Big Push vs. Unbalanced Growth: A Development Historian's Tour

Two Visions for the Poorest Countries

In the years after World War II, a new subdiscipline emerged in economics. Development economics — the study of how poor countries become rich — had no established orthodoxy, no canonical models, and no consensus on basic questions. Its practitioners were inventing the field as they worked, drawing on classical economics, Keynesian macroeconomics, wartime planning experience, and direct observation of economies that bore little resemblance to the industrial West.

Among the many debates that defined the field’s early decades, none was more consequential than the argument between two visions of how industrialization should proceed. One vision, associated most closely with Paul Rosenstein-Rodan, held that poverty was a trap that could only be escaped through a massive, coordinated investment effort — a “big push” that would simultaneously create the industries, infrastructure, and demand needed for self-sustaining growth. The other, associated with Albert Hirschman, held that the big push was not only impractical but unnecessary. Development, Hirschman argued, proceeds through imbalance — through the deliberate creation of shortages and bottlenecks that provoke the problem-solving, investment, and institution-building that make further growth possible.

The debate was not merely academic. It shaped the policies of newly independent nations across Asia, Africa, and Latin America. It informed the lending decisions of the World Bank and the planning ministries of governments. And it continues to reverberate in contemporary arguments about industrial policy, foreign aid, and the role of the state in economic development.

Rosenstein-Rodan’s Big Push

Paul Rosenstein-Rodan, an Austrian-born economist who spent much of his career at MIT, laid out the big push theory in a 1943 paper written while he was at the Royal Institute of International Affairs in London. The paper, “Problems of Industrialisation of Eastern and South-Eastern Europe,” addressed a concrete policy question: how could the underdeveloped regions of Europe be industrialized after the war?

Rosenstein-Rodan’s answer began with a diagnosis. Poor countries are poor because they are trapped in a low-level equilibrium. The domestic market is too small to justify investment in modern industry. The market is too small because incomes are low. Incomes are low because there is no modern industry. Each firm, contemplating investment individually, sees insufficient demand for its product and stays out. But if many firms invested simultaneously, the workers they employed would spend their wages, creating demand for each other’s products. The problem is not that investment is unprofitable but that it is unprofitable for any single firm acting alone. It would be profitable for all firms acting together.

This is a coordination failure. The individually rational decision — do not invest, because the market is too small — produces a collectively irrational outcome — everyone stays poor. The solution is coordinated investment: a big push that simultaneously launches many industries, creating the demand that justifies each individual investment. The state, or an external agency like the World Bank, would play the role of coordinator, ensuring that the investments are made together rather than piecemeal.

Rosenstein-Rodan’s argument rested on two economic concepts that would become central to development theory. The first was complementarities: the returns to one investment depend on whether other investments are made simultaneously. A shoe factory is more profitable if there is a steel factory (whose workers buy shoes) and a textile factory (whose workers also buy shoes) and a railroad (which transports shoes to distant markets). The investments are complementary — each one raises the returns to the others.

The second was increasing returns to scale. Modern industrial production has fixed costs — machinery, infrastructure, training — that make it unprofitable below a certain output level but highly profitable above it. In a small, poor economy, no individual industry can reach the scale needed for profitability. But a coordinated investment program can push many industries past the threshold simultaneously, launching the economy onto a growth path.

Nurkse and the Vicious Circle

Ragnar Nurkse, an Estonian-born economist at Columbia University, extended and refined Rosenstein-Rodan’s argument in his 1953 book Problems of Capital Formation in Underdeveloped Countries. Nurkse described the predicament of poor countries as a “vicious circle of poverty.” Low income leads to low savings, low savings leads to low investment, low investment leads to low productivity, and low productivity leads back to low income. Each link in the chain reinforces the others, trapping the economy at a low level.

Nurkse’s contribution was to emphasize the demand side of the vicious circle. Even if a country could mobilize savings (through foreign aid, forced saving, or inflation), the question remained: what should the savings be invested in? If the domestic market is small, investment in any single industry faces limited demand. Nurkse’s solution was “balanced growth” — investment spread across many industries simultaneously, so that each industry’s output creates demand for the others. This was Rosenstein-Rodan’s complementarity argument repackaged as a development strategy.

Nurkse was careful to note that balanced growth did not mean investing in everything at once. It meant investing in industries whose outputs were consumed by the workers in other industries — a pattern of complementary investment that would create a virtuous circle of demand and supply. The strategy required planning, because the market signals that would guide private investment were absent: the demand for shoes would not appear until the steel factory was built and its workers had income to spend.

Hirschman’s Counterargument: The Strategy of Unbalanced Growth

Albert Hirschman — German-born, European-educated, a veteran of the French Resistance and the Marshall Plan — came to development economics through direct experience in Colombia in the 1950s, where he served as an economic advisor. His 1958 book The Strategy of Economic Development was both a theoretical treatise and a polemic against the big push.

Hirschman’s objection was not that Rosenstein-Rodan’s logic was wrong. The complementarities were real, the coordination problem was real, and the poverty trap was real. His objection was that the big push was a fantasy. It assumed that underdeveloped countries possessed the one resource they most conspicuously lacked: the ability to plan, coordinate, and implement a massive, simultaneous investment program. If a country had the state capacity, administrative skill, and political will to execute a big push, it would not be underdeveloped in the first place.

In place of balanced growth, Hirschman proposed unbalanced growth. The idea was counterintuitive: instead of trying to develop everything at once, development should proceed by deliberately creating imbalances — investing heavily in a few sectors while leaving others underdeveloped. The resulting bottlenecks, shortages, and pressures would create incentives and political pressure for further investment, which would in turn create new imbalances and new pressures. Development was not a coordinated leap but a chain reaction of disequilibria, each one provoking the next.

The key concept was linkages. Hirschman distinguished between backward linkages (an industry creates demand for its inputs, stimulating upstream investment) and forward linkages (an industry creates supply of inputs for downstream industries, stimulating their development). A steel plant, for example, has strong backward linkages (demand for iron ore, coal, and transport) and strong forward linkages (supply of steel for construction, machinery, and consumer goods). Investing in steel — even if the economy is “not ready” — creates pressures that induce investment in related sectors.

Hirschman was explicit about why he preferred imbalance to balance. Balanced growth, he argued, was a conservative strategy that spread resources thinly across many sectors, reducing the pressure on any one of them to develop. Unbalanced growth concentrated resources where they would create the strongest inducement for further investment. The shortages and bottlenecks that balanced-growth advocates saw as costs, Hirschman saw as signals — price signals, profit signals, political signals — that would mobilize resources and decision-making capacity that the balanced-growth approach assumed away.

He also introduced the concept of the “hiding hand” — a deliberate inversion of Adam Smith’s invisible hand. Entrepreneurs and planners, Hirschman argued, routinely underestimate the difficulties of a project. If they knew in advance how hard it would be, they would never start. But having started, they are forced to solve problems they did not anticipate, and in the process they develop capabilities they did not know they had. The hiding hand converts ignorance and overconfidence into creative problem-solving. This is a strikingly different view of the relationship between error and progress than the one offered by rational choice theory.

The Planning Debate of the 1950s and 1960s

The big push vs. unbalanced growth debate was part of a larger argument about the role of planning in development. In the 1950s and 1960s, the dominant view in development economics was that poor countries needed comprehensive economic planning — not Soviet-style central planning (though some advocated that), but indicative planning of the kind practiced by France, Japan, and South Korea. The state would identify priority sectors, allocate investment, coordinate infrastructure, and manage trade to protect infant industries.

This view was shared, in different forms, by Rosenstein-Rodan, Nurkse, W. Arthur Lewis, Gunnar Myrdal, and Raul Prebisch. They disagreed on many specifics, but they agreed that the market alone could not deliver development. The coordination failures were too severe, the market signals too weak, and the private sector too small and too risk-averse to drive industrialization without state leadership.

Hirschman accepted much of this diagnosis but was skeptical of the planning apparatus. His experience in Colombia had shown him that plans rarely survived contact with reality. Implementation was the binding constraint, not analysis. The question was not “what should we do?” but “what can we actually get done, given the political constraints, bureaucratic limitations, and informational gaps that characterize every government in the developing world?”

On the other side of the debate, P.T. Bauer and the early neoclassical development economists argued that planning was the problem, not the solution. Government intervention distorted prices, created opportunities for corruption, and suppressed the entrepreneurship that drives growth. The market, if left alone, would allocate resources efficiently. This view gained influence in the 1970s and 1980s and became the basis for the Washington Consensus.

Murphy, Shleifer, and Vishny: Formalizing the Big Push

For three decades after Rosenstein-Rodan’s paper, the big push remained an intuition — a persuasive story without a formal model. Mainstream economists dismissed it as vague and unfalsifiable. The increasing returns and coordination failures at its core were difficult to model with the tools of general equilibrium theory, which assumed perfect competition and constant returns.

In 1989, Kevin Murphy, Andrei Shleifer, and Robert Vishny published “Industrialization and the Big Push,” which gave the big push its first rigorous formalization. Their model featured an economy with many sectors, each of which could be operated with a traditional, constant-returns technology (think cottage industry) or a modern, increasing-returns technology (think factory). The modern technology was more productive but required a fixed cost of adoption. Whether it was profitable to adopt the modern technology in any one sector depended on how many other sectors had adopted it — because the workers in modern sectors earned higher wages and spent those wages across the economy, creating the demand that justified the modern technology.

The model had multiple equilibria. In the “bad” equilibrium, no sector industrializes because the demand from other sectors is too low to justify the fixed cost. In the “good” equilibrium, all sectors industrialize because the aggregate demand from industrialized workers makes the modern technology profitable in each sector. The economy can be stuck in the bad equilibrium even though the good equilibrium is available — a poverty trap in the formal sense.

Murphy, Shleifer, and Vishny’s paper showed that Rosenstein-Rodan’s intuition was logically coherent: coordination failures can trap an economy in poverty, and a coordinated investment push can move the economy to a better equilibrium. But the paper also exposed the fragility of the argument. The existence of a poverty trap depends on the strength of the complementarities and the size of the increasing returns. If the complementarities are weak or the increasing returns are modest, there is no trap and no need for a big push.

Modern Echoes: Industrial Policy, SEZs, Duflo vs. Sachs

The big push vs. unbalanced growth debate has not been settled. It has been transformed into contemporary arguments about industrial policy, foreign aid, and the design of development interventions.

The industrial policy revival. After decades of neoliberal skepticism about state intervention, industrial policy has made a comeback. Economists like Dani Rodrik, Justin Lin, and Mariana Mazzucato have argued that successful development has almost always involved active state promotion of specific industries — through subsidies, trade protection, infrastructure investment, and technology transfer. The debate echoes the 1950s: advocates of industrial policy invoke coordination failures and increasing returns; critics invoke government failure, corruption, and rent-seeking.

Special economic zones. China’s spectacular growth after 1980 was driven in part by special economic zones (SEZs) — geographically limited areas with favorable tax treatment, infrastructure, and regulatory frameworks. SEZs are a compromise between the big push (concentrated investment in a limited area to overcome coordination failures) and Hirschman’s unbalanced growth (deliberately creating a pocket of modernity that generates linkages to the rest of the economy). The success of SEZs in China, and their more mixed record elsewhere (in sub-Saharan Africa and South Asia), suggests that the big push logic is valid but context-dependent.

The Sachs-Easterly-Duflo debate. Jeffrey Sachs’s Millennium Villages Project was an explicit attempt at a big push at the village level: simultaneous investment in health, education, agriculture, and infrastructure in a dozen African villages, with the goal of demonstrating that coordinated intervention could break the poverty trap. William Easterly criticized the project as repeating the mistakes of midcentury planning — overestimating what external coordination could achieve and underestimating the importance of local institutions and incentives.

Esther Duflo and the randomized controlled trial (RCT) movement offered a third path: instead of big pushes or grand theories, evaluate specific interventions one at a time, rigorously, and scale the ones that work. This approach — piecemeal, evidence-based, skeptical of grand narratives — is closer in spirit to Hirschman’s pragmatism than to Rosenstein-Rodan’s vision. But critics of the RCT approach argue that it misses the complementarities and general equilibrium effects that the big push literature emphasizes. A bed net that prevents malaria in one village may have no effect on the structural transformation of an economy. The question is not whether individual interventions work but whether they add up.

What the Debate Teaches

The big push vs. unbalanced growth debate is, at bottom, a disagreement about the binding constraint on development. If the binding constraint is coordination — the inability of individual actors to internalize the benefits of simultaneous investment — then the big push is the right strategy. If the binding constraint is capacity — the inability of governments and institutions to plan and execute complex programs — then unbalanced growth, with its reliance on market-generated pressures and decentralized problem-solving, is more realistic.

Both constraints are real. Which one binds more tightly depends on the country, the period, and the sector. Countries with strong state capacity (South Korea in the 1960s, China after 1978) have executed big-push-style strategies with remarkable success. Countries with weak state capacity have often wasted the resources devoted to big pushes, producing white-elephant projects, corruption, and debt.

Hirschman’s deepest insight was not about economics but about epistemology. Development, he argued, is a process that cannot be fully planned because it cannot be fully understood in advance. The information needed to design a big push — which industries to invest in, at what scale, in what sequence — is not available to planners. It is generated by the process of development itself, through the errors, surprises, and adaptations that occur when people grapple with concrete problems. The role of the economist is not to design the optimal plan but to understand the mechanisms by which societies solve problems — and to avoid getting in the way.

This is a humbler vision of development economics than the one that Rosenstein-Rodan and Nurkse proposed. It is also, arguably, a more honest one. The history of development over the past seventy years is full of planned pushes that failed and unplanned breakthroughs that succeeded. Understanding why requires not just models of coordination failure but models of institutional capacity, political will, and the messy, nonlinear dynamics of economic change.