The dependency map and mechanisms that transcend national borders

By Emiliano López and Facundo Barrera Insua

Our previous note reconstructed Marxist dependency theory as the cornerstone of a tricontinental political economy. That was the theoretical foundation. This note builds the empirical architecture.

If dependency is real, then it should be measurable. However, mainstream economics provides inadequate concepts and tools for measuring dependency. GDP per capita, debt ratios, and current account balances do not capture the multidimensional structure of subordination described by dependency theory, which operates across trade, technology, finance, production, global networks, and surplus distribution.

Three puzzles illustrate this point. Although Chile has a manageable fiscal deficit and per capita income that place it among “high-income” countries, it has one of the highest Structural Dependency Indexes (SDI) in the world. For its part, Brazil is the world’s tenth-largest economy, but it is more subordinate than South Korea. Finally, India has built a globally recognized technology services sector, yet it scores higher on technological dependency than Brazil.

Mainstream institutions cannot see these contradictions because they are not looking for them. For example, the IMF’s Article IV consultations routinely praise Chile for “macroeconomic stability” and “sound institutions,” yet our index places it in the peripheral range of structural dependency, with elevated scores across multiple dimensions. Peru and Colombia, which the World Bank often cites as reformist success cases with “improving business environments,” fall within the peripheral and hyper-peripheral range. South Korea, a country that supposedly graduated from the periphery, an OECD member, and the world’s twelfth-largest economy, registers a mean SDI of 0.390, well below the peripheral threshold but still far from the hegemonic centers rather than among the autonomous powers. The conventional toolkit cannot explain why a country with Samsung and Hyundai has not achieved structural autonomy, its network dependency score (0.678) remains among the highest in the sample, revealing that corporate scale does not equal systemic centrality. On the contrary, as we shall see below, our framework can explain this because it measures not the size of firms but the economy’s position within global networks of value extraction. Doing Business rankings, the Human Development Index, and per capita GDP at purchasing power parity: these indicators obscure the asymmetries because they measure symptoms, not structure.

A tricontinental political economy cannot remain at the level of theoretical denunciation. It must produce analytical instruments capable of distinguishing between economies that exploit multipolar conditions to construct genuine autonomy and those in which dependency mechanisms persist or intensify despite a favorable conjuncture. What follows is the construction and application of the SDI: a composite measure that captures dependency across six dimensions, applied to 31 countries over 28 years (1996–2023). The result is a cartography of the global hierarchy, built from 856 country-year observations.

The six chains: how we built the SDI

Classical dependency theory understood that peripheral subordination does not operate through a single mechanism. Marini’s analysis of the dependent capital circuit (M–C...P...C’–M’) showed that external control is exercised at every phase of the accumulation cycle, from initial capital formation to production and finally to the realization and appropriation of surplus value. We translated this insight into a measurement framework. Each of the SDI’s six dimensions corresponds to a specific phase of the circuit and captures a distinct mechanism that constrains autonomy.

Regarding the first phase of the cycle (M-C), we present two dimensions. The technological dependency question asks: Do you innovate, or do you license? The Fourth Industrial Revolution has created new hierarchies of innovation capacity in which peripheral countries face systematic exclusion from research, design, and governance of technological systems. The issue is not the import of machinery per se, but the structural barriers to autonomous technological development that reproduce center-periphery relations in novel forms. We measure the extent to which an economy’s productive structure remains biased toward primary activities rather than technology-intensive ones, and the share of the value embodied in its exports that originates in foreign innovation.

Financial dependency measures an economy’s position within what Bonizzi, Kaltenbrunner, and Powell have called “subordinate financialization”: integration into global financial systems in a position of structural disadvantage. The currency hierarchy that places the dollar at the apex and peripheral currencies at the bottom enables the extraction of what Musthaq terms “new imperialist rent,” surplus appropriation through financial arbitrage made possible by the subordinate status of peripheral monies. This is the most complex dimension of our index, requiring multiple indicators that range from the degree of foreign ownership of domestic assets and the composition of external liabilities to the volatility of the exchange rate and the extent to which sovereign debt must be denominated in foreign currencies. One component deserves mention: the Exchange Rate Deviation Index, the gap between GDP at purchasing power parity and GDP at market exchange rates, which reveals the structural undervaluation embedded in the monetary hierarchy.

In the second phase of the cycle (C...P...C’), we will distinguish two dimensions: Productive and Network dependency. The first one captures the constraints on autonomous accumulation, and it is, as we shall see, the dimension that most sharply distinguishes economies that have transformed their structural position from those that have not. The issue goes well beyond “technological backwardness.” Even countries that have achieved significant industrial development continue to exhibit high productive dependency because the conditions under which peripheral industry operates (subordinate integration into transnational production networks, external determination of technological trajectories, financial pressures that divert surplus from fixed investment) structurally limit autonomous accumulation. We track cost competitiveness, productivity gaps relative to the global frontier, the degree of wage compression, and, crucially, investment deficiency: whether an economy is channelling its surplus into expanding its productive apparatus or diverting it elsewhere. That last indicator will prove decisive. As the data will show, the distinction between economies that invest and those that disinvest is the sharpest fault line in the global hierarchy.

Regarding Network dependency, we ask whether you are a central node or an appendage. Peripheral countries provide inputs and assembly services, while central countries control high-value activities and technological development. We measure this through the asymmetry of a country’s position in global trade networks and its capacity to function as an intermediary rather than as a terminal node. Most countries in our sample remain subordinated within global trade networks regardless of their other characteristics. The promise that “inserting into global value chains” would promote convergence has not materialized; insertion reproduces hierarchy.

The mechanisms that drain surplus from the periphery do not require visible plunder. Emmanuel demonstrated that unequal exchange operates through apparently equal market transactions: when wage differentials between countries exceed productivity differentials, an hour of metropolitan labor exchanges for multiple hours of peripheral labor, creating systematic value transfer even when commodities trade at equivalent prices.

Finally, in the third phase of the cycle (M’–D’), we distinguish between Distributive and Commercial dependency. Distributive dependency operationalizes these mechanisms of surplus appropriation concretely. We track the outflow of realized surplus through profit repatriation, interest transfers abroad, and the gap between the value a country creates in global markets and the value it captures domestically. We also measure the social reproduction deficit: how far public spending on health, education, and social protection falls short of levels needed to sustain the labor force. Contemporary patterns of superexploitation extend beyond traditional wage compression to encompass what Marini identified as the three tactics of peripheral capital: increasing labor intensity, extending the working day, and compressing real wages below the value of labor power, that is, wages insufficient to reproduce workers’ productive capacity, a structural feature of the dependent cycle of capital.

Commercial dependency answers the question: what do you sell, and what do you buy? This dimension captures external control over the realization phase of the capital cycle, where surplus value extraction depends on market conditions determined by core economies. It goes beyond the classical focus on commodity concentration. A country can achieve apparent export diversification while remaining trapped in low-value segments of global production networks, what recent scholarship calls the “peripheralization of industrial activities.” This involves manufacturing export growth coexisting with systematic exclusion from strategic value creation. We track how much of a country’s exports consist of primary commodities versus high-technology goods, how much of the value-added in its exports is domestic, and how dependent its productive apparatus is on imported capital goods.

A methodological note. The six dimensions are aggregated into a single composite score using a weighting method that preserves theoretical consistency: higher values always indicate greater dependency. All indicators are normalized to a common scale, ensuring the index ranges from 0 (maximum autonomy) to 1 (maximum dependency). A country scoring 0.20 retains substantial control over its accumulation process; a country scoring 0.80 has that process determined, in most dimensions, from outside. The detailed methodology, including variable construction, aggregation procedures, and robustness tests, will be presented in the forthcoming academic paper. Our dataset draws on TiVA-OECD, BACI, Penn World Tables, IMF IFS, World Bank WDI, and UNCTAD databases. The sample comprises 31 countries spanning 28 years and three major crises.

Table 1. SDI Dimensions and Operationalization. Source: own elaboration.

Four worlds in the global economy

The global mean SDI is 0.494, indicating that most countries exhibit moderate dependency. The mean, however, conceals a dramatic architecture of inequality. A standard deviation of 0.201 reflects a world structured into qualitatively different positions, not one converging toward a common standard.

Figure 1. Country Ranking — Structural Dependency Index, Average 1996–2023.

Source: own elaboration based on World Bank, Penn World Table, BACI, and OECD TiVA.

The empirical results organize themselves into the four configurations we mapped in our previous note, defined not by income levels or geographic location but by the interaction between structural dependency and the state’s capacity to mediate it. The SDI captures the first axis of that interaction: the depth and breadth of structural subordination. The second axis, state mediating capacity, requires its own index, currently under construction. A comprehensive mapping of the transitions between quadrants will require both instruments to work in tandem. For now, we begin where the data are ready: the economic-structural plane. What follows is a reading of the data through that framework.

Figure 2. Structural Dependency in the Global Economy, 1996–2023.

Source: own elaboration based on World Bank, Penn World Table, BACI, and OECD TiVA. Countries are classified using the four-quadrant framework developed by López (2026).

At the end of the spectrum, the United States registers an SDI consistently near zero throughout the entire period (mean 0.029). Germany sits higher but still in the low-dependency range (mean 0.220). Japan, the United Kingdom, and France sit in the semi-hegemonic range. Their low scores reflect imperial positioning rather than developmental achievement, a distinction we drew in the first note of this series. What the data add is temporal precision.

Japan’s SDI has increased over the period (range: 0.086–0.520), the only major economy in this quadrant to exhibit rising dependency. The erosion is concentrated in the technological and financial dimensions, reflecting the loss of manufacturing leadership to China and South Korea and the consequences of three decades of stagnation. A position at the apex of the hierarchy is not a permanent entitlement. The data show it can erode.

The United States, by contrast, shows almost no variation. Its autonomy is structural, embedded in the architecture of the global system itself: dollar hegemony, technological control, network centrality. This is not a finding that requires interpretation. It is the baseline against which peripheral dependency is measured.

China registers a mean SDI of 0.334 over the full period. But the static position matters less than the trajectory. China’s SDI dropped from 0.528 to 0.164 over 28 years. Its standard deviation is among the highest in the sample (0.129), reflecting not instability but directional change, the dismantling of dependency across multiple dimensions simultaneously.

The productive dimension is the driving force behind this transition, with the investment component playing a decisive role. China’s gross fixed capital formation as a share of GDP has consistently exceeded 40%, reaching peaks above 45%. These are levels that are unmatched by any other economy in our sample. Our productive dependency dimension captures this directly: as China’s investment rate rose, its productive dependency score fell, dragging the overall SDI downwards. This is because fixed capital accumulation generates the technological capabilities, productivity gains, and domestic market expansion that reduce commercial, technological, and network dependencies in tandem. Investment in productive capacity is not just one dimension among six. It is the lever that moves the others.

The conditions for China’s transformation were historically specific. What the data add to the theoretical argument is precision: GFCF rates above 40%, sustained over two decades, combined with declining foreign value-added content in exports and rising domestic value capture. These are not abstractions. They are measurable policy outcomes.

Malaysia’s SDI trajectory shows the long-term dividend of rejecting IMF conditionality in 1997: preserved state control over the banking system and strategic enterprises translated into lower network dependency scores than Thailand, which accepted structural adjustment. But Malaysia’s greater openness to foreign capital in electronics manufacturing keeps its technological dependency scores elevated, the highest in the entire sample (0.946), a pattern the data make visible.

For its part, the contested periphery is where the data become most damning. In Brazil, the SDI is 0.602, with a standard deviation of 0.043. That number deserves attention: it is the world’s tenth-largest economy, with the largest industrial base in Latin America and a domestic market of 200 million people. Over the past 28 years, its structural position has barely changed. That low variation across 28 years of data is not stability. It is stagnation with a decimal point.

Regarding Argentina, the SDI is 0.763, placing it firmly in the peripheral range. But the headline number obscures a more important finding. Argentina’s distributive dependency score of 0.783 is among the highest in the sample and the dominant dimension in its profile. The surplus drainage is not cyclical. It is a defining structural feature of the economy, and it explains why a country with historical industrial capacity, an educated workforce, and abundant natural resources reproduces crisis after crisis without altering its position in the hierarchy.

At the same time, India’s SDI is 0.562. Here, the explanation is strongly related to technological dependence (0.624), which remains high despite its globally recognized IT sector. The specialization in services has not resolved technological subordination; it has coexisted with it.

At the hyper-peripheral extreme (SDI above 0.80), Thailand, Colombia, Peru, and Mexico exhibit elevated dependency in virtually all six dimensions. Chile, with a mean SDI of 0.711, occupies a position that is revealing precisely because it is counterintuitive: the standard-bearer of neoliberal reform in Latin America, a country classified as high-income, registers a level of structural dependency higher than both Brazil and India. The “model” produced macroeconomic stability and structural subordination simultaneously.

The contrast with China on the productive dimension is devastating. While China channeled surplus into fixed capital at rates above 40% of GDP, the major Latin American economies exhibit investment rates between 15% and 22%. These are insufficient to close productivity gaps, build autonomous industrial capacity, or develop the endogenous technological capabilities that would reduce dependency across the other five dimensions. Their productive dependency scores remain persistently high, not because they lack an economic surplus, they generate one, but because that surplus is systematically diverted away from productive accumulation. The question that our next notes will address is where this surplus goes: financial speculation, profit repatriation, luxury consumption, and capital flight.

The geographic pattern is unmistakable. Latin America is the most uniformly dependent region in our sample. Every Latin American country in the dataset except Costa Rica falls into the peripheral or hyper-peripheral range, and Costa Rica itself sits in the upper semi-peripheral band. This consistency, persisting despite significant differences in size, resource endowments, and policy orientation, suggests that dependency operates through systemic mechanisms that transcend national policy choices.

The contested semi-periphery is where the political question becomes sharpest. Within our sample, India and Brazil occupy this space most clearly: economies with significant state capacity and large domestic markets whose SDI scores nonetheless remain stubbornly high.

Their dimensional profiles tell the story. Brazil combines moderate technological scores with elevated productive, financial, and distributive dependency, the empirical signature of an economy whose surplus is drained through the very channels our distributive dimension measures: profit repatriation, interest transfers, compressed domestic value capture. The institutional capacity exists. The class configuration that would redirect it does not, or not yet. In turn, India presents a different combination: lower financial dependency but high productive and technological scores, reflecting an economy where the services-led growth model has not translated into autonomous industrial accumulation. In both cases, the question is the same: why does state capacity coexist with structural subordination? The SDI identifies the constraint. However, it does not explain the political choices that perpetuate it.

What each dimension reveals

Disaggregating the SDI into its six components shows how different mechanisms of dependency operate with varying intensity and how they cluster.

Network dependency is the most universal constraint, with a mean of 0.624; most countries remain subordinated within global trade networks regardless of their other characteristics. This is the empirical refutation of the promise that global value chain integration would promote convergence. Productive dependency follows closely at 0.525, reflecting widespread constraints on autonomous industrial development even in countries with significant manufacturing bases.

Commercial and financial dependency show moderate mean values (0.400 and 0.337, respectively) but with high variation: these are the dimensions where national profiles diverge most sharply, reflecting distinct modalities of insertion into global markets and financial circuits. Technological dependency sits at an intermediate level (0.423), while distributive dependency (0.423) shows the greatest variation among the dimensions (standard deviation of 0.207), reflecting heterogeneous patterns of surplus appropriation across the hierarchy.

Figure 3. Maps of Dependency by Each Dimension, 1996–2023.

Source: own elaboration based on World Bank, Penn World Table, BACI, and OECD TiVA.

The dimensional maps reveal distinct geographic patterns. Commercial dependency clusters strongly in resource-rich countries. Technological dependency affects both resource exporters and some manufacturing-oriented economies. Financial and network dependencies exhibit more dispersed patterns, reflecting the complex architecture of contemporary global economic integration and the waning of US hegemony amid hyper-financialization in the US and other countries in the Global North.

At this level of disaggregation, the central finding is that dependency is never unidimensional. Countries rarely exhibit a single-dimensional dependency. Instead, it tends to cluster across related dimensions, creating systemic patterns of subordination that reproduce themselves across different historical periods and policy orientations. This is precisely why our framework insists on measuring all six dimensions simultaneously: partial analyses that focus on trade alone, or finance alone, or technology alone, will always underestimate the depth of the constraint. China’s trajectory suggests that overcoming these multidimensional constraints requires comprehensive transformation strategies that address productive, technological, and network dependencies simultaneously. This explains why, despite sustained efforts to grow and diversify, most countries remain locked in subordinate positions.

What the data settle, and what they leave open

Our data settles the question. The first is empirical: dependency is not a metaphor from the 1960s but a measurable structure that persists across 28 years, three global crises, and radically different national policy orientations. The SDI hierarchy, from hegemonic centers to the hyper-peripheral extreme, has been remarkably stable for almost three decades.

The second is about possibility: structural transformation is possible. China’s trajectory from 0.528 to 0.164 is the proof. But the productive dimension tells us how it happened, through sustained investment in fixed capital at rates no Latin American economy has approached, generating cascading reductions in technological, commercial, and network dependencies. The lever is identifiable. The conditions under which it can be pulled are specific and demanding.

The third issue is the question that the SDI opens but cannot answer: why do some ruling classes reinvest the surplus into productive accumulation, while others divert it into finance, luxury consumption, and capital flight?

The index measures this constraint. However, it does not explain the decisions that reproduce it. To achieve this, we need a different kind of analysis: one centered on class configurations, domestic political coalitions, and the institutional arrangements through which surplus is allocated. We will develop that analysis in the next note, following the surplus, to see where it goes.

We have the map. Now, we will follow the surplus.

---

Emiliano López is a researcher at CONICET–Universidad Nacional de La Plata and Chief Economist at Tricontinental: Institute for Social Research.

Facundo Barrera Insua is a researcher at CONICET–Universidad Nacional de La Plata and an economist member at Tricontinental: Institute for Social Research.