In the middle decades of the twentieth century, a group of economists and sociologists working in Latin America built a body of theory around a single unsettling proposition: the poverty of the poor economies is not a starting condition that growth will gradually wash away, but a product, manufactured continuously by the same world economy that enriches the rich ones. Dependency theory, as the body of work came to be called, holds that underdevelopment is actively made. The regions supplying coffee, copper, sugar, tin and oil to the industrial centres were not simply behind, waiting at an earlier station on a line every economy travels. Their labour systems, their landholding patterns, their ports and railways and tax codes had been shaped, over centuries, by the demands of buyers elsewhere, and that shaping was still going on. Rich and poor were not two separate stories. They were one story told from two ends.

What a newcomer most needs to know, before any names are attached, is that this body of theory was never one position. It contains at least three, and they contradict one another on the question outsiders assume it answers with a single voice. One strand holds that development inside dependency is impossible, because the relation itself produces underdevelopment as its necessary other face. A second holds that dependency bends development into distorted shapes without preventing it, so that dependent economies can industrialise, grow and urbanise while remaining subordinate. A third, more technical strand argues that the injury runs through trade itself: where capital moves freely and labour cannot, the prices at which goods exchange quietly siphon value from low-wage economies to high-wage ones. Any summary that merges these into a slogan about rich countries making poor countries poor has lost the theory in the act of introducing it.

This article separates the three positions, gives each its evidence and its strongest critics, and keeps the versions apart, because most arguments about dependency theory are really arguments at cross purposes: a refutation aimed at the strongest version is triumphantly applied to the whole family, while defences quietly retreat to the weakest. The Marxist ancestry of the theory, in the work of Paul Baran, and its relation to the broader Marxist account of imperialism are both part of the story; the wider theory of imperialism has its own treatment in the series guide to Marxism and imperialism, and this article links to it rather than repeating it. What follows concentrates on the dependency writers themselves: the structuralist economists who prepared their ground, the three positions in turn, the super-exploitation argument that grew alongside them, and then the criticisms, from the Marxist objection that the leading dependency theorists were not Marxists at all to the East Asian economies whose industrialisation put the strong version under a pressure it could not comfortably bear.

The claim at the centre of dependency theory

The canonical claim can be stated in one sentence: the underdevelopment of the peripheral economies is actively produced by their subordinate integration into a world economy dominated by industrialised centres, rather than being an original condition that development will overcome. Every word in that sentence carries weight. Produced, because the theory is historical before it is anything else: it asks how particular regions came to export what they export and import what they import, and it answers with conquest, slavery, colonial administration, and the commercial policies of the centuries in which the world market was assembled. Subordinate, because integration as such is not the accusation; every economy trades, and the dependency writers were not preaching autarky in their theoretical work, whatever their political admirers later made of them. The accusation is about the terms on which a region is integrated: what it is encouraged to produce, who owns the enclaves that produce it, where the surplus accumulates, and which decisions about its future are taken in boardrooms and ministries elsewhere. Centres, finally, because the theory is relational. A centre is not simply a rich place; it is a place whose development feeds on the relation, whose industries grew up processing the periphery’s materials and selling back finished goods, whose working class was fed and clothed partly by transfers the periphery never voted for.

Against whom was this claim aimed? Its first target was the modernisation theory dominant in North American social science after the war, which arranged societies along a single path from tradition to modernity and read the poor economies as earlier versions of the rich ones, delayed by internal obstacles: the wrong values, the wrong institutions, a shortage of entrepreneurs. Dependency theory inverted the direction of explanation. The obstacles, it argued, were not internal heirlooms but external impositions maintained by ongoing relations. A plantation economy did not lack entrepreneurs because its culture discouraged enterprise; it lacked diversified industry because its land, credit and infrastructure had been organised, by identifiable people for identifiable reasons, around exporting one crop. Where modernisation theory saw a queue, dependency theory saw a division of labour with a history and with beneficiaries.

The second target was closer to home and cut deeper, because it stood inside the Marxist tradition itself. Orthodox communist parties in Latin America, following a stage theory of history, held that countries like Brazil or Chile still contained feudal remnants, that a national bourgeoisie existed whose interests clashed with both feudal landlords and foreign imperialism, and that the correct strategy was a two-stage revolution: first a bourgeois-democratic stage in alliance with that national capital, then socialism later. Dependency theory, in its most famous strand, denied the premise. Latin America, it argued, had been capitalist since the sixteenth-century conquest, its mines and plantations built for the world market from the first day; there was no feudalism left to overthrow and no progressive national bourgeoisie waiting in the wings, because the local possessing classes were woven into the export structure itself. A dispute about periodisation thus became a dispute about whom to march with, and it is impossible to understand the passion of the dependency debates without grasping that a theory of trade was simultaneously a theory of political alliance.

Why do dependency theorists reject the idea of an original backwardness?

Because the historical record shows the periphery being made, not merely found. Regions drawn into the world market as suppliers of bullion, sugar or coffee had their labour systems, landholding and trade routes reorganised around external demand. What modernisation theory treated as a starting point, dependency writers treated as a product: an outcome with authors, dates and mechanisms.

The rejection matters because of what follows from each starting point. If backwardness is original, development is a matter of importing the missing ingredients: capital, technology, skills, attitudes. Aid and investment then appear as remedies generously supplied from outside, and their failure can be blamed on the recipient’s deficiencies. If backwardness is produced, the remedies themselves come under suspicion, since the channels through which capital and technology arrive are the same channels through which the producing relation operates. The dependency writers did not deny that machinery, agronomy and engineering cross borders; they denied that the crossing was innocent. A mine is a transfer of technology and also a structure of ownership, a labour regime, a claim on the host state’s police, and a destination for profits, and the theory insists that the second list cannot be wished away by celebrating the first.

There is a third element in the canonical claim that summaries usually drop, and dropping it flattens everything else in this article. Dependency is a relation between economies mediated by relations within them. The external chain lands somewhere: on a landlord class that prefers exporting to feeding the home market, on a merchant class that profits from importing what might be made at home, on a state whose revenues depend on customs duties from the export enclave. The dependency writers called this the internalisation of the external, and it is what separates their work from a simple geography of blame. A purely external theory could not explain why similarly placed economies responded differently to the same world market pressures, nor why liberation from a colonial power so often changed the flag over the customs house and nothing else. The internal class structure is where the relation reproduces itself, which is also why the tradition’s best writers spent as much effort on landlords, merchants and generals at home as on bankers abroad.

The definition wars: what counts as a dependent economy?

Before the positions can be separated, the school’s central noun needs its own scrutiny, because the dependency writers never agreed on what dependency is, and the disagreements were not pedantry: each definition smuggles in a theory. At one pole stand definitions by trade structure, in which a dependent economy is one whose exports concentrate in a few primary commodities and whose imports cover its manufactures, a definition that risks circularity, since it makes dependence a synonym for the trade pattern it is meant to explain. At another stand definitions by ownership, counting the foreign share of the commanding heights, which mistake a symptom for the relation when foreign firms operate as the local bourgeoisie’s chosen vehicle. Dos Santos’s conditioning definition, treated in its own section below, was crafted to escape both traps by defining dependence through the structure of the dependent economy’s own reproduction: its accumulation is conditioned by another economy’s expansion because its decisive sectors, its technology and its finance answer to decisions taken there. Cardoso’s situational usage escapes the traps from the other side, refusing definition by checklist altogether and treating each situation’s configuration of external linkage and internal alliance as the object of study. The definitional spread explains a standing puzzle of the secondary literature, why dependency theory can be simultaneously so easy to summarise and so hard to examine: the examiners keep testing one member’s definition against another member’s conclusions. A careful reader fixes the definition in force before grading the argument built on it, and the three positions grid later in this article should be read with that discipline in mind, since Frank’s dependence is a structure of surplus drainage, Cardoso’s a situation of conditioned development, and Emmanuel’s a property of the wage structure, three referents sharing one word.

The foil: stages, take-off and the modernisation promise

To feel the force of the dependency inversion, a reader needs the doctrine it inverted in full view, because modernisation theory was not a straw target but the governing common sense of the postwar development establishment, armed with foundations, ministries and money. Its most influential statement was Walt Rostow’s The Stages of Economic Growth, published in 1960 with a subtitle that announced the contest openly: A Non-Communist Manifesto. Rostow arranged all economies along one road: the traditional society, the preconditions for take-off, the take-off into self-sustaining growth, the drive to maturity, and the age of high mass consumption, whose exhibit was the United States. The scheme’s kindness was also its condescension. Every society was travelling; some had merely started late, and their present was the rich countries’ past, readable in advance. Poverty, on this view, is a condition of beginnings: a shortage of capital, of technology, of the modernising elite attitudes that set accumulation going. Remedy follows diagnosis. Supply the missing inputs, train the elites, wait for the take-off, and the queue advances.

The doctrine had a foreign policy, and the policy reveals the stakes the dependency writers saw in what looked like an academic scheme. The Alliance for Progress, launched in 1961, poured North American money into Latin America on modernisation premises: fund the preconditions, strengthen the reforming middle classes, and the continent would take off while being inoculated against the Cuban example. When the money arrived and the take-off did not, when the reforming classes allied instead with militaries and the militaries with the export sectors, the modernisation framework had no internal resource for explaining the failure except to blame the recipients’ traditions more loudly. Dependency theory was built in the space that failure opened. Its founding gesture was to refuse the queue: to insist that the Latin American present was not the North American past, that the continent had been inside world capitalism for four centuries and bore the marks of that membership, and that a theory of development which cannot see the relation between the developer and the developed is not an innocent simplification but an ideology with a research budget.

The sociological wing of modernisation made the same move in cultural dress. Traditional values, ascriptive ties, religious fatalism: the obstacles were located inside the poor society’s head, and development became a pedagogy in which the periphery would be taught to want what the centre already was. Against this the dependency writers set a methodological rule that outlasted their conclusions: explain a society’s condition by its insertion into structures larger than itself before reaching for its psychology, because the psychology itself has usually been formed by the insertion. A planter class that will not industrialise may be displaying not a traditional mentality but a perfectly modern rationality within a structure that rewards exporting and punishes manufacturing. Rostow’s stages survive now mostly as a foil in textbooks, which is the dependency school’s doing more than anyone’s: it was the Latin American critique that made the single-road picture untenable in serious company, whatever became of the critique’s own positive claims.

The structuralist groundwork in Latin America

Dependency theory did not spring from Marxism directly. Its immediate precursor was the structuralist economics developed at a United Nations regional commission in Latin America from the late 1940s, and the debt matters, because the structuralists supplied the vocabulary, the statistical habits and the founding empirical claim on which the later, more radical theory was built. The Economic Commission for Latin America, known by its Spanish initials as CEPAL, was founded in 1948 and seated in Santiago de Chile. Its executive secretary from 1950 was the Argentine economist Raúl Prebisch, who arrived with an unorthodox past: he had run his country’s central bank, had watched the Depression destroy the export model on which Argentine prosperity rested, and had drawn heterodox conclusions from the wreckage. Under Prebisch the commission became something unusual in the history of economic institutions, a place where a region’s economists studied their own economies collectively and produced, within a few years, a coherent alternative account of how the world economy worked.

How did a United Nations commission become the theory’s first home?

The Economic Commission for Latin America, founded in 1948 and headquartered in Santiago, gathered the region’s economists inside one institution with a mandate to study its economies as a group. Under Raúl Prebisch it produced the centre and periphery vocabulary and the terms of trade argument, giving structuralism staff, statistics and a printing press.

The vocabulary came first. Prebisch divided the world economy into centres and peripheries, terms he chose deliberately against the older language of advanced and backward countries. The centre, in his usage, is where technical progress originates and where its fruits are retained: productivity rises in the centre’s industries, and because its workers are organised and its firms oligopolistic, the gains appear as higher incomes rather than lower prices. The periphery supplies primary commodities, food and raw materials, produced under competitive conditions with abundant labour; there, productivity gains leak abroad as lower export prices. The periphery thus exports part of the fruit of its own technical progress to the centre’s consumers, a mechanism Prebisch described with a dry fury that still reads freshly. The schema was not Marxist. Prebisch wanted industrialisation within a reformed capitalism, not its abolition, and his commission’s policy child was import-substitution industrialisation: tariffs and planning to build, behind protective walls, the industries the market had failed to summon, so that the periphery could manufacture what it had been importing and keep its productivity gains at home.

The founding empirical claim was the terms of trade thesis, advanced independently by two economists around 1950. Prebisch, working from British trade statistics reaching back to the 1870s, argued that the prices of primary commodities relative to manufactures had declined over the long run: a given basket of peripheral exports bought steadily less of the centre’s machinery and finished goods. Hans Singer, a German-born economist working at the United Nations in New York, published the same conclusion in 1950 from the distribution of gains between investing and borrowing countries. The joint result entered the literature as the Prebisch-Singer thesis. Its logic fitted the centre-periphery schema like a key: if the centre keeps its productivity gains as income while the periphery surrenders its own as cheaper exports, then the exchange of primary products for manufactures is a slow leak, and specialisation according to comparative advantage, the classical prescription, condemns the periphery to finance the centre’s rising living standards with its own. Free trade between structurally unequal partners was, on this account, not a neutral rule but a mechanism of transfer.

Structuralism was always more than the thesis, and its internal richness explains why so many dependency writers began as CEPAL economists before breaking away. The Brazilian Celso Furtado gave the tradition its historical depth, reading Brazil’s economy as a sequence of export cycles whose internal structures each left behind, and treating underdevelopment as a specific historical formation rather than a stage. The Chilean Osvaldo Sunkel developed the paired concepts of transnational integration and national disintegration: the modern export sector integrates upward into the world economy while the society around it fragments, so that growth and disintegration proceed together. These were diagnoses with policy implications that stopped well short of the dependency school’s later conclusions, and the distance matters. The structuralists identified the leak and proposed to plug it with industrialisation sponsored by the state. The dependency writers asked why, after two decades of import substitution, the plugging kept failing: why industrialisation behind tariffs had produced new dependencies on imported machinery, foreign patents and external finance, so that the periphery now imported the means of importing less. That question, pressed hard in the 1960s, is the hinge on which structuralism turns into dependency theory.

The structuralists also modelled a method the dependency writers inherited: thinking of the world economy as a single structured whole whose parts are defined by their positions in it, rather than as a collection of national economies at different stages. Once that move is made, a great deal follows. National statistics stop being self-explanatory. A boom in export earnings can be a symptom of deeper subordination. A fall in the terms of trade is not bad luck but the signature of a structure. And the relevant comparison for a peripheral economy is never its own past alone but the relation that binds its past to the centre’s. The Marxist dependency theorists took this holism from Santiago, married it to a theory of surplus and class that CEPAL never held, and turned a commission’s research programme into an indictment.

Two economists, one result: the making of the terms of trade thesis

The Prebisch-Singer thesis is often taught as a statistical finding, and its biography shows it was closer to a conversion experience repeated independently in two lives. Prebisch came to it through the Argentine collapse. As the young manager and then governor of his country’s central bank in the 1930s, he had administered an economy whose prosperity rested on selling beef and grain to Britain and buying manufactures back, and he watched the Depression snap the arrangement: export prices fell off a cliff while the prices of imported machinery did not, debts denominated in the centre’s currency grew heavier as earnings shrank, and the orthodox prescriptions his institution applied deepened the contraction they were meant to cure. His heresy began as bookkeeping. When, after a spell in political exile from the Banco Central he had founded, he took up the Santiago commission’s leadership, he ordered the trade statistics of the previous seventy years read as a single story, and the story they told was of a slow, compounding transfer: the periphery’s exports bought less machinery per ton with each passing decade, because the fruits of technical progress were priced differently at the two ends of the exchange. Hans Singer reached the same destination from the refugee’s road, a German economist driven out by the Nazis, employed on the economic side of the young United Nations, studying the terms on which capital had flowed to borrowing countries. His 1950 paper framed the asymmetry as the distribution of gains between investing and borrowing nations: the investor’s country kept the productivity gains of its industries as rising incomes at home, while the borrowing country’s gains were exported, quite literally, in the form of cheaper commodities.

That both men published within a year of each other, on different continents, from different data, is the strongest reason the thesis commanded attention even from economists who disliked its implications. It was not a fitted curve in search of a patron. It also had a precision its slogan form loses. The claim concerned relative prices over long periods between two classes of goods, primary products and manufactures, and it rested on an institutional mechanism, the differing organisation of labour and product markets at centre and periphery, not on a conspiracy and not on any single decade’s misfortune. When Prebisch carried the argument to the first conference of the United Nations Conference on Trade and Development in 1964, as its secretary-general, the thesis became the intellectual charter of the Group of 77, the coalition of developing countries that pressed the claim into diplomacy. A statistic had become a bloc, which is why the subsequent econometric quarrels over trend lines, treated later in this article, were never only about econometrics.

The policy child: import substitution and its crisis

Structuralism’s theory had a policy, and the policy’s career is the bridge between the commission’s economics and the dependency school’s radicalism, because dependency theory was in large part the theory of why the policy failed. Import-substitution industrialisation proposed that the periphery manufacture at home what it had been importing, behind tariffs, quotas and an active state, beginning with the easy stage of consumer goods whose markets obviously existed: textiles, shoes, processed food, then the harder stages of intermediates, steel, chemicals, capital goods. Through the 1950s the larger Latin American economies did exactly this, and for a time the numbers obliged. Brazil and Mexico industrialised at speed; São Paulo became one of the world’s great manufacturing cities; a generation of economists could believe the leak had been plugged by the state standing at the customs house.

The crisis of the model, when it came in the 1960s, was diagnosed from inside the structuralist camp before the dependency writers generalised it, and the internal diagnosis is worth following because it taught the younger generation where the bodies were buried. The easy stage exhausted itself: once local factories supplied the home market with consumer goods, further industrialisation required importing the machinery, components and patents to make them, so each increment of industrial growth raised the import bill it was meant to reduce, and the external bottleneck tightened rather than relaxed. Agriculture, neglected and squeezed to feed the cities cheaply, stagnated, generating the food imports and rural exodus that swelled the urban margins. Inflation, which the structuralists against the monetarists read as the monetary expression of these bottlenecks rather than a simple excess of demand, became endemic and politically corrosive. And the industrialisation that did occur was strikingly undemocratic in its fruits: the protected factories sold to the minority with incomes, employed relatively few, and deepened the income concentration that CEPAL’s own studies documented with mounting alarm. Furtado drew the darkest conclusion, a stagnation thesis for Brazil in which the very structure of demand under extreme inequality condemned the economy to stall once the easy substitutions ran out.

Into this crisis of the policy stepped its two radicalisations, and the dependency school can be defined as their intersection. One line of criticism, running through Frank, concluded that the failure was written into the insertion itself: a periphery integrated as a supplier could not industrialise its way out on the system’s own terms, and the tariff walls had merely changed the form of subordination from importing goods to importing the means of producing them. The other line, running through Cardoso and the Brazilians who debated Furtado’s stagnationism, Maria da Conceição Tavares and José Serra prominent among them, corrected the stagnation thesis from the left: Brazil was not stalling, it was growing, ferociously and exclusionarily, under a military regime that had crushed wages and opened the commanding heights to foreign firms, and the task was to theorise growth-with-dependence rather than mourn its absence. Between those two lines, one denying that dependent industrialisation was happening and one insisting on analysing it, the impossibility split of this article was born; it is no accident that it was born in Brazil, the economy large enough to make both readings arguable from the same statistics.

The inflation war: structuralists against monetarists

Before the dependency school existed, its parent tradition fought the battle that taught Latin America’s economists to think structurally, and the battle’s theatre was the region’s chronic inflations. The monetarist diagnosis, carried by the International Monetary Fund’s missions and by economists of the Chicago connection, read inflation as a monetary fever: deficits financed by the printing press, cured by austerity however sharp the recession it induced. The structuralist answer, built by Sunkel, Furtado and the commission’s economists from the Chilean and Argentine records, relocated the fever’s source in the economy’s anatomy: an inelastic food supply from the unreformed countryside meeting urban demand swollen by industrialisation, a foreign-exchange bottleneck that devalued the currency and passed import prices into every cost, fiscal systems incapable of taxing the propertied, so that inflation was the mechanism by which real conflicts over income, landowners against workers against the state, were temporarily settled by the printing press in the absence of any settlement of the conflicts themselves. On this reading the Fund’s medicine treated the settlement mechanism while aggravating the conflicts, and the political history of the era, stabilisation plans collapsing governments from Buenos Aires to Santiago, gave the structuralists their evidence in headlines. The inflation war matters here because it was the school’s methodological nursery: the habit of refusing the surface variable, money, prices, trade balances, and demanding the structure beneath it was formed over the price indices of the 1950s before it was ever turned on the world economy, and the dependency theory’s characteristic gesture, the mechanism behind the magnitude, is the inflation war’s gesture generalised.

Furtado in office: the Northeast as an internal periphery

Structuralism’s one season in executive power came in Brazil, and it taught the generation that followed more than any book. In 1959 Celso Furtado, returned from his years at the Santiago commission, took charge of a new development agency for Brazil’s Northeast, the SUDENE, charged with the region whose sugar economy and droughts made it the standing refutation of Brazilian national averages. Furtado’s diagnosis treated the Northeast as an internal periphery in the commission’s own sense: its export earnings and its savers’ funds drained toward the industrialising Southeast, whose protected factories then sold it manufactures at prices its planters’ incomes could not match, so that one country’s development was manufacturing another’s underdevelopment inside a single flag. The remedy proposed was structuralism applied at regional scale, public investment, industrial incentives, land and water reform in the drought polygon, against the resistance of the sugar oligarchy whose political machine the plan threatened. The experiment lasted until the coup of 1964 abolished its author into exile, and its afterlife in the dependency debates is double. For the school’s radicals it proved that even a reforming technocracy armed with the correct diagnosis would be destroyed by the class structure it challenged, a lesson in the internalisation of the external written in decrees. For its subtler members it proved something they used against Frank: that the structures were national and regional as much as international, and that a theory of the world chain which could not see Brazil draining its own Northeast had mistaken its scale.

Prebisch’s last turn: the critique of peripheral capitalism

The founder lived long enough to indict his own policy’s fruits, and his last book belongs in this record because it dissolves any simple picture of structuralism as the dependency school’s complacent parent. Peripheral Capitalism: Crisis and Transformation, published in 1981 when its author was eighty, surveyed the Latin America that import substitution had built and found the structure wanting in terms his younger radical critics could have drafted. The periphery’s industrialisation, Prebisch argued, had been captured by a privileged consumption society: the upper and middle strata imitated the centre’s consumption patterns with the centre’s products, absorbing the surplus that development required, while the mass of the population remained outside the modern economy’s benefits, and the system’s dynamic condemned it to recurrent crisis between the pressure of that consumption and the claims of accumulation. His remedy had travelled as far from 1950 as his diagnosis: a transformation in the use of the surplus, democratic in form, in which the state would discipline privileged consumption and redirect resources to the excluded, a programme he knew belonged to no party that existed. The convergence is striking: the octogenarian of the commission and the exiled Marxists of the dependency school, arriving from opposite directions, met on the ground of the surplus and its misappropriation by a consumption structure copied from the centre. Intellectual genealogies in this field are rarely trees; at the end they are rivers rejoining.

Mexico’s stabilising development and its discontents

Mexico supplied the structuralist era with its most successful pupil and its most instructive disillusion. Under the formula its finance ministers called stabilising development, the Mexican economy from the middle 1950s to about 1970 combined rapid industrial growth with price stability and a fixed exchange rate, a performance the international institutions exhibited as proof that the periphery could develop within the rules. The dependency writers who examined the case, several of them Mexicans working in the universities the miracle had expanded, read the same decades as the assembly of a deeper dependence: the industrial plant foreign-owned in its dynamic branches, the countryside squeezed until its crisis arrived as migration, the stability itself purchased by wage restraint and by borrowing against the future, and the political settlement, a single party’s long hegemony, resting on the exclusion that 1968 exposed when the state answered its own students with massacre in the plaza of Tlatelolco. The North American political scientist Roger Hansen’s counter-account, The Politics of Mexican Development, conceded much of the description while disputing the moral: Mexico’s rulers, he argued, had made of the revolution’s inheritance a machine for distributing just enough to every organised interest to keep the system legitimate, and its failures were failures of that machine’s arithmetic rather than of the world market. The Mexican case thus became the debate in miniature, the same growth read as dependence perfected, as development achieved, and as a political technology of inclusion wearing out, and a student who masters the three readings owns the tradition’s whole grammar of evidence.

Argentina: the rich periphery that declined

Every general theory of the periphery must face the case of the country that was rich. Around the First World War Argentina stood among the half-dozen wealthiest societies on earth per head, its pampas feeding Europe, its capital a byword for Belle Époque opulence, and across the following century it declined, relatively and then absolutely in its worst decades, through a sequence of coups, inflations and defaults that made it the dependency debate’s home puzzle: the school’s founder-generation included Argentines, Prebisch above all, for whom the puzzle was autobiography. The dependency reading of the decline ran through the landowning oligarchy’s alliance with British capital: the pampas elite, enriched by the very trade that should have financed diversification, bound the economy to its role as Britain’s farm, formalised as late as 1933 in the Roca-Runciman pact that guaranteed the beef trade at the price of concessions to British firms, and blocked every project, industrial, fiscal, agrarian, that threatened the arrangement; decline was thus the long bill for a century of successful insertion on the oligarchy’s terms. The Argentine economist Marcelo Diamand, working in the structuralist line, formalised the economy’s stop-go rhythm as a pendulum between the export sector’s exchange rate and the industrial sector’s, each coalition’s victory planting the next crisis, a mechanism more precise than the drain and compatible with it. Liberal historians answered that Argentina’s institutions, its politics of distribution and its macroeconomic habits, explain the decline without residue, pointing to settler economies with similar endowments, Australia above all, that made the opposite journey under comparable insertion. The Argentine case therefore performs the same office for the whole tradition as Korea performs for its strong version, in the opposite direction: it is the exhibit each side must carry, the declinists’ proof that insertion can impoverish the rich, the institutionalists’ proof that nations are responsible for their own trajectories, and no verdict this article could deliver would be honest that did not leave the case, like the country, magnificently unresolved.

The Marxist ancestry: Hobson, Luxemburg and Lenin

Dependency theory presented itself as a correction of classical Marxism from the periphery, and a reader should know which ancestors it was correcting, since the correction defines the novelty. The classical Marxist literature on imperialism, written before and during the First World War, explained the expansion of capitalism outward from the centre’s internal pressures. John Hobson, the liberal whose Imperialism of 1902 the Marxists mined, blamed underconsumption at home: maldistributed income starved domestic demand, surplus capital sought outlets abroad, and the flag followed the portfolio. Rosa Luxemburg, in The Accumulation of Capital of 1913, gave the outward drive a sterner logic: capital, she argued, cannot realise its surplus value within a closed capitalist system and must perpetually seize non-capitalist milieux, peasants, artisans, colonies, consuming them as it goes, so that imperialism is the political form of accumulation’s need for an outside. Lenin’s pamphlet of 1916 relocated the cause in monopoly: competitive capitalism had become monopoly capitalism, banks and industry fused into finance capital, and the export of capital, the division of the world among combines and powers, defined an epoch whose wars were its bookkeeping. All three ancestors share a standpoint the dependency writers would abandon: they stand in the centre and look out, asking why the centre expands. The periphery appears as the field of expansion, its fate a consequence deduced from the centre’s needs.

The dependency inversion was to stand in the periphery and look at the relation. Its writers had read the canon, Frank and dos Santos closely, and they judged that a theory built from the centre’s motives could not explain the periphery’s structures: why this landlord class, that enclave, those terms of trade, the specific deformations that two centuries of incorporation had deposited. Where Lenin’s scheme treated the export of capital as the centre’s overflow seeking yield, Baran and his successors asked what the inflow did to the society receiving it, and found the receiving structure, not the sending pressure, to be the active ingredient in underdevelopment. Luxemburg’s outside came closest to their concerns, and several dependency writers acknowledged the kinship: her non-capitalist milieu, consumed by accumulation, is the ancestor of the satellite drained by the metropolis. But even Luxemburg’s periphery is passive, a reservoir and a market, while the dependency writers populated theirs with classes, states and strategies of their own. The school’s claim to originality rests substantially on that relocation of the analytical standpoint, and its characteristic errors, the underweighting of the centre’s internal dynamics among them, are the relocation’s price.

The forgotten ancestor: the price scissors of the Soviet debate

The mechanism at the heart of unequal exchange, a transfer effected through prices rather than theft, has an ancestor in a debate the dependency writers rarely cited: the Soviet industrialisation controversy of the 1920s. Evgeny Preobrazhensky, theorist of the Left Opposition, proposed that a socialist state ruling a peasant country must finance its industry by a primitive socialist accumulation, deliberately setting the terms of trade between state industry and peasant agriculture against the countryside, buying grain cheap and selling manufactures dear, so that the surplus for the factories is pumped, his verb, through the price system from the pre-socialist sector. Nikolai Bukharin answered that the scissors would kill the goose: squeeze the peasantry and it withdraws from the market, sows less, and the towns starve, a prediction the collectivisation catastrophe wrote in famine after both debaters were dead. The episode matters here for two reasons. It establishes that the transfer-through-prices mechanism was theorised, with full awareness of its costs, a generation before Emmanuel, inside the Marxist tradition the dependency school inherited, and that the tradition’s own history contained both the mechanism and its most famous refutation-by-anticipation. And it explains a certain reticence: dependency writers building a case against the world market’s prices had little incentive to advertise that the socialist state’s founding economists had operated the same scissors on purpose, a silence the historian should note as evidence that the school’s Marxism, like all Marxisms, selected its ancestors. The scissors remain the cleanest illustration in the whole literature of this article’s recurring distinction, between a transfer of value, a fact about prices, and a judgement about the transfer, which depends on what the squeezed sector’s surplus builds and what it destroys.

A sympathetic outsider: Hirschman among the pioneers

One figure stands at the edge of this story as its most instructive witness, because he shared the pioneers’ formation, quarrelled with both their orthodox and their radical successors, and left the fairest account of what the whole development enterprise, dependency included, had been. Albert Hirschman, a German economist of the same refugee generation as Singer, worked in Colombia in the 1950s as an adviser, wrote The Strategy of Economic Development out of that experience, and spent a long American career anatomising the rhetoric by which societies persuade themselves that change is impossible or compulsory. Hirschman’s sympathies were with the Latin American structuralists, whose company he kept and whose historical sense he shared against the stage theorists; his famous essay on the rise and decline of development economics mourned a field that had begun, in Santiago and elsewhere, with the conviction that poor economies were different in kind and had ended by surrendering that conviction to a universal economics with better mathematics. Toward the dependency radicals his attitude was the affectionate severity of a man who thought their diagnosis half right and their politics of despair wholly wrong: his lifelong theme, the bias for hope, the systematic underestimation of a society’s latent problem-solving capacities, was aimed at every doctrine, Frank’s above all, that converted a structural analysis into a sentence of impossibility. A reader who wants the tradition’s inner tensions held in one humane intelligence could do worse than end with Hirschman, who managed the rare feat of being claimed by neither camp because he had been fair to both in print.

Baran and the Marxist source

If structuralism supplied the map, the Marxist source of dependency theory is a single book: The Political Economy of Growth, published in 1957 by Paul Baran, a Russian-born economist who had worked in wartime Washington and ended his career at Stanford as nearly the only avowed Marxist holding a chair in an American economics department. Baran’s book did for the periphery what his joint work with Paul Sweezy would later do for the centre, in the account of monopoly capital treated in the series article on Baran and Sweezy: it built the argument around the concept of the economic surplus, and it drew from that concept a conclusion that scandalised both liberal development economics and the communist orthodoxy of its day. Underdevelopment, Baran argued, is not the residue of a pre-capitalist past. It is the obverse of development in the centres, produced by the same historical process, and in the colonial and semi-colonial world it was still being produced in his own lifetime.

What did Baran mean by economic surplus?

Baran defined the economic surplus as the gap between what an economy produces and what it consumes in essentials. His decisive move was to distinguish the small surplus actually saved and invested under monopoly capitalism from the far larger potential surplus wasted on luxury, advertising, finance and military spending, which planning could redirect to development.

The distinction between actual and potential surplus does the analytical work. A poor economy, on Baran’s accounting, is not poor because it generates nothing above subsistence. Plantations, mines and export agriculture generate considerable surplus. The problem is what happens to it: it is drained abroad as repatriated profit and interest, squandered at home on the luxury consumption of landed and merchant classes whose tastes were formed by the centre’s department stores, absorbed by bloated commercial and financial intermediaries, or frozen in speculative landholding. Meanwhile the potential surplus, the output an economy could achieve if its resources were rationally employed, dwarfs the actual product: peasants work fragments of exhausted land with primitive tools while fertile estates lie idle as status holdings. Underdevelopment, in this framing, is a failure of surplus use imposed by a class structure, and that class structure was itself installed and maintained by the penetration of foreign capital. The comprador landlord and the import merchant are not survivals from a feudal past; they are the local management of the world market.

Baran gave the argument a historical edge that the dependency writers sharpened into their signature claim. The classic case was India, whose textile industry, competitive with Britain’s at the end of the eighteenth century, was dismantled under colonial rule while Lancashire’s mills, protected and privileged, took the market. Whether every detail of that older deindustrialisation debate survives archival scrutiny is a question historians continue to argue over; what matters for the lineage is the form of the claim. Development in one place and underdevelopment in another are not parallel processes but coupled ones, two products of a single expansion. Baran also broke with stage theory as firmly as Frank later would: since the periphery’s ruling classes owed their positions to the existing insertion into the world economy, they could not be expected to lead a national development project against it, and the parties waiting for them to do so were waiting for their own gravediggers to unionise. The Political Economy of Growth is thus the bridge text of the whole tradition, structuralist in its holism, Marxist in its categories of surplus and class, and dependency theory in embryo in its refusal to treat underdevelopment as an original condition.

The older drain: Naoroji and the genealogy of the transfer idea

The claim that integration transfers wealth from periphery to centre is older than Marxism in the periphery, and its nineteenth-century form deserves a section, because the dependency writers inherited an argument Indian nationalism had already perfected. Dadabhai Naoroji, the Indian liberal who sat in the British Parliament and presided over the Indian National Congress, spent decades computing what he called the drain: the portion of India’s revenue and export surplus that left the country without any equivalent return, as home charges for the colonial administration, pensions, interest and the profits of the ruling connection, financed by taxing the Indian peasantry. His Poverty and Un-British Rule in India, published in 1901, assembled the case with an accountant’s patience. The drain mattered analytically because it reframed the empire’s central boast. British rule claimed to develop India, railways, irrigation, law; Naoroji answered that a development financed by unrequited exports from one of the world’s poorest populations was a transfer wearing a hard hat, and that India’s poverty was not despite the connection but substantially through it. Every element of the later dependency argument is present in embryo: the unrequited transfer, the coupling of the centre’s accumulation to the periphery’s impoverishment, the insistence that the accounts be kept whole rather than item by gleaming item.

The genealogy runs from Naoroji through the Indian nationalist economists to Baran, who cited the Indian case as his historical warrant, and from Baran to the Latin Americans, who universalised it. Acknowledging the line does the dependency school no harm; it clarifies what was new in 1960 and what was not. Not new: the transfer mechanism and the moral bookkeeping. New: the relocation of the mechanism inside a theory of the world economy as a single structure, the refusal of the feudal-survival diagnosis, and the insistence that the drain operates not only through colonial tribute, Naoroji’s channel, but through the ordinary price system of a formally sovereign periphery, Emmanuel’s channel, so that independence changes the drain’s plumbing without switching it off. A reader who sees the genealogy also sees the school’s vulnerability in long perspective: the transfer idea had been contested for a century on the same grounds the econometricians would revive, the counterfactual problem of what the periphery’s resources would have earned in some other arrangement, which no ledger of actual flows can by itself answer.

Precursors in the region’s own historiography

Latin America, too, had reached the dependency intuition by its own road before Frank codified it, and the school’s official genealogy, which begins with CEPAL, understates the debt. The Brazilian historian Caio Prado Júnior, a communist writing in the 1930s and 1940s, argued in his Formation of Contemporary Brazil that the colony had been organised from the first as an enterprise for supplying European commerce, its slavery, its monoculture and its settlement pattern all intelligible only as functions of that external meaning, and that the inheritance structured the republic that followed. Here, a generation before the metropolis-satellite language, is the core Frank thesis in Portuguese: the continent’s economy was born facing outward, and its internal deformities are the sediment of that orientation. In Argentina, Sergio Bagú’s Economy of Colonial Society, published in 1949, dissected the colonial order and concluded that its regime was a colonial capitalism, neither feudal nor a mere extension of Spain’s, built on coerced labour harnessed to world markets, a formulation that anticipated the Laclau-Frank exchange by two decades and sided, in advance, against the feudal diagnosis. Within the communist parties these historians were heterodox voices; the official line needed its feudal remnants. The dependency school of the 1960s can thus be read as the moment when the region’s own revisionist historiography, armed with CEPAL’s statistics and Baran’s categories, broke the parties’ theoretical monopoly and made the colonial-capitalist reading the common sense of a generation. Frank’s originality lay less in inventing the reading than in systematising it into a portable model and carrying it, with a pamphleteer’s genius for the exportable formula, into the universities of the North.

Stavenhagen’s seven theses: the manifesto before the model

The region’s revisionism found its pamphlet form in 1965, in a short text that cleared the ground Frank’s model would occupy. Rodolfo Stavenhagen, a Mexican anthropologist who had worked among the indigenous communities of his own country, published a newspaper essay dismantling, one by one, seven theses he identified as the common sense of Latin American social science: that Latin American societies are dual, a modern sector beside a feudal one; that development will come by diffusing capital and technique into the feudal sector; that a progressive national bourgeoisie exists and leads; that the middle classes are the natural agents of development; that the region’s problem is the integration of marginal populations into national society; and their companions. Each thesis, he argued, inverted the reality: the so-called archaic sector is the product of the modern one, the national bourgeoisie is bound to the metropolitan system it supposedly opposes, and the marginal are not outside the society but produced by its manner of including them. The essay’s importance for this article is diagnostic. It shows the dependency sensibility fully formed, in a newspaper, from fieldwork, before its famous models existed: the refusal of dualism, the internalisation of the external, the suspicion of the national bourgeoisie, all present in a Mexican anthropologist’s polemic. Stavenhagen also carried the argument into the countryside he knew, where the ladino towns exploiting the indigenous hinterland of Chiapas reproduced, inside one Mexican state, the metropolis and satellite relations Frank was mapping at continental scale, an early statement of the internal colonialism his colleague Pablo González Casanova was naming in the same years for the relation between Mexico’s mestizo centres and its indigenous regions.

Frank and the development of underdevelopment

The strand that gave dependency theory its fame, its slogans and most of its enemies was built by Andre Gunder Frank. Frank was a German-born, American-trained economist who arrived in Latin America in the early 1960s, taught in Brazil and Chile, and wrote with the urgency of a man who believed the seminar room and the struggle outside it were the same room. His essay of 1966, The Development of Underdevelopment, compressed the argument into a title that became the tradition’s most quoted phrase, and his book Capitalism and Underdevelopment in Latin America, published the following year, applied it to the histories of Chile and Brazil. Frank’s claim radicalised Baran’s coupling into a law of the system: the development of the metropolitan centres and the underdevelopment of the satellites are two faces of one process, the expansion of world capitalism, and they have been generating each other since the satellites were first incorporated.

The mechanism Frank proposed was a chain. The world economy, he wrote, is organised as a constellation of metropolises and satellites, each metropolis appropriating the economic surplus of its satellites and appropriating it upward. The chain runs unbroken from the peasant village through the provincial town, the national capital, the regional metropolis, to the world centres in Europe and North America. At every link a local metropolis faces both ways: satellite to the level above, extracting terms set by it; metropolis to the level below, extracting in its turn. The structure explains, in Frank’s hands, a paradox that runs through Latin American history: the regions most tightly bound to the metropolis in the past, the old mining zones and sugar plantations of the colonial centuries, are the most underdeveloped regions of the present, because their surpluses were siphoned away most completely and their structures deformed most thoroughly around the siphoning. Conversely, the episodes of autonomous industrial growth in the periphery, Argentina and Brazil during the world wars and the Depression, occurred precisely when the chain slackened, when war or crisis weakened the metropolis’s grip and the satellites were briefly left to their own resources. History, read this way, convicts integration itself.

How do metropolis and satellite chains work?

In Frank’s model the world economy is a single chain of metropolises and satellites. Each metropolis appropriates surplus from its satellites, which are themselves metropolises to satellites further down: the world centre draws from national capitals, which draw from provincial towns, which draw from the countryside. Development and underdevelopment are generated together at every link.

Frank’s second claim was the historical one, and politically it was the explosive one. Latin America, he insisted, has been capitalist since the conquest. The mines of Potosí, the sugar plantations of the Brazilian northeast, the haciendas that grew up supplying them: none of these were feudal institutions in any analytically serious sense, because all were created by and for a world market, produced commodities for sale, and were reorganised whenever market conditions changed. Scholars who described the continent as feudal, or as dual, a modern coastal sector alongside a feudal interior, were mistaking the effects of capitalist underdevelopment for the survivals of an earlier mode of production. The so-called feudal landlord was a capitalist landlord whose backward techniques reflected his market position: labour was cheap and land abundant, so extensive methods paid. To call his estate feudal was to misunderstand both feudalism and the estate.

Why did the feudalism question matter politically?

Because strategy followed from the diagnosis. If Latin America remained feudal, then a bourgeois revolution came first and alliances with a progressive national capitalist class made sense. If the continent had been capitalist since the conquest, those allies were part of the structure producing underdevelopment, and the alliance strategy rested on a misreading of five centuries.

The conclusion Frank drew was uncompromising, and it is the conclusion that defines the strong version of dependency theory. Since the satellite’s underdevelopment is generated by its participation in the system, and since the local bourgeoisie is a link in the chain rather than a victim of it, no development worthy of the name is possible within the relation. The satellites cannot develop by integrating more efficiently, by exporting more, by educating their elites at the centre’s universities, because every such remedy strengthens the mechanism. The only exit is rupture: breaking the chain, which in the political language of the 1960s meant socialist revolution, and in the economic logic of the argument meant withdrawing from the circuits through which the surplus drains. Frank wrote as the Cuban revolution’s prestige stood high and as guerrilla movements multiplied across the continent, and his theory gave that moment a systematic warrant. It also handed his critics a clean target, for a claim that development under dependency is impossible can be tested against any dependent economy that plainly industrialises, and the 1970s were already preparing the test.

The impossibility split

Here the article reaches the distinction on which everything else in it depends, and it deserves a name of its own: the impossibility split. Dependency theory’s reputation, in lecture halls and in the polemics of its opponents, rests on the claim that development under dependency cannot occur, that integration into the world economy is a trap with no internal exit. That claim is real, argued, and historically important; Frank held it and built the tradition’s most famous work on it. But the theory’s most politically consequential exponents held the opposite. The Brazilian sociologists whose study of 1969 became the school’s most durable work argued that dependent economies do develop, unevenly, exclusionarily, under foreign tutelage, but develop nonetheless, and one of those authors went on to govern the largest dependent economy in the hemisphere. The split between these two positions is not a nuance within a shared doctrine. It is a disagreement about the central question, running through the tradition from its first decade, and it rearranges every other judgement a reader makes. Criticisms that demolish the strong version glance off the weaker one; evidence paraded against the whole theory often turns out, on inspection, to be evidence one dependency theorist had already used against another.

Did dependency theory say development was impossible?

One strand did. In its account, underdevelopment is manufactured by the same process that develops the centre, so no escape exists inside the system. A second strand, set out in a major 1969 study, argued that dependence bends development into distorted shapes without preventing growth altogether. The strands are persistently run together.

The split also explains a curious feature of the secondary literature: dependency theory is the rare school that is routinely refuted by quoting its own members. The East Asian industrialisations are deployed against Frank by writers who appear not to notice that Cardoso’s framework accommodates growth of exactly that kind, indeed that Cardoso constructed his position partly in response to Brazilian growth that Frank’s categories strained to describe. Conversely, defenders of the school under pressure retreat from Frank to Cardoso without announcing the move, so that the theory seems to absorb every fact: stagnation confirms it, growth confirms it, and a theory confirmed by everything is suspected of explaining nothing. That suspicion, the charge of unfalsifiability through internal variety, is the strongest objection to the tradition as a whole, and it can only be answered by holding the versions apart and grading each on its own claims, which is the discipline this article follows. Strong version: development under dependency is impossible; its evidence is the historical record of incorporation and drainage. Weak version: development under dependency is possible but conditioned, its forms set by the alliance of foreign capital, the state and local classes; its evidence is the industrialisation of the larger Latin American economies in the postwar decades. Technical version: whatever growth occurs, trade between high-wage and low-wage economies transfers value systematically; its evidence is a mechanism, not a chronicle. Three claims, three different fates when tested.

The three positions grid

The three positions grid below compresses the comparison into the form in which it can be taught, memorised and argued with. Each row is a position stated in its proponents’ terms; the final column registers the objection that has done it the most damage, which in two cases of three was first pressed by other Marxists.

Position	Central claim	Whether development is possible	Mechanism identified	Political conclusion	Principal objection
The development of underdevelopment, associated with Frank	Underdevelopment is manufactured by the same expansion that develops the centre	No, not while the satellite remains inside the chain	Metropolis and satellite links drain surplus upward from village to world centre	Rupture with the system; no progressive national bourgeoisie exists to ally with	Defines capitalism by market exchange, so it cannot explain why incorporated regions diverged
Associated dependent development, associated with Cardoso and Faletto	Dependence is a concrete situation that conditions growth rather than forbidding it	Yes, in distorted, exclusionary and externally steered forms	An alliance of foreign firms, the state and local capital industrialises from within the relation	Politics proceeds through historical situations and shifting alliances, not iron laws	A framework compatible with growth and stagnation alike risks explaining everything and nothing
Unequal exchange, associated with Emmanuel	Trade between high-wage and low-wage economies itself transfers value	The argument ranks transfer, not growth, as the primary injury	Profit rates converge across borders while wages do not, so prices repay embodied labour unequally	Wage gains in the centre rest partly on cheapened imports, dividing workers internationally	Wages are treated as set independently of productivity and of relations in production

Read down the third column and the impossibility split is visible at a glance: two of the three positions say development can happen, and the tradition’s fame rests disproportionately on the one that says it cannot. Read across any row and the price of each position is visible too, since every mechanism strong enough to do explanatory work attracts an objection aimed at exactly its strongest joint. The sections that follow take the second and third positions in turn, at the length their influence warrants, and then turn to the strand that tried to repair the first position from inside the Marxist tradition: the theory of super-exploitation, which relocated the whole argument from the sphere of exchange to the sphere of production, where its ablest critics had been demanding it go.

How the argument crossed languages

The school’s geography includes a geography of publication, and a short account of it explains why the Anglophone world received the tradition in the distorted order it did. The founding texts were written in Spanish and Portuguese for Latin American readers, in journals like El Trimestre Económico and Desarrollo Económico and through the region’s university presses, and they crossed into English late, selectively and often through movement publishers rather than academic ones. Monthly Review Press in New York, the house Baran and Sweezy built, was the principal bridge: Frank’s books appeared there, and the Monthly Review itself had carried the development of underdevelopment essay in 1966, which is why the Anglophone debate began from Frank’s strongest formulations rather than from the structuralist groundwork or the Brazilian corrections, which arrived in English a decade behind, Cardoso and Faletto not until 1979, Marini effectively a generation behind. Penguin’s education list put a reader of Latin American pieces into paperback for British students, and New Left Review served as the arena where the school’s fate in Marxist theory was decided, carrying Laclau’s critique, Brenner’s verdict and the replies. The sequence of translation was thus a selection mechanism on the theory itself. English speakers met the law before the situations, the slogan before the statistics, and spent years refuting or defending a school whose subtler half was still in transit. Any modern assessment written in English should discount its own verdict accordingly, a caution this article has tried to practise by grading the positions separately rather than the reputation jointly.

The institutional geography: faculties, journals and the exile circuit

Ideas travel in institutions or they do not travel at all, and the dependency tradition’s infrastructure deserves a paragraph of its own. Its laboratories were the region’s new social science faculties: the Latin American Faculty of Social Sciences, founded in Santiago in 1957 under UNESCO’s auspices, which trained a generation of the continent’s sociologists in a single building; the Council of Latin American Social Sciences, founded in 1967 to link them; the research centres attached to the universities of Chile, Mexico and São Paulo; and a journal circuit, from the commission’s own review to the Mexican and Chilean quarterlies, in which a Santiago argument could be common property in Buenos Aires within the year. The coups of the 1970s smashed the infrastructure and, by smashing it, continentalised it: the exile circuit that carried Cardoso to Cambridge and the exiles of Santiago and Buenos Aires to Mexico City, where a publishing house founded by exiles issued the debates of a continent from a single address, created the first genuinely Latin American social science, its members having learned each other’s countries the hard way. When the school declined, it declined institutionally as much as intellectually: the faculties depoliticised by dictatorship, the journals thinned by exile’s end and by the market reforms that starved the public universities, and the tradition’s afterlife in textbooks and in the commission’s neo-structuralism is the residue of an ecosystem, not only of an argument. A reader who wants to know why these particular books were written should picture the buildings as well as the doctrines: a commission’s library in Santiago, a centre’s seminar room in Mexico City, a borrowed office in Cambridge, and the letters flying between them.

Dos Santos and the forms of dependence

Between Frank’s law and Cardoso’s situations stands the work of Theotonio dos Santos, the Brazilian economist who gave the school its most quoted definition and its most careful periodisation. Dos Santos, exiled like his contemporaries, worked in Chile at the Centre for Socio-Economic Studies of the University of Chile, the CESO, which in the years before the Chilean coup of 1973 formed with the Santiago commission the densest concentration of dependency thinking on earth. His definition, formulated in an essay of 1970 that Anglophone journals carried to the North, describes dependence as a conditioning situation in which the economy of one group of countries is conditioned by the development and expansion of another economy to which it is subjected, the dependent economy able to expand only as a reflection of the dominant one’s expansion, with positive and negative effects on its immediate development. The definition’s care is its point: dependence conditions rather than mechanically determines, and the conditioning works through the internal structures of the dependent formation, which is why dos Santos belongs to neither pole of the impossibility split but supplies the vocabulary both poles borrowed.

His periodisation divided dependence into historical forms, and the scheme remains the fastest way to see that dependency was never imagined as one frozen relation. Colonial dependence rested on the export of precious metals and agricultural goods under a commercial monopoly enforced by the colonial power, with labour coerced in mines and plantations. Financial-industrial dependence, consolidated at the end of the nineteenth century, worked through the great export economies financed by metropolitan loans and serviced by raw material exports, the era of the railway bond and the nitrate field. Technological-industrial dependence, the postwar form, rests on the multinational corporation investing inside the dependent economy itself, producing for its internal market while commanding its technology, its finance and its strategic decisions from abroad. Each form generates its characteristic class structure and its characteristic politics, and confusing them produces the errors the school’s critics most enjoyed exposing: arguments framed for the era of the enclave applied to the era of the subsidiary. Dos Santos also pressed, against Frank, an insistence that dependence must be studied through the dependent economy’s own accumulation process and class struggles, a methodological demand that pulled the whole school toward the production relations its critics accused it of ignoring. That his name is less famous in English than Frank’s or Cardoso’s is an accident of translation and academic fashion; within Latin America his formulation was the working definition of a generation.

Cardoso, Faletto and associated dependent development

The second position was born as an internal correction, written in Santiago in the middle 1960s by two sociologists close enough to the dependency debates to share their premises and reject their most famous conclusion. Fernando Henrique Cardoso, a Brazilian sociologist exiled by his country’s military coup, and Enzo Faletto, a Chilean historian, composed their study in the orbit of the same commission that had housed Prebisch, and published it in Spanish in 1969 as Dependencia y desarrollo en América Latina. Its English translation, Dependency and Development in Latin America, appeared a decade later and carried the argument into the North American academy, where it was promptly misread as a softer version of Frank. It is not a softer version of anything. It is a different kind of theory, built on a methodological refusal: Cardoso and Faletto declined to treat dependency as a formal relation between abstract economies, deducible once and for all, and insisted instead that dependence exists only in concrete historical situations, where external structures are internalised through specific alliances of classes and states. There is no theory of dependency in the abstract, they argued, there are only analyses of concrete situations of dependency, each with its own political possibilities.

What does associated dependent development mean?

It names growth that is real but bent. In Cardoso and Faletto’s account, foreign capital, the state and local capital ally to industrialise a dependent economy from inside the dependency relation: factories, wages and cities grow, while the pattern of production answers to external markets and imported technology, and the benefits settle unevenly across classes and regions.

The empirical spur was Brazil. The strong version implied that a satellite economy could not industrialise within the world system, yet Brazil, the largest satellite of them all, was visibly industrialising: steel at Volta Redonda, automobiles around São Paulo, a consumer durables sector, a swelling industrial working class. Cardoso’s response was not to abandon the dependency framework but to relocate its object. The question was never whether the periphery could grow; it was what kind of growth dependency produced, who carried it, and who was left outside it. His answer named the configuration the associated-dependent development of the grid: an association among multinational firms bringing technology and finance, state enterprises commanding infrastructure and basic inputs, and local private capital occupying the interstices, jointly industrialising the economy from inside the dependent relation. The result is growth with a characteristic deformation. Production is geared to the consumption patterns of a privileged minority whose tastes mirror the centre’s, because that is where purchasing power sits; the mass of the population, whose labour is cheap and whose votes the authoritarian state of the period suppressed, figures in the model as a cost to be minimised rather than a market to be cultivated. Development and exclusion are not successive stages. They are the same process read from different floors of the building.

Two features of this position separate it from Frank’s as sharply as any theoretical difference can. The first is the role of the state and of politics. For Frank, the chain determines outcomes with something like mechanical necessity, and the satellite bourgeoisie is a transmission belt. For Cardoso and Faletto, structures constrain but do not write the script: the same dependent situation can house populist industrialisation, bureaucratic-authoritarian industrialisation, or stagnation, depending on the alliances struck and the battles won. Dependency analysis therefore has to be historical and political all the way down, and its predictions are wagers, not deductions. The second feature is the verdict on the local bourgeoisie. Where Frank dissolved it into the chain, Cardoso restored it as an actor, compromised and associated with foreign capital, yet real, capable of bargaining, investing, and occasionally defying its senior partners. The national bourgeoisie whose existence the strong version denied re-enters the analysis, not as the hero of a two-stage drama but as a partner in a three-sided firm whose junior share is still a share.

It is a matter of plain biographical record that Cardoso’s later career carried him far beyond the seminar room: the exiled sociologist returned to Brazil, entered electoral politics as the military regime opened, served as a senator and as finance minister, and held the presidency of Brazil from 1995 to 2002. This article records the fact because the brief of a reference work is completeness about its subjects, and because the trajectory matters to the theory’s history: the most influential statement that dependent development was possible came to be associated with a man who later governed the hemisphere’s largest dependent economy. What his administration did with that inheritance is a question for the political history of Brazil, and it is deliberately not assessed here; the theory must be judged by the arguments of 1969, which say what they say regardless of the offices their author later held.

The strongest objection to the second position is the mirror image of the objection to the first. Frank’s version is falsifiable and was, in important respects, falsified; Cardoso’s version fits so many outcomes that critics have asked what could possibly count against it. If growth confirms dependency and stagnation confirms it too, if democracies and dictatorships both express it, the concept risks becoming a redescription of Latin American reality in dependency vocabulary rather than an explanation of it. Defenders reply that this mistakes a framework for a law: the approach specifies mechanisms, the internalisation of external interests through class alliances, and mechanisms can be looked for and found wanting in particular cases. The exchange between these charges, unfalsifiability on one side and mechanical determinism on the other, is the internal tension the whole tradition lived inside, and no honest account of it should pretend the tension was resolved.

The Brazilian miracle as the school’s courtroom

Between 1968 and 1973 the Brazilian economy grew at rates among the fastest ever recorded, under a military regime of exceptional brutality, with foreign firms pouring into the industrial Southeast and income concentration rising sharply enough to be documented by the regime’s own statisticians. The episode, advertised by the government as a miracle, became the dependency school’s courtroom, because every position in the family had to plead before the same facts. For the strong version the miracle was an embarrassment to be explained away: growth of that speed, in the largest satellite, flatly contradicted the impossibility claim, and Frank’s treatment of it, as a fever of the export enclave destined to collapse, sounded increasingly like a refusal of the evidence. For Cardoso the miracle was the framework’s vindication in real time: here was associated dependent development exactly as specified, the triple alliance forming before the analysts’ eyes, growth real and deformation real, the industrial working class expanding while its share of the product shrank, an outcome his categories had described in advance and Frank’s had forbidden. The Brazilian economist Paul Singer, whose name the reader should distinguish from Hans Singer of the terms of trade thesis, supplied the courtroom’s finest testimony, dissecting the miracle’s demand structure: an economy growing on the consumption of its richest decile, on state investment and on exports, with the wage squeeze not an accident of the model but its flywheel. When the miracle faltered after 1973 on the oil shock and the debt it had accumulated, each side claimed the sequel as well, stagnation for the strong version, the costs of association for the weak, and the episode closed with the school’s internal verdict already delivered by its own members’ practice: the living research programme was the one that had predicted growth and argued about its form.

The triple alliance tested: Evans, Hymer and the multinational question

The second position received, a decade after its statement, the empirical monument it needed, and the monument is worth pausing over because it shows what dependency analysis looked like when practised as fieldwork rather than proclaimed as law. Peter Evans, an American sociologist, took Cardoso’s configuration to the Brazilian record in Dependent Development, published in 1979, and documented the triple alliance in operation: multinational corporations in the dynamic sectors, state enterprises holding petrochemicals, steel and infrastructure, local private capital in the supplying and finishing trades, the three bound by joint ventures, shared directorates and a common interest in excluding the popular sectors from the table. Evans’s Brazil industrialised exactly as the framework said a dependent economy could, through association, and his study gave the weak version of dependency theory something the strong version never acquired, a body of evidence assembled by a scholar testing mechanisms case by case. It also quietly transformed the research agenda: after Evans, the interesting questions were about the terms of association, the bargaining between states and firms, the sectors where local capital held its own, and the dependency debate in the North American academy became a literature on alliances and bargaining, nearer to Cardoso’s intentions than to Frank’s.

Behind Evans stands an earlier and stranger figure whose work connects the dependency school to the theory of the multinational firm. Stephen Hymer, a Canadian economist at Yale whose doctoral dissertation of 1960 founded the modern study of foreign direct investment, wrote in his last years as an avowed Marxist and sketched, in an essay published after his early death, a law of uneven development under the multinational corporation: the firm’s internal hierarchy, headquarters strategy concentrated in a few world cities, routine production dispersed to the periphery, reproduces within its own organisational chart the centre-periphery structure of the world economy, so that the geography of corporate command becomes a map of global inequality with the head offices as its capitals. Hymer’s death in a road accident in 1974 cut the project short, and his Marxist writings were collected only posthumously, but the line from his seminar at Yale to the dependency debates is direct: Evans was among his students. The episode matters for the school’s history because it shows dependency thinking fertilising a neighbouring field, the economics of the firm, at the moment its home discipline was beginning to dismiss it.

The school’s most famous member, finally, lived to complain about his own reception, and the complaint deserves recording because it diagnoses how theories decay in transit. In an essay of the later 1970s on the consumption of dependency theory in the United States, Cardoso watched North American scholars convert his historical method into a checklist: dependency operationalised as a variable, foreign penetration measured by an index, hypotheses tested against cross-national statistics, his situations of dependency flattened into a theory of external determination with regression coefficients. His protest was methodological before it was political. A concept forged to insist that external structures exist only as internalised through concrete class alliances had been consumed as its opposite, a doctrine in which the external determines and the internal merely registers. Readers of the Anglophone literature should hold the warning in mind: much of what North American journals tested and partly confirmed, partly refuted, under the name of dependency theory was a formalisation its best-known living author had publicly disowned, and the refutations landed on a target he had helped to vacate.

The bargaining school: Vernon and the obsolescing bargain

Running parallel to the dependency literature, and sharing more of its questions than either side enjoyed admitting, stood a North American political economy of the multinational firm whose founder supplied the era’s most portable mechanism. Raymond Vernon of Harvard, whose Sovereignty at Bay appeared in 1971, proposed that the bargain between a multinational company and a host government obsolesces: before the investment is sunk, the firm holds the power, shopping among hosts for concessions; once the plant is built, the technology installed and the workers trained, the power migrates to the host, which can renegotiate taxes, ownership and local content from strength, and the cycle then repeats as the technology ages and new suitors call. The mechanism is a liberal’s mechanism, indifferent to value theory, and its author drew liberal conclusions about the maturing of host states. Yet the questions are the dependency school’s questions, the shifting balance between foreign capital and the national state, and the bargaining literature’s case studies, of copper in Chile and Zambia, of oil in Venezuela, of bauxite in Jamaica, constitute an archive the dependency analysts themselves mined. The two literatures divided the era’s problem between them: the bargaining school explaining the variance among dependent economies by the skill and leverage of their states, the dependency school asking why the game itself, however skilfully played, left the structure standing. Evans’s Brazil sits precisely at their junction, which is why his book was read with respect in both camps, a rare passport in a quarrelsome decade.

The dependent state: O’Donnell and bureaucratic authoritarianism

Dependency analysis acquired, in the 1970s, a political science of the first rank, when the Argentine scholar Guillermo O’Donnell asked why the most industrialised countries of Latin America, the very ones where dependent development was proceeding fastest, produced the era’s most systematic military dictatorships. His answer, set out in Modernization and Bureaucratic-Authoritarianism in 1973, tied the coups of Brazil in 1964 and Argentina in 1966 to the exhaustion of the easy stage of import substitution. Deepening industrialisation into intermediates and capital goods required, on his account, levels of investment, wage discipline and political exclusion that the populist democracies could not deliver; the bureaucratic-authoritarian state, a regime of armed forces as institutions, technocrats and transnational capital, arose to impose the deepening by terror, normalising the economy for the triple alliance Evans was documenting in the same years. The thesis mattered beyond its cases because it fused the dependency school’s economics to a theory of the state: the form of rule in the dependent economy was derived from the phase and structure of its dependent accumulation, not from the cultural traits modernisation theory blamed. O’Donnell’s later self-criticisms, entered as Argentina’s own bureaucratic authoritarianism collapsed in failure and as the Southern Cone’s transitions complicated the story, are part of the record and part of the school’s honour: its best members treated their frameworks as revisable wagers on history, which is the Cardoso discipline operating in political science. The section of the dependency legacy that political scientists still teach descends from this work more directly than from any of the trade arguments, a reminder that the tradition’s centre of gravity was always the internal articulation of the dependent formation, exactly where its subtlest members said it was.

Emmanuel and the technical argument from trade

The third position abandoned narrative history for mechanism. Arghiri Emmanuel, a Greek-born economist working in Paris, published L’échange inégal in 1969, and its English edition of 1972, Unequal Exchange: A Study of the Imperialism of Trade, carried the argument into the Anglophone debates with a remarkable passenger: a long critical commentary by Charles Bettelheim, printed in the same volume, to which Emmanuel replied. The book is the most technically ambitious work in the dependency family, the only one that attempts to prove its central claim from the categories of Marxian value theory rather than illustrate it from the record, and it begins from a puzzle in the classical theory of international prices. When capital is mobile between countries, competition presses profit rates toward equality across them, since capital deserts low-return employments for high-return ones. When labour is immobile, fenced in by immigration law and language and the sheer cost of movement, wages face no such pressure, and they differ between countries not by percentages but by multiples. Emmanuel asked what happens to prices when these two facts are combined, and his answer was that an entire transfer mechanism hides inside ordinary, competitive, equivalent exchange.

How can an exchange of equivalents transfer value?

Emmanuel’s answer turns on what competition equalises. Capital moves across borders chasing returns, so profit rates converge internationally, while workers cannot move and wages stay far apart. Goods are then priced by costs plus the common profit rate, and a low-wage economy’s exports systematically embody more labour than the imports its earnings command.

The reasoning runs in words as follows, and the brief of this article requires it to run in words: no equations, and a single concrete illustration using generic goods. In Marx’s theory, commodities sell not at their individual values but at prices of production, formed by adding to each industry’s costs the profit due to its capital at the general rate. Extend that formation across borders, with one world rate of profit and many national wage levels, and a systematic skew appears. Wages enter costs. A producer paying very low wages has low costs, but competition does not let him pocket the difference as an extraordinary profit, because the price of his product is set by the common rate on his capital, not by his wage bill. The low wage therefore shows up as a low price. His goods leave his country priced below what their embodied labour would command if his workers were paid at the wage level of his trading partner, and the goods he buys in return are priced by costs swollen with those high wages. Hours of labour embodied in exports thus purchase fewer hours of labour embodied in imports, year after year, through no fraud, no monopoly and no gunboat: the transfer is executed by the price system itself, at the going rate, in exchanges where equivalents change hands by every legal and commercial standard.

Take the illustration in its simplest generic form. Imagine two economies trading woven cloth and machine tools, each advanced the same capital, each setting its capitals to work under the discipline of the same world profit rate. In one economy the weavers are paid a fifth of what the toolmakers abroad earn, because that is the wage its history has settled. The cloth reaches the world market priced from its low costs; the tools arrive priced from their high ones. When the cloth exporter counts what a season of its weaving has bought, it finds that the labour woven into its bales commands, through the prices realised, a markedly smaller quantity of the labour embodied in the tools than an hour-for-hour reckoning would have returned. Nothing has been stolen in any sense a court could name. Yet the ledger of embodied labour, the ledger Marx’s theory teaches its readers to keep, shows a balance paid by the low-wage side, and Emmanuel’s point is that this balance is not an accident of bargaining but a structural consequence of unequal wages under equalised profits. He drew the further, scandalous conclusion himself: the workers of the high-wage economies are among the transfer’s beneficiaries, since cheapened imports raise their real wages, and the solidarity of labour across borders is therefore divided against itself by the mechanism, a conclusion that detonated inside a tradition committed to international working-class unity.

What did Bettelheim object to in Emmanuel’s argument?

Writing in the same volume, Charles Bettelheim objected that wages cannot be treated as an independent lever, since they reflect productivity, class struggle and the value of labour power in each formation. He also charged that explaining transfers by wage gaps alone inverted the theory: exploitation is rooted in production relations, with exchange redistributing what production creates.

Bettelheim’s commentary is the hinge between the Emmanuel debate and the greater storm that broke over the whole dependency family later in the decade, because his objections rehearse, from within a shared Marxism, the charge that the theory of production had been quietly replaced by a theory of circulation. If exploitation happens where surplus labour is extracted, in the mine and the mill, then an economics that begins from prices and wage differentials has started the story in the middle, however elegant its arithmetic. Emmanuel answered with spirit: his critics, he said, confused the question of where value is produced with the question of how it is distributed among nations, and the second question does not vanish because the first has been answered correctly. The exchange between them remains the finest internal document of the tradition, two Marxists of formidable learning disputing, in one binding, whether the world market merely distributes an exploitation accomplished elsewhere or itself constitutes a machinery of transfer. Readers who work through it will understand the neo-Smithian storm before it breaks, because Brenner’s later critique is, in large part, Bettelheim’s production-first objection generalised into a verdict on the entire school.

Shaikh’s rejoinder: unequal exchange under a second examination

Emmanuel’s construction attracted a second internal examiner whose objections differ from Bettelheim’s and cut at a different joint. Anwar Shaikh, a Pakistani-American economist formed in the classical Marxian tradition, subjected the foreign trade doctrines of the Marxist schools to a long audit at the end of the 1970s and concluded that unequal exchange, as Emmanuel built it, rests on a confusion between the regulation of prices and the transfer of value. In Shaikh’s counter-argument, international differences in wages and productivity produce systematic deviations of prices from values, as they do inside any economy, but a deviation is not a tribute: the low-wage country’s capitalists are not thereby donating value to their trading partners, and the real mechanism by which wealthy economies dominate poor ones runs through the production structures themselves, the monopolisation of technique, the capital movements that discipline peripheral states, the direct extraction within foreign-owned production. Shaikh’s audit extended beyond Emmanuel to the whole neo-Ricardian atmosphere in which value was being computed from price systems, and his verdict, that the law of value operates through the turbulence of real competition rather than through elegant price identities, aligned him unexpectedly with Brenner’s production-first demand while rejecting Brenner’s historiography. The episode completes this article’s picture of the third position’s fate: attacked from the production side by Bettelheim, from the value-theoretic side by Shaikh, and from the empiricists by the immeasurability of its magnitudes, unequal exchange survived less as a theorem than as a question, whether the wage hierarchy of the world labour force is itself a mechanism of appropriation, which subsequent generations have continued to pose in new vocabularies. That a refuted theorem keeps regenerating as a live question is a familiar pattern in this tradition’s history, and the reader will meet it again in the afterlife of the whole school.

The plumbing of the drain: transfer pricing and technological dependence

The dependency tradition is sometimes accused of dealing in atmospheres, and its most concrete researchers answered by following the invoices. In the early 1970s the Greek economist Constantine Vaitsos, working with the Andean Pact’s industrial programme, examined the accounts of foreign subsidiaries in the region’s pharmaceutical and other technology-intensive industries and documented a mechanism of transfer with the great merit of being visible in a ledger: transfer pricing. A subsidiary importing intermediates or technology from its own parent pays prices the parent sets, and Vaitsos’s comparisons against world market benchmarks found overpricing on a scale that converted the subsidiary’s apparent accounts into an instrument for moving profits out of the host economy untaxed and unrecorded as remittances, while royalty and technical-assistance contracts levied further charges on the same captive buyer. The finding mattered to the wider debate out of proportion to its sector, because it located the drain inside the ordinary bookkeeping of the world’s most modern firms, in the technological-industrial form of dependence dos Santos had periodised, and it had policy offspring: the Andean Pact’s Decision 24 of 1970, the era’s most ambitious code on foreign capital, with its fade-out requirements and its screening of technology contracts, was drafted in the light of such evidence. The same research programme generalised the point into the concept of technological dependence: the periphery did not merely import machines but imported the capacity to choose machines, its engineering imagination colonised by the suppliers’ catalogues, so that even an honest price for a technology could purchase the wrong development. Sceptics replied that the Andean code deterred investment its members needed and was quietly dismantled by its own signatories within the decade, a fair point about the remedy that leaves the diagnosis, the plumbing, standing in the record.

Amin and accumulation on a world scale

The dependency argument had an African-Egyptian architect whose version rivals Emmanuel’s in ambition and exceeds it in the territory claimed. Samir Amin, trained in Paris, formed by work in the planning ministries of Egypt and Mali, and settled for most of his career in Dakar, published Accumulation on a World Scale in 1970, a two-volume argument that the accumulation of capital had to be theorised at the level of the world system or not at all, national economies being moments of a single process whose laws operate globally. Amin’s centre-periphery scheme shared the structuralists’ geography but hardened its dynamics: formations at the centre are autocentric, their departments of production answering to one another and to a home market constituted by rising wages, while formations at the periphery are extraverted, their productive structures assembled to serve external demand, disarticulated internally, and therefore incapable of reproducing the centre’s dynamic even when their growth rates impress. The adjectives did durable work. Extraversion named, in one word, the condition Marini analysed through super-exploitation and Cardoso through alliance: an economy whose parts face outward, importing their coherence from abroad.

Amin’s version of unequal exchange followed Emmanuel’s while generalising it: within a single world market for commodities and capital, labour is rewarded at wildly different rates for comparable productivity, and the difference constitutes a transfer from periphery to centre embedded in the price structure, which Amin treated as the modern form of the primitive accumulation Marx had located at capitalism’s dawn, now permanent rather than preliminary. From this followed his distinctive strategy, pressed in the debates of the 1970s and after: delinking. Delinking, as Amin defined it against its caricatures, is not autarky and not withdrawal from trade; it is the subordination of external relations to the logic of an autocentric national development, refusing to let the criteria of the world market, profitability at world prices, comparative advantage as given, dictate the internal pattern of accumulation. A delinked economy still trades; it declines to be governed by its trade. The proposal carried the strong version’s instinct for rupture into a subtler register, rupture not with exchange but with the law of value as the world market imposes it, and it kept the dependency tradition’s political question alive after the strong version’s economics had been damaged: not whether the periphery may trade, but on whose criteria its economy is organised.

Marini and the super-exploitation strand

Alongside the three positions of the grid there runs a further argument, underrepresented in English for decades, which many Latin American readers regard as the tradition’s most original contribution. Ruy Mauro Marini, a Brazilian sociologist formed in the same exile circuits as Cardoso and driven abroad by the same coup, set out his account in Dialéctica de la dependencia, published in 1973, after years of teaching and organising in Mexico and Chile. Marini accepted the transfer diagnosis: dependent economies lose value abroad through the mechanisms the unequal exchange literature describes. His question was what happens next, inside the dependent economy, when its capitalists confront the drain. His answer was super-exploitation, and it relocates the entire dependency argument from the market to the workplace, which is why this strand survives the neo-Smithian criticism better than its neighbours.

How does super-exploitation compensate for value transferred abroad?

Marini’s mechanism works inside production. Capital in the dependent economy, squeezed by the value drained away through trade, restores its margin by driving its own workers harder: lengthening the working day, intensifying labour, and cutting wages below the cost of reproducing labour power. The loss suffered in exchange is clawed back from the workforce at home.

Super-exploitation, in Marini’s precise usage, is not a synonym for very low wages or for harsh conditions generally. It names a specific configuration in which labour power is remunerated below its value, below the cost of reproducing the worker at the historically given standard, through three devices that can operate separately or together: extending the working day beyond its normal limits, raising the intensity of labour so that more is expended in the same hours, and expropriating part of the worker’s consumption fund, the wage share that covers the necessities of life, converting it into a fund of accumulation. Ordinary exploitation, in Marx’s scheme, pays labour power at its value and extracts surplus from the hours beyond necessary labour; the worker is reproduced, the system is stable, the extraction is in a grim sense sustainable. Super-exploitation eats the worker. It consumes labour power faster than wages restore it, which is why Marini treated it as a structural feature of dependent capitalism rather than a abuse correctable by inspection: the compensation mechanism is not a policy choice of greedy employers but the dependent economy’s way of settling its external accounts, transferring the cost of the world’s transfer onto the bodies of its own working class.

From this foundation Marini derived two further claims that gave the strand its reach. The first concerns the internal market. A working class paid below reproduction cost cannot buy back the consumer goods a full industrialisation would need to sell, so dependent industry is steered toward luxury goods for the minority who can pay and toward exports, which is a production-side explanation of the same deformation Cardoso described from the side of class alliance; the two accounts, written by rivals in the same exile, converge on the shape of Brazilian industry from opposite directions. The second claim named a phenomenon Marini called sub-imperialism: the largest dependent economy, Brazil, reaching the stage where its own capital exports and its regional projection reproduce, at a smaller scale, the centre’s behaviour toward its neighbours, without ceasing to be dependent on the world centre itself. Dependency, on this account, is not a wall but a ladder, and a country can climb several rungs of it while the structure stands.

It should be recorded, since the brief of this article is completeness, that Marini’s work travelled badly in English through no fault of its quality. The Dialéctica circulated in Spanish and Portuguese, was argued over in the universities and parties of two continents, and reached Anglophone readers whole only long after its moment, a delay that skews the English-language picture of the whole tradition toward Frank and Wallerstein. The strand matters disproportionately to the verdict a careful reader should reach, because it is the one version of dependency theory that grounds its claims squarely in relations of production: in how labour is paid, driven and consumed inside the dependent economy. When the great criticism arrived, charging the school with theorising exchange while neglecting production, Marini’s work stood as the standing counterexample, the dependency theory that had been about production all along.

The French detour: articulation and the lineage subsidy

The super-exploitation mechanism found its most rigorous anthropological form in France, in the Marxist anthropology of the 1970s, and the detour is worth the reader’s time because it closed the tradition’s most obvious theoretical gap: why capital tolerates, and indeed requires, the survival of the pre-capitalist worlds it was supposed to dissolve. Claude Meillassoux, studying the lineage societies of West Africa, argued that imperialist capital draws labour from communities whose own organisation reproduces that labour at no cost to the employer: the worker is raised, fed in sickness and old age, and his family maintained, by the kinship economy, so the wage need cover only the worker’s own subsistence during his working years, a subsidy from the domestic community to capital that Meillassoux, and Pierre-Philippe Rey in his work on the Congo, formalised as the articulation of modes of production, the capitalist mode conserving the lineage mode precisely because it feeds on it. The parallel with the Latin American arguments is exact enough to constitute independent invention of the same theorem: de Janvry’s peasant plot subsidising the seasonal wage, Marini’s consumption fund cut below its value, the Caribbean plantation’s provision grounds, are one mechanism in three vocabularies. The French school gave the theorem its most abstract form and its sternest internal critics, who charged the articulation literature with a functionalism in which everything that survives is explained by capital’s need for it, the same objection the dependency writers heard from Brenner in another register. That the mechanism was discovered independently on three continents suggests it names something real about how the world economy actually assembled its labour forces; that each discovery attracted the same objection suggests the real thing is easier to glimpse than to model without teleology.

Sub-imperialism examined: Brazil beyond its borders

Marini’s most provocative coinage deserves its own examination, because it tested whether dependence could be squared with power. Sub-imperialism named the condition of a dependent economy that has reached the monopoly and finance stage of its own accumulation and therefore projects itself outward, exporting capital, manufactures and political weight into its region, while remaining subordinate to the world centre whose firms and finance share its commanding heights. Marini’s exhibit was the Brazil of the miracle years: Brazilian construction firms and banks expanding into neighbouring economies, Brazilian manufactures displacing the centre’s in regional markets, the Brazilian state underwriting an energy and arms diplomacy from Asunción to Luanda, all resting on the super-exploitation at home that kept its costs low and its internal market narrow. The concept attracted a durable scholarly quarrel. Critics objected that it moralised a normal stage of capitalist development, that every industrialising country projects power regionally, and that Brazil’s external accounts remained those of a debtor, not a creditor power. Defenders answered that the concept’s whole point was the combination, expansion abroad built on dependence at the summit, which the older dichotomy of imperialist and dependent could not house. The quarrel matters for the article’s central discipline because it is the impossibility split resurfacing in geopolitics: where the strong version’s logic required the dependent state to be a transmission belt, the Brazilian state of the 1970s was visibly an actor with projects of its own, and Marini, the tradition’s most uncompromising production-side theorist, bent his categories to admit it rather than deny the evidence, a flexibility from the school’s sternest quarter that its caricaturists rarely mention.

The plantation school of the Caribbean

The Caribbean produced the tradition’s most distinctive regional variant, and its relative neglect in the Anglophone survey literature mirrors the neglect of Marini. Around the University of the West Indies in the 1960s, the economists Lloyd Best and Kari Polanyi Levitt, working between Montreal and Mona, built what they called the plantation economy model, an account of the Caribbean as a set of economies created whole by the plantation: an institution imported entire, capital, labour, crop and market all external in origin, which organised production for export and left behind, at emancipation and again at independence, societies whose every circuit, monetary, fiscal, dietary, ran through the metropole. George Beckford’s Persistent Poverty, published in 1972, gave the model its canonical statement, tracing how the plantation’s legacy organised landholding, starved domestic agriculture, and reproduced poverty as a system property rather than a residue. The Caribbean school matters to the wider tradition in two ways. It supplied the purest case of the development of underdevelopment anywhere, islands whose plantation centuries left them importing food they could grow in sight of the fields; and it modelled dependence at a scale where the mechanisms are visible to the naked eye, which is why its texts remain the most teachable in the whole literature. Its economists also drew the practical conclusion their title implied, arguing in the New World Group’s journals for a regional economics of self-reliance years before delinking had a name.

The demographers of Santiago: CELADE and the population question

One wing of the structuralist ecosystem deserves its sentence in this history because it carried the school’s method into the era’s most technical policy science. The Latin American Demographic Centre, CELADE, founded in Santiago in 1958 within the United Nations system and directed for its formative period by the Panamanian demographer Carmen Miró, trained the region’s population specialists and read the demographic transition structurally: the periphery’s explosive urban growth, its mortality decline imported as medical technique ahead of any transformation in production, its fertility responding to land hunger and the collapse of the peasant economy rather than to the diffusion of modern attitudes. When the international population establishment arrived in the 1960s and 1970s with its targets and its alarm, the Santiago demographers supplied the region’s governments with the counter-argument that echoed through the 1974 Bucharest population conference, where the developing world’s delegates insisted that development, not contraception imposed from outside, was the solvent of high fertility. The episode is dependency analysis at its most characteristic even in the absence of its name: a surface variable, the birth rate, refused as an explanation and traced to the structure beneath it, by technicians whose work was the more effective for being unanswerable on its own statistical ground.

The school’s vocabulary at work: enclave, comprador, internal colonialism

A tradition’s durability can be measured by the words it leaves in other people’s mouths, and the dependency school’s vocabulary colonised the language of development more thoroughly than its doctrines did. Enclave, from the mining and plantation sectors walled off from their host economies, generalised into the standard term for any production complex whose linkages run outward rather than inward. Comprador, borrowed from the China coast’s vocabulary for the native agent of a foreign house, named the local class fraction whose income is a commission on the external relation itself. Internal colonialism, formulated by Pablo González Casanova for the relations between Mexico’s mestizo cities and its indigenous regions, proposed that the metropolis-satellite structure reproduces itself within the dependent nation, a proposal Stavenhagen’s Chiapas fieldwork and the Brazilian Northeast’s history both sustained, and one that quietly transformed the theory’s scale: dependence is not only something nations suffer from other nations but a fractal pattern recurring at every level where a centre organises a hinterland’s surplus for its own consumption. Extraversion, Amin’s adjective for economies whose parts face outward, travelled from his work into African political science generally. Even the school’s failures left vocabulary behind: dualism survived as the name of the error, the picture of two sealed sectors, that the whole tradition existed to refute. A student who learns to hear these words as arguments rather than descriptions has absorbed the school’s method more deeply than one who can recite its conclusions, because each term compresses a claim about structure that can be checked, and several of them are checked daily by writers who have never opened Frank.

Liberation theology: the school’s categories in the pulpit

No account of the tradition’s reach is complete without the institution that carried its analysis furthest beyond the universities. When the Peruvian priest Gustavo Gutiérrez published A Theology of Liberation in 1971, the book that named a movement, its social analysis was the dependency school’s, acknowledged in its pages: the poverty of the continent read not as backwardness but as the underside of a history of incorporation, the poor not as objects of charity but as a class produced by structures the theology named sinful in their organisation. The Latin American bishops’ conference at Medellín in 1968 had already absorbed the vocabulary, speaking of institutionalised violence and of liberation from dependence, and through the base communities of the following decades the school’s grammar of centre, periphery and structural causation reached parishes that had never heard of CEPAL. The episode matters to a history of the theory for two reasons. It demonstrates that dependency analysis functioned, in its home region, as a general social common sense and not only as an economics, the framework through which a generation of priests, teachers and organisers parsed their world. And it imported the theory’s tensions into the church, where the Vatican’s doctrinal authorities of the 1980s pressed exactly the objection the Marxist critics had pressed from the other side, that a social analysis defining history by structural conflict would swallow the faith it claimed to serve. That the same accusation arrived from the Congregation for the Doctrine of the Faith and from New Left Review, in different vocabularies, measures the reach of the categories Gutiérrez borrowed.

The cultural wing: reading Donald Duck in Allende’s Chile

The framework had a cultural front, and its most famous dispatch came from Santiago in the same years as the copper nationalisation. In 1971 the Chilean writer Ariel Dorfman and the Belgian sociologist Armand Mattelart published a short, furious reading of the Disney comics then saturating Latin America’s newsstands, arguing page by page that the duck universe taught its child readers the ideology of dependence: a world of treasure hunts in exotic lands whose natives surrender their gold for trinkets, of wealth acquired by luck and inheritance rather than production, of the metropolitan hero whose appetites define adventure and the peripheral fool whose function is to be robbed with a smile. The book became a continental bestseller, was burned after the coup of 1973, and founded, with the wider Latin American literature on cultural imperialism, a field that outlived its parent: media studies and the political economy of communication still begin where Dorfman and Mattelart began. Its place in this article is as evidence of reach. A theory of the world economy that can generate, in the same city and season, a nationalisation decree and a classic of comic-book criticism is not a doctrine but a civilisation’s way of arguing with itself, and the cultural wing’s central claim, that dependence is reproduced in the imagination before it is reproduced in the accounts, is the one part of the inheritance its successors in cultural studies never thought to return.

The Cuban test: dependency debated by a revolution in office

One country ran the dependency experiment as a government, and its internal debates repay attention because they show the theory’s categories at work under the pressure of actual administration. Cuba after 1959 was the school’s founding Exhibit A of satellite deformation, a monoculture economy whose sugar, land, utilities and refineries answered to the United States, and whose revolution Frank and others read as the chain breaking. Inside the revolutionary government, the question dependency theory could not answer for it quickly emerged: breaking the chain was an event, but what was an economy to do the morning after? The Great Debate of the middle 1960s, conducted in the journals of Havana between Ernesto Guevara’s partisans and the advocates of economic calculus associated with Carlos Rafael Rodríguez, with Charles Bettelheim advising from Paris, disputed whether a transition economy should govern itself by the law of value, enterprise profitability and material incentives, or by budgetary finance, moral incentives and the conscious suppression of commodity relations. The debate is usually filed under socialist economics, but its terms were set by dependency: a sugar island, blockaded, trading its crop to the socialist bloc at politically fixed prices, asking whether the world market’s discipline could be refused without refusing arithmetic itself. Cuba’s subsequent decades, sugar dependence reconstituted with a new metropolis and then exposed by that metropolis’s collapse in 1991, became the standing cautionary exhibit for every version of the argument: delinking from one centre by attaching to another reproduces dependence in a second language, a lesson the school’s subtler members drew explicitly and its slogans never could.

The argument travels: a Chilean laboratory and an African history

Dependency theory was produced in institutions, and two of them deserve attention because they shaped what the theory became. The first is the Santiago of the 1960s and early 1970s, where an accident of exile concentrated the school’s talents in one city: the UN commission with its library and printing shop, the Latin American Faculty of Social Sciences training the region’s postgraduates, the university centre where dos Santos and Marini taught, Brazilian exiles of the 1964 coup arguing with Chilean hosts, and Frank passing through with his manuscripts. When Salvador Allende’s Popular Unity government took office in 1970, the laboratory acquired a state. Dependency economists advised ministries; the copper nationalisation of 1971, aimed at the American companies whose Chilean mines were the textbook enclave, was argued in the school’s own vocabulary of surplus drainage; and the experiment’s destruction in the coup of September 1973 scattered the laboratory across the world, its members fleeing to Mexico, Europe and North America carrying the theory in their luggage. The diaspora globalised the school even as it ended the conditions that had produced it, a pattern the theory itself, with its attention to how structures travel through people, was well built to appreciate.

The argument’s most famous single application came from the other side of the Atlantic. Walter Rodney, a Guyanese historian teaching at the University of Dar es Salaam, published How Europe Underdeveloped Africa in 1972, and the title announced the method: Frank’s development of underdevelopment, rebuilt from African archives. Rodney’s book remains the most readable monument of the tradition, and its argument deserves the space its influence earned. Africa on the eve of the slave trade, he insisted, was not a stagnant continent awaiting history but a field of developing societies with metallurgy, long-distance trade and state formation under way. The slave trade then removed, across four centuries, a vast share of the continent’s young labour, while organising coastal states for capture and warping the politics of the interior around the supply of captives. Colonialism, arriving after, built railways that ran from mine to port and nowhere else, taxed peasants into the money economy on terms that forced them into migrant labour, and suppressed the crafts, the cloth industries above all, that competed with metropolitan manufactures. Underdevelopment in Africa, the book concluded, was not the absence of a European-style past but the presence of a European-made one, a conclusion whose force survived the qualifications later historians entered against particular chapters. Dar es Salaam in those years was itself a dependency seminar, with Tanzanian socialism attempting its own delinking and scholars like Issa Shivji dissecting the class structure the ujamaa state was creating; the school’s African career confirms that its categories travelled because they answered a structurally similar question on a second continent: why incorporation into the world economy had so consistently coincided with impoverishment.

Ujamaa: delinking attempted under another name

Tanzania offered the era’s most deliberate attempt to govern by something like the tradition’s prescription, and its record belongs in the evidence. The Arusha Declaration of 1967, issued under Julius Nyerere’s leadership, proclaimed self-reliance as national doctrine: development would be built on the country’s own resources, foreign capital and aid treated as temptations that bend a nation’s priorities, and the peasantry organised in ujamaa villages, communal settlements meant to modernise agriculture on African socialist principles without dissolving it into a market in land and labour. The experiment’s economists and its critics, Issa Shivji among them, debated in real time whether the new state was escaping dependence or bureaucratising it, and the external shocks of the 1970s, oil, war with Uganda, collapsing sisal and coffee prices, ground the model down until the government of Nyerere’s successor accepted the Fund’s programme in 1986. The Tanzanian case resists both camps’ morals. Against the strong dependency reading, Tanzania’s poverty at the experiment’s end owed as much to its own coercions, the forced villagisation that disrupted the harvests it meant to raise, as to any drain, a reminder that the internal relations the weak version insists on studying include the state’s violence against its own peasantry. Against the liberal reading, the dependence Nyerere diagnosed was no fantasy: a mono-export economy whose prices were set in markets it could not influence had indeed been delivered, by its insertion, to the mercy of the 1970s. Amin, who advised and argued across the continent, treated Tanzania as delinking attempted with inadequate means and excessive hurry; the historians have mostly treated it as the dependency era’s noblest policy failure, and the distinction between those judgements is the distinction between the tradition’s politics and its analysis, which this article has kept separate throughout.

Dependency in office: the Chilean road and the copper question

Between 1970 and 1973 the dependency framework held ministerial office, and the experiment deserves separate treatment from the laboratory section that follows because its economists governed in the theory’s own vocabulary. The government of Salvador Allende took power in the country where the school’s Latin American wing was institutionalised, staffed its economic team from the commission’s orbit, Pedro Vuskovic at the economy ministry the emblematic appointment, and read its programme directly off the diagnosis: the enemy was the structure of dependence, in the mine, the latifundio and the bank. Its signature act, the nationalisation of the great copper mines in 1971, carried the Chilean congress unanimously, and its signature controversy was pure dependency arithmetic: compensation to the American owners, Kennecott and Anaconda, was computed after deducting the excess profits the companies were judged to have extracted since 1955, a deduction that reduced the bill, on the government’s figures, below zero, Naoroji’s ledger enforced as law. The reply from the centre rehearsed the other half of the theory: credits dried up in what the Chilean foreign minister called an invisible blockade, the copper price sagged, and the economy’s internal class war, the truck owners’ strikes, the empty shelves, the parliament and the executive in constitutional combat, destroyed the experiment before its economics could be judged on their own terms, ending in the coup of September 1973. The school’s survivors drew the lesson their framework dictated: a dependent formation’s attempt to break the relation by decree, without first transforming the alliances that reproduce it, invites the relation’s defenders, at home and abroad, to close ranks. Sceptics drew the opposite one, that the blockade’s effectiveness proved how much worse than dependence the attempted cure could be, and the Chilean road remains, like its country for the Argentine question, the case every general statement about the tradition must carry and none can settle.

The counter-laboratory: Santiago’s second experiment

Santiago holds a distinction no other city can claim: it hosted both the dependency school’s headquarters and the laboratory of the doctrine that replaced it. The Chilean economists trained at the University of Chicago under the exchange agreements of the 1950s, the group the press named the Chicago Boys, returned with a complete counter-theory of the periphery’s ills: not dependence but distortion, not the world market but the state that had strangled it, and after the coup of 1973 they were handed the economy their rivals had been analysing. The shock programme of 1975, the privatisation of the nationalised firms, the opening to imports that swept away the import-substitution industries street by street, constituted the dependency framework’s controlled experiment in reverse: here was a dependent economy deliberately reinserted into the world market on the market’s own terms, by economists who had read the school’s diagnosis as a list of errors to be corrected by their opposites. The exiled dependency writers, watching from Mexico City and a dozen other refuges, read the results as their framework predicted its enemies would never read them: deindustrialisation in the exposed sectors, a consumption boom for the upper strata financed by foreign borrowing, the debt accumulating toward the crash of 1982 that wrecked the experiment’s own claims. The liberal economists read the same years as a painful transition whose fruits arrived later. The city itself delivered no verdict, but the symmetry is the historian’s gift: within a few streets, the two great theories of the Latin American economy were each given a state to practise on within a single decade, and the arguments of seminar rooms from Berkeley to Dakar were thereafter arguments about Santiago twice over.

The countryside inside the theory: de Janvry and functional dualism

The dependency school’s treatment of the countryside contains the production-side mechanism that most completely answers, in advance, the objection that the school theorised only exchange. Alain de Janvry, a French economist long resident in California who worked across Latin America, gave the argument its mature form in The Agrarian Question and Reformism in Latin America in 1981. Latin American agriculture, he showed, is not a feudal residue awaiting dissolution, the stage theorists’ diagnosis, nor a simple victim of the export enclave, but a structure functionally articulated with dependent capitalism. The peasant economy, semi-proletarianised, retaining its plots while supplying seasonal wage labour, reproduces the workforce at a cost partly borne by the peasants’ own subsistence production, which allows urban and export capital to pay wages below the full cost of reproducing labour power: the countryside subsidises the factory and the mine with cheap food and cheap workers whose upbringing it finances itself. De Janvry called this functional dualism, deliberately against the older dualism of two sealed sectors: the peasant sector is not separate from capitalism but harnessed to it, its apparent archaism a working component of the accumulation model. The parallel with Marini is exact and was acknowledged on both sides: super-exploitation names the settlement from the side of the wage, functional dualism from the side of the household that makes the low wage survivable. Together they form the tradition’s most durable contribution to agrarian political economy, a demonstration that the dependent economy’s backward forms persist not because modernisation has failed to reach them but because the modern sector’s costs are quietly invoiced to them. Any verdict on the school that omits this literature has graded the pamphlets and skipped the scholarship.

The marginality debate: the school’s urban frontier

The dependent city posed the tradition a problem its categories had to stretch to hold, and the stretching produced one of its most instructive internal debates. The explosive urbanisation of the postwar decades filled the periphery’s capitals with populations the formal economy could not employ: street vendors, casual builders, domestic servants, the inhabitants of the favelas and barriadas that modernisation writing treated as a transitional embarrassment. The Argentine sociologist José Nun proposed that Latin America was generating not a reserve army of labour in Marx’s functional sense, workers held in readiness for capital’s expansions, but a marginal mass, a superfluous population relative to the monopoly stage of dependent accumulation, permanently outside the wage relation and therefore outside the classical drama of exploitation altogether. Aníbal Quijano, the Peruvian sociologist, answered from within the same tradition that the marginalised were no outside: the informal economy’s cheap services and goods subsidised the formal sector’s wage bill, reproducing its labour power at prices the formal economy could not have paid for itself, a functional integration echoing de Janvry’s countryside argument in the city. The North American anthropologist Janice Perlman’s fieldwork in Rio’s favelas, published as The Myth of Marginality in 1976, then entered evidence against the premise both sides shared: her favela residents were employed, aspiring, integrated and exploited, marginal, she argued, chiefly in the eye of a theory that needed them to be. The debate matters beyond its case because it is the school’s method examining itself: three dependency-shaped minds, given the same streets, produced the outside, the subsidy and the myth, and the disagreement was settled, insofar as fieldwork settles anything, by the one who went and lived there. It also marks the point where the tradition’s questions pass into the modern literature on informality, which inherited the favela and dropped the world system, keeping the street and losing the chain.

The gender critique: the household the school could not see

A later generation of critics, working largely from inside the development field the school had helped to create, charged the tradition with a blindness at the centre of its own favourite mechanism. Dependency analysis, the charge ran, counted the surplus extracted from mines, plantations and factories while treating the household, where labour power is daily and generationally reproduced, as a free good of nature; the unpaid labour of women, cooking, cleaning, child-rearing, nursing the broken worker back to the gate, appears nowhere in the chains of drainage, although it subsidises every wage the system pays. The economist Lourdes Benería, whose work on women’s labour and economic crisis in Latin America grew from the same regional soil as the school itself, showed how the adjustment programmes of the debt years were balanced on precisely this invisible subsidy, as households absorbed the cuts through intensified female labour. The critique connects, for a reader following this article’s threads, with the production-side strands the school did produce: Marini’s worker paid below the value of his labour power is reproduced, in fact, by a household whose unpaid work covers the difference, and de Janvry’s peasant plot performs the same office for the seasonal labourer. Feminist political economy took those hints further than their authors had, rebuilding the reproduction of labour power as a theoretical centre rather than a residual. The episode belongs in the record as the strongest internal evidence for the article’s recurring verdict: the tradition’s exchange-side versions invited correctives, and its production-side concepts, extended past their authors’ intentions, kept supplying them.

The Indian echo: the mode of production debate

The tradition’s categories travelled furthest, and were worked hardest, in a countryside its authors never studied. Through the 1970s the pages of India’s Economic and Political Weekly carried a celebrated controversy, the mode of production in Indian agriculture debate, in which the Frank-Laclau question, what does market involvement make of a social formation, was fought over the Indian village with a ferocity and a scholarship that repaid the borrowing. If the Indian peasant sold his crop, hired labour and bought manufactures, was Indian agriculture capitalist, as Frank’s criterion implied? Was it feudal, semi-feudal, colonial, in Utsa Patnaik’s influential formulation a formation dominated by a landlord class whose rents blocked accumulation, or, in Ashok Rudra’s contrary evidence from Punjab’s green revolution farms, already capitalist in its relations whatever its vocabulary? The Indian debate reproduced the Latin American one with uncanny fidelity, down to the political stakes, since each classification implied a revolutionary strategy for the subcontinent’s communist parties, and it added what the original lacked, a dense empirical literature of village studies against which the categories could be broken and reforged. Its inclusion here is a point about how theories live: the dependency framework’s most durable export was not any of its conclusions but its central question, the demand to classify a formation by the relations that organise its surplus rather than by the goods that cross its markets, and that question was as serviceable in Ludhiana as in Lima.

The neo-Smithian criticism

The most damaging criticism the tradition received came from inside Marxism, and it was aimed not at any conclusion but at the definition silently doing the work beneath all of them. In 1977 the American historian Robert Brenner published, in New Left Review, a long essay on the origins of capitalist development whose subtitle named the charge: a critique of neo-Smithian Marxism. Adam Smith had explained development by the extent of the market: widen trade, deepen the division of labour, and growth follows. Brenner’s accusation was that the dependency theorists, and the world-systems analyst Immanuel Wallerstein alongside them, had rebuilt Marxism on Smith’s foundation. In their work, capitalism is defined by production for the market, by trade, by the pursuit of profit through exchange. But for Marx, Brenner insisted, capitalism is a set of relations of production: a historically specific arrangement in which direct producers, separated from the means of production, must sell their labour power to live, and in which surplus is extracted in the wage relation itself. Define the system by exchange and its motor becomes trade; define it by production relations and its motor becomes class. Everything else in the critique unfolds from that fork.

Why does the definition of capitalism decide the whole dispute?

Because the two definitions locate exploitation in different places. Define capitalism by production for the market and exploitation travels through exchange, so chains of trade become the villain. Define it by how surplus labour is extracted from direct producers, and the same trade appears as a surface over class relations that have to be examined country by country.

Brenner’s positive demonstration was comparative, and it was chosen to hurt. Early modern Europe offered regions integrated into the same expanding markets with opposite results: in the west, and supremely in England, market involvement accompanied the dissolution of serfdom, the rise of tenant farmers employing wage labour, and an agricultural productivity revolution; east of the Elbe, the same grain markets accompanied the second serfdom, the tightening of coerced labour on noble estates producing for export. Identical integration, antithetical outcomes. If trade were the motor, the outcomes should rhyme. They diverge, Brenner argued, because the prior class relations diverged: where peasants had won free tenure they could be separated from the land and become the raw material of a labour market; where lords held jurisdictional power they could answer rising prices by chaining their peasants harder. The market amplifies a class structure; it does not substitute for one. Against Frank specifically, the charge was that the satellite mechanism, by treating incorporation as sufficient explanation, could not say why incorporation industrialised one region and enserfed another, and that a theory unable to explain divergence within its own central process has described a pattern, not a mechanism. The reader who wants the theoretical root of this dispute should consult the series treatment of the mode of production, since the quarrel is, at bottom, about what that concept contains.

Brenner had a forerunner inside the Latin American debate itself. Ernesto Laclau, the Argentine theorist, had taken Frank to task in 1971 for exactly the definitional slip: Frank called the colonial economies capitalist because they produced for a market, while the labour relations inside them, slavery, serf-like obligations, debt peonage, were anything but free wage labour. Laclau’s remedy was the concept of a social formation articulating several modes of production, feudal and capitalist elements combined under the dominance of the world market’s logic. The exchange that followed, Laclau’s critique and Frank’s rejoinder, is one of the tradition’s set pieces, and it matters here because it shows the neo-Smithian charge was not an ambush from outside. The school had been warned, by one of its own region’s subtlest minds, that its central category was doing unearned work, six years before Brenner generalised the warning into a verdict.

The dependency replies were several and unequal. Wallerstein answered that the world-system is a single division of labour within which varied labour controls coexist, and that insisting on free wage labour as the criterion repeats the error of defining a whole by one of its parts. Frank, in later work, simply widened his claim until the world system was five thousand years old, a manoeuvre that traded Brenner’s question rather than answering it. The strongest reply, though rarely made in so many words, is Marini’s strand: a dependency theory built on super-exploitation does not define capitalism by exchange at all, since its entire mechanism is the extraction of surplus from workers paid below the value of their labour power, and on that strand Brenner’s dichotomy loses its target. The fair verdict, and the one this article adopts, is the one the brief of the strongest objection prescribes: concede that the objection defeats the exchange-based versions, the Frank of the chains and any reading of Emmanuel that stops at prices, while recording that the production-based strand stands to answer another day. A criticism that kills the famous version and spares the rigorous one has performed a service the school’s defenders were too proud to perform for themselves.

The East Asian challenge at full strength

Every theory has the evidence it would rather not meet, and for dependency theory that evidence sits on the western rim of the Pacific. In the decades after the dependency school’s founding texts appeared, South Korea and Taiwan, economies as peripheral on paper as any in Latin America, poorer than most of them at the start of the 1960s, without natural resources, war-torn within living memory, industrialised at a speed that outran every centre economy of the same period. They did it not by withdrawing from the world market, the strong version’s prescribed exit, but by plunging into it: exporting first textiles and plywood, then steel, ships, semiconductors and cars, until economies that had exported wigs and sugar stood among the world’s industrial powers. If integration into world capitalism manufactures underdevelopment, these economies should not exist. Their existence is the standard empirical challenge to the strong version, and intellectual honesty requires stating it at full strength before any reply is attempted.

How did the East Asian economies industrialise inside the world market?

They combined reforms no Latin American government of the period assembled together. Land reform broke the landlord class and equalised the countryside; the state directed credit to chosen industries, protected the home market while forcing firms to earn their protection by meeting export targets, suppressed consumption, and schooled the workforce at extraordinary rates. Growth was state-led, export-driven and ruthlessly monitored.

The full-strength statement has several layers, and each layer cuts at a different joint of the strong version. Begin with the starting conditions: by the dependency account, these economies carried every mark of the satellite. Korea had been a colony, annexed by Japan, its industry built as an appendage of the coloniser’s war economy and then shattered by partition and war. Taiwan had passed from Japanese rule to a refugee regime that arrived with the defeated mainland government in 1949. Both were primary exporters with dense peasantries and tiny industrial sectors. Neither possessed the internal market, the capital stock or the technological base that modernisation theory listed as prerequisites, and both were inserted into the world economy at its subordinate edge, as recipients of American aid and military protection. Then consider the sequence: land reforms in both economies, carried out under occupation pressure and revolutionary threat from the mainland example, destroyed the landlord class as a political force, the very class whose survival the dependency writers treated as the internal lock on the system. Consider the states that followed: authoritarian, interventionist to a degree Latin America’s liberal economists would have called dirigiste, allocating credit by decree, picking industrial sectors, disciplining firms that failed export targets by withdrawing their licences and loans, holding wages down while pouring resources into technical education. Consider finally the orientation: these states did not seal their economies behind the import-substitution wall and wait; they used protection as a nursery with a clock on it, and they sold to the world market with a ferocity that turned export performance into the regime’s legitimating creed. Within a generation the structure of both economies had been transformed, real wages had risen far beyond anything the super-exploitation model would predict as permanent, and the transformation had occurred through integration, not against it.

Against this record the strong version has no adequate answer, and the replies its defenders offered mostly concede the point while changing the subject. It is true that the East Asian states enjoyed geopolitical rents no Latin American country could command: vast American aid, privileged access to the American market, military spending that pumped demand through local industry, all extended because these were frontline states in a cold war and their success was a showcase the hegemon was willing to subsidise. It is also true that the Japanese colonial period, for all its brutality, left infrastructure and an administrative cadre that muddy any simple satellite portrait. But these are explanations of why the exception occurred, and a theory of impossibility is not entitled to exceptions explained by contingency; contingency is exactly what the impossibility claim ruled out. The honest conclusion is the one the impossibility split predicts: East Asia refutes the strong version, the claim that development under dependency cannot occur, and it does not touch the weaker claim that dependency conditions the form development takes, since the East Asian form, state-brokered, export-disciplined, politically authoritarian for most of the ascent, was itself nobody’s picture of autonomous liberal development. Cardoso’s framework absorbs the cases with room to spare. Frank’s cannot absorb them at all.

A later scholarly wave hardened the challenge into a rival orthodoxy. Studies of the Korean and Taiwanese ascents by economists and political scientists in the 1980s and after documented the state machinery in detail: the deliberation councils, the export targets reviewed monthly at the highest level, the policy loans at negative real interest rates for favoured sectors, the public enterprises in steel and shipbuilding that anchored private suppliers. The World Bank’s own report on the East Asian miracle, published in 1993, acknowledged the interventions while labouring to present them as market-friendly exceptions; less apologetic readers concluded that the region’s success was a triumph of governance over the price mechanism, the opposite moral from the one free-trade doctrine required. For dependency theory the lesson of that literature is double-edged and worth keeping whole. On one side it buries the strong version beyond exhumation. On the other it buries, just as deep, the idea that the periphery’s fate is written by its insertion alone: the Korean state and the Brazilian state faced the same world market and built different machines, and the difference was made by class coalitions, land reform and administrative will at home. That is the conclusion Cardoso and the internalists had been urging since 1969, and it arrived, vindicated, from the far side of the Pacific.

The dependency replies to East Asia, assessed

The replies the school’s surviving members made to the East Asian record, once the first shock passed, are better than the slogans and deserve a fair hearing before the verdict of the previous section is allowed to stand as the article’s last word on the subject. The strongest reply came from the weak version’s own logic. Cardoso’s framework had never predicted uniformity: it predicted that outcomes would be settled by the historical configuration of state, classes and external linkage, and the East Asian configuration, land reform imposed under occupation, a landlord class destroyed by war and reform rather than entrenched by exports, a state built for siege conditions, differed from the Latin American one at exactly the joints the framework marked as decisive. On that reading the Pacific cases are not anomalies but confirmations of the situational method, and the North American political scientist Bruce Cumings, writing on Korea’s political economy, supplied the reading with scholarship: Korea’s development, he argued, cannot be told as a national story at all, since its industrial structure was formed inside a Northeast Asian regional system, first under Japanese imperial direction and then under American hegemony, so that the case exhibits dependence managed by an unusually capable state inside an unusually favourable regional hierarchy. Other dependency writers pressed the point about the hegemon’s subsidy: no Latin American state was offered the American market, American aid and American military demand on the Korean scale, because no Latin American state sat on the frontier of the Cold War’s hottest border. These replies are genuinely responsive to the facts, and a scrupulous reader should grant them their force. What they cannot do is resurrect the strong version, since every one of them proceeds by specifying conditions, and a claim of impossibility that survives by accumulating conditions has already surrendered. The replies are therefore best understood as the school’s centre of gravity completing its migration, from Frank’s law to Cardoso’s situations, under the tuition of the Pacific evidence.

Evans’s second act: embedded autonomy and the Pacific cases

The tradition’s most consequential afterlife inside scholarship runs through the student of Hymer who tested the triple alliance, because Peter Evans spent his later career explaining the very cases that had refuted his teachers. His Embedded Autonomy, published in 1995, compared the industrial transformations of Korea, Brazil and India and asked what kind of state could broker development rather than merely preside over dependence. His answer named a configuration: a meritocratic, internally coherent bureaucracy, Weberian in its insulation, yet densely connected to the industrialists it promoted, embedded in the society it steered. Korea’s economic planning apparatus approximated the type; Brazil’s state, brilliant in pockets, fragmented across its vastness, approximated it partially; India’s, coherent but disconnected from capital, produced coherence without transformation. The book’s place in this article’s story is its genealogy. The questions are Cardoso’s questions, about the alliances through which external structures are internalised and the political configurations that decide what dependence does; the East Asian case, which entered the debate as the dependency school’s executioner, exits Evans’s study as the occasion for its subtlest descendant to explain why one dependent formation’s state could discipline capital while another’s could only host it. Intellectual traditions rarely get to grade their own refutations. The dependency tradition, through Evans, came close: the evidence that killed its strong version became, in its ablest heir’s hands, the proof of its weak version’s method, that everything depends on the historically built state and the alliances it can strike.

The developmental state: the tradition’s grandchild in the policy schools

Evans’s synthesis had a sibling literature whose descent from the dependency debates is equally plain. The developmental state literature, launched into prominence by Chalmers Johnson’s study of Japan’s industrial ministry in 1982 and extended by Amsden’s and Wade’s Pacific studies, asked the question the dependency school’s survivors were asking, under what political conditions does a late developer’s state steer accumulation rather than merely host it, and answered with the bureaucracy’s quality, its alliances and its instruments. In Africa the Malawian economist Thandika Mkandawire carried the argument back onto the dependency school’s home continent, insisting, against both the structural adjustment consensus and the older dependency fatalism, that developmental states were as possible in Africa as in Asia, that the continent’s post-colonial history contained real if defeated attempts, and that the impossibility theorems, whether of dependency or of the Washington doctrines, served chiefly to excuse the analyst from studying the cases. Mkandawire’s formulation closes the circle this article has been drawing since its account of the split. The impossibility claim, in every generation, is the tradition’s temptation and its error; the study of concrete formations, their states, classes and alliances, is its method and its residue. The developmental state literature is not dependency theory, and its practitioners are not obliged to the school’s vocabulary, but the family resemblance is the kind intellectual history exists to record: a grandchild who has changed the surname and kept the eyes.

The pioneer argument: development defended from within Marxism

The East Asian evidence demolished the strong version empirically; a book published in 1980 tried to demolish the whole dependency outlook theoretically, from a position claiming to be more Marxist than the Marxists it attacked. Bill Warren, a British scholar who died before his book appeared, argued in Imperialism: Pioneer of Capitalism that the dependency school had inverted Marx. Marx, Warren reminded his readers, had expected capitalism’s expansion to be brutal and progressive at once, dissolving stagnant social orders, creating industry and a proletariat even in the colonies, preparing the ground for socialism precisely by developing the productive forces. The dependency writers, in Warren’s telling, had replaced this with a moralised picture in which imperialism only blocks, only drains, only deforms, and in doing so had become the theorists of a romantic anti-capitalism whose practical politics was nationalism in socialist dress. The pioneer thesis deserves its hearing at full strength, since it is the pro-development argument in its most considered Marxist form.

What did Warren’s pioneer thesis claim?

Warren argued that imperialism was doing in the postwar periphery what Marx expected capitalism to do: spreading capitalist relations, building industry, and dissolving the precapitalist orders that blocked progress. Manufacturing across the so-called Third World was expanding substantially, he pointed to the statistics of the postwar decades, and the obstacles to development lay in internal structures and state policies, not in integration itself.

Warren’s evidence was the industrial statistics of the postwar decades, which genuinely did show manufacturing output growing across Asia, Africa and Latin America at rates the stagnationist picture could not accommodate, together with the spread of wage labour, urbanisation and capitalist farming that stage theorists had once predicted and dependency writers had ceased to expect. His polemical targets were chosen with care: Frank’s satellites, Baran’s blocked surplus, the whole architecture of externally induced paralysis. Where the dependency school saw a periphery prevented from developing, Warren saw a periphery developing unevenly and painfully, held back where it was held back by the survival of precapitalist relations and by states that protected them. The remedy implied was the opposite of rupture: deeper capitalist development, the completion of the bourgeois transformation the dependency writers had pronounced impossible or worthless.

The debate that followed Warren’s posthumous book was one of the sharpest the British Marxist journals hosted, and its outcome instructs. Critics answered that Warren had confused the growth of capitalist production somewhere in the periphery with development of the periphery as a whole, that enclave industrialisation under foreign ownership could raise output statistics while leaving the structure of dependence intact, and that his figures aggregated the exceptional performers with the stagnant many. Others noted that Marx’s own writings on colonialism were more ambivalent than the pioneer reading allowed, celebrating the dissolution of the old order and cataloguing its horrors in the same pages. Yet Warren’s core empirical point survived the exchange dented but standing: the postwar periphery was industrialising, unevenly, really, and under foreign capital’s heavy presence, and any theory that had staked itself on the impossibility of that outcome had to be revised or retired. The pioneer thesis is best read as the Marxist form of the same verdict East Asia delivered in the indicative mood. Between them they mark the boundary of the dependency tradition’s first life: the strong version died of evidence, the school survived by becoming its weaker, subtler members.

The new international division of labour

In 1977 three German sociologists, Folker Fröbel, Jürgen Heinrichs and Otto Kreye, published a study whose title announced a mutation in the world economy: a new international division of labour was forming, in which manufacturing itself, not only primary production, was being relocated to the periphery. Their evidence was the world market factory: the export-processing enclaves, from the Mexican border programme of maquilas established in 1965 to the free production zones of East and Southeast Asia, where young women assembled electronics, sewed garments and wound components for the centre’s brands under wage regimes a fraction of the home factories’ costs, the finished goods re-entering the centre’s markets as imports from its own firms. The finding cut across the dependency debate at an angle neither side had prepared. Against the strong version it showed industrialisation proceeding apace in the periphery, through integration, in the newest sectors. For the dependency framework it posed the subtler question the weak version existed to ask: what kind of industrialisation is this, whose technology, whose design, whose marketing, whose profits, and what does it deposit in the host economy besides wages? The world market factory became the successor image to the mine and the plantation in the tradition’s iconography, the enclave rebuilt for the age of the container ship and the semiconductor, and the arguments over whether it represented dependence overcome or dependence modernised rehearsed, with new evidence, the school’s founding quarrel. Fröbel and his colleagues, it should be noted, drew no dependency moral themselves: their frame was the centre’s capital, fleeing its own organised working classes toward cheaper and more governable labour, a Brenner-compatible story of class relations driving geography, and the German study’s capacity to be cited by all camps is the surest sign it had located a real joint in the world economy’s skeleton.

The critics of the market’s other flank: Bauer and Lal

The dependency school fought on two fronts, and the second front is too often forgotten because the Marxist critique was louder. From the liberal side, the school’s most relentless opponent was Peter Bauer, a Hungarian-born economist at the London School of Economics who had studied rubber smallholders in Malaya and traders in West Africa and emerged convinced that the development establishment had the causal arrows backwards. Bauer’s Dissent on Development, collected in 1971 from essays written across the postwar decades, argued that foreign trade and investment had historically been the great engines by which poor regions acquired skills, capital and connection: the trading minorities and expatriate firms the dependency writers indicted were, in his fieldwork, the people who had built the marketing networks and taught the techniques. Underdevelopment, where it persisted, he blamed on the policies of governments, and he turned the school’s moral charge around: the picture of the periphery as victim, he argued, was an alibi manufactured by its rulers and flattering to Western guilt, while the actual poor, the traders and farmers of his studies, got on with exchange whenever officials let them. Bauer’s pupil Deepak Lal sharpened the brief a decade later in The Poverty of Development Economics, an assault on what he called the dirigiste dogma shared, in his telling, by structuralists and dependency writers alike: the belief that markets in poor countries fail so comprehensively that planning must replace them. Between Brenner on one flank and Bauer on the other, the school’s exchange-based arguments were squeezed from both directions at once, the Marxist charging that trade was a surface, the liberal that it was a blessing, and both agreeing, against the middle where most dependency writing lived, that the theory of trade could not carry the weight the tradition had loaded onto it.

Dependency theory and world-systems analysis

No confusion in this territory is more persistent, or more understandable, than the merging of dependency theory with the world-systems analysis of Immanuel Wallerstein. The two emerged in the same years, shared heroes and enemies, and eventually shared personnel, since Frank himself migrated into the world-systems camp and spent his late career extending its timescale. Wallerstein’s The Modern World-System, whose first volume appeared in 1974, proposed that the proper unit of analysis is not the national economy at all but the single capitalist world-economy born in the long sixteenth century, divided by an axial division of labour into core, periphery and a third zone his predecessors lacked: the semiperiphery. States, on this account, are not containers of development but instruments and prizes inside one system, strong in the core, weak in the periphery, and the whole structure reproduces itself through unequal exchange between zones producing different kinds of goods with different kinds of labour control: free wage labour characteristically in the core, coerced cash-crop and mining labour in the periphery, sharecropping and tenancy in between.

What does the semiperiphery do in Wallerstein’s scheme?

It is the system’s shock absorber and its political stabiliser. Semiperipheral economies mix core-like and peripheral activities, exploit the periphery beneath them while being exploited from above, and thereby divide the world’s discontents: opposition to the system is fragmented because a substantial zone has something to lose. The category also houses the industries the core sheds as their profitability fades.

The differences from dependency theory are structural, and listing them cleanly is the surest way to keep the schools apart. The unit differs: dependency analysis begins from a national economy in its concrete situation, Wallerstein from the world-economy as the only self-contained unit, nations being its administrative fictions. The mechanism differs in emphasis: the dependency writers foreground alliances, states and the internalisation of external interests, Wallerstein the division of labour between zones and the interstate system that polices it. The middle category differs most visibly: where dependency writing tends toward a binary of centre and periphery, tempered by Cardoso’s situations, world-systems analysis builds the semiperiphery into the architecture, giving Brazil, Mexico or Spain a structural address rather than an anomaly. And the fate of states differs: for Cardoso the state is the terrain on which dependency is negotiated and occasionally bent; for Wallerstein state strength is itself a product of position in the system, so the lever Cardoso wants to pull is, in the rival scheme, part of what needs explaining. Brenner’s critique, it should be noted, refused to distinguish the schools at all, classing both as neo-Smithian for defining the system by its exchanges, a lumping that dependency partisans resented and that the record of Marini’s production-side strand, once again, complicates.

A rival Marxist account of backwardness deserves mention here because it answers the same phenomena with the opposite machinery, and readers meeting the dependency tradition should know the alternative exists. The theory of uneven and combined development, originating with Trotsky and treated in full in the series article on uneven and combined development, holds that backwardness under the whip of external competition confers a privilege: late developers skip stages, import the most advanced techniques whole, and combine archaic and ultramodern forms in unstable amalgams whose contradictions drive revolution. Where Frank reads incorporation as the manufacture of underdevelopment, Trotsky’s heirs read it as the forced manufacture of modernity in deformed and explosive shapes. The two accounts share a world market and little else, and much of the Marxist argument about the periphery in the twentieth century can be organised as a contest between them.

Frank’s last turn: five thousand years and the Asian age

The strong version’s author outlived his school’s decline by decades and spent them enlarging its canvas until the canvas burst. Having migrated into world-systems analysis, Frank concluded that even Wallerstein’s five centuries were a parochialism: the world system, he argued in a series of works culminating in ReOrient in 1998, is some five thousand years old, a single Afro-Eurasian economy of accumulation in which Asia, and China above all, held the centre until about 1800. Europe’s rise, on this telling, was late, brief and financed by the periphery’s own treasure: American silver, mined by the satellites of Frank’s first books, bought Europe its ticket into Asian markets it could not otherwise have penetrated, and the industrial revolution followed from a temporary vacancy at the top of an ancient system rather than from any European essence. The irony is rich enough to support a seminar. The metropolis-satellite model began as an account of how Europe’s expansion underdeveloped the rest; its author ended by arguing that Europe had been, for most of recorded history, the margin of somebody else’s system, and that the Asian economies’ ascent at the end of the twentieth century was less a miracle than a restoration. Critics replied that a system five thousand years old explains everything and nothing, that the specificity of capitalism dissolves in it entirely, Brenner’s objection returning at maximum scale. Sympathetic readers took the narrower lesson, which the whole tradition had been teaching since Prebisch: the hierarchy of centre and periphery is a historical product with dates, not a fact of nature, and what history made in one configuration it can unmake in another. Frank died in 2005, having been, in succession, the dependency school’s prophet, its most refuted member and its most indefatigable reviser, and the obituaries’ difficulty in summarising him is the measure of how much of the era’s argument passed through one restless mind.

The fine print of the founding statistic

Because the terms of trade thesis carries so much of the structure, its evidentiary fine print deserves a hearing in the body of the article rather than a footnote’s exile. Prebisch’s original series was built from British trade returns stretching from the 1870s to the eve of the Second World War, inverted: Britain’s imports, overwhelmingly primary products, stood as a proxy for the periphery’s export prices, and Britain’s exports, overwhelmingly manufactures, for the centre’s. The construction contained an elegant trap that critics identified early. British import prices were recorded inclusive of freight and insurance, export prices free on board; as transport costs fell across the period, as they dramatically did, the inclusive series would register a decline belonging to the freight market rather than to the commodities, and part of the periphery’s apparent loss was the world’s gain in cheaper shipping, misread as a transfer to the centre. Defenders answered that the trap, while real, could not account for the pattern’s size or its breaks, and that alternative series built from the periphery’s own records showed the same long deterioration in the eras that mattered most. The dispute then migrated into the quality problem: a unit of manufactures in 1938 embodied capabilities no 1876 unit possessed, so comparing their prices flatters the centre’s side of the ledger by treating improvement as mere price. Each generation of econometricians inherited the quarrel with better data and sharper methods, unit roots, structural breaks, cointegration, and the quarrel survived them all, which is itself the finding a reference article should report: the thesis is neither the established fact its founders believed nor the artefact its first critics announced, but a durable provocation that forced trade theory to take distribution between nations seriously. The school’s founding statistic thus ends in the same posture as the school, strongest as a question, contestable as an answer.

The founding statistic received a second life in the econometric era that illustrates the pattern perfectly. In 1988 two World Bank economists, Enzo Grilli and Maw Cheng Yang, published a carefully constructed index of non-fuel commodity prices reaching back to 1900, built to answer the Prebisch question with modern data hygiene, and found a statistically significant long-run decline: the founding tendency, measured by the centre’s own technicians, stood. The subsequent literature then proceeded to qualify the result toward dissolution, finding the decline concentrated in particular commodities and sub-periods, sensitive to structural breaks and to the treatment of the war years, with some series better described as stationary around occasional level shifts than as trending downward. The pattern is the one this section described at series level: a real signal, robust enough to survive hostile reconstruction, wrapped in enough construction-dependence that no two honest technicians report the same number. Prebisch’s thesis thus ended the century in the condition of many founding claims in economics, neither established nor refuted but absorbed into the field’s permanent background, cited by commodity exporters’ negotiators and doubted by their counterparts, awaiting the next reconstruction.

The measurers: dependency theory in the journals

A quieter fate than refutation befell the school in the North American social science journals, where during the 1970s and 1980s a generation of researchers tried to find out whether dependency could be measured. The attempt had been invited early: the British economist Sanjaya Lall, in a careful stocktaking of 1975, asked whether dependence was a useful concept in the analysis of foreign direct investment and concluded that the word bundled together too many distinct mechanisms, ownership, technology, market structure, external vulnerability, to test as one thing, a judgement that split the subsequent literature into those who tested the pieces and those who defended the bundle. The testers produced the era’s most cited findings. Volker Bornschier and his collaborators, working with cross-national statistics, reported a pattern subtle enough to please nobody completely: flows of new foreign investment were associated with faster growth in the short run, while the accumulated stock of foreign capital penetration was associated with slower growth and greater inequality over the longer run, as if the medicine and the disease were the same substance at different dosages. Christopher Chase-Dunn, from the world-systems side, found investment dependence associated with slower growth in the periphery across the postwar decades. Critics of the whole enterprise replied that the measures begged the question, that a stock of foreign capital is a strange proxy for a relation Cardoso insisted was historical and political, and that the regressions’ average effects erased exactly the situations the theory existed to distinguish. The measurers had hold of something real, the time-shape of investment’s effects, and the school’s historians had hold of something real too, the irreducibility of situations, and the two literatures mostly passed each other in the night, which is why the empirical verdict on dependency theory in the journals remains, decades on, a hung jury rather than an acquittal or a conviction.

What the numbers can and cannot show

A reference article owes its readers candour about the evidence base, and in this field the evidence base is a battlefield. Begin with the thesis that started everything. The claim that primary commodity prices decline relative to manufactures across the long run has been tested econometrically for generations, and the results refuse to settle into a verdict either camp can quote without embarrassment. One widely used index of commodity prices, constructed in the late twentieth century from the records of the postwar decades, reported a long decline punctuated by collapses; critics answered that the index’s basket, its endpoints and its treatment of quality change in manufactures, where a machine of 1970 and a machine of 1930 share a name and nothing else, build the conclusion into the arithmetic. Time-series specialists then disputed whether the series trends downward at all or merely breaks at a few catastrophic dates and otherwise wanders. The honest summary is the contested-figures protocol this article follows throughout: the terms of trade of commodity exporters deteriorated sharply in particular eras, especially the Depression and the 1980s, and whether a smooth secular law lies beneath the episodes remains a live econometric quarrel, not a fact available for citation in either direction.

Why do estimates of value transfer disagree so widely?

Because the answer is manufactured by the assumptions. Measuring a transfer of embodied labour requires choosing whose wage counts as the benchmark, how to convert currencies at something other than market rates, which sectors to include, and how to treat productivity gaps between economies. Vary those choices in defensible ways and the estimated transfer swings from modest to enormous. The disagreement is methodological before it is political.

The same discipline applies to the modern literature that attempts to quantify unequal exchange directly. A family of estimates built from global trade and input-output data, produced by economic anthropologists and political economists in the world-systems lineage, concludes that the net appropriation of embodied labour by the high-income economies from the rest is very large, large enough, its authors argue, to constitute a central mechanism of global inequality rather than a curiosity at its margin. Other economists dispute the wage data, the price conversions and the counterfactual at the heart of the method, which asks what goods would cost if labour were paid uniformly across the world, a benchmark whose meaning is itself contested between those who see it as the theory’s natural measuring rod and those who see it as an assumption that delivers its conclusion. This article takes no side in that arithmetic, because the arithmetic cannot be settled here and because the protocol of the series forbids presenting contested estimates as findings. What can be said without controversy is narrower and still significant: the questions the dependency tradition posed about the terms on which labour is rewarded across borders have migrated from pamphlets into datasets, and the datasets have not made the questions go away.

Bandung’s children: the political atmosphere of the school’s birth

The theory was written inside a political weather system larger than itself, and no reader should finish this article without it. In 1952 the French demographer Alfred Sauvy coined the term Third World for the nations that, like the Third Estate of the old regime, were counted as nothing and aspired to be something. Three years later the leaders of twenty-nine Asian and African states met at Bandung in Indonesia, Nehru, Zhou Enlai, Nasser, Sukarno among them, and proclaimed a collective refusal of the Cold War’s two camps, a refusal institutionalised in the Non-Aligned Movement founded at Belgrade in 1961. The dependency writers were this generation’s economists: the school’s characteristic premise, that the periphery’s nations share a structural situation beneath their enormous differences, is Bandung’s premise translated into political economy, and its characteristic hope, that the periphery could act collectively on its situation, is Bandung’s hope, carried into the trade negotiations of UNCTAD and the commodity diplomacy of the 1970s. The translation changed the premise in one decisive respect. The Bandung leaders spoke of nations reclaiming sovereignty; the dependency writers, formed by Marxism, asked which classes inside each nation the sovereignty would serve, and their suspicion of the national bourgeoisie is the sound of the Bandung generation arguing with itself. When the collective project collapsed in the debt crisis, the theory built in its image collapsed with it, a dependence, the historians cannot resist noting, that the theory itself would have predicted: ideas, too, are conditioned by the structures that carry them.

The diplomatic hour: commodity power and the new international economic order

The dependency school’s ideas had one season in which they were not merely read but wielded, and the season deserves its place in the record because it tested what the theory’s categories could do as instruments of statecraft. The vehicle was the United Nations Conference on Trade and Development, created in 1964 with Prebisch as its first secretary-general, where the developing countries organised themselves as the Group of 77 and translated structuralism into a negotiating programme: stabilise commodity prices through international agreements, win preferential access for peripheral manufactures, expand aid toward agreed targets. A decade later, emboldened by the oil exporters’ demonstration in 1973 that a commodity cartel could move the world price of the centre’s most vital import, the coalition carried the programme to its high-water mark in the demand for a New International Economic Order, adopted as a declaration by the UN General Assembly in 1974: permanent sovereignty over natural resources, indexation of commodity prices to the prices of manufactures, a common fund to finance buffer stocks, codes of conduct for multinational firms. For a few years the vocabulary of centre and periphery was the vocabulary of the General Assembly’s majority, and dependency’s diagnosis, that the rules of the world economy were themselves a mechanism of transfer, was being asserted by foreign ministers.

The hour passed. The commodity agreements mostly failed or were never funded; the oil weapon proved unrepeatable for exporters of coffee and copper, whose markets punished solidarity with substitution and stockpiling; the debt crisis of the 1980s converted the Group of 77’s members from negotiators into supplicants, each bargaining separately with creditors for rescheduling; and the intellectual climate in the centre swung to the market doctrines that treated the entire UNCTAD project as a costly misunderstanding. The episode nonetheless belongs in any honest account of the tradition, for two reasons. It demonstrates that dependency theory was never only an academic construction: its categories organised the largest diplomatic coalition of the twentieth century’s second half. And its failure demonstrates something the theory’s own subtler members had predicted, that the periphery’s bargaining weakness was structural rather than rhetorical, a matter of what its members sold and owed, which no resolution could repeal. The diplomatic hour is thus evidence entered for both sides: for the school’s reach, and for the hardness of the structure it described.

The debt crisis: the drain in the age of finance

In August 1982 Mexico announced that it could not service its foreign debt, and the periphery’s long argument with the world economy entered the phase that tested every framework at once. The borrowing of the 1970s, recycled oil money lent by the centre’s banks at floating rates, had seemed to many governments an escape from the old constraints: development financed without the multinational’s equity or the aid donor’s conditions. When the centre’s monetary authorities raised interest rates at the decade’s turn to break their own inflation, the escape inverted. Through the 1980s the flow of resources between Latin America and its creditors reversed sign, and the region, on the accounting of its own commission, transferred outward, in interest and profit remittances beyond new lending, sums comparable to a Marshall Plan running in reverse, a fact the dependency writers did not fail to set beside Naoroji’s ledgers. The lost decade that followed, of falling incomes, slashed imports and the structural adjustment programmes by which the creditors’ institutions supervised the austerity, divided the school’s survivors. Some read the crisis as the drain perfected for the financial age, dependence administered no longer by the enclave or the subsidiary but by the debt contract and the letter of intent. Others conceded what the episode exposed in the tradition: dependency theory had theorised production and trade with great sophistication and money hardly at all, and a framework surprised by the decisive weapon of the decade had a hole at its centre where a theory of finance should be. Both readings survive in the literature, and the second, the self-critical one, is the tradition’s most useful bequest to any successor framework: the drain changes its plumbing with each era, and a theory of it must be rebuilt for the era of the bond as it was built for the era of the bill of lading.

The heir in Santiago: neo-structuralism

The commission that started the story also supplied its epilogue, and the epilogue is instructive about how research programmes survive their defeats. After the debt crisis of the 1980s and the collapse of the region’s faith in both protection and its opposite, the Economic Commission for Latin America, renamed with the Caribbean added, produced a self-consciously renovated doctrine that the literature calls neo-structuralism. Its founding document, Changing Production Patterns with Social Equity, published in 1990, accepted the market disciplines its predecessors had resisted, open trade, macroeconomic restraint, export competitiveness, while insisting on the structuralist inheritance: competitiveness must be genuine, built on technical progress diffused through the production system, rather than spurious, built on cheap labour and depleted resources, and growth must be reconciled with the equity the old industrialisation had sacrificed. Behind the document stood the Chilean economist Fernando Fajnzylber, whose earlier comparative study had coined the tradition’s most haunting pair of images: an empty box for the Latin American economies that had achieved competitiveness without equity or equity’s rhetoric without competitiveness, and the contrasting Asian pattern the region was urged to study. Neo-structuralism is dependency’s heir in the strict genealogical sense, descended through CEPAL from Prebisch, and its moderation measures the distance the century had travelled: the grandson doctrine preaches insertion into the world market on better terms, where the grandfather’s radical readers preached rupture. Whether that moderation is wisdom purchased by defeat or the original structuralist insight finally freed of its revolutionary escort is a judgement the literature continues to argue, and this article leaves it where the participants left it, noting only that the family resemblance, the insistence that structure governs what policy can achieve, survived both the school’s triumph and its scattering.

The grandchild’s revolt: post-development

The tradition’s most radical descendant rejected the family’s founding noun. In the 1990s a current of writers, the Colombian anthropologist Arturo Escobar’s Encountering Development their best-known text, argued that development itself, the enterprise whose terms dependency theory had disputed, was the problem: a discourse invented at the inauguration of the postwar order, when the American president’s programme of 1949 christened two billion people underdeveloped, which constituted the periphery as an object of expert intervention and licensed the apparatus of agencies, missions and plans that governed it. On this reading the dependency writers, for all their fury, remained prisoners of the discourse, quarrelling over development’s distribution while accepting its desirability and its measures. The post-development writers proposed instead the defence of the local knowledges and communal economies the discourse pathologised. The revolt is instructive about the parent tradition in both directions. It confirms dependency theory’s deepest methodological claim, that the categories through which the periphery is known are themselves instruments of the relation, a claim the post-development writers radicalised from economics into epistemology. And it exposes the parent tradition’s unexamined inheritance, the productivism it shared with its enemies, the shared faith of structuralist, Marxist and moderniser alike that the periphery’s future was more industry, which the grandchildren, writing after the ecological turn, were the first in the lineage to doubt. A family that produces its own heretics on schedule is alive by the only test that matters in the history of ideas.

The material drain: Bunker and the extractive economy

One scholarly bridge carried the tradition’s mechanism into territory its founders never surveyed. Stephen Bunker, an American sociologist who worked in the Brazilian Amazon, published Underdeveloping the Amazon in 1985 and rebuilt the dependency argument around matter and energy rather than value alone. Extractive economies, he argued, differ structurally from productive ones: they export not only commodities but the organised energy of ecosystems, soils, forests, mineral stocks accumulated over geological time, while the value they receive in return is priced by labour and capital costs elsewhere; the tapping of each new extractive frontier reorganises the region around the removal, railways from pit to port, labour regimes of boom and abandonment, and leaves behind a landscape and a polity deformed by the sequence of rushes. The Amazonian record, rubber, timber, minerals, cattle, gave the argument its archive. Bunker’s work seeded what later scholarship systematised as ecologically unequal exchange, the study of how trade drains biophysical stocks from periphery to centre beneath the monetary accounts, and it demonstrates the tradition’s generative pattern at its best: a mechanism stated by the founders in the vocabulary of value, translated by a fieldworker into the vocabulary of matter, and found to hold in both. The plantation school had glimpsed the point in the exhausted soils of the sugar islands; Bunker gave it a method.

The supercycle test: a commodity boom judged by the old questions

The tradition received one final examination in the first decade of the twenty-first century, when a long commodity boom, driven above all by Asian industrial demand, reversed the terms of trade in the periphery’s favour for the longest sustained interval since the founders wrote. South America’s exporters grew, reduced poverty and accumulated reserves on the windfall, a decade that might be entered as the Prebisch-Singer thesis’s empirical funeral. The commission in Santiago, staffed by the structuralists’ heirs, declined to celebrate, and its reasoning was pure neo-structuralism: the boom, it warned, was reprimarising the region’s exports, hollowing the manufacturing built across the protected decades, and leaving the economies more exposed to the price cycle than the generation that had sworn never again. When the cycle turned in the middle of the following decade, the hollowing was visible in the statistics and the political crises of the commodity states followed the old script with a fidelity no satirist would risk. The episode is recorded here as history, its actors’ choices closed, and its lesson is methodological: the school’s central questions, what does the periphery sell, who owns the selling, and where does the surplus settle, proved capable of organising the evidence of a boom as readily as of a slump, which is the mark of a framework rather than a mood. Frameworks survive their funerals when the questions outlive the answers, and by that test the tradition this article has surveyed remains, in the strict sense, unburied.

The paradigm’s decline, explained in its own terms

A school that taught a generation to explain ideas by the structures carrying them invites the same treatment for its own disappearance, and the explanation holds up well. The dependency programme flourished inside a specific configuration: a long postwar boom that made national development projects plausible, developmentalist states that employed its graduates, universities expanding faster than dictatorships could police them, and a world politics, Bandung to the NIEO, in which the periphery acted as a bloc and needed a doctrine. Every element of the configuration dissolved within a decade. The coups destroyed its institutions and scattered its personnel; the debt crisis replaced its question, how to develop, with the creditors’ question, how to adjust; the Pacific results retired its strongest theorem; the bloc politics that had been its atmosphere evaporated as the commodity coalition failed and the Cold War’s end removed the audience for a third position between the camps. Even the school’s successes hastened the dissolution, since its best students carried its method into successor fields, value chains, developmental states, the new economic sociology, that no longer needed its name. Intellectual historians sometimes treat the 1980s as the decade dependency theory was refuted; the school’s own method suggests the colder verdict that it was decommissioned, its questions redistributed to fields with better funding and worse memories, which is why its history reads less like a death than like a firm broken up for its assets.

The Northern classroom: dependency in the Latin American studies boom

The theory’s migration north reorganised a field. The Latin American studies establishment in the United States, expanded after the Cuban revolution had made the region a strategic anxiety, had been built on modernisation assumptions and area expertise; the dependency literature arrived in translation, through the Monthly Review and the journals, as an indictment written by the area’s own scholars, and by the middle 1970s no syllabus in the field could omit it. The Latin American Research Review, founded in 1965 as the boom’s house journal, carried the controversy in its pages, and a generation of North American graduate students learned Spanish, in part, to read arguments their own professors could not summarise safely. The Northern reception performed the customary operations on its import: it systematised the school into a theory with axioms its authors had never signed, taught it as a doctrine beside modernisation theory in the textbooks’ paired chapters, and thereby manufactured the simplified dependency theory that Brenner, and Cardoso himself, then criticised. The point is worth the researcher’s attention because the object of half the subsequent literature, the dependency theory refuted in the journals, is this classroom artefact, a position-set assembled for teaching, while the originals went on being subtler and sadder in their own languages. Fields remember what they can examine on, and the school’s Northern fate is the standing warning this article has applied throughout: before refuting a tradition, establish which of its positions the refutation touches.

Palma’s settlement: a methodology, not a formal theory

The scholarly argument over what dependency theory had been was settled, insofar as scholarship settles anything, by a Chilean economist’s long review article in 1978. Gabriel Palma, examining the whole corpus for the journal World Development, proposed a disjunction that reorganised the field’s self-understanding: dependency analysis should be read not as a formal theory of underdevelopment, a set of general laws in the manner its strong version aspired to, but as a methodology for the analysis of concrete situations, a way of posing questions about how external structures are internalised in specific national formations at specific moments. On that criterion, Palma argued, Cardoso and Faletto’s book was the tradition’s exemplar and Frank its cautionary tale: the one generating questions that new evidence could refine, the other a closed system that evidence could only embarrass. The settlement’s durability is its best advertisement. Four decades of retrospective literature have largely operated inside Palma’s disjunction, praising the tradition as a method and a sensibility while declining its laws, and the present article’s own discipline, its insistence on separating the positions and counting the impossibility claim as one strand’s signature rather than the school’s creed, is Palma’s distinction at work. There is a fitting modesty in the outcome: a school that began by denouncing the universal stages of Rostow ended by accepting, from one of its own reviewers, that it too had produced not a universal theory but a way of asking questions, and that the questions, unlike the answers, had not dated.

The theory’s afterlife and the questions that remain

Dependency theory as a unified research programme had a lifespan, roughly from the middle 1960s to the early 1980s, after which its institutional world dissolved: the exile circuits dispersed as the dictatorships fell, the journals changed their agendas, the development profession reorganised itself around other frameworks, first the neoliberal counter-revolution that treated the state as the problem, then the institutional and microeconomic turns that dissolved structures into incentives and experiments. It would be tidy to report that the tradition died of its refutations, but the record is less flattering to everyone. The strong version was indeed refuted, by East Asia and by the industrial statistics Warren marshalled. The school as a whole was abandoned as much as answered, overtaken by the same political defeats that scattered its authors, and the questions it had forced onto the agenda, about the terms of trade, the distribution of gains from trade, and the politics of insertion into world markets, were set aside rather than settled. Setting aside is not refutation, and the vocabulary of the school survives wherever those questions resurface: in the literature on global value chains and commodity chains that studies how lead firms govern production across borders, in debates about resource curses and commodity dependence conducted by institutions that would not cite Frank in a footnote, and in the lecture halls of three continents where the three positions are still the fastest route into the problem of why development is so unevenly shared.

What replaced dependency theory in the seminar room is a revealing list. Globalisation theory took the integration and dropped the asymmetry. Institutionalist development economics took the history and dropped the world market, relocating the causes of poverty inside national borders, in courts and property registries, where the dependency writers had always said the external relation was also at work. The global value chains literature, the school’s closest living relative, restored the cross-border production structures and the lead-firm power that dependency writers called the internalisation of the external, usually without the value theory and always without the politics. Each successor kept a limb of the original animal. None reassembled it, and the animal’s central nervous system, the claim that the wealth and poverty of regions are produced together by one process and must be explained together or not at all, remains available for any generation that finds the successors’ partial accounts insufficient. Traditions in social theory rarely die outright; they subside into the discipline’s groundwater, and the test of this one is that its questions keep rising to the surface in other people’s vocabularies.

For students and exam candidates

Build every answer on the impossibility split, because it is the distinction examiners are testing when they set this topic. State the three positions separately before comparing them: the claim that incorporation itself produces underdevelopment, the claim that dependence conditions a real but deformed development, and the technical claim about value moving through trade between unequal wage levels. Keep the mechanisms apart in your paragraphs, the declining terms of trade, unequal exchange and super-exploitation are three different engines, and conflating them is the commonest way strong candidates lose marks. Handle the East Asian evidence with precision: it defeats the strong version and is partly absorbed by the weak one, and an answer that says simply that dependency theory was refuted will be marked as a misunderstanding. Place the school inside the wider argument about empire by working through the companion guide to Marxism and imperialism alongside this article, and fix the grid in memory with a revision card for each position, claim, mechanism, verdict on development, objection, using the VaultBook study notes tool for this series. Practise one sentence until it is automatic: dependence, for the subtlest of these writers, was a situation to be analysed country by country, not a fate to be pronounced over a continent.

For teachers and seminar leaders

The three positions grid is built to be taught from. Open the seminar by distributing the grid with its political conclusion column blanked, and have students complete it from short extracts, a chapter of Frank against the Cardoso and Faletto conclusion, before revealing the original; the disagreements that surface in the exercise are the seminar. Follow with the Emmanuel and Bettelheim exchange as a lesson in how Marxists argue from shared premises, assigning half the room to reconstruct each side’s definition of capitalism and requiring every intervention to name its definition. A second session can stage the East Asia dossier: one team argues refutation from the industrial record, the other argues conditioning from the land reforms, the security subsidy and the disciplined states, and the class votes not on which team won but on which version of the theory each piece of evidence touches. Close by returning to the grid and asking where Warren, Marini and Amin would have to be placed, and what new rows their placement forces. Students leave with the durable skill the topic exists to teach: the ability to take a school of thought apart into positions that evidence can separate, instead of inheriting it as a slogan.

For researchers, historians, and journalists

The primary corpus is larger in Spanish and Portuguese than in English, and the untranslated remainder, Marini’s essays, dos Santos’s later work, the commission’s internal debates, still rewards anyone who reads it. The commission’s archive in Santiago and its published review are the first ports of call for the structuralist prehistory, and the run of the Monthly Review across the 1960s and 1970s is the best single record of how the argument was received and fought over in the North. Two continuations deserve a researcher’s priority. The first is the literature that followed production as it fragmented across borders, for which the series study of imperialism and the organisation of global supply chains gathers the entry points; the second is the country literature in which the framework was applied, contested and revised on Latin American evidence, gathered in the companion study of dependency theory in Latin America. Observe the contested figures protocol throughout: the terms of trade series have been revised repeatedly, the transfer estimates rest on contestable conventions, and a citation that does not name its convention is a claim about the literature rather than about the economy.

Frequently Asked Questions

Q: What is dependency theory?

Dependency theory is the name given to a family of accounts, developed chiefly in Latin America in the decades after the Second World War, which explain the persistence of poverty in the world’s periphery as a product of the way poorer economies were incorporated into the world economy rather than as a survival from an earlier age. Its shared core has two parts. The structure of the international division of labour, who sells what to whom and on what terms, systematically shapes what a dependent economy can become; and that external structure is reproduced inside each country by the alliances of the classes and institutions that profit from it. Beyond the core, the family disagreed about almost everything important, above all about whether genuine development was possible without breaking the relation itself, and any summary that omits the disagreement has not understood the school.

Q: What is the Prebisch-Singer thesis?

The Prebisch-Singer thesis holds that the prices of primary commodities tend, over long periods, to fall relative to the prices of manufactures, so that countries specialising in primary exports must ship ever larger volumes abroad to purchase the same quantity of industrial goods. It was advanced independently around 1950 by Raúl Prebisch, from the trade statistics of a United Nations commission, and by Hans Singer, from an analysis of how the gains from technical progress are distributed. Their shared mechanism was an asymmetry of labour markets: in the industrial centres, strong unions and concentrated firms convert productivity growth into higher incomes, holding prices steady, while in the periphery surplus labour lets productivity gains escape as lower export prices. The conclusion drawn was that the periphery transfers the fruits of its own technical progress to its customers, and the evidence for the tendency has been fought over by statisticians ever since.

Q: What is the development of underdevelopment?

The development of underdevelopment is Andre Gunder Frank’s formula for the claim that the poverty of the periphery is not an original condition awaiting correction but a condition actively manufactured by the same historical process that enriched the centres. Frank traced chains of metropolises and satellites, each link monopolising the surplus of the one below, running from the world centres down through national capitals and provincial towns to the remotest village, so that incorporation into the world market from the conquest onward organised the satellites’ economies around extraction. His prize evidence was a historical inversion: regions that had been most tightly bound to the world market by colonial booms, the sugar Northeast of Brazil, the mining districts, the plantation islands, were afterwards among the poorest on their continents. The formula was aimed at the picture of two worlds, one feudal and one capitalist, awaiting contact; for Frank there had only ever been one world, and its development and its underdevelopment were a single product.

Q: What is unequal exchange?

Unequal exchange, in Arghiri Emmanuel’s technical usage, names a transfer of value that occurs through ordinary trade at market prices when two conditions hold together: capital moves freely enough between countries for profit rates to tend toward equality, while labour cannot move and wages remain set at persistently different national levels by history and institutions. When those conditions hold, the prices at which goods settle on world markets embody systematically less labour from the high-wage country and more from the low-wage country in each hour exchanged, so value flows toward the high-wage side inside transactions that are exchanges of equivalents by every commercial test. The transfer is invisible in any single bargain and appears only in the structure of prices as a whole. Critics within Marxism, Charles Bettelheim first among them, objected that the argument takes wages as given where a theory of production should explain them, and the dispute remains the most formally exact quarrel the school produced.

Q: What is super-exploitation?

Super-exploitation is Ruy Mauro Marini’s term for the systematic payment of labour power below its value in dependent capitalism, and it names a mechanism rather than a moral judgement. Where the value of labour power is the cost of reproducing the worker at a historically given standard, super-exploitation cuts beneath that standard by three routes: stretching the working day, driving the intensity of labour beyond its normal level, and appropriating part of the consumption fund the worker requires to live and raise a family. Marini argued that dependent capital resorts to these routes to compensate for the value it loses abroad through deteriorating prices and remitted profits, and that the arrangement is made survivable by the non-capitalist production around it, the family plot and the unpaid household, which carry part of the cost of reproducing the worker. Because the claim is about production relations rather than about trade, it is the strand of the theory least damaged by the objection that dependency writers confused exchange with exploitation.

Q: Did dependency theory say development was impossible?

The question cannot be answered for the school as a whole, because the school disagreed with itself on precisely this point, and the disagreement is the key to the whole literature. The best-known strand, associated with Frank, did teach impossibility: satellites could not develop within the capitalist world system, since their underdevelopment was the system’s product, and only rupture with it opened a future. A second strand, set out by Fernando Henrique Cardoso and Enzo Faletto in their study of 1969, rejected the impossibility outright: dependence, they argued, conditions the form development takes, steering it through alliances between foreign capital, the state and local firms, but growth under dependence is real growth, distorted and exclusionary rather than illusory. A third position, Emmanuel’s, implied that development was blocked less by incorporation itself than by the wage and price structure of trade. Most textbook summaries attribute the first strand’s answer to all three, which is the error this article’s grid exists to prevent.

Q: What is the neo-Smithian criticism?

The neo-Smithian criticism is Robert Brenner’s charge, delivered in 1977, that dependency theory and world-systems analysis define capitalism by the wrong criterion. Where these writers treat market involvement, production for exchange and the pursuit of profit through trade as the marks of capitalism, a usage Brenner traced to Adam Smith’s account of the division of labour spreading with the extent of the market, Brenner insisted that capitalism exists only where producers have been separated from the means of production and must sell their labour power to survive. The charge has teeth because the definition decides where exploitation is located: in circulation, as the dependency writers’ mechanisms of drains and unequal prices imply, or in production, where the Marxist theory of surplus value places it. On Brenner’s criterion, much of what Frank called capitalist since the conquest was nothing of the kind, and the stagnation of such formations followed from their own class structures rather than from any satellite relation to a metropolis.

Q: Does East Asia refute dependency theory?

East Asia refutes the strongest version and disciplines the rest. The industrial transformations of South Korea and Taiwan were achieved in economies deeply inserted into the world market, exporting to the centres and dealing with their firms, which is exactly the configuration the impossibility claim declared sterile; no reading of the record can restore that claim. The weaker versions fare differently, since they never denied that dependent development could occur and can point to the conditions that made these cases exceptional, sweeping land reforms, an American security subsidy on a generational scale, and states unusually able to discipline both capital and labour, as evidence that insertion is everywhere mediated by national structures, which was their thesis. The careful verdict is therefore double: as a law of impossibility the theory died in the Pacific, and as a method for asking how external structures are internalised it can claim the Pacific cases among its better exhibits.

Q: What is the difference between dependency theory and world-systems analysis?

The difference is the unit of analysis. Dependency writers worked at the level of the national formation: their question was how the world economy’s structures are internalised, differently in each country, through its classes, its state and its history, which is why their best work is comparative and concrete. Immanuel Wallerstein’s world-systems analysis, launched in 1974, demoted the nation to a component and made the whole system, its core, semiperiphery and periphery bound in a single division of labour since the sixteenth century, the only legitimate object of study, with states and classes intelligible as expressions of position within it. Consequences follow in every direction: the dependency writers periodised dependence into historically distinct forms, colonial, financial, technological, while the systems framework holds one structure across five centuries; and Andre Gunder Frank’s migration from the first camp to the second, and eventually beyond it, is the biographical trace of the quarrel. The two shared personnel, journals and enemies, which is why they are so often mistaken for one theory.

Q: Is dependency theory still used?

As a unified school, no: the research programme that flourished from the middle 1960s dissolved by the early 1980s under the weight of the East Asian evidence, its critics’ objections and the debt crisis that changed the questions. As equipment, constantly. Its vocabulary, core, periphery, enclave, extraversion, is standard issue in development studies, often used by writers unaware of its provenance. Its institutional descendant, the neo-structuralism of the Santiago commission, carried its core questions about the quality of a country’s insertion into the policy mainstream of its region. Its method, the insistence that external structures must be traced through internal alliances before they explain anything, survives inside the global value chains literature and the comparative study of developmental states. And its mechanisms are reached for whenever trade between unequal partners is debated, by participants who frequently believe they are improvising them. The school’s ideas are used the way groundwater is used, invisibly and everywhere.

Q: How does the impossibility split separate the three positions?

Ask each position a single question, whether a dependent economy can develop as a capitalist economy while remaining dependent, and the three answers fall into a row. The first answer is no: incorporation into the world system is itself the cause of underdevelopment, so development and dependence exclude each other by definition. The second answer is yes, at a price: development occurs under dependence, steered by the alliance of foreign firms, the state and domestic capital, and its price is paid in exclusion, in wages held below the economy’s possibilities and decisions taken elsewhere. The third answer changes the variable: development is obstructed not by incorporation as such but by the wage structure frozen into world prices, so the obstacle could in principle be moved by moving wages and the terms on which goods exchange. The split separates the positions because the three shared a diagnosis of transfer and divergence while drawing incompatible conclusions from it, and because later debates, over East Asia above all, judged each answer separately rather than the school wholesale.

Q: What should a student separate between unequal exchange and unequal development?

The two phrases answer different kinds of question, and keeping them apart is the first discipline of the topic. Unequal exchange is a mechanism: a precise proposition about how value moves between countries when goods trade at prices formed under equalised profit rates and unequal wages. A mechanism can be specified, modelled and argued over on its own terms, and accepting or rejecting it settles nothing by itself about any country’s past or future. Unequal development is an outcome: the observed fact that productive structures, technologies and living standards diverge between regions over time. Outcomes have many causes in any mixture, and unequal exchange is at most one candidate among super-exploitation, state capacity, class structure, war and geography. The characteristic error of weak answers is to slide from the outcome to the mechanism as if naming the divergence proved the transfer, or to assume that refuting the mechanism dissolves the outcome, which remains on the table demanding explanation either way.

Q: Why did the East Asian cases become the debate’s hardest evidence?

They were the hardest evidence because they met the strong version on its own ground and beat it there. South Korea and Taiwan were not sheltered exceptions to the world market but among its most exposed participants, selling to the centres, importing their technology, hosting their firms and taking their aid, and they nevertheless achieved within a generation the deep industrial transformation that the impossibility claim, applied consistently, forbade any satellite to achieve. Other problem cases could always be talked away: a stagnating economy could be called too dependent, a growing one not dependent enough, a deformed one proof of the rule. The Pacific cases resisted every rescue of that kind, since each qualification conceded that national conditions decide what insertion does, which was the opposing position’s thesis. Evidence becomes decisive in a theoretical quarrel not when it is dramatic but when the losing side’s replies strengthen the winner, and by that standard the Korean and Taiwanese records are the debate’s hardest exhibits.

Q: How does the three positions grid order the replies to stagnation?

The grid orders the tradition’s answers to a stagnating periphery by the depth at which each locates the cause. Its first row places the cause in incorporation itself: a satellite economy stagnates because stagnation is what the metropolis-satellite chain produces, and the reply is rupture, since no rearrangement inside the relation can cure a condition the relation defines. Its second row places the cause in the national configuration: stagnation is one possible outcome of dependence, selected by the alliances in command of the state, as Brazil’s industrialisation drive showed when it found its coalition, and the reply is political, a change in the coalition and the form of the state. Its third row places the cause in the price structure of trade: stagnation persists because low wages are frozen into the prices at which the periphery sells, and the reply is to raise the wage level and the terms of insertion. Reading down the grid replays the school’s internal argument in order, each row keeping the evidence above and refusing its conclusions.

Q: What separates a transfer of surplus from a failure to invest?

A transfer of surplus is a movement across a boundary: value produced within one economy and appropriated within another, whether through prices that undervalue its exports, profits remitted by foreign subsidiaries, interest on external debt or royalties on imported technology. Its measure is a flow between territories. A failure to invest is a use, or rather a non-use, of surplus that never leaves home: the landlord’s imported luxuries, the merchant’s hoard, the state project’s marble, the industrialist’s foreign bank account fed from domestic earnings. Paul Baran’s categories hold the two apart because they can coexist in any proportion: an economy may suffer a heavy drain and invest its remainder furiously, or retain everything and consume it all, and the policy conclusions differ in each combination, delinking for the first, a domestic reckoning for the second. Summaries that merge them into a single complaint about wealth leaking away lose the school’s founding distinction, and with it the ability to say what, concretely, should be done differently.