Dependency Theories Andre Gunder Frank’s most influential book, Capitalism and Underdevelopment in Latin America (Frank 1979a, cited below as CULA), starts with a sentence which sums up his position admirably: I believe, with Paul Baran, that it is capitalism, both world and national, which produced underdevelopment in the past and which still generates underdevelopment in the present. (CULA: xi) It is essential to realize how drastic a break with the classical Marxists this is. Compare it with Marx: National differences and antagonisms between peoples are daily more and more vanishing, owing to the development of the bourgeoisie, to freedom of commerce, to the world market, to uniformity in the mode of production and in the conditions of life corresponding thereto. (Marx, Manifesto: 49) Frank’s claim that capitalism causes underdevelopment is characteristic of the dependency theories which dominated Marxist (and radical non-Marxist) thinking on the world economy from the late 1960s to the late 1970s. These theories see the world capitalist system as divided into a centre and a periphery (terminology varies; metropolis and satellite, or core and periphery, are alternatives). The normal processes of the system cause the gap between centre and periphery to widen, as the centre develops at the expense of the MARXIST THEORIES OF IMPERIALISM 162 periphery, while the periphery is reduced to a state of dependence. Imperialism, in the usual sense of political and military dominance, plays a secondary role in dependency theories, which were intended to explain what was seen as a continued failure of development in the Third World, in the era of decolonization. In so far as dependency theorists discussed the history of underdevelopment, one can distinguish a strong form of the theory (Europe found countries that were developed for the time, and made tham underdeveloped), and a less common weak form (Europe prevented underdeveloped countries from developing) (Griffin and Gurley 1985). During the 1970s, dependency theory came under increasing fire, and by the 1980s those writers still working in the dependency framework were visibly on the defensive. I shall concentrate on a few of the most influential writers in the dependency tradition, primarily Andre Gunder Frank, Immanuel Wallerstein, and Samir Amin, with the emphasis on the works which were most discussed and emulated. By contrast with the earlier periods surveyed in this book, a great variety of work in the dependency framework exists, so the discussion has to be selective. This book is about Marxist theories of imperialism, not about modern theories of economic development, and topics for discussion have been selected accordingly. Dependency theory emerged from debates on development economics, and by no means all of its adherents claimed to be Marxists; were I writing the history of dependency theory in its own right, I would have more to say about writers like Cardoso and Dos Santos. Frank is sometimes treated as the inventor of dependency theory (Simon and Ruccio 1986), though others could dispute the honour with him. His work has typically taken the form of essays, subsequently collected and published in book form. The most important collections are Capitalism and Underdevelopment in Latin America (CULA, 1969a, first published in 1967) and Latin America: Underdevelopment or Revolution (LAUR, 1969b). Lumpenbourgeoisie: Lumpendevelopment (LL, 1972) was a reply to his critics, and represents a minor modification and restatement of his position. Dependent Accumulation and Underdevelopment (DAU, 1978) represents a further shift on some issues, and was influenced by other writers, notably Amin. DAU contains a historical account of the world economy interspersed with more theoretical essays. Of Frank’s many subsequent publications, three DEPENDENCY THEORIES 163 (1981, 1983, 1984) deserve mention. I shall concentrate on the earlier works, since they have become part of the intellectual history of the period, even where Frank has subsequently changed his views. Frank was criticized by Laclau (1971), among others. Wallerstein’s The Modern World System (MWS, 1974a) is the first instalment of a promised four-volume analytical history of the capitalist world economy, continued in The Modern World System II (1980); between them they cover the period from 1450 to 1750. A more accessible introduction to his views is to be found in Historical Capitalism (1983), and in an article (1974b) which is included, with others, in a useful collection, The Capitalist World Economy (CWE, 1979; see also 1984). Wallerstein’s views (and those of Frank) have been critically analysed by Brenner (1977). Amin’s major works are Accumulation on a World Scale (1974, cited below as AWS), Unequal Development (1976, cited as UD), and Class and Nation: Historically and in the Present Crisis (1980), which all cover very similar ground. Imperialism and Unequal Development (1977, cited as IUD) is a collection of essays, one of which, ‘The End of a Debate’, sets out Amin’s main ideas very clearly. He has also written a considerable amount about particular areas, notably North Africa (1966) and West Africa (1971). Addo (1984) contains useful essays by Frank and Wallerstein as well as Amin. Amin has been criticized for eclecticism (Barone 1982; cf. Griffin and Gurley 1985), though it is not clear why that should be a vice. 8.1 FRANK Frank identified capitalism with a system of (world-wide) exchange, characterized by monopoly and by exploitation. He also (implicitly) argued that any part of the world which is affected in any fundamental way by ‘capitalism’ (exchange) is ‘capitalist’. He had little difficulty in showing that the effects of the capitalist world economy have penetrated so deep into Latin America that no part of the continent has been unaffected. Even areas substantially devoted to self-sufficient, subsistence farming (such as north-east Brazil – see CULA: 153-4) are the products of the decay of earlier export industries, while latifundias (large estates worked by MARXIST THEORIES OF IMPERIALISM 164 peasant cultivators) originated as a response to commercial opportunities even where they have subsequently declined into almost self-sufficient isolation. Incorporation into the world capitalist system leads to development, in some areas, and to the development of underdevelopment elsewhere. Underdevelopment, as Frank uses the word, is not an original state; he coined the name undevelopment for the state of affairs before capitalist penetration (though neither he nor his followers have used it much, on the grounds that all vestiges of older forms have long since been reconstructed by capitalism). The ‘development of underdevelopment’ occurs because the world capitalist system is characterized by a metropolis–satellite structure. The metropolis exploits the satellite, surplus is concentrated in the metropolis, and the satellite is cut off from potential investment funds, so its growth is slowed down. More important, the satellite is reduced to a state of dependence which creates a local ruling class with an interest in perpetuating underdevelopment, a ‘lumpenbourgeoisie’ which follows a ‘policy of underdevelopment’ (LL, passim). I shall first discuss Frank’s conception of the structure of the world capitalist system, the ‘chain of metropolis–satellite relations’, then the transfer of surplus from satellite to metropolis and its consequences, and finally the economic and political structures that result. Consider first the idea of a chain of metropolis–satellite relations. The monopoly capitalist structure and the surplus expropriation/appropriation contradiction run through the entire Chilean economy, past and present. Indeed, it is this exploitative relation which in chain-like fashion extends the capitalist link between the capitalist world and national metropolises to the regional centers (part of whose surplus they appropriate) and from these to local centers and so on to large landholders or merchants who expropriate surplus from small peasants or tenants, and sometimes even from these latter to landless laborers exploited by them in turn. At each step along the way the relatively few capitalists above exercise monopoly power over the many below, expropriating some or all of their DEPENDENCY THEORIES 165 economic surplus, and to the extent that they are not expropriated in turn by the still fewer above, appropriating it for their own use. Thus at each point, the international, national and local capitalist system generates economic development for the few and underdevelopment for the many. (CULA: 7-8. Similar statements may be found throughout Frank’s work, e.g. CULA: 146-8, about Brazil) This ‘chain’ of metropolis-satellite relations has existed, according to Frank, since the sixteenth century; changes since then represent only changes in the forms of dominance and exploitation of the satellite, not changes of substance. This is the principle of ‘continuity in change’ (cf. Addo 1986). As a description, Frank’s ‘chain of metropolis–satellite relations’ is plausible. Relations of dominance and surplus extraction exist not only between the direct producers and their immediate exploiters, but at all levels in the world system. The idea is essential to Frank, since the ‘chain’ serves both to channel surplus to the metropolis and to create the class interests that sustain underdevelopment. As an analysis, however, it raises more questions than it answers. I shall criticize Frank primarily for conflating very different kinds of relations on the basis of purely superficial similarities. In particular, I shall argue that merchant capital and modern monopoly capital are quite different. The basic idea is of an exchange relationship in which the metropolis has a monopolistic position because each metropolis has several satellites, while each satellite confronts only one metropolis. The concept of monopoly used here is familiar from economics textbooks: a single seller facing a multitude of small buyers or, conversely, a single buyer facing many sellers. The monopolist can set the terms of exchange, and capture any surplus controlled by the other party. Frank, however, generalized the idea to cover any exploitative or unequal relationship: ‘the source or form of this monopoly varies from one case to another’ (CULA: 147). One can say, for example, that landlords monopolize access to the land, capitalists monopolize the means of production, and so on. But if one does not distinguish between different forms of monopoly, and in particular between class monopolies and individual monopolies, MARXIST THEORIES OF IMPERIALISM 166 between monopoly control of the means of production and monopoly in exchange, then the assertion that exploitation is the result of monopoly becomes an empty tautology. At the international and inter-regional levels of Frank’s hierarchy, there are two distinct kinds of monopoly involved. The first, and the one that fits Frank’s account best, is the monopolistic system of merchant capital, established in Latin America following the Spanish and Portuguese conquests and dominant into the twentieth century. Merchants collect products for export and for inter-regional trade, and distribute foreign and urban products. They are usually not involved directly in production; where they are, their activities in organizing production are secondary. Mercantile monopoly may be associated with either pre-capitalist or (small-scale) capitalist relations of production. The second kind of monopoly, which Frank does not distinguish clearly, is modern monopoly capital, characterized by large-scale capitalist production. In underdeveloped countries, it typically appears in the form of multinational companies, though national monopolies also exist. In contrast to merchant capital, modern monopoly capital exercises direct control over production and normally introduces fully capitalist production relations and the most modern technology. To confuse the two is a major mistake. It is true that some cases are hard to categorize, for example where multinational firms are engaged in buying agricultural products from small-scale producers, or where they organize production on a relatively primitive technical basis (e.g. plantation agriculture), but the existence of borderline cases does not invalidate the classification. At the local levels of the hierarchy, closer to the direct producers, Frank mentions merchants, landlords and (occasionally) capitalists. These must be clearly distinguished from each other. Frank argues that it is often difficult to disentangle these relationships in practice, but this surely makes it even more essential to be clear in analysis. I am not arguing that these different exploitative relations exist independently of each other. In each different historical period they have interacted and reinforced each other – this is what gives Frank’s account its descriptive verisimilitude. The next major point to note is that Frank identifies an economic hierarchy of individuals or classes (the ‘relatively few capitalists DEPENDENCY THEORIES 167 above’ exploiting ‘the many below’) with a spatial or geographic hierarchy (world and national metropolises, regional centres, local centres). This coincidence of economic and spatial relations is characteristic of some, but not all, systems of exploitation. Merchant capital, for example, sometimes creates a geographical hierarchy, when the output of scattered production units is gathered in regional centres for export in bulk, or for shipment to urban markets. This is the case that fits Frank’s picture best: the economic and geographical hierarchies coincide. Modern monopoly capital is often administered through a superficially similar geographical hierarchy. The head office is in a major world centre, regional offices or local subsidiaries are established in large cities, while productive activities are located wherever manpower, markets and raw materials supplies dictate (see Hymer 1972; Chandler and Redlich 1961). This hierarchy of administration is not in any real sense a ‘chain of metropolis– satellite relations’. The intermediate levels of management are no more than agents of the corporation, with no independent economic base. There is a direct wage relation between the workers and the corporation as a unit of capital. Another kind of hierarchy is created by the existence of subcontractors and the like, subordinate to the large corporations. This was discussed by Hilferding and is still important (Friedman 1977). As with merchant capital, it involves exchange relations, but the units involved are producing units, and are often located within a single urban area. All of these hierarchical structures differ from competitive capitalism, which is characterized by a ‘chain’ with only one link: the relation between workers and capitalists. There are superficial similarities between Frank’s ‘chain’ (and also Wallerstein’s core–periphery relation) and the classical Marxist theories of imperialism. Lenin said that finance capital ‘spreads its net’ over the world, and Bukharin wrote of a ‘few consolidated, organised economic bodies’ confronting an agrarian periphery. However, for Bukharin and Lenin, this was part of a process of internationalization that was transforming the world system by concentrating power and wealth at the centre, but simultaneously developing production and creating a true proletariat in the periphery. Where Frank and Wallerstein saw an essentially static system of redistribution persisting for centuries, the classical MARXIST THEORIES OF IMPERIALISM 168 Marxists saw a process of development that was transforming the world. The essential difference between these views is to be found in the classical Marxist emphasis on the relations of production. Frank is only able to argue that the ‘chain’ has remained essentially unchanged by ignoring the real changes in the relations of production which followed the displacement of merchant capital by modern monopoly capital. The ‘chain of metropolis–satellite relations’ is, according to Frank, the cause of the ‘development of underdevelopment’. This latter phrase is not very clearly defined, but part of its meaning seems to be a quantitative retardation in the growth of output, employment and productivity. (The other aspect of its meaning is a qualitative deformation of the system.) I now turn to the argument (used, in rather different forms, by Baran, Frank, Wallerstein and others) that the transfer of surplus from the satellite to the metropolis leads to a retardation of development in the satellite. First, consider the concept of surplus. Frank referred to Baran’s definitions (chapter 7 above), but there are additional difficulties in using this definition in Frank’s framework. Baran defined surplus as the difference between output (actual or potential) and consumption (actual or necessary). For a self-sufficient producing unit, the surplus is simply the physical excess of what is produced over what is consumed. Frank, however, was explicitly concerned with production units that are not self-sufficient, but are involved in a network of exchange relations. In this case, the goods that are consumed are not produced within the unit concerned, so production and consumption must be valued in comparable units before the difference between them can be calculated. The problem of valuation was a major preoccupation of the English classical economists; Marx tackled it by using labour values. Neither Frank nor Baran suggested any solution. It is not a mere technicality; to argue that there is exploitation in exchange, it is necessary to compare the prices that are actually paid with some reference set of ‘correct’ prices. At the prices actually paid, obviously, producers get the ‘value’ of what they sell. I know of no simple or universal solution to this problem, which greatly reduces the usefulness of the concept of surplus. Labour values make sense only in a relatively homogeneous system, in which competition leads to a certain DEPENDENCY THEORIES 169 uniformity in levels of technique by eliminating inefficient producers. This is the context implied by Marx in his discussion of value. As Frank himself insisted, the world economy is not now, and has never been, homogeneous in this sort of way, so the concept of surplus can only be used in a rather loose and qualitative fashion. If we accept a concept of surplus, we must still ask: what are the effects of a transfer of surplus from satellite to metropolis? Here Frank’s conflation of the economic and geographical dimensions of exploitation adds powerfully to the confusion. To say that one region extracts surplus from another suggests strongly that some physical goods are being seized from one place and shifted to another. Within a system of exchange, however, what is happening is that an exchange of goods of unequal value (in some sense) is taking place, and that control over the use of certain sums of value is being transferred from individuals (or groups) who live in one place to individuals (or groups, or corporate bodies) who live (or have their head offices) in another. Without analysis of the use of the surplus, this tells us nothing about the geographical location of new investment. The whole point of Baran’s concept of surplus is that surplus represents potential investment and hence economic growth. As a starting point, consider what determines the geographical pattern of investment in a fully developed (ideal) capitalist system. Here, investment will be directed into the activities that yield the highest profits (this is what leads to the formation of a general rate of profit, see Marx, Capital, III, ch. 10). In geographical terms, investment will be located where costs are lowest (allowing for transport costs), and thus, other things being equal, investment will go to low wage areas, i.e. underdeveloped areas. I do not argue that these ideal conditions have ever existed, nor that an allocation of investment according to profitability would be desirable. The real point is that the geographical pattern of investment need not correspond at all to the geographical location of the owners of the surplus. If anything, the super-exploitation of underdeveloped countries should mean more rapid development, as the classical Marxists expected. A large part of Frank’s argument is therefore misdirected; the transfer of surplus from satellite to metropolis cannot in itself explain the lack of development in the satellite. The MARXIST THEORIES OF IMPERIALISM 170 pattern of development depends on the factors that govern the use of the surplus. Part of Frank’s answer to this criticism would no doubt be to refer to his discussion of the distortion of the satellite’s economy as a result of economic dependence, to be discussed below. He could also argue that the structure of the chain of metropolis–satellite relations impedes the return flow of investment from metropolis to satellite. Here, the distinctions between merchant capital and modern monopoly capital are relevant. In the case of a mercantile hierarchy, there are very real barriers to the productive use of the surplus in the satellite. The production units are separate from the capitalist enterprises that exploit them. If mercantile profit is ploughed back into the expansion of the mercantile enterprise itself, this does nothing to expand actual production. Where production is in the hands of pre-capitalist producers, there may be no way in which investment funds could be channelled into production, at least without transforming the social relations of production, a task which merchant capital may have neither the means nor the desire to undertake. Where production is organized in small capitalist firms or in pre-capitalist units which can absorb money capital productively (for example, slave plantations), the problem is that investment funds will not be forthcoming unless production is profitable, and production is not profitable if the potential profits are being creamed off by mercantile middlemen. In the case of modern monopoly capital, especially in the form of multinational companies, the case is different. It has been argued by many commentators (e.g. Hymer 1972; Adam 1975) that multinational companies look over the whole world to select sites for investment, potentially profitable markets, and so on. They have no special reason to concentrate investment in their home countries, indeed they are multinational precisely because they have not done so in the past. The investment decisions of multinational companies should approximate more closely to the ‘pure’ capitalist pattern than those of any preceding form of capital. Frank cited evidence showing that the outflow of profit from Latin America to the USA in various forms greatly exceeds the return flow of investment funds from the USA. This is not really surprising, since it is generally true that only a small part of profit is DEPENDENCY THEORIES 171 reinvested, whether the profits are generated in advanced or underdeveloped countries, by local or foreign capitalists. In all these cases, a large fraction of profit is consumed. This may be an indictment of capitalism, but not particularly of foreign capital. Baran argued that monopoly capital is particularly prone to waste the surplus. I have criticized this view, and in any case it is a charge directed against monopoly capital in general, not foreign or multinational corporations as such. The relevant question is whether multinational companies reinvest less than locally owned capitalist enterprises. In Frank’s treatment of foreign firms, nationalist rhetoric often supplants sober analysis. My conclusion, then, is that the transfer of (control over) surplus to the metropolis is not, in itself, an explanation for lack of development in the satellite, which must be explained by looking at the uses of surplus. Of course, the fact that surplus is in different hands may be relevant to determining its use, but it is not the only factor. In particular, the relations of production and of exchange are crucial factors. Where production is in the hands of multinational corporations, the scope for productive use of surplus is far greater than it is where production is pre-capitalist and surplus is captured by merchant capital. In all cases, however, the incentives and opportunities provided by the economic environment are critically important. The transfer of surplus is not the only cause of the ‘development of underdevelopment’; Frank also claimed that participation as a satellite in the capitalist world system leads to a distorted and dependent economic structure. This idea was developed independently by a number of writers, and led to them becoming known as the dependency school. The point is that even a nationalist government cannot succeed in promoting capitalist development because of the constraints imposed by the international environment. Dos Santos (1970) defined dependence as follows: By dependence we mean a situation in which the economy of certain countries is conditioned by the development and expansion of another economy to which the former is subjected. The relation of interdependence . . . assumes the form of MARXIST THEORIES OF IMPERIALISM 172 dependence when some countries (the dominant ones) can expand and be self sustaining, while other countries (the dependent ones) can do this only as a reflection of that expansion. The story runs as follows: underdeveloped countries produce a narrow range of staple raw materials for export (this is the result of earlier stages of development). Incomes are very unequal and much of the surplus flows out of the country, so the mass market for consumer goods is limited. ‘Import substituting’ industrialization involves capital-intensive techniques, so employment is low, as are wages, leaving a large part of the population ‘marginalized’, either unemployed or in low productivity traditional activities. (The concept of ‘marginalization’ has much in common with dualistic theories, except that here it is the lack of dynamism of the modern sector that excludes people, forcing them back into low productivity activities, where theories of dualism originally saw the modern sector as held back by the traditional sector.) The market remains narrow, further constricting development. Furtado (1973) suggested a further factor: consumption patterns among the elite are copied from those of more advanced countries, and the result is to bias demand towards imports or towards goods produced by capital-intensive methods, reinforcing the problem. Modern production methods require imported capital goods, imported components and materials, while multinational companies remit profits abroad (openly or by devious means). The balance of payments is therefore a constant problem, halting growth, and compelling the retention of traditional export industries as foreign exchange earners. Consider, first, the problem of narrow markets. There are two issues: the absolute size of the market and its rate of growth. A small market limits the opportunities for use of modern large-scale techniques; there is no doubt that small underdeveloped countries are at a disadvantage here, though larger underdeveloped countries (Brazil, India, etc.) are not. Production for the (potentially almost unlimited) export market can overcome both this problem and the problem of a slowly growing market. Relatively slow growth in the market for consumer goods is the result not of a high rate of DEPENDENCY THEORIES 173 exploitation, but of a rising rate of exploitation. Correspondingly, the market for means of production could expand relatively rapidly. This is the standard argument against under consumptionism, an argument Frank accepted (DAU, chapter 5). It is argued, however, that means of production are predominantly imported, so the growth of local industry is limited by the slow growth of the market for consumer goods. The case thus hinges on assumptions about the pattern of international specialization. Exports (which generate income, and hence demand) are assumed to be confined to traditional exports. Production for the local market is assumed to be limited to consumer goods. An explanation of the pattern of specialization is needed to give these arguments a solid foundation. What of the balance of payments? Suppose a country exports a limited range of traditional staples, and cannot easily expand its exports. Growth in export earnings is limited, so import growth must be limited too (since imports must, in the end, be paid for by exports). If imports, say of capital equipment or materials, are essential for growth, the growth of the whole economy can be limited by the balance of payments constraint. The argument is familiar to development economists (it is a form of ‘structuralism’; see Little 1982), and is perfectly coherent. The problem is that it, like the ‘narrow markets’ argument, rests on the assumption that nothing can be done about the pattern of specialization; indeed, it is just a variant of the ‘narrow markets’ argument, for the special case in which export markets are narrow. What is needed to add some substance to these arguments is some explanation of why a country should be so limited in the range of goods it can produce and export. All the evidence is that many Third World countries have, in fact, widened the range of goods they produce, either to replace imports of manufactures (this is called ‘import substituting industrialization’ in the jargon) or by exporting manufactures themselves (‘export-oriented industrialization’). Some countries which shifted to exporting manufactures became net importers of primary products themselves, creating markets for exporters of primary products. The benefits have not been evenly spread, of course, but what is needed is an explanation of why some countries have succeded while others, as yet, have not. To deny the possibility of something which is visibly happening is not very helpful. MARXIST THEORIES OF IMPERIALISM 174 Finally, Frank stressed the political consequences of dependency. The ruling classes in underdeveloped countries owe their position to their place in a ‘chain’ that runs from the countryside to the imperialist metropolis, and thus have an interest in maintaining it. This colonial and class structure establishes very well defined class interests for the dominant sector of the bourgeoisie. Using government cabinets and other instruments of the state, the bourgeoisie produces a policy of underdevelopment in the economic, social and political life of the ‘nation’ and the people of Latin America. (LL: 13) This part of Frank’s argument is at its best in his historical analysis of a crucial turning point in Latin American history, the period after independence in the nineteenth century. At this time there was a bitter conflict between the ‘Europeans’, the advocates of free trade, and the ‘Americans’, the advocates of protection for domestic industries. The ‘Europeans’ were led by merchants who handled the export and import trades and by agricultural export interests, and they were the stronger party precisely because the preceding centuries of dependence had created an economy dominated by the groups that stood to gain from continuation of the system. Free trade made imported manufactured goods available cheaply to the export agriculturalists, and the weakness of the local currency increased the value of exported products in terms of the depreciated currency, transferring income to those who sold goods for export. Local manufacturing industry was unable to compete with imports without protection, so perpetuating the imbalance (LL: 61). State policy was geared to the needs of the export sector in other ways as well: taxes, distribution of land, immigration policy, ports, railways, and so on. Frank summarizes: ‘the “European” lumpenbourgeoisie built “national” lumpenstates which never achieved real independence but were, and are, simply effective instruments of the lumpenbourgeoisie’s policy of lumpendevelopment’ (LL: 58). By contrast, the United States did not offer suitable conditions for export agriculture except in the south. DEPENDENCY THEORIES 175 Consequently, the class structure which developed there, based at the start on small farmers, did not present any obstacle to a development policy which permitted the Northern bourgeoisie to become strong enough to use independence to promote integrated development, to defeat the planter/exporters of the South in the Civil War, to impose a policy of industrialisation and arrive at their own industrial ‘take-off’ point. (LL: 58-9) Frank was surely right to argue that state policy is a critical element in economic development, that state policy is the outcome of conflict between classes and fractions of classes with conflicting interests, and that classes and fractions which benefit from an existing economic structure have an interest in perpetuating it and are in a strong position to succeed. However, his political analysis was radically impoverished by the presumption that international trade is bad for development, and self-sufficiency good, so classes and fractions were classified by whether they were for or against trade. The issues in development policy are far too complex to be captured by this simple opposition. In any case, the idea that state policy since the Second World War has been hostile to industrial development is simply ridiculous. In the majority of Third World countries, levels of industrial protection have been very high (absurdly high in some cases), and state investment has been poured into infrastructure and ‘heavy’ industries. It can be argued that some countries, at least, have lost out by neglecting agriculture and primary production, not the reverse. The interesting question is: how did industrial capital succeed in enlisting such wholehearted state support, when it was, to begin with, so weak? This, and related issues, will be taken up in chapters 10 and 11. A final comment: Frank’s arguments were presented in a distinctive way, which seems to me to be an important source of weakness. His normal procedure was to make brief, sloganistic assertions, and then to justify and expand on them by giving a series of historical examples, frequently quoting at length from other writers or from original sources. The problem with this style of argument is that it leaves no room for systematic theoretical exposition; one is left repeatedly saying: yes, it happened like that in those cases, but why, and must it be the same everywhere? In MARXIST THEORIES OF IMPERIALISM 176 addition, crucial terms (development, underdevelopment, metropolis, satellite, capitalism, and so on) are never explicitly defined. The reader is left to infer their meaning from the essentially descriptive uses they are put to. They often have a spectrum of meanings rather than a single precisely defined sense, blurring the logic of some of his most important assertions. 8.2 WALLERSTEIN I shall summarize Wallerstein’s arguments briefly, since they have a great deal in common with those of Frank. He insisted that any social system must be seen as a totality. Nation states, in the modern world, are not closed systems and cannot be treated as if they were. We take the defining characteristic of a social system to be the existence within it of a division of labour, such that the various sectors or areas within are dependent upon economic exchange with others for the smooth and continuous provisioning of the needs of the area. (CWE: 5) The only kinds of social system that have existed are ‘minisystems’ (closed local economies), ‘world empires’ (defined by the extraction of tribute by a central authority) and ‘world economies’ (formed by a market exchange). A ‘world’ system does not necessarily have to cover the whole globe; it is a ‘unit with a single division of labour and multiple cultural systems’. A world economy, then, is a world system without a single central authority. The modern world system is capitalist, since it is a world economy (as defined). Capitalism and a world economy (that is, a single division of labour but multiple polities) are obverse sides of the same coin. One does not cause the other. We are merely defining the same indivisible phenomenon by different characteristics. (CWE: 6) The capitalist world system is divided into three tiers of states, those of the core, the semi-periphery and the periphery. The DEPENDENCY THEORIES 177 essential difference between them is in the strength of the state machine in different areas, leading to transfers of surplus from the periphery to the core, which further strengthen the state machines in the core. State power is the central mechanism since ‘actors in the market’ attempt to ‘avoid the normal operation of the market whenever it does not maximise their profit’ by using the state to alter the terms of trade. The link with imperialism in the traditional sense, the dominance of one state over others, is obvious. At least in origin, the core–periphery division is explained through a sort of technological determinism. Western Europe specialized in manufacture and animal raising, activities which require relatively high skills, and are better carried out by relatively well-paid free wage labour. The resulting social structure is the foundation of relatively strong ‘core’ states, able to manipulate markets to their advantage. Hispanic America (mining) and Baltic east Europe (grain) specialized in activities requiring relatively little skill, hence capitalists chose (through state intervention) forms of coerced labour, and a difference of interests emerged between manufacturing and primary product export interests. Local states were weak, and readily subjugated by the core, so these areas became ‘peripheral’. Once in existence, the core–periphery division is maintained by the ability of the core states to manipulate the workings of the system as a whole to suit their needs. They deliberately weaken peripheral states or eliminate them altogether by conquest, and also alter the workings of markets by imposing monopolistic restrictions, protecting their own industries, forbidding corresponding protection in the periphery, and so on. The ‘semi-periphery’ is a sort of labour aristocracy of states or geographical areas. Without it, a world system becomes polarized and liable to revolt, while an intermediate tier diffuses antagonisms. This argument is hard to accept as an explanation of the existence of a semi-periphery. Was the creation of a semi-periphery deliberate? The particular cases Wallerstein cites (Italy in the sixteenth century, Russia later, including the USSR) do not seem to have been deliberately created by the core states. In any case, if the core is divided into distinct national states (by definition), who is there to oversee the interests of the system as a whole? The notion of a semi-periphery is rather convenient for the theory because it provides, so to speak, a site for change. New core states can emerge MARXIST THEORIES OF IMPERIALISM 178 from the semi-periphery, and declining core states can sink into it. On the other hand, the notion of a semi- periphery can easily become an excuse for ad hoc explanations; core states are expected to succeed, so any core state that does badly can be redescribed as semi-peripheral, and so on. There is a risk of leaving the theory with no substance at all. At this stage, one might well ask what has happened to relations of production and to classes in the ordinary Marxist sense. Wallerstein seems to count anybody who produces for profit in the market as a capitalist. Labour-power is indeed a commodity but ‘wage labor is only one of the modes in which labor is recruited and compensated in the labour market. Slavery, coerced cash-crop production, sharecropping and tenancy are all alternative modes’ (CWE: 17). Marx’s concept of capitalism in terms of a relation between free labour and capital is ruthlessly ditched. ‘Class analysis’ amounts, in Wallerstein’s view, to the analysis of the interests of ‘syndical groups’ within particular states, and is legitimate provided we look at the ‘structural position and interests in the world economy’ of these groups. At the same time, classes have no permanent reality and are no more fundamental than ‘ethnonations’. That, at least, is my interpretation of some more than usually opaque passages in Wallerstein’s writings. (See CWE: 24, 224-6.) The central point, which has been the subject of much debate, is that ‘modes of labour control’ (wage labour, slavery, etc.) are secondary results of the functioning of a world system defined by the existence of market links. The situation in the core is such that free wage labour tends to be chosen (by the ruling class, with state support) while in the periphery more coercive systems are used. Overall, Wallerstein’s primary assertion is that a world system must be analysed as a whole. Few are likely to disagree. Beyond this, I find little more than a series of definitions and phrases, with a mass of detailed historical material that often seems to have little connection with his overall generalizations. What is lacking is a level of theory that would connect the two, by specifying exactly how, and under what conditions, the structure produces the results he claimed for it. DEPENDENCY THEORIES 179 8.3 LACLAU’S CRITIQUE Both Frank and Wallerstein defined capitalism as a system of exchange relations. They did so, quite deliberately, to include relations of exploitation such as those between landlords and peasants, which do not involve wage labour in the strict sense. This approach has been criticized by a number of Marxist writers. Laclau, in an article which has been the starting point for much debate (1971, reprinted in Laclau 1977), stated his essential point as follows: Of course, Frank is at liberty to abstract a mass of historical features and build a model on this basis. He can even, if he wishes, give the resulting entity the name of capitalism. . . . But what is wholly unacceptable is the fact that Frank claims that his conception is the Marxist concept of capitalism. Because for Marx – as is obvious to anyone who has even a superficial acquaintance with his works – capitalism was a mode of production. The fundamental economic relationship of capitalism is constituted by the free laborer’s sale of his labourpower, whose necessary precondition is the loss by the direct producer of ownership of the means of production. (Laclau 1977: 23) Marx regarded exchange, and the development of merchant capital, as perfectly consistent with the persistence of pre-capitalist modes of production. I have already described this aspect of Marx’s analysis. Laclau distinguished between ‘modes of production’ and ‘economic systems’. A mode of production is ‘an integrated complex of social productive forces and relations linked to a determinate type of ownership of the means of production’ (1977: 34). The feudal and capitalist modes are the only two that are relevant: The feudal mode of production is one in which the productive process operates according to the following pattern: 1. the economic surplus is produced by a labour force subject to extraeconomic compulsion; 2. the economic surplus is privately MARXIST THEORIES OF IMPERIALISM 180 appropriated by someone other than the direct producers; 3. property in the means of production remains in the hands of the direct producer. In the capitalist mode of production, the economic surplus is also subject to private appropriation, but as distinct from feudalism, ownership of the means of production is severed from ownership of labour power. (Laclau 1977: 35) An ‘economic system’ consists of ‘the mutual relations between the different sectors of the economy, or between different production units, whether on a regional, national or world scale’ (1977: 35), so ‘an economic system can include, as constitutive elements, different modes of production.’ (Similar ideas were developed by a number of writers in the 1970s, and will be discussed further in chapter 10; Laclau’s version is discussed here, because it was presented as a critique of Frank and Wallerstein.) According to Laclau, Feudal production for exchange is not ruled out, so Frank’s evidence that supposedly feudal areas of Latin America have been deeply influenced by exchange becomes irrelevant, and ‘to affirm the feudal character of the relations of production in the agrarian sector does not necessarily involve maintaining a dualist thesis’ (1977: 32). Wallerstein responded: The substantive issue, in my view, concerns the appropriate unit of analysis for the purpose of comparison. . . . Sweezy and Frank better follow the spirit of Marx if not his letter’ (CWE: 9). The essence of his reply is that the system must be viewed as a totality, and that the only totality that actually exists is the world economy. This brings out the key point very well. Both Frank and Wallerstein were looking for descriptive generalizations, based directly on the observed facts. Marx, by contrast, insisted on the necessity of abstraction. The capitalist mode of production was not, for Marx, a directly observable empirical thing (like the capitalist world economy); it was instead a conceptual object, the product of thought. The aim is to pick out key relationships and examine them in isolation before elaborating the analysis to deal with the complexities of the real world. Wallerstein’s reply, therefore, misses its target, since Marx was not looking for a totality that really exists. However, an appeal to the authority of Marx only settles the issue for dogmatists (cf. Foster-Carter 1979). Marx did indeed define a mode of production in terms of relations of production, but it DEPENDENCY THEORIES 181 remains to be shown that this approach gives better results in explaining reality than any other. I have argued that major weaknesses in Frank’s analysis follow directly from his neglect of relations of production. If this is accepted, then it strengthens Laclau’s case. On the other hand, Laclau’s own explanation of underdevelopment and of the persistence of pre-capitalist modes is most unconvincing. Laclau wrote ‘[Frank] shows us how the advanced countries have exploited the peripheral countries; what he at no time explains is why certain nations needed the underdevelopment of other nations for their own process of development’ (Laclau 1977: 35-6). And again: ‘if we want to show that . . . development generates underdevelopment, what we have to prove is that the maintenance of pre-capitalist relations of production in the peripheral areas is an inherent condition of the process of accumulation in the central countries’ (Laclau 1977: 37). There is a c lear l ogical f allacy h ere. L aclau claims t hat if development requires underdevelopment, then underdevelopment will happen, but not otherwise. This is the crudest kind of functionalism; whatever is necessary for capitalism will happen. It cannot be sustained. If underdevelopment were necessary for development, it might still fail to happen, either because owners of metropolitan capital, not having understood Marx, fail to realize the necessity, or because they lack the means to enforce their wishes. On the other hand, underdevelopment might not be necessary, but could happen all the same, just as an innocent bystander might be killed in a shoot-out without his death being necessary to anybody. Then again, underdevelopment might contribute to development without being necessary to it, it could be the jam on the bread. I stress this point, perhaps too heavily, because this logical fallacy is all too common in Marxist writings. Laclau tried to show the necessity of underdevelopment to development through the theory of the falling rate of profit, which I have already criticized (chapter 2). The organic composition of capital, he claimed, rises in the advanced countries, bringing down the rate of profit, which must be offset by expansion into areas where the organic composition is low. Even if true, however, this would not explain the persistence of pre-capitalist modes; rather, one would expect investment in underdeveloped areas to lead to the replacement of pre-capitalist modes by capitalism (as Lenin, for example, expected). MARXIST THEORIES OF IMPERIALISM 182 Brenner (1977) developed Laclau’s critical points and added a very thorough critique of Wallerstein, while advancing quite different positive arguments. His central point is that capitalism is unique, above all, for the way in which it promotes technological development, increased productivity and hence increased profits (through relative surplus value). This tendency to develop the forces of production is one of Marx’s main conclusions. Brenner showed that other modes of production (‘modes of labour control’ in Wallerstein’s terms) do not have the same dynamic. Equally, ‘modes of labour control’ are not freely chosen by ruling classes. They are the result of class struggle. Different parts of the world economy, therefore, have their own tendencies that derive from the modes of production installed there. This does not mean that they evolve independently of each other, but that analysis should start from the workings of distinct modes of production, and then go on to analyse how they interact with each other. This line of argument will be developed further in chapter 10. Note, incidentally, that if increasing productivity is the essence of capitalist development, the prosperity of the ‘core’, or the ‘metropolis’, does not have to be at the expense of anybody else; development is not necessarily the ‘other side of the coin’ of underdevelopment (which is not to deny that international flows of surplus can and do take place). This is not a new idea; it is a reassertion of the classical Marxist perspective. 8.4 AMIN Amin’s guiding vision is summed up by the titles of his two main works: Accumulation on a World Scale and Unequal Development. The process of accumulation, of development, must be analysed as a single process on a world scale, but it takes place in a world divided into many distinct national social formations, containing different modes of production. Accumulation does not tend to create uniformity between these social formations, but divides them into two categories: those of the centre and those of the periphery. Accumulation at the centre is autocentric (self-centred); it is governed by its own internal dynamic, as analysed by Marx. In the DEPENDENCY THEORIES 183 periphery, by contrast, accumulation is dependent or extroverted, constrained by the centre–periphery relation. At the heart of Amin’s analysis, as I read it, is an explanation of unequal specialization, to be discussed in detail below. Put very simply, he argues that the pattern of international specialization is determined by absolute cost levels (not by comparative advantage, as Ricardo thought), and that cost levels depend on productivity and on wages. The countries of the centre developed capitalism earlier, or under especially favourable conditions, and got a huge lead in productivity during a period in which wages were held down to something close to physical subsistence levels in both centre and periphery. This established a pattern of unequal specialization. Later, wages started to rise at the centre, but the centre’s lead in productivity remained enough to ensure lower costs at least in most sectors of industry. Unequal specialization is thus both cause and consequence of unequal development; both are anchored in the conditions of production and reflected in exchange relations. Given this pattern of unequal specialization and development, capitalism at the centre evolved as the classical Marxists predicted, driving out pre-capitalist modes of production, while capitalist development in the periphery was blocked, since the periphery could compete only in resource-based activities (minerals, tropical agriculture), and there could only be limited development oriented to the narrow domestic market. The larger part of the population was excluded from the capitalist sector, so pre-capitalist modes of production were not eliminated. This provides an explanation for the typical structure of underdeveloped countries, ‘peripheral capitalism’. After a certain stage of development, wages started to rise at the centre, while massive unemployment and the persistence of precapitalist modes of production held wages down in the periphery. Amin’s explanations of increased wages at the centre are both obscure and inconsistent; I shall not go into detail here; see Brewer (1980a; 1980b: 250-6). High wages at the centre and low wages in the periphery lead to unequal exchange, as analysed by Emmanuel (discussed in chapter 9 below). Very briefly, Emmanuel argued that low wages in the periphery mean low prices for its products, while the products of the high wage centre sell at high prices. Some care is needed here; according to Amin, the centre has advantages in MARXIST THEORIES OF IMPERIALISM 184 productivity that more than outweigh its higher wages, so one might expect its prices to be low. However, he argued that although the centre has relatively high productivity in most lines of (industrial) production (thus maintaining unequal specialization despite the wage gap), the periphery has relatively high productivity in the few lines of production in which it specializes. In its export industries, therefore, the periphery combines high productivity with low wages, the recipe, given international equalization of profits, for unequal exchange. Amin criticized Emmanuel for treating wages as an ‘independent variable’, without adequately explaining them. In Amin’s words: ‘Which is cause and which effect: the international prices, or the inequality in wage levels? The question is pointless. Inequality in wages, due to historical reasons (the difference between social formations) constitutes the basis of a specialisation and a system of international prices that perpetuate this inequality’ (UD: 151). There are thus no independent variables, only a self-perpetuating process (see also Amin 1977: 185). The social formations of the periphery are characterized by disarticulation between sectors, since most of the industries producing means of production are absent. The links between industries producing means of production and consumer goods, analysed by Marx in his schemes of reproduction, exist at the world level but are incomplete within the national economy. The periphery is also characterized by unevenness of productivity; export sectors operated by foreign capital on the most modern lines coexist with primitive pre-capitalist sectors. The capitalist mode of production is dominant but does not tend to become exclusive, and the tertiary sector becomes overexpanded as a result of the pattern of demand and the lack of opportunities for investment in industry. These features of peripheral capitalism are not at all characteristic of ‘traditional’ pre-capitalist societies, they are the product of the ‘development of underdevelopment’. How does Amin’s account of the world system fit in with those of other writers? He took the analysis of international prices from Emmanuel, adding his own account of unequal specialization to complement it. Together these theories amount to the first serious analysis of international trade in the Marxist tradition. Other DEPENDENCY THEORIES 185 writers have, of course, described the destruction of industries in the periphery by foreign competition (Marx on the destruction of Indian handloom weaving, etc.), but Amin developed the idea and made it the centre of his analysis. Unequal specialization provides a foundation for an analysis of peripheral capitalism that has much in common with Baran and with dependency theories. On the other hand, Amin analysed social formations in terms of the interaction of different modes of production, like a number of writers discussed in chapter 10. Monopoly, as such, plays very little part in his analysis, but the mobility of capital is of central importance. Amin argued that economic laws of the sort elaborated by Marx in Capital only apply to a pure capitalist system, while capitalism, in fact, coexists in the world system with other modes of production. An ‘economistic’ analysis which deals only with quantitative relations between narrowly economic variables can only be a subordinate part of the story. Real history can only be understood by the analysis of concrete social formations and cannot be reduced to a preordinated succession of modes of production. A major part of his work is therefore devoted to historical analysis (UD, chapters 1, 5). I cannot hope to give an adequate account of this aspect of Amin’s writing here, and I will not try to do so; I will only pick out some points of theoretical interest and try to indicate the main outlines. Amin’s discussions of the Arab world and of Africa are especially notable for their combination of wide-ranging knowledge and analytical insight. He defined five modes of production, of which four are familiar: the primitive-communal, slave owning, simple petty-commodity and capitalist modes. The one that stands out as unusual is: the ‘tribute-paying’ mode, which adds to a still existing village community a social and political apparatus for the exploitation of this community through the exaction of tribute; this tributepaying mode of production is the most widespread form of precapitalist classes, and I distinguish between (a) its early and (b) its developed forms such as the ‘feudal’ mode of production, in which the village community loses its dominium eminens over the soil to the feudal lords. (UD: 13) MARXIST THEORIES OF IMPERIALISM 186 The ‘tribute-paying’ mode is clearly an old friend, the Asiatic mode, under another name. It covers a spectrum of social structures, from African societies little removed from the primitive-communal mode, though the great Asiatic cultures, to feudal Europe. The allinclusive nature of this concept rather reduces its usefulness, though it has to be said that this is an area in which it is hard to make any very firm distinctions. Pre-capitalist societies can incorporate several modes of production and, as a result, a very complex class structure. A formation dominated by the tribute-paying mode may contain commodity production and exchange, though the pure form of the tribute-paying mode does not. Different social formations interact, for example by trade, which can form the basis for societies living on a surplus produced elsewhere. The key to the analysis of any social formation is the production and circulation of surplus, defined, much as in Baran, as ‘an excess of production over the consumption needed in order to ensure the reconstitution of the labour force’ (UD: 18). This amounts to a rich and flexible framework for historical analysis. Amin described a pre-capitalist world made up of three ‘central’ tribute-paying formations (China, Egypt, India), with a ‘periphery’ around them which was much less stable and was influenced more by the centre than vice versa. The Mediterranean periphery produced a society dominated by the slave-owning mode, and dependent on its own periphery (Europe) for supplies of slaves. The collapse of this society faced with barbarian invaders produced European feudalism and, finally, capitalism. On the far eastern periphery of China, Japan evolved a feudal society and thus an indigenous capitalism. All this bears a suspicious resemblance to Marx’s (now usually derided) judgement that Asiatic societies have no real history, but with the reversal that these Asiatic societies are now called ‘central’. One point is that progress took place on the periphery; Amin argued that today the transition to socialism must start from the periphery. It is an interesting thought, though the feudalism–capitalism transition is so different from the transition to socialism that the analogy cannot really be regarded as a proof of anything. Amin’s treatment of feudalism is somewhat inconsistent. It is the most developed form of the tribute-paying mode: ‘when well DEPENDENCY THEORIES 187 developed [the tribute-paying mode] nearly always tends to become feudal (this happened in China, India and Egypt)’ (AWS: 140), so that the ‘central’ tribute-paying formations are described as moving towards feudalism. On the other hand, feudalism is peripheral because it is a borderline case analytically (IUD: 16) and because it develops on the borders (geographically) in areas where natural conditions were less favourable and centralizing tendencies weaker (UD: 55-6). This may represent a shift in position between the two works cited, but I suspect that it is a reflection of the unsatisfactory state of the definitions. The point, of course, is that capitalism emerges from ‘peripheral’ feudalism, so the erstwhile periphery becomes central. The emergence of capitalism from its feudal origins is a familiar story, explained both by the formation of a landless proletariat and by mercantile accumulations of loot; which is the determining factor is not clear. It is clear, however, that capitalism emerged where it did because it was preceded by feudalism. The ‘new’ capitalist centres of the USA, Canada, and the like, were by-products of proletarianization at the centre, which led to emigration and to the formation in areas of settlement of societies which had an unusually large element of petty-commodity production, and thus an unusually favourable context for capitalist development. Amin divided the development of the capitalist world economy into three (familiar) stages. The first, mercantilist stage, is marked by the emergence of capitalism in its homelands and by the establishment of a net of exchange relations connecting capitalist with pre-capitalist formations. The next stage is that of developed pre-monopoly capitalism (‘competitive capitalism’ for short, though relations with the periphery were often monopolistic), lasting from about 1800 to about 1900. This stage, according to Amin, was characterized by approximately equal exchange between centre and periphery (since wages were still low at the centre). Productivity increases were passed on as price reductions under competitive conditions. It was a ‘pause’ of a century while ‘Europe and the United States withdrew into themselves’ (UD: 187). Again, there are inconsistencies in Amin’s phraseology, if not in his story, in that his ‘pause’ was also a period in which ‘external extension of the capitalist market was . . . of prime importance as a MARXIST THEORIES OF IMPERIALISM 188 means for realising surplus value’ (p.188, i.e. the following page). During this period, in any case, the foundations of a new pattern of international specialization were laid, and the dividing line between centre and periphery was established. Amin was, of course, mainly interested in the imperialist stage (from 1900). Wages started to rise (with productivity) at the centre, capital became fairly mobile, and world markets became closely integrated, creating the conditions under which unequal exchange takes place. In the emerging periphery, capitalist development was blocked by the competitive strength of the centre, and pre-capitalist modes of production survived. Peripheral social formations developed in a variety of ways, according to the pre-existing social structure, the date of capitalist penetration, the opportunities offered by natural conditions and so on. It is in the analysis of these different paths to peripheral capitalism that Amin’s skills as a writer of analytical history are most evident, and I will not try to summarize. All of these varied peripheral formations, however, move towards a common pattern of peripheral capitalism. There is some ambivalence in Amin’s account; on the one hand peripheral capitalism is defined by the persistence of pre-capitalist modes of production while, at the same time, pre-capitalist modes have become a mere ‘shell, whose content has become the sale of labour power’ (IUD: 191). We too often confine ourselves to looking for the capitalist relation at the ‘microeconomic’ level, that of the firm. . . . In peripheral capitalism . . . the petty commodity production mode may appear to be integrated within the capitalist market, but in reality capital dominates the direct producer. The latter is not a petty commodity producer. . . . In fact he is very like the cottage industry proletarian as he formerly existed in Europe; that is exploited by capital to which, in fact, he sold his labour power rather than his product. Here the failure to see that it is the sale of labour power which gears the system is a failure to understand the unity of the world system. (IUD: 90-1; see also UD: 361) DEPENDENCY THEORIES 189 The absorption of pre-capitalist modes by capitalism (as opposed to an articulation between distinct modes) has developed very rapidly recently. Amin also conceded that the rise of multinational companies could change matters drastically, as they shift their activities to sources of low-cost labour. Taken together these hints could be the basis for a prediction that ‘peripheral capitalism’ is in the process of changing very fundamentally and rather rapidly. At the heart of Amin’s argument is an economic process: the development of capitalism in the periphery is blocked by the superior competitive strength of the industries of the centre, manifested in an ability to undercut the industries of the periphery or to establish a price level which prevents new industries emerging at all. This mechanism works, according to Amin, at all stages of development, both of the centre and of the various social formations of the periphery, at least from the industrial revolution onwards. The distortion towards export activities (extraversion), which is the decisive one, does not result from ‘inadequacy of the home market’, but from the superior productivity of the centre in all fields, which compels the periphery to confine itself to the role of complementary supplier of products for the production of which it possesses a natural advantage: exotic agricultural produce and minerals. When, as a result of this distortion, the level of wages in the periphery has become lower, for the same productivity, than at the centre, a limited development of industries focussed on the home market of the periphery will have become possible, while at the same time exchange will have become unequal. The subsequent pattern of industrialisation though import-substitution, together with the (as yet embryonic) effects of the new international division of labour inside the transnational firm, do not alter the essential conditions of extraversion, even if they alter the forms that it takes. (UD: 200) The destruction of craft production and the blockage of capitalist industrialization forced the population into pre- capitalist agriculture, in the first instance, and later into an overexpanded MARXIST THEORIES OF IMPERIALISM 190 tertiary (service) sector, forcing down wages and reinforcing the hold of landlords and other pre-capitalist exploiting classes. Both low wages and the persistence of pre-capitalist modes follow from the absence of a fully fledged industrial sector. There are other mechanisms at work as well. Capital never scorns extra-economic coercion when it is the most cost-effective means to maximize profits, and coercion is absolutely necessary to break into societies where commodity production is not well established. State power, and the economic power of monopoly, also play a role at all stages of development. However, Amin stressed the convergence of the formations of the periphery towards a common path, despite their very different histories; the economic mechanism summarized above is his general explanation of the persistence of underdevelopment. Amin’s line of argument can be related to the theory of ‘comparative advantage’ or ‘comparative costs’. There are two elements to it. First, for simple geographical reasons, the peripheral formations have a comparative advantage in the production of ‘exotic agricultural produce’ and (certain) minerals. This is not difficult to accept. Second, the employment created is not sufficient to absorb the whole labour force. Historical experience suggests that this may be true, but the exact mechanisms involved are not clear from Amin’s account, and need further examination. A preliminary point must be dealt with at the start. In mainstream economics, the productivity of labour in different branches of production in different countries is taken as given. Amin, rightly, criticized this. The periphery is not ordained by nature to be a supplier of raw materials; the productivity of labour in different activities is the result of a historical process of development. What a theory of specialization can do is to show how unequal development manifests itself in trade relations. These relations, in turn, modify the subsequent pattern of development. Standard post-Ricardian theories of international specialization are based on an assumption of full employment. This is not so obvious in the usual Ricardian example with only two goods, since if one country specializes in one, the other country must specialize in the other. In practice, of course, there are a multiplicity of commodities. Suppose there are two countries, A and B. If one, say A, has twice the labour force of the other, then it must specialize in DEPENDENCY THEORIES 191 the production of commodities that absorb two thirds of the total labour of the two-country system. If we think of a spectrum of commodities, arranged in order, with those in which A has the greatest comparative advantage at one end, a dividing line must be established between A’s exports and its imports, at a point that allows full employment in both countries. How could this happen? If there is unemployment of resources in A, then factor rewards (wages, profits, etc.) must fall there, until A is able to undercut B in some additional branches of production. The dividing line is moved along the spectrum of activities until full employment (or at least equal unemployment) is established everywhere. In Amin’s account, by contrast, the industries assigned (by the ‘invisible hand’) to the periphery are not enough to absorb the productive resources available there, so there is massive unemployment and, at the same time, money capital is diverted away from productive uses. It seems that his argument is incomplete; falling wages (and other factor rewards) should restore the periphery’s capacity to compete in enough areas to offset the fall in demand. In fact, his analysis is in an even worse position. He claimed that trade with a more advanced economy will create unemployment. Ricardo’s classic analysis (1951, chapter 7) is intended to show that trade will raise incomes in the less advanced as well as the more advanced economy, so there is no need to force down wages in order to restore full employment. The point of the Ricardian argument is simple: exchange of commodities can only transmit a structure of relative prices. If one country can produce every commodity with, say, a hundred times lower real costs than the other country this makes no difference at all provided that relative costs are the same. I will examine the analysis, to begin with, for the case where both countries are wholly capitalist; the basic point applies to any commodity-producing system. There are two distinct cases in which Amin’s account can be rescued. First, there is the case in which the comparative advantages are such that the less developed area is excluded from activities which are relatively labour-intensive given the techniques used in that area. This is exactly Amin’s case; the periphery was excluded from ‘industrial’ activities, which were carried out by labourintensive (craft) methods, and forced to specialize in agriculture MARXIST THEORIES OF IMPERIALISM 192 (land-intensive). The result is increased pressure on the land, raising rents and either forcing down wages or creating unemployment. (This is an application, in different circumstances, of the Stolper- Samuelson (1941) argument.) Here, it is the pattern of comparative costs that matters. The second case is quite different. The Ricardian argument assumes that only commodities are mobile between different countries, so only (relative) commodity prices are equalized. Suppose, instead, that capital is mobile as well. Capitalists set up production wherever costs are lowest. Costs depend largely on unit labour costs, and hence on wages relative to productivity. This is still not enough to ensure that there will be de-industrialization and unemployment in the periphery. If the wage differential matches the (assumed) productivity differential, there is no reason for the periphery to be uncompetitive. Amin considered this possibility (new ‘light’ industries are established in the periphery once wages in the centre have risen far enough), but it plays only a subsidiary part in his argument. We therefore have to assume that wages are, at least to some extent, fixed in real terms; for example, in the nineteenth century wages were close to subsistence in both centre and periphery, so wages could not be cut in the periphery. (On the theoretical argument underlying this paragraph, see Brewer 1985.) It follows, then, that trade can cause unemployment, as Amin claimed it did, when capital is mobile and productivity differences outweigh differences in wages. The more advanced country has the higher rate of profit (lower costs), so capital flows out of the less developed country, leaving unemployment. Specialization is determined by absolute and not comparative costs. Ricardo commented on this possibility in a passage cited by Amin (Ricardo 1951: 136). Amin referred to the outflow of capital from the periphery to the centre, which fits in with this analysis, though he also argued that profits are higher in the periphery than in the centre as a result of super-exploitation, which is not consistent with the analysis presented here. The level of productivity in different activities, and its evolution, are clearly crucial to the argument as I have presented it. Amin discussed the pattern of relative ad vantages under the guise of a discussion of ‘sectoral unevenness of productivity’ (in the periphery): DEPENDENCY THEORIES 193 It is not, of course, possible to compare productivities in the strict sense of the word except between two enterprises that produce the same product. . . . Between one branch and another one can speak only of different profitabilities, as Emmanuel has reminded us. All the same, if, with a given price structure, conditions are such that labour, or capital, or both, cannot be rewarded in one branch at the same rate as in another, I say that productivity is lower in that branch. In the capitalist mode of production . . . the effective tendency is for labour and capital to be rewarded in all branches at the same rates. If, however, this price structure [of the centre] . . . is transmitted to the periphery, the result will be that factors cannot be rewarded at the same rate in the different branches if the technical conditions (and so the productivity) are distributed otherwise than at the centre. (UD: 215-16) Note the assumption here that the price structure of the centre determines world prices. This rules out unequal exchange in Emmanuel’s sense, since the point of unequal exchange is precisely that low wages in the periphery are incorporated in low prices for the periphery’s products (see chapter 9 below). Amin could, perhaps, argue that prices are intermediate between those corresponding to the costs of the centre and of the periphery, giving both unevenness of productivity (as defined) and unequal exchange. In IUD, Amin went much further in rejecting Emmanuel’s analysis, asserting that the same products are produced at the centre and in the periphery and even denying that ‘the productions exchanged on the world market are specific, that they have irreducible use values’ (IUD: 209). The point seems to be that it doesn’t matter what is produced, since a demand can be created for anything, say for plastic flowers. However, even if the goods produced are constantly changing they must still be distinct from each other, or there is no basis for exchange or for a social division of labour. In particular, if use values are not ‘specific’, there is no room for unequal specialization, and Amin’s analysis evaporates, along with the theory of unequal exchange. If the periphery produces the same range of goods as the centre, they must sell at the MARXIST THEORIES OF IMPERIALISM 194 same prices, and super-exploitation must appear as super-profits for firms producing in the periphery, rather than through unequal exchange. Super-profits and unequal exchange could coexist if capital were only partly mobile. Amin did not specify the assumptions involved clearly. Levels of productivity, and their evolution, remain to be explained. Amin largely took them for granted. At an early stage in the evolution of the world economy, when commodity exchange was the only integrating factor, productivities in different areas would differ according to the level of development reached. Specialization would depend on comparative (not absolute) costs. Given a pattern of costs such that the periphery was excluded from the main lines of industry, the pattern would tend to persist, or to be reinforced, since there would be no opportunity to gain experience in industry. Once capital becomes mobile, it is difficult to see why technology, and hence productivity levels, are not transferred, along with capital. They clearly have been in some cases, as Amin knew (IUD: 212). A major part of the answer must surely lie in external economies: conditions of production that an individual enterprise either cannot provide for itself or need not provide where the industry is already well established. Examples are a skilled labour force, a network of suppliers, suitable transport services, and so on. These are, of course, all reasons why it is difficult (i.e. in a capitalist context, costly and hence unprofitable) to establish production in a new location. In the earlier parts of the imperialist stage, say around 1900, when wages in the centre were still fairly low and technology was still largely in the hands of skilled workers rather than being systematized and brought fully under the control of capital (cf. Braverman 1974), it is easy to see that there was little incentive for established capitalist firms in the centre to shift production to the periphery. In the present stage, however, wage differentials are large, and multinational firms have great experience of transferring technology, so it is much harder to see why a productivity gap should persist. Without it, Amin’s main arguments would collapse. He admitted that transfer of industries from the centre to the periphery is indeed the tendency, but gave a number of reasons why DEPENDENCY THEORIES 195 this tendency does not dominate. First, it takes time, and is only beginning. This is true, but why should it take much more than, say, the turnover time of capital equipment? Second, capitalism needs the high wages of the centre to provide a market (IUD: 213). Even if this were true (it is yet another version of under-consumption, which I have already criticized at length), it would still not explain why individual companies should continue to produce where wages are high. Third, he referred to the need for balance of payments equilibrium (IUD: 213). This is no argument: large-scale capital outflows from the centre would not disturb its balance of payments, if accompanied by exports of capital goods. Once industries have been transferred, the centre’s capacity to pay for imports would be reduced, but so would the general level of activity and hence the demand for them. Amin was willing to accept this argument for the periphery; why not for the centre in its turn? Fourth, development in the periphery must be blocked and distorted in order ‘to reproduce its own conditions of existence’ (IUD: 218). This simply restates the problem. Why and how does it reproduce itself? He finally, rather desperately, asserted that if, say, Mexico were to become a fully developed province of the USA, ‘the contradiction would shift from the economic to the cultural and political domains’ (UD: 381). To sum up, Amin provided a plausible account of the evolution of a periphery which is integrated, by stages, into a world market, while retaining a distinct wage level, a distinct social structure (persistence of pre-capitalist modes) and a lagging productivity level, at least in some sectors. He did not, however, have an adequate explanation of the evolution of productivity, especially in the era of multinational companies. De Janvry (1981: 9) treated Amin as part of the revival of classical Marxist ideas (‘to some extent’); his overall framework, however, was recognizably in the dependency tradition, even if his use of the concept of a mode of production was more sophisticated than most. MARXIST THEORIES OF IMPERIALISM 196 8.5 DEPENDENCY THEORY: THE BALANCE SHEET The essential elements of the dependency approach seem to be: (1) the (capitalist) world system is divided into a centre and a periphery (or equivalent terms); (2) the societies of the periphery are ‘dependent’, while those of the centre are not; (3) dependency restricts or distorts development in the periphery in some harmful way. I will consider these three elements in turn. Dependency theory divides the world system into centre and periphery. Frank, it is true, allowed for a chain of satellites, satellites of satellites, and so on, while Wallerstein introduced a ‘semiperiphery’, but the terminology clearly shows the underlying bipolar model. If such a classification is to be useful, there must be important features shared by all peripheral economies which differentiate them from all the centre economies. It is very hard to see what they are. Consider, for example, Korea, Ethiopia, India, Singapore, and Argentina. All would, I think, be counted as peripheral by most dependency writers, but what is gained by lumping them together? The only way to justify classifying them all as ‘peripheral’, is to claim that all are ‘dependent’. It is surprisingly hard to discover what ‘dependence’ means. The root meaning seems to be that the countries of the centre are in some sense masters of their own fate, while the dependent countries of the periphery are not. In any simple sense, this is clearly ridiculous; in the modern world, no country is unaffected by external events. Size matters, of course; the USA clearly relies less on external markets (for example) than Singapore does, but equally, Belgium is more ‘dependent’ (in that sense) than India. Something more must be meant. One possible meaning is that the countries of the centre gain from their involvement in the world system, while the countries of the periphery do not. I postpone discussion of this possibility for a moment. Another element in the idea of dependence, I suspect, is the notion that the world system is somehow shaped by the centre countries to suit their own purposes. This might reasonably be said of the political role of the major countries; imperialism in the traditional sense was exactly the reshaping of the world to suit the dominant imperial powers, and it is entirely possible to claim that imperialist powers continue to exert political dominance after formal independence, or over countries that have never been formally subjugated. The point of dependency theory, however, DEPENDENCY THEORIES 197 seems to be that the impersonal mechanisms of the market maintain the dependence of some countries on others, without the need to use state power overtly (Wallerstein is an exception to this rule). If this is what is meant, it betrays a failure to understand how capitalism works. There is no central planning agency in a market system, no overall purpose. Huge numbers of individual businesses and households, in many different countries, make their separate plans, which interact through the market. In the absense of state intervention, countries are not relevant units, at this level of analysis. The crucial element in dependency theory, then, has to be the claim that development in ‘dependent’ countries is restricted or distorted in some way, while development in the centre countries is not. This is not an easy claim to assess, since it presupposes some standard rate and pattern of development which the dependent countries are unable to attain. Early writers in the dependency tradition were fairly clear about what they meant; the periphery would lag even further behind the centre, would be unable to develop a significant industrial sector, and would continue to depend on primary product exports to pay for their imports of manufactures. This is not what has happened. I postpone fuller discussion to chapter 11, merely noting here what any standard statistical source will show; output has grown more rapidly in less developed countries than in more developed countries. Population has also grown faster, but even so per capita output has at least kept pace. Industrial output and exports have grown even more rapidly; a widespread process of (capitalist) industrialization is underway. The point is not that capitalism is doing a good job, in some abstract or ideal sense; there is poverty, working conditions are often very bad, and so on. The point is that dependency theory gives no useful guidance to the analysis of these problems. There were some attempts to define a concept of ‘dependent industrialization’, but this impressed few apart from those with an intellectual investment in dependency theory to protect. An even more telling indictment of dependency theory is the gross unevenness of capitalist development in the Third World. A few countries have experienced extremely high rates of growth, and will clearly join the centre countries soon. The most widely discussed examples are in east Asia, though there are examples in southern Europe and elsewhere. By contrast, other parts of the MARXIST THEORIES OF IMPERIALISM 198 Third World have grown very slowly, and in some cases per capita incomes have actually fallen substantially over decades. Growth of manufactured exports has been even more heavily concentrated into a few exceptionally successful places. To force these very different cases into the single category of ‘dependence’ is simply not helpful. It is slightly difficult to understand how dependency theory came to dominate radical thought in the way it did, given its visible weaknesses, and even harder to explain why Marxists accepted a framework of analysis so alien to the mainstream of Marxist thought. It can be argued (Harris 1986, ch. 7) that dependency theory, and related ideas, served the interests of the emerging middle classes of the Third World. Certainly, a system of thought which blamed all ills on foreigners, and legitimized state intervention to support domestic industry, was very convenient both for the personnel of the state apparatuses and for emerging industrial capital. Dependency theory emerged at a time when classical Marxism was very weak, and had its roots in development economics, from which it inherited the concepts of national interests and national development. It has come to look like a blind alley, and attention has switched to examination of the differences in the internal structures of different ‘peripheral’ countries. 8.6 SUMMARY Both Frank and Wallerstein identified capitalism with a network of exchange relations, on a world scale, that channel surplus from satellite (periphery) to metropolis (core). Both insisted that the internal structure and development of different parts of the world economy is primarily determined by their place in the whole, and that the organization of production at a lower level (enterprise, sector, nation state) is secondary. Both asserted that development and underdevelopment are opposite sides of the coin, that one is the result of the other. My main criticism of both writers is that there is little connection between their grandiose general statements and their (sometimes illuminating) discussion of particular historical cases. What is lacking is real theory. Frank and Wallerstein did, however, make an important contribution by insisting on the DEPENDENCY THEORIES 199 importance of underdevelopment and the necessity of analysing it in terms of the development of a world system. Amin argued that the capitalist world economy is divided into two distinct types of social formation, those of the centre and those of the periphery. In the centre, the capitalist mode of production eliminates other modes, and generates a process of development of the sort analysed by the classical Marxists. In the periphery, capitalist development is ‘blocked’ by the competition of the more advanced industries of the centre, so pre-capitalist modes persist for a long time, and an economic and social structure quite distinct from that of the centre arises. Amin’s explanation of unequal specialization hinges on relative levels of productivity in the centre and the periphery, which he did not explain fully. His is the most sophisticated version of dependency theory, and contains many valuable insights. On a more general level, dependency theory fails, because it cannot explain the burst of industrialization in the so-called periphery in the decades since the Second World War, nor can it help to account for the extreme unevenness of capitalist development in the Third World.