was to advance the cause of free trade by showing that customs unions were a damaging distraction for a number of political and economic reasons, including the possibility of losses from trade diversion.
Examining the reception of Viner’s work by international economists revealed all manner of historiographical sins. However, rather than burning in the fires of historiographical hell, the practitioners produced good scientific fruit, advancing the theory of customs unions. The rhetoric of scientific introduction increasingly pushes practitioners to commit these historiographical sins, for only if they can demonstrate a contribution relative to the previous literature can a paper be published. As a work is repeatedly used in this way by practitioners, it becomes a classic, largely ceases to be read, and instead is used as an emblem, marking and justifying a field of inquiry. After attaining classic status, it comes to be exploited differently by historians, for contextualization and deconstruction.
Future work could test the hypothesis of a relationship between bad contextual history and scientific fruitfulness in other areas of economics, and consider other causal mechanisms. The rhetoric of refereeing and introductions (and thesis literature reviews, which function in a similar way) in the practitioner literature could be further explored. Does the reading cycle identified here of practitioner use (of use, then attaining classic status, then being exploited differently by practitioners and historians) apply to other classic works of economics?
