1Although the unfolding of the Global Recession since late 2007 and its lingering repercussions have strongly eclipsed the trade and development policy debate, the literature that pertains to the latter has, surely and steadily, continued to grow. (See, for some of the latest ideas, the edited volumes by Stiglitz and Lin, 2013, and Stiglitz, Lin and Patel, 2013). In the meantime, the deadlock in the Doha Round of multilateral trade negotiations held under the auspices of the World Trade Organization also partly serves as a reminder that many of the underlying issues in that debate are still very much alive.
2Over the course of the last few centuries when different countries at different points in their history sought to ignite or speed up their industrial development, a variety of arguments for state support for industry promotion have been advanced. Invariably but not surprisingly, they have also triggered counter-arguments. Given that the support requested by those who have considered state actions desirable often entail some deviations from ‘free trade’ (although other policy actions have often been additionally recommended), this long-lasting and ongoing debate has, for better or for worse, been labeled by many as ‘free trade’ versus ‘protectionism’. Many of such arguments for state support, though far from all, are based on some form of conceptualization of external economies, and in other instances, more broadly of market failures. It is the diverse efforts at this conceptualization, even if crude at times, that have yielded the different arguments.
- 1 This distinction has rarely been fathomed by researchers. Napoleoni (1975, 32) is one of the few ex (...)
3This paper seeks to clarify some aspects of this ongoing debate by contrasting the treatment of external economies, which in some cases are slightly broadened to include related market failure considerations, in two distinct analytical frameworks in the development literature, namely neoclassical and Hirschman’s linkage model. Both of them have roots in Smith’s Wealth of Nations (hereafter WN), specifically in his espousal of the benefits of division of labor, although this paper contends that there are two types of such division, and hence the two different frameworks.1
4Historically, many have grappled with the broad question of whether unfettered interactions of self-interested agents in the market place would promote the greatest socially beneficial effects (Medema, 2009, ch. 2), of which the debate on the desirability of a ‘free trade’ policy is but one, albeit important, aspect. Though by no means the first to do so, Smith’s discussion of that subject has certainly attracted as much attention as any other who has attempted it. In that regard, it could be a very revealing exercise to trace the analytical roots back to him, and then to illustrate how situating external economies in the two frameworks result in different appreciations of their roles in economic processes and of the policy implications that arise therefrom.
- 2 Those who have considered so have only made indirect connections through relating Hirschman to Youn (...)
5In the manner that both frameworks have been developed, both allow for situations of market failures and thus the possibility for state intervention to improve the social outcome of freely operating market forces. However, the root of the neoclassical framework in Smith’s WN has, in some sense, burdened its subscribers with an obligation to defend free trade. As Krugman (1987, 131) puts it in a way that cannot be more telling, “If there were an Economist’s Creed, it would surely contain the affirmations ‘I understand the Principle of Comparative Advantage’ and ‘I advocate Free Trade’.” And he would later add, “even brash young theorists are hesitant to risk excommunication by challenging these tenets directly” (Krugman, 1992, 424). As will be seen in the ensuing discussion, this ‘creed’ has, especially in the postwar era, shaped the refinement of the theory of commercial policy in ways that would rescue the case for free trade from almost all situations of market failures. The situation is quite different with Hirschman’s linkage model, granted that very few researchers have considered relating his works to Smith’s.2 But once one sees the connection between successful activation of linkages with increasing returns, and that some countries have historically been more successful in this regard than others, then free trade can precipitate and aggravate uneven development. For the lagging countries the challenge that they face is how to possibly manage trade, incoming foreign investment, and technology acquisition in order to close the development gap with the leading countries. While part of this challenge again involves addressing market failures, the nature of the linkage model dictates how, at least for some of these failures and the relationship among them, they should be viewed differently than in the neoclassical framework.
- 3 For a good doctrinal history of neoclassical theories of international trade, see Gomes (1990).
6To start from Smith, but yet to bring the discussion up to the present, obviously necessitates some descriptions of the evolution of the respective analyses. Given the enormous size of the trade and development literature, however, such descriptions can only be selective within the scope of a paper like the present one.3
7The neoclassical framework is the focus of section 1. A useful way to view the mainstream trade and development literature is to think of it as looking at two sides of a coin. One side contains the message that under some ‘standard’ assumptions, one of which is that there is no market failure, free trade is the most efficient policy. It is rooted in the two-party specialization (in the production of final products) and exchange situations that Smith considered in the WN. It is most famously illustrated by the tailor–shoemaker–farmer ‘division of labor’ example, which yields conclusions on the benefits of unfettered two-person exchange that he generalizes to the benefits of free trade between “great kingdoms” (Smith,  1979, 456-457). This two-party exchange/trade model is the rudimentary building-block of more elaborate general equilibrium (GE) frameworks that constitute the conventional mainstream trade models. The evolution of this aspect of the neoclassical framework is outlined in subsection 1.1.
- 4 Medema (2009) is in a sense a culmination of research that that author has conducted on this subjec (...)
8Part of the other side of the same coin contains the converse of the aforementioned message, that is, under those same assumptions (including no market failure), any deviation from free trade will be inefficient. The classical economists had thus all warned against the misallocation of resources that would result from protection, though Smith himself allows in the WN for a few situations where protection may be advantageous (Smith, ibid., 463-468). As is well-known, Mill, while issuing the same warning, broke ranks by offering a qualified endorsement of protection in the so-called infant-industry situation (Mill,  1965, 918-919). This endorsement was subsequently echoed by Sidgwick, Marshall and Pigou. All these are illustrated in subsection 1.2. The discussion there benefits from, and thus parallels, a literature that has emerged during the last couple of decades that re-examines the emergence of Cambridge welfare economics (hereafter the CWE literature) especially through the hands of Sidgwick, Marshall and Pigou (O’Donnell, 1979; Aslanbeigui, 1990; Medema, 2009;4 Groenewegen, 2010; Caldari and Masini, 2011; Backhouse and Medema, 2012; Aslanbeigui and Oakes, 2012 and 2014; Collard, 2014).
9Then the remaining portion of this side of the coin, which considers what happened essentially after Pigou, is discussed in subsection 1.3. The key figures here are Haberler, Meade, Bhagwati, Ramaswami, Johnson, and Corden. They significantly contributed to the construction and/or refinement of the theory of domestic divergences/distortions (or sometimes simply the postwar theory of commercial policy). In applying it, different types of externalities or arguments for infant industry protection (or market failures in general) are first identified and separated. Then alternative policy instruments that can conceivably address each market failure under consideration are enumerated. But can these instruments be rank-ordered in ‘efficiency’ terms? Thanks to certain additional assumptions—a pivotal one is that various failures can be compartmentalized for analytical treatment—they can indeed be unambiguously rank-ordered. These researchers all recognized that in the presence of different types of market failures (Corden, 1974, chs. 6-9; Bhagwati et al., 1998, chs. 21-30), there is no doubt that a laissez-faire policy would no longer be efficient. But the key question is whether some deviation from free trade is the best or most efficient way of addressing those failures. Invariably in their theory, with monopoly power in trade being the most prominent exception, trade interventions are ranked inferior to non-trade interventions in the compartmentalized treatment of each failure. That is, a non-trade policy instrument is almost always more efficient in correcting a domestic ‘divergence’ to attain optimal equilibrium, while externally, the (small) country can maintain a free trade stance and reap all benefits that presumably come with it. It thus appears that the case for free trade remains intact.
10The theory referred to in subsection 1.1 rose to dominance between the last third of the 19th and the first third of the 20th century, and enabled neoclassical economists to stake what Hirschman (1981) dubs a ‘monoeconomic’ claim. However, that claim was seriously challenged with the onset of the Great Depression and the birth of Keynesian economics. The two decades or so after WWII was a period that was extremely fertile for the germination of development ideas that rejected that very claim (Hirschman, ibid.). Space only allows the consideration of one such idea here and that is Hirschman’s own linkage model (Hirschman, 1958; 1966; 1977; see also Scitovsky, 1954; Fleming, 1955).
11If the emergence of neoclassical welfare analysis was significantly facilitated by the incorporation of utilitarianism and marginalism into economics (Medema, 2009, ch. 2), the linkage framework, which is outlined in section 2, is based on neither. And while neoclassical welfare analysis is nested within equilibrium frameworks, the linkage model views the development process as a moving chain of disequilibria. Indeed, it will be argued in section 2 that that model can be regarded as being rooted in the other type of division of labor found in the WN. The best representation of this second type of division, albeit at the factory level, is the manufacturing of pins described in Bk. I, Ch. I of that classic. At the economy level, which is more relevant to the present discussion, this is illustrated by the woolen coat example found in the latter part of the same chapter. In both examples, the production of a product (and in the case of the coat, its delivery as well) is broken down into phases of operations. In fact, in the latter example, even some of the inputs into the woolen coat are so broken down, which is at least partly a manifestation of what Young (1928) describes as industrial differentiation in the course of reaping increasing returns. But while the passages related to these examples reflect a state of such division that has been attained, with regard to the woolen coat example the linkage model can be interpreted as an attempt to map out, how, through the activation of production and other linkages, that state of division (and beyond) can come to be realized through a dynamic process where “each move in the sequence [that leads away from equilibrium] is induced by a previous disequilibrium and in turn creates a new disequilibrium that requires a further move” (Hirschman, 1958, 66, italics added).
12The activation of linkages thus enables further division of labor at the economy level to occur. The further such division takes place, the more would increasing return and higher productivity be realized. Historically speaking, some nations have enjoyed a head start over the others in this dynamic process. But for the present paper the pertinent question to pose is: for those nations that were/are trailing, would free trade enable the latent development-inducing linkages in their economies to be activated or would it rather impede their activation? At the risk of simplification, the ‘protectionists’—including the two usual prime suspects, Hamilton and List—generally did not believe so and there are passages in their works that indeed convey early glimpses of linkage considerations. Neither did most of the early postwar generation of development economists like Hirschman (see also UNCTAD (2001)). Due to various external economies (e.g., the technological as well as pecuniary external economies in Scitovsky (1954)) and other market failures, the activation of some linkages simply does not result from the play of spontaneous market forces. However, it remains to be clarified what exact criterion would be employed to support policy actions considering that the model does not involve utilitarianism. What is clear is that unlike the neoclassical framework, the interconnectedness of most linkages precludes many market failures from being compartmentalized for treatment, with the consequence that the unambiguous rank-ordering of instruments that is feasible in the neoclassical framework simply cannot be performed as readily and cleanly. Indeed, the policy message from this analytical framework is currently relatively general, possibly involving packages that contain combinations of trade as well as non-trade instruments that depend on the country and the time concerned. For the latent linkages that are conceivable for activation are obviously not entirely the same from one country to another, and for any given country, from one point in time to another.
13Some of the researchers contributing to the aforementioned CWE literature have carefully illustrated how, contrary to the characterization by their critics, the Cambridge welfare economists—Sidgwick, Marshall and Pigou—had all expressed an awareness of the possible limitations and pitfalls of state formulations and implementations of policies that supposedly are intended to correct market failures. Section 3 briefly considers similar expressions of caution on the part of some ‘protectionists’ and supporters of import substitution. It takes note of how the neoclassical research into rent-seeking was apparently partly driven by the disappointingly low estimate of the cost of protection in many empirical works. Last but not least, it draws attention to a new literature that questions the conclusions of the neoclassical rent-seeking literature. Section 4 concludes.
14In the WN one type of division of labor is best represented by the tailor–shoemaker–farmer example (Smith,  1979, 456-457), although discussions of that type are already laid out in the first few chapters of Bk. I. It is from this example that Smith famously draws the conclusion of the mutual benefits of specialization (in the direct production of final products) and free trade. In his Principles Ricardo spices this up into the shoemaker–hat-maker example by adding the principle of comparative advantage, which example is found in a footnote on a page right after he introduces his famous England–Portugal trade model (Ricardo,  1951, 136n). First in Essay I in his Some Unsettled Questions (Mill,  1974) and then subsequently in a slightly elaborated version in his own Principles (Mill,  1965, Bk. III, Ch. XVIII), Mill further develops this into his England–Germany trade model by incorporating demand considerations so as to solve for international values. Samuelson was so impressed that he wrote, “Mill’s chapters on international values provide the first complete model of general equilibrium—one essentially isomorphic with Jevons’ chapters a generation later on competitive exchange and with Walras’s equilibrium” (Samuelson, 1982, 36).
15But Mill’s model still had to go through some significant reshaping and refinement before it would be turned into a full-blown GE one, like the Heckscher–Ohlin–Samuelson model. While space does not allow a detailed description of this evolution process, the role played by Marshall, Edgeworth, Haberler, and Meade should be briefly mentioned.
- 5 This work of Marshall was initially privately printed in 1879.
16By means of geometrical techniques (e.g., offer curves), Marshall systematically developed and refined Mill’s analysis of international values. He also attempted to transform Ricardo’s labor cost theory into a ‘real cost’ theory by adding capital and other production costs, thereby beginning the modification of the supply-side of Ricardo’s England–Portugal model. At least as significant to the refinement of GE analysis was his investigation into the question of stability in situations of multiple equilibria (Marshall,  1974;5 Gomes, 1990, 8, 28, 34).
17Then it was Edgeworth, whom Pigou allegedly dubbed the ‘toolmaker’ (Gomes, ibid., 46), who extended the demand side analysis to something that mainstream theorists are familiar with today, and that was the introduction of ‘trade indifference curves’, his box diagram, and the contract curve (Edgeworth, 1894). By relating Marshall’s offer curves to the underlying utilities of individuals in each trading country, and with the aid of the contract curve concept, these added up to a pivotal step in setting the stage for conducting some sort of trade and welfare analysis. However, for his indifference curves to capture the entire country’s (community’s) preference, simplifying assumptions have to be additionally inserted that effectively treat each country as a single individual (Gomes, ibid., 45).
18Returning to the supply side, it was Haberler (1930) who, in an effort akin to that made by Marshall to move beyond the labor theory of value, introduced the ‘production substitution curve’ (today’s production possibility curve). Then Meade ( 1966), through the innovation dubbed the ‘international trade butterfly’, achieved for the first time an integrated treatment of production, consumption and trade (Gomes, 1990, 46).
- 6 For early contributions to the trade literature by Cournot and Mangoldt that contain certain neocla (...)
19This is how, in an abbreviated manner, the lineage of neoclassical GE analysis as applied to trade can be traced.6
20The natural flipside of the classical economists’ support of free trade is their rejection of trade protection, although allowances were granted by Smith ( 1979, 463, 465) for protection under national defense considerations and for equalizing the tax treatment of domestic and foreign products, and, in a sense, also for commercial reciprocity (ibid., 467-468). A significant reason behind this rejection rests on static efficiency considerations (ibid., 453), under the implicit assumption of full employment of resources and given technology. This type of reasoning was later religiously echoed by Ricardo ( 1951, 314) and Mill ( 1965, 914).
21 That Mill also broke ranks and endorsed infant-industry protection is well known. It is in Book V of his Principles ( 1965) which is titled, ‘On the Influence of Government’, where he examines what he considers to be the often erroneous theoretical basis of government interference (Ch. X), and yet also the grounds and limits of the laisser-faire or non-interference principle (Ch. XI). According to Medema (2009, 27-28), among the classical economists “Mill was instrumental” in the “elaboration of a greatly expanded theory of the failure of the systems of natural liberty.” It is in the former chapter where he provides a qualified endorsement for infant-industry trade protection.
- 7 Sidgwick recognizes that there could be situations where, within a country, the “original introduce (...)
22Medema (ibid., p. 28) further suggests that that expansion of the theory of ‘market failure’ “continued at the hands of Sidgwick” and he “took center stage” in moving to “a much more markedly positive assessment of the possibilities of corrective policy actions undertaken by the state than … in the classical tradition”. Chapter V (‘Free Trade and Protection’) in Bk. III (‘The Art of Political Economy’) of Sidgwick’s Principles of Political Economy is where he echoes Mill’s qualified endorsement of infant-industry protection ( 2011, p. 488).7 But what indicator or measurement might be feasible and sensible to employ to determine whether corrective policy actions had indeed improved the ‘welfare’ outcome? He considered the possibility of using some money measure of well-being but maintained serious reservations (O’Donnell, 1979, 590).
- 8 For a brief but useful discussion on this last application, see Groenewegen (2010, 29-31).
- 9 But Groenewegen (2010, 37-38) adds that, “In the privacy of his study, he [Marshall] showed himself (...)
23Marshall conceptualized the consumers’ surplus, which he put to use in analyzing tax policies (e.g., indirect taxes on certain commodities), monopolies and the varying welfare implications of different types of returns to scale.8 But as Medema (2009, 56) observes, after establishing the theoretical case for certain classes of market failure and their associated government corrective measures, he “did not take the next step and advocate for a wide-ranging set of policies” to address the failures. That is because of his concern with the incentive for corruption, and other limitations of the political and bureaucratic processes when efforts are made to implement those measures (ibid., 57).9
- 10 In a different context Marshall lamented, “Public opinion in England acted like the savage monarch (...)
- 11 Writing more than 10 years before this passage, he declared, “Taking account therefore of the polit (...)
24When it specifically comes to the subject of trade analysis and policy, the previous subsection has pointed out Marshall’s theoretical contributions to extending Mill’s England–Germany model, in manners that allowed Edgeworth to subsequently take a step that set the stage for welfare considerations of policies. But in practice he made distinctions between “new”/“undeveloped” countries and the older industrial ones. He praised that “[t]he bold upright character of Mill is conspicuous” when he courageously offered his endorsement for protecting “young” industries in the former group of countries (quoted in Whitaker, 1975, 56).10 At one point in his Money, Credit & Commerce he echoes Mill by saying that, “a Protective tax, which helps a young industry to develop its latent strength, may be in the interest of an undeveloped country”. However, he also cautioned that “the tax must inevitably do some hurt to those few of her industries which are manufacturing for exportation” (Marshall,  1991, 218). Indeed, even for a ‘new’ country of his time like the US, which he visited in 1875 “to study the problems of national industry and international trade from the American point of view”, he warned that “however simple the plan on which a protective policy started, it was drawn on irresistibly to become intricate; and to lend its chief aid to those industries which were already strong enough to do without it. In becoming intricate it became corrupt, and tended to corrupt general politics” (Marshall,  1926, 394).11 This kind of caution and other related considerations rhyme with his general reservations against state actions outlined at the end of the previous paragraph. It is then not surprising that when he reluctantly became involved in the Tariff Reform debate in 1903, with all these cautions duly considered, he decided that a free trade policy would be better for an old industrialized country like England. Indeed, even in his Presidential Address to the Economic Science and Statistics Section of the British Association in 1890, he declared, “I believe that many of those arguments that tell in favour of Protection as regards a new country tell against it as regards an old one. Especially for England a Protective policy would, I believe, be an unmixed and grievous evil” (Marshall, , in Pigou, 1956, 263).
- 12 In WW and the first edition of The Economics of Welfare (EW), Pigou considers an additional criteri (...)
25Pigou adopted Marshall’s analytics, especially marginal analysis, but not his consumers’ surplus framework. Instead, there is a certain general welfare criterion—the size and distribution of the ‘national dividend’ in his Wealth and Welfare (hereafter WW)—that is based upon utilitarian ethics that he sought to maximize.12 In doing so, he can be regarded as modernizing Sidgwick’s thoughts (O’Donnell, 1979, 589). As Medema (2009, 60) characterizes it, “he put Sidgwick’s ideas into a Marshallian theoretical framework.” While bringing these elements together, he was led to introduce the concept of marginal net products (MNP) of resources in different uses. Under free market forces, private MNPs in different uses are brought into equality. This would maximize the national dividend if there were absolutely no divergence between private and social MNPs. But if there exists a (some) divergence(s) in certain resource use, then that national dividend achieved under free market forces would not be a maximized magnitude. There is then a possibility that the state could adjust the outcome to augment the national dividend.
- 13 Although he did not particularly devote himself to developing and extending Ricardo’s England–Portu (...)
26Like Sidgwick and Marshall, one of the possible areas for considering state intervention is foreign trade. Like Marshall, Pigou was involved in the Tariff Reform debate and that was also when he began to construct his welfare analysis. In two short books and a few articles published between 1904 and 1906 (Pigou, [1904a] 1975; 1904b; 1904c; 1906a; [1906b] 1968), he systematically evaluated the main elements contained in Chamberlain’s proposal and responded to each. In doing so he demonstrated that he was certainly well versed in the principle of comparative advantage (Pigou, [1904a] 1975, 20-22, 25).13 Focusing specifically on “the effect of Protection upon Labour”, Pigou (1906a, 12-13) spells out what he then considered to be the “correct method” of estimating the impact of a policy program as involving a three-stage economic inquiry:
- 14 This first stage was discussed earlier in Pigou (1904c).
The first stage is an inquiry into the effect of that policy upon the National Dividend as a whole.14 For primâ facie, anything that enlarges that dividend is likely to be advantageous, and anything that diminishes it disadvantageous, to all the agents of production in the country. … The second stage is an inquiry as to the effect of Protection upon the distribution of the dividend among the various agents. … The third and final stage concerns the effects of Protection on the manner in which Labour receives its share. … If … Protection would lessen either the irregularities of employment, or the proportion of people engaged in sweated industries, the consequent improvement in the men might be well worth purchasing even at a cost of some reduction in their earnings.
- 15 Pigou (1906b, Part II) addresses the subject of preferential import duties, which is not part of th (...)
- 16 Pigou ([1906b] 1968, 12-14) again brings up Hamilton and List, and even quoted from the latter’s Na (...)
27Like Marshall, he made distinctions between new (relatively unindustrialized) and old (industrialized) countries. It was for an old country like England that he methodically applied this three-stage economic analysis to arrive at the conclusion that Chamberlain’s proposals stood on weak grounds at best (Pigou, 1906a, 18-31; 1906b, Part I).15 Then it is in considering the new countries where he echoed Mill’s endorsement for infant-industry protection and stated, “It is, however, possible that, in certain circumstances, this loss [from protection] may be more than counterbalanced if it leads indirectly to the training and development of new national aptitudes” (Pigou, [1904a] 1975, 36). Mentioning Hamilton and List, Pigou drew attention to the “more remote effects” of protection when it stimulates “the development … of ‘productive powers,’ … [and] may lead to an ultimate gain more than commensurate with the immediate loss” (Pigou, 1906a, 18, italics original).16 Therefore, “a perfectly good case can be made out for the Protection of particular industries under special circumstances and for a limited time” (Pigou, [1904a] 1975, 37, italics original).
28Similar to Marshall, Pigou expressed practical reservations toward the ability of the government to improve on market failures in general (Medema, 2009, 65-72). He harbored similar reservations specifically with regard to protectionist policies. He took these into consideration when he explained his opposition to Chamberlain’s proposals in the Tariff Reform debate, stating in The Riddle of the Tariff that,
No economist would maintain that abstract theory alone is decisive against every possible form of special and limited tariff manipulation on the part of the State. … [F]rom the difficulty of securing in any actual Government sufficient wisdom, strength, and singleness of aim to introduce Protection only so far as it is advantageous to the community, and withdraw it inexorably so soon as the public interests require its withdrawal, it is practically best for a statesman to adhere to the broad and simple rule of “taxation for revenue only—at any rate in a free community where habits of commercial enterprise are fully developed” (Pigou, [1904a] 1975, 47-48).
29Similar cautions were voiced in Pigou (1906a, 19-20) and again in Pigou (1906b, 80-82), where in the latter he declares that “economic considerations, though scientifically interesting, are not of first-rate importance. It is upon practical considerations that the issue raised by these proposals really depends” (ibid., 80).
- 17 At a later point in his life Pigou did again express his agreement with List, although he again ind (...)
30The reader who is familiar with Pigou’s WW that was published in 1912 will readily recognize that the three-stage economic analysis introduced in Pigou (1906a) and quoted above is reproduced as the headings of Parts II, III and IV in that book. There are thus strong grounds to the suggestion made by Aslanbeigui and Oakes (2014) that his contributions to the Tariff Reform debate can be regarded as the genesis of welfare economics. However, there are two unanswered questions that pertain to the subject matter of the present paper. First, while in the Tariff Reform debate he applied his three-stage analysis to draw policy conclusions for England the old country, he did not do the same for any of the new countries; he merely echoed Mill’s endorsement. To be sure, his publications during that debate would understandably focus on England. But to the extent that he outlined a framework of welfare analysis, it would be of interest to see, even if only hints of, whether it would be equally applicable to other countries (i.e. in other contexts). For instance, is it possible to incorporate something like the cultivation of List’s ‘productive powers’, a notion which Pigou had alluded to, into his fledgling welfare analysis? If not, then is that a reflection of the incompleteness, if not limitation, of that analysis? Second, and this is more of a puzzle than a question, if the genesis of welfare economics took place in the context of a major trade policy debate, then why is there absolutely no trace of consideration of trade policies for both old and new countries in the WW (and, for that matter, also none in The Economics of Welfare)?17
- 18 Marshall did not apply his analytical tools of internal and external economies laid out in his Prin (...)
31Any attempt to answer this latter question could be speculative. In any case, it falls outside of the scope of the present paper. With regard to the first question, it should be noted that neither Marshall nor Pigou (and certainly not Sidgwick) made any effort to dissect and develop Mill’s original endorsement.18 Could it be that, as Singer (1950, p. 477) puts it, “the ‘infant’ argument for protection is but a sickly and often illegitimate offspring” of the static theory of comparative advantage? There was thus a void in applying some sort of ‘welfare analysis’ to treat various cases for protection, though it did not last long.
32With the evolution of neoclassical GE trade analysis outlined in subsection 1.1, and with the emergence of the CWE literature sketched out in subsection 1.2, it is only a matter of time that the two sub-disciplines would feed on each other to yield a more formal ‘trade and welfare’ analysis. This came to fruition in the 1950s, initially chiefly through Harberler (1950) and Meade (1955), where the latter’s work is precisely subtitled Trade and Welfare. Apparently they made independent discoveries and these gave rise to two branches of subsequent development, the priority between which has been subject to some dispute (Panagariya, 2006). One branch has chiefly evolved through the works of Bhagwati and Ramaswami (1963), Johnson (1965), and Bhagwati (1971). It clearly employs a GE framework and is currently best epitomized by Bhagwati et al. (1998). The other branch employs both partial equilibrium and GE analysis. It is best represented by Corden (1974).
33In Harberler (1950) and Meade (1955), one finds early attempts to distinguish between different cases of market failures, including externalities and situations that possibly warrant infant-industry promotion, the corrections of which trade interventions could conceivably be invoked. Non-trade interventions were also mentioned. However, at this stage the attempts at policy-ranking were rather tentative and not rigorous. Hence, the message that emerged was not strongly affirmative: when there are market failures, a laissez-faire policy is not optimal and it could not be sure that free trade would still be the best policy. Not coming out with a definite ‘yes’ on free trade was regarded as agnostic. That was indeed the reaction to a paper that Bhagwati delivered at the International Economic Association Conference at Brissago, Switzerland in 1961, and which was published in 1963. Therein he conveyed such tentativeness and the lack of clarity regarding policy-ranking that would arise. “[T]he ranking of trade policies,” he lamented, “is affected by the income-distribution chosen” (Bhagwati, 1963, 25, italics original). He also showed some awareness of what would later be dubbed ‘second-best’ complications (ibid.) and separately of the problem of ‘imperfect foresight’ (ibid., 29). When Sir Roy Harrod refers to this paper in his Introduction to the conference proceedings, which apparently also captures the feelings of many conference participants, he describes it as reflecting a “chilling agnosticism”. It appeared to him that “in the present phase of static theory, its exponents can no longer derive from it confident recommendations for helping humanity. The first fresh flush of zeal has departed” (Harrod, 1963, xiii).
34As is always the case in economic analysis, definite ‘conclusions’ can readily be bought with bold assumptions. That is precisely what has happened with the construction of the theory of domestic divergences. And as is also typically the case, when these assumptions become repeatedly used, then what initially appeared as bold would start to appear as standard, and might at some point not even be stated. Indeed, as Corden (1974, 43) alarmingly discovers, “in spite of the surfeit of articles in the economic journals on the ‘theory of domestic distortions’, no one other than Meade  has found it necessary to remove or review these assumptions [upon which the theory is based].” He then proceeds to offer a very careful scrutiny, devoting a whole chapter in his Trade Policy and Economic Welfare to examining four assumptions upon which he regards the conclusion of the theory to be dependent. These assumptions maintain that: [A] subsidies can be financed by ‘non-distorting’ taxes; [B] taxation involves no collection costs; [C] there are no costs of disbursement of subsidies; and [D] the income distribution effects of various policies can be neglected (Corden, ibid., 42).
35When applied to a situation where there are external economies accruing to the production of a certain importable, a non-trade intervention in the form of a production subsidy is concluded to be superior to a tariff (Corden, 1974, 9-14; Bhagwati et al., 1998, 302-305). The production subsidy directly targets at the divergence, which is a production externality, and it does not create any consumption distortion loss as the tariff does. Externally, free trade is thus maintained as the tariff is ruled inferior and not imposed, although domestically it is no longer laissez-faire given the deployment of the production subsidy.
- 19 The infant-industry case has at least as often been dissected as situations that involve dynamic in (...)
- 20 Knowledge creation and diffusion is a probable situation of market failure that is quite similar to (...)
36External economies have also been associated with certain situations that resemble possible infant-industry cases for protection.19 An oft-cited situation involves general labor training and the question of who would pay for it (Corden, 1974, 260-262). Firms that invest in doing so cannot force the workers that they train to remain employed with them (the investing firms). Indeed, once these workers have acquired the skills and experience through training, they can fetch higher wages and may readily be poached by other firms that have not invested in such training. It has thus been suggested that these workers should pay for their own training by accepting suitably low wages during the training period, where if such low wages do not allow them to socially subsist at a certain level, then they would have to borrow on the capital market to finance their training. In situations where the capital market has not been adequately developed, like in many less-developed countries, the first-best policy is then to improve that market so that such financing becomes feasible. The second-best policy involves a subsidy that encourages firms to provide worker training. Notice that these are again non-trade interventions. They are regarded as far superior to a tariff, which, while it renders the domestic production and sale of the product concerned to be more profitable, will at best generate a remote impact on labor-training decisions. Besides, it generates a consumption distortion loss in the market for the product concerned.20
- 21 See Bhagwati and Ramaswami (1963, 50) for an example of a zealous assertion of this.
- 22 Johnson (1965, 7-8) shifts the onus of proof to those who support the use of trade taxes, suggestin (...)
37It is not a coincidence that in both of the applications of the theory of domestic divergences above, some sort of non-trade subsidy (or more generally, non-trade intervention) is ranked more efficient than a trade intervention. Thanks to the assumptions that Corden enumerated, the theory is ‘subsidy-biased’ (Corden, 1974, 43, 54). Thus, assumption [A] solely reserves ‘by-product’ distorting effects for trade taxes while any other tax that finances non-trade interventions is assumed not to exhibit such effects.21 When it comes to the collection costs of various taxes, assumption [B] appears to level the playing field. However, as is typically the case in most less-developed countries, trade taxes incur relatively lower collection costs, though that advantage is here conveniently assumed away.22 Assumption [C] is one that according to Corden (1974, 43), the subscribers to the theory do not bother “even to state it”. It negates an important advantage that tariffs possess over subsidies in general, in that a tariff simultaneously taxes consumers and subsidizes producers without having to actually collect and distribute funds. It is thus not surprising that this feature is sometimes dubbed a ‘two birds with one stone’ argument for using tariffs (Keesing, 1974, 210; see also Corden, ibid., 49). However, to buttress the subscribers’ case for the ‘subsidy-biased’ approach, subsidy disbursement costs have been implicitly assumed away.
38Corden’s own scrutiny of the four assumptions led him to the verdict that “the central argument is unshaken” by removing assumptions [A] and [D]; discarding assumption [B] “slightly dents it”, while abandoning assumption [C] “does affect it” (1974, 33). Perhaps because of his leniency, the issues raised by that review have been belittled or neglected. Thus, for instance, in their exposition and application of the theory Bhagwati et al. (1998, 286) simply declare that all issues which pertain to income distribution, to the infeasibility of lump-sum transfers, or to the direct costs of raising revenues and disbursing them will not be considered. Similarly, most other applications of the theory are continued to be made that are oblivious to, or show an unawareness of, the assumptions, not to mention the consequences of their removals. In the meantime, the apparently clean rank-ordering and powerful conclusions have emboldened the subscribers to forcefully state their case against infant-industry tariff protection (Baldwin, 1969), and has led prominent development economists like Myint to declare that the theory “cuts the ground under all the arguments for tariff protection which purport to correct the misallocation of resources” (Myint, 1987, 114, italics added).
- 23 In the later years of his life Mill expressed regret in some private correspondence over his endors (...)
- 24 See, for instance, Ho (2013, 561-563).
- 25 See Meade (1955, 102), and Lipsey and Lancaster (1956-1957, 11).
- 26 Such compartmentalization is neatly reflected in how each chapter in Bhagwati et al. (1998, chs. 21 (...)
39While in Mill’s endorsement passage he did make his case by referring to a protecting duty ( 1965, 919),23 a careful re-examination of the works of Hamilton and List (and for that matter other leading suspects of protectionism) indicate that they had invariably considered non-trade instruments as well, and occasionally made sensible comparisons of the desirability of alternative instruments.24 More usually, they discussed the deployment of instruments on multiple fronts to spur the development of the countries that concerned them. (See section 3 below in the case of List.) That is a reminder of a fifth assumption behind the theory of domestic divergences, a very implicit one, the presence of which some subscribers to the theory would admit when confronted, but otherwise would rather not bring it up. And that is: market failures do not simultaneously occur in manners that render the policy-ranking which otherwise applies to correcting each failure in isolation ambiguous. For as Stern (1989, 617) puts it, “if several failures occur simultaneously it can be misleading to analyse them one-by-one”. This assumption is specifically intended to steer clear of the complications associated with the Theory of Second Best25 and to enable unambiguous rank-orderings of alternative instruments that correct each failure in isolation. In practice, it allows the analytical treatment of each failure to be compartmentalized from the treatment of others as if no two or more such failures simultaneously occur.26 Its presence is also reflected in another assumption that occasionally is explicitly made, which states that any revenue that is raised from a trade tax, should the latter be deployed, is simply distributed back to consumers in a non-distorting manner. For where is the need to consider the possibility of using trade-tax revenues to fund the correction of any additional failure, when it is assumed that there is none aside from the one under consideration? There is absolutely no chance that a trade tax could earn extra ‘brownie points’ in these policy-ranking exercises.
40While it was in the 1950s when the trade and welfare literature emerged as outlined in subsection 1.3 above, it was also in the 1950s when Scitovsky drew attention to an important alternative approach to thinking about external economies. “[I]t is becoming increasingly clear,” he (1954, 143) thus declares, “that the concept of external economies does duty in two entirely different contexts.” One of these is “equilibrium theory”, which was the focus of Section 1. The other he simply calls “the theory of industrialization in underdeveloped countries”. And it is in the late 1950s that a variant of this latter theory took more concrete shape in Hirschman’s linkage model (Hirschman, 1958; see also 1968 and 1977).
41To Scitovsky, external economies “stand for very different things in the two contexts” (1954, 143). Those defined in the theory of industrialization are “much wider” and “include, but go far beyond, the external economies of equilibrium theory” (ibid.). Specifically, in GE theory one finds ‘technological external economies’, which involve direct/nonmarket interdependence among producers. On the other hand, in the theory of industrialization there are in addition ‘pecuniary external economies’, which involve interdependence among producers through the market mechanism (ibid., 145–6). He illustrates the latter by considering a situation where profits are treated as “a sign of disequilibrium”. Thus, in a freely competitive industry the existence of profits attracts investment in that industry. As that tends to eliminate the profits that have called it forth, that initial investment can be regarded as bringing equilibrium in that industry nearer. However,
The same investment … may raise or give rise to profits in other industries; and to this extent it leads away from equilibrium. … The profits of industry B created by the lower price of factor [that is the output of] A, call for investment and expansion in industry B, one result of which will be an increase in industry B’s demand for industry A’s product. This in turn will give rise to profits and call for further investment and expansion in industry A (ibid., 148-149).
42This passage caught the attention of Hirschman (1958, 66), who regards it as “a most pertinent portrayal of how development is set and kept in motion”, even though he was a little disappointed that Scitovsky ‘short-circuited’ his description of this process in the ensuing part of the above passage by “reach[ing] in a single jump a new point of equilibrium where the ‘elimination of investment’ has been accomplished.”
- 27 Hirschman (op. cit., ch. 6) refers to input-output analysis to aid his introduction of the producti (...)
43 Hirschman attempted to formulate this development mechanism more carefully and the results from his first attempt were published in his Strategy of Economic Development (hereafter Strategy). He observed that “development is a lengthy process during which interaction of the kind described by Scitovsky takes place not only between two industries, but up and down and across the whole of an economy’s input–output matrix, and for many decades” (1958, 66).27 He deduced from Scitovsky’s example that the external economies that the latter emphasized “are essentially caused by production complementarities of one type or another, and we are thus returning to the complementarity effect of investment … the investment-promoting character of investment” (ibid., 67). Further borrowing from Scitovsky his insight that development involves a chain of disequilibria, Hirschman decided that “[a]n ideal situation obtains when … one disequilibrium calls forth a development move which in turn leads to a similar disequilibrium and so on ad infinitum” (ibid., 71–2). Put differently,
[T]he sequence that ‘leads away from equilibrium’ is precisely an ideal pattern of development from our point of view: for each move in the sequence is induced by a previous disequilibrium and in turn creates a new disequilibrium that requires a further move. This is achieved by the fact that the expansion of industry A leads to economies external to A but appropriable by B, while the consequent expansion of B brings with it economies external to B but subsequently internal to A (or C for that matter), and so on. At each step, an industry takes advantage of external economies created by previous expansion, and at the same time creates new external economies to be exploited by other operators (ibid., 66-67, italics added).
44Then he rephrased this process in terms of production linkages. After defining backward and forward production linkages and how ‘satellite industries’ can be established through such linkage effects (ibid., 100-102), he pointed out that the ‘joint production linkage effects’, that is, the linkage effects of two or more industries considered in combination, “are larger than the sum of the linkage effects of each industry in isolation”. Thus, for instance,
When industry A is first set up, its satellites will soon follow; but when industry B is subsequently established, this may help to bring into existence not only its own satellites but some firms which neither A nor B in isolation could have called forth. And with C coming into play some firms will follow that require the combined stimuli not only of B and C but of A, B, and C. This mechanism may go far toward explaining the acceleration of industrial growth which is so conspicuous during the first stages of a country’s development (ibid., 104, italics original).
45Indeed, this “helps to account for the cumulative character of development” (ibid., italics added).
46 As indicated in the Introduction to the present paper, successful activation of these linkages over time means increasing realization of extended division of labor of the kind represented by the woolen coat example in the WN. To Smith, this kind of division of labor is limited by the extent of the market (Smith,  1979, Bk. I, Ch. III). This point is very much echoed by Hirschman in his Strategy when, in considering the activation of backward linkages, he appeared at this stage to take for granted that as soon as domestic demand passes the threshold of minimum economic size, domestic production would be undertaken (ibid., 118; see also Hirschman, 1968, 15-16). But if it is such a spontaneous process, then should a country’s development simply be left under the guidance of Smith’s invisible hand?
47 Hirschman certainly did not think so. To more fully appreciate this one should realize that his framework can readily be applied to interpret the historical process of uneven development between developed and less-developed countries. Thus, historically some countries (e.g., England) were able to introduce the kind of division of labor that is the focus of this section sooner than other countries, and then were able to extend it rapidly as their markets grew and as they also exported to overseas markets. They enjoyed rapid productivity enhancements and technological improvements, and reaped many ‘first-mover’ advantages. But many others lagged behind these leading industrial powers and a situation of uneven development emerged.
48Incidentally, this is exactly what Marx observed at one point in Capital, vol. I. For the leading industrial countries, “as soon as the general conditions of production appropriate to large-scale industry have been established, this mode of production acquires an elasticity, a capacity for sudden extension by leaps and bounds”. Although securing expanding sales outlets and sources of raw materials could be barriers to their sustained rapid growth, the increasing cheapness of their machine-made articles, coupled with revolutions in the means of transport and communication, empowered them to conquer foreign markets, ruining the handicraft production in many lagging countries and forcibly converted them into producers of food and raw materials. A new international division of labor had thereby been created that is suited to the requirements of the leading industrial countries, where one part of the globe had been converted into a chiefly agricultural field of production for supplying the other part that was a pre-eminently industrial field (Marx,  1977, pp. 579-580). For the agricultural countries, the pressing question then becomes: Can the invisible hand rescue them and enable them to close the development (including technological) gap?
- 28 The mutual benefits of specialization and trade that supposedly follow from Smith’s tailor–shoemake (...)
49 A very important distinction that Hirschman (1958) makes is how different products bring with them different richness or paucity of production linkages. The parallel with Smith is again noteworthy. “The nature of agriculture, indeed,” it is thus stated in the WN, “does not admit of so many subdivisions of labour, nor of so complete a separation of one business from another, as manufactures.” This is “perhaps the reason why the improvement of the productive powers of labour in this art, does not always keep pace with their improvement in manufactures” (Smith,  1979, 16).28 To Hirschman, “agriculture certainly stands convicted on the count of its lack of direct stimulus to the setting up of new activities through linkage effects: the superiority of manufacturing in this respect is crushing. This may yet be the most important reason militating against any complete specialization of underdeveloped countries in primary production” (1958, 109-110).
50This is where the works of some so-called ‘protectionists’ become relevant. Hamilton ( 1966) and List ( 1983;  1991), for instance, consider the disadvantages of countries that are predominantly agricultural. They had read the WN and appreciated the significance of achieving this second type of division of labor so that America, France and the German states of their respective times could increase their national wealth accordingly. But with reference to what England had achieved in that regard in those times, their question was how the lagging countries could replicate such type of division. Although they did not employ the concept of linkages that Hirschman was to later popularize, Hamilton and List, with minds not then interfered by Marginalism and by assumptions like perfect information, conceived of development processes simply as ‘one thing leading to another’ (or, more generally, some things leading to some others, and so on, ad infinitum).
- 29 List’s acute awareness of the linkages and interconnections between various “occupations” or units (...)
51For instance, in his Natural System List refers at one point to an iron-master who makes pig iron. As if he had backward and forward linkages in mind, he claims that this person “cannot make a profit if he is unable to secure supplies of cheap fuel because the local collieries have not been sufficiently developed. Nor can he make a profit unless there are in existence enough foundries, steelworks, and engineering plants to buy the pig iron that he produces.” Generalizing from this, he concludes that, “Every factory needs to operate in association with countless other enterprises which supply raw materials, buy the finished product, or construct and maintain machinery. No factory reaches its maximum efficiency unless all the factories with which it is linked have also reached their maximum efficiency” ( 1983, 72-73, italics added). In his National System he similarly considers at one point a “machine manufactory”. For its success, he observes, “it is necessary that the mines and metal works should furnish it with the necessary materials”. Equally important, “all the hundred different sorts of manufactories which require machines should buy their products from it” ( 1991, 151). When viewed as part and parcel of development processes, that means “by the establishment and continuance of industry one branch of trade originates, draws after it, supports and causes to flourish many others” (ibid., 298).29 But instead of expressing faith in the invisible hand, he, as well as Hamilton, believed that the state would have a role to play in igniting the process through both trade and non-trade measures, where the latter would include technology policies that are intended to help close the technological gap.
52What about Hirschman? What are the possible reasons why an agricultural and/or primary producing country, left on its own, could not spontaneously diversify into industrial activities? He offered some clearer answers in the couple of decades following the publication of his Strategy. While market-size considerations are important for the activation of linkages, “the backward linkage dynamic may be held back at some point simply by ‘technological strangeness’.” Specifically, “[w]hen a backward linkage effect points to an industry which is technologically quite distinct from the one requiring the input, the input-utilizing industrialist is less likely to be attracted to the input-producing industry than if the latter is closely related to processes and techniques with which the industrialist is already familiar.” As examples, he contrasted the “spontaneous vigor” that would likely exist in the linkage dynamic between the ‘inbred’ metalworking and chemical industries on the one hand, with the relationship of ‘technological strangers’ between the textile and chemical industries on the other. “This point is of particular importance for the machinery industry,” he concluded, “since machinery is usually a technological stranger to the industry in which it is utilized.” Hence, “[t]o identify and then to remove this sort of bottleneck should be a principal task of public agencies concerned with industrial development” (Hirschman, 1968, 24).
53Applying this new consideration to agricultural and primary activities, Hirschman observed that “the size of the technological jump is almost always big for backward linkage since, under modern conditions, the technique of growing wheat or coffee or sugarcane has very little in common, except for the seeds or cuttings, with the technique of fabricating the tools or synthetic fertilizers that are used as inputs by the growers.” The same can be said for many of the forward linkages that are crucial elements of staple-based economic development. That is, “[t]he milling, refining, and other processing industries are usually technological strangers to the staple. … Hence, for a very long time these industries are carried on abroad, where, moreover, they are usually protected through the well-known tariff pattern making for stiff effective rates” (Hirschman, 1977, 78).
54Hirschman further argued that “technological strangeness is often combined with other obstacles to backward or forward linkages being taken up by the producers themselves”. For instance, there is a need for large amounts of capital due to scale requirements to utilize certain technologies. Separately, there is typically a lack of marketing access and knowledge (Hirschman, 1977, 78, n. 17).
55Based on his consideration of technological strangeness, should the state act to remove such a bottleneck for some sectors using combinations of trade and non-trade policy instruments, then investments that “lead away from equilibrium” would be undertaken, initially in those sectors. But these would generate pecuniary external economies, which when harnessed by certain firms (in those and/or other sectors) and/or by the state with new investments, would generate yet more external economies, and so on in chains of disequilibria. It is through the activation of backward and forward production linkages along such chains over time that an initially less-developed country could gradually replicate the relatively intricate network of division of labor that one typically finds in the advanced countries.
56It should be noted that in Hirschman’s later works (e.g., Hirschman, 1977), he broadened the linkage concept beyond production linkages to include consumption linkages, fiscal linkages, banking linkages, and (subtly) institutional linkages. Limited by space, these are only mentioned here in passing.
57In referring to the CWE literature in subsection 1.2 above it was indicated that both Marshall and Pigou expressed reservations toward the government’s ability to improve market outcomes in general (Medema, 2009, 56-57, 65-72). The discussion there also provides some textual evidence that illustrates that they held the same kind of reservations when it specifically concerned employing trade policies to achieve certain outcomes.
58Leaving aside the Cambridge welfare economists, even the so-called protectionists had expressed some cautions. Thus, in considering the “drawbacks of tariffs”, List laments in his Natural System that “they create thousands of customs officials in our frontiers; they restrict complete freedom of trading … And to some extent tariffs lower the moral standards of those who live near a frontier by putting temptation in their way to break the law.” “If anyone could suggest how the advantages of a tariff can be secured without the drawbacks,” he then offers, “we would be delighted to support such a proposal” ( 1983, 105). Before then, unfortunately, only if a nation is prepared to put up with such drawbacks can the advantages of a tariff be secured (ibid., 106). What is interesting about these passages is that right after them, he describes, to the probable embarrassment of those who thought that all List knew was to advocate tariffs, the “many ways in which industrialisation can be promoted.” To quote him at length:
- 30 This passage ends with the sentence, “But all these measures will have little effect unless they ar (...)
They include the establishment of technical schools; the granting of financial assistance to enable scientists to travel abroad; the holding of industrial exhibitions at which new inventions can be shown; the establishment of companies to promote industry and internal commerce by regulating rivers and by constructing highways, canals, railways and steamships; the granting of honours to those who advance scientific knowledge or establish new industries; the payment of generous subsidies for new industries, new processes, new plants, and new efficient factories; and the granting of state loans to industrial enterprises” (ibid., 106, italics added).30
59This list was essentially reproduced in his National System, with the additions of “offers of prizes” and “patent laws” (List,  1991, 308). In fact, it was because he wanted to guard against complacency on the part of the protected domestic manufacturers that he proposed the state to offer awards and prizes. He believed that these could serve to encourage product quality improvements, efficiency in production, consistency in such quality and efficiency performance, and technical knowledge acquisition and sharing (List,  1983, 120). Nor was it the case that when it came to tariffs, List wanted them to be prohibitive. Instead, he maintained that “under all circumstances partial but carefully limited foreign competition is really beneficial to their [the Germans’] own manufacturing progress” ( 1991, 188).
60 Into the postwar (WWII) decades, supporters of import substituting industrialization are often grouped together with protectionists. But again, the more prominent suspects had all warned of the possible drawbacks of protection, especially if pursued excessively. That is the case for Prebisch and Singer (Ho, 2012, 874-879). The same is true of Myrdal, who discussed the ‘soft state’ and problems of corruption (Myrdal, 1970; Ho, 2010, 282-285). Seen in that light, the cautions voiced against rent-seeking and/or ‘directly-unproductive profit-seeking’ (DUP) activities by the neoclassical trade and development researchers since the 1970s are not entirely new, except that by employing public-choice theoretic modeling they have advanced that area of research in a formal way.
61And there is a side-note to the drive behind the neoclassical literature on rent-seeking. Given the analytical thrust of the body of theories described in section 1, it is only obvious that it would spur a huge volume of empirical work that sought to determine the cost of protection and/or how protection may have hurt certain aspects of economic performance. Especially prior to the 1970s, most estimates were based on partial equilibrium frameworks to determine the size of the conventional ‘Harberger–Johnson–Meade triangle’ of production and consumption distortion losses that resulted from protection. However, in the words of Bhagwati (1989, 29), these estimates “have been dismally low and singularly unfit as a weapon to fight protectionism on the ground that it is expensive”. Similarly, Krueger (1984, 544) echoes that “a number of economists have expressed unease because of the small size of the estimates”. It is then partly because of this disappointment that focus started to be placed on the deadweight costs of tariffs and quantitative restrictions, with rent-seeking being one such possibility (Krueger, ibid.). Other possibilities include the costs associated with various DUP activities (Bhagwati, 1989, 29-30). In other words, rent-seeking and DUP activities became a popular subject for analysis in the economics mainstream partly because of the frustration that the traditional static costs estimated by various cost-of-protection calculations have come out to be too small; additional evidence is needed so that there can be a stronger case to attack trade protection. Regardless of what drove it in the first place the research that has sprung up has raised some important questions.
62But in recent years there has emerged a new body of research that casts some doubt on the principal claims of this literature. In particular, it has drawn attention to the fact that the neoclassical rent-seeking literature has focused its attention almost exclusively on the social costs of the resources used up in the rent-seeking process. But at the same time very little attention has been paid to the different types of rents which rent-seeking and the DUP activities have created in different contexts, and how they have been put to different uses with varying effects. Put differently, the neoclassical rent-seeking analysis typically focuses on the ‘input’ side of the rent-seeking story, while generally neglecting the ‘output’ side of it. In contrast, this new body of research investigates into how the overall effect of rent-seeking depends on both the cost incurred and the rent created (Khan and Sundaram, 2000).
63The findings of this new line of research can also have implications for re-examining the neoliberal push for good (market-enhancing) governance, which is distinct, though not necessarily mutually exclusive from, ‘growth-enhancing’ governance. The latter type of governance relates to capacities in the acquisition of productivity-enhancing technologies, maintaining political stability in the course of social transformation, and more generally overcoming various types of market failures (Khan, 2007).
64In some ways this brings the discussion back to yet one more revelation in the CWE literature. Thus, while Sidgwick, Marshall and Pigou were all skeptical about what the government could achieve during their times, they were cautiously optimistic about the longer-term possibilities for more effective intervention (Backhouse and Medema, 2012, 990-993). Development economists who still have faith in the governments of the countries that they are researching into can almost take a page out of the theory of domestic divergences: if the principal reason for one to doubt the effectiveness of interventions to correct market failures is government failure, then should the first-best response be doing something to ameliorate that government failure? Indeed, as Lall (1999, pp. 16-17) puts it when he concludes his own study of the ‘high-performing Asian economies’ minus Japan, “if government capabilities to intervene properly are insufficient, the first remedy must be to improve upon those capabilities”. That requires some cautious optimism of the kind shared by the Cambridge welfare economists.
65This paper contrasts the treatment of external economies in the neoclassical framework with that in Hirschman’s linkage model, and compares the differences and similarities in policy messages. To put these differences and similarities in better perspective, it outlines the doctrinal historical roots of both frameworks, tracing the roots to Smith’s WN, though with each framework attached to one of two types of division of labor that is considered in that classic. The discussion underscores the importance of critically appraising the role of concepts like external economies, or market failures in general, against the analytical frameworks in which they are situated.
66The neoclassical trade literature has enjoyed a tremendously long period of development and refinement. It is firmly rooted in Smith’s example of the benefits of specialization (in the direct production of final products) and trade. Because of that root, throughout the long period of development and refinement of his example into full-blown GE trade models, a creed has become deeply ingrained in the mind of neoclassical economists, and that is to advocate and defend free trade. Hence, assumptions that ensure that the free trade conclusion will result from deductive reasoning have become standard attachments to the models. Although allowances are made for market failures, additional assumptions are made to ensure that under almost all circumstances, free trade is still regarded as the most efficient policy as long as some non-trade interventions are implemented to remedy the domestic divergences upon which the failures are based.
67It is more possible to disassociate oneself from this creed if one turns attention to the other type of division of labor in Smith, where the production operation in procuring different products is increasingly divided up into phases. This paper suggests that Hirschman’s linkage model can be regarded as associated with this other type of division of labor. While it has attracted much less interest in its extensions and refinements than the neoclassical framework, its distinctiveness from the latter is very noteworthy. By incorporating into his framework various obstacles to the activation of linkages, he had brought together what Scitovsky regarded as elements of the ‘theory of industrialization’, namely both technological and pecuniary external economies, while leaving room for other market failures as well. To be sure, neoclassical economists will recognize some of these as failures by the same label that they have treated under their framework. However, to better allow the conclusion that free trade is the best policy to remain intact, that treatment hinges on the compartmentalization of various market failures. In contrast, the holism that is characteristic of Hirschman’s framework—which puts strong emphasis on inter-sectoral connections—implies that there is at most limited room for compartmentalizing the treatment of market failures. In the neoclassical model, market failures are treated within an equilibrium framework. In Hirschman’s model, wherever the initial points of intervention might be (e.g., to overcome technological strangeness for certain ‘strategic’ sectors), the belief is that the initial rounds of investment and intervention would “lead away from equilibrium” and generate pecuniary external economies that are to be harnessed by some firms or by the state itself in subsequent rounds of investments and interventions, thereby triggering a chain of disequilibria. The investments undertaken in the process are typically lumpy, which contrasts sharply with the marginal changes in the neoclassical framework. Instead of rank-ordering alternative individual policy instruments that would each correct any given market failure with different degrees of exactitude, this framework would call for actions on multiple fronts, with adjustments as (and if) development progresses. And given that the policy instruments deployed on these multiple fronts involve both non-trade and trade instruments (including how to overcome technological strangeness), it does not set itself up so much as a debate between free trade and protectionism.
- 31 For instance, Krugman (1993, Appendix) claims to have modeled forward and backward linkages in “min (...)
- 32 Very recently, the conceptual relevance and significance of linkage processes find some expressions (...)
68However, whether this linkage model will attract more interest remains to be seen. It is hard to imagine that there is much room for formal modeling if one genuinely seeks to capture the richness of various linkage mechanisms as originally spelled out by Hirschman.31 But while it may possess some conceptual appeal, there are also reasons why it has largely remained unattractive, at least to neoclassical economists. For instance, the exact meaning of pecuniary external economies has been questioned and their importance doubted (Mishan, 1965, 12-15; 1971, 6). They are commonly associated with ‘problems’ of indivisibilities. There is then the consideration of how investment decisions are very much interdependent in an economy with less than perfect information (Chenery, 1959). What these considerations suggest is that, to extend and refine this model, much conceptual clarification remains to be done. In the course of such an exercise the emergence and the role of firms as nodes of ‘knowledge’ in the linkage network, as well as how alternative industrial structures may differentially impact on the speed and patterns that linkages may spread, would have to be investigated. All this also serves as a reminder that there is a void in the model’s ‘welfare’ criteria: what criteria can conceivably be employed to evaluate the efficacy of alternative policy packages, that is, different programs of actions on multiple fronts? Where should the state first intervene to ignite the unleashing of the chain of disequilibria and why? To move beyond stipulations of policy prescriptions at very general levels, there is then a call, more to historians than econometricians, to carry out careful and meticulous historical case studies. Last but not least, how dense a linkage network would it be sensible for each country, or each regional grouping of small countries, to attempt to build over time? How does that depend on the size of each country? At what point would it be deemed too much duplication across countries and regions?32 Put differently, how much room would be left for some specialization and trade akin to Smith’s first type of division of labor after some allowance has been made for each country to establish their respective linkage networks?
69In the meantime the neoclassical framework in the form of the theory of domestic divergences appears to have approached maturity. Back in the 1970s when there was what Corden characterized as a “surfeit of articles” on this theory, the famous trade and development economist, Díaz-Alejandro, shrewdly remarked that “any bright graduate student, by choosing his assumptions regarding distortions and policy instruments carefully, can produce a consistent model yielding just about any policy recommendation he favored at the start” (Díaz-Alejandro, 1975, 97). Occasionally in recent years, new articles composed along that line have been published in journals in the area of trade and development, although the flow of such articles has diminished compared with the heyday. It is worth noting that Díaz-Alejandro appended a caution to the above remark: “Algebra and consistent models can prove nothing about the real world, but perhaps the major contribution of these models, and of those sure to follow them, is to force a discussion of the realism of assumptions that are crucial to theoretical demonstrations” (ibid.).
70Hopefully, the discussion in this paper has contributed to that end.
The author would like to thank two anonymous reviewers for their helpful comments that helped his revision of the first draft of this paper. All remaining errors are, however, solely his own.