Editorial perspective · Part 26 of 28
The Birr and the Pretence of Knowledge · VI — Costs, comparisons, and conclusions
The Strongest Case for the Statist Path — Steelman and Rebuttal
Article 11 examined Meles Zenawi's intellectual project on its own terms, in his own voice. This article does the larger job: it presents the strongest version of the *developmental-state literature*'s defence of the kind of monetary arc…
An argument by Zef Telahun
This is an editorial perspective — signed opinion, not the site's neutral analysis. Factual claims are footnoted; the synthesis, emphasis, and judgement are the author's.
The argument the series owes the reader
Article 11 examined Meles Zenawi’s intellectual project on its own terms, in his own voice. This article does the larger job: it presents the strongest version of the developmental-state literature’s defence of the kind of monetary architecture Ethiopia ran, as articulated by its most serious living advocates, and then engages it on its strongest grounds. The advocates are Ha-Joon Chang of Cambridge, Joe Studwell whose How Asia Works has been the most influential popular treatment of the East Asian model, Robert Wade whose Governing the Market remains the canonical academic study of state-led development, and Justin Lin whose “new structural economics” is the most institutionally significant contemporary articulation of the position. The series’s editorial position is against their conclusion. The article’s job is to present their case at its full intellectual strength before saying why.
The case to be made is genuinely powerful. It has decades of empirical work behind it. It is supported by the historical record of East Asian development in ways that no honest critic can dismiss. The Ethiopian implementation of the developmental-state framework was, in many specific respects, not what these thinkers were actually advocating — the Ethiopian errors were errors, in some cases, against the developmental-state literature’s prescriptions rather than expressions of them. The article will say so.
What the article will also say is that, even taking the developmental-state literature at its strongest, the monetary leg of the package — the specific use of the central bank to finance industrial-policy expenditure — is the weakest part of the case, and the part that the most successful East Asian states actually did not implement in the form Ethiopia did. The series’s argument against Ethiopian monetary practice is, in this sense, consistent with rather than opposed to the strongest developmental-state literature. The argument is that Ethiopia did not implement the developmental-state model. It implemented something else and called it that.
The developmental-state case at its strongest
The strongest case for state-led development rests on five claims, each substantially supported by the historical record.
First, market failures in late development are pervasive and identifiable. Late-developing economies suffer from coordination failures (industries that would be profitable if they existed together but unprofitable in isolation), information failures (capital markets that cannot price new industries because no one has experience with them), and externalities (skills, supplier networks, technology spillovers) that the market alone does not internalise. The state has a legitimate role in addressing these failures through targeted interventions.
Second, the historical record of successful late development is overwhelmingly state-led. From Britain in the eighteenth century through the United States, Germany, Japan, South Korea, Taiwan, Singapore, and China, no country has achieved the transition from poor to rich through pure market mechanisms. Every successful catch-up has involved substantial state direction of investment, protection of infant industries, and active management of the trade regime. Ha-Joon Chang’s Kicking Away the Ladder documents this case with extensive evidence; the empirical claim is, on the evidence, hard to dismiss.1
Third, export discipline is the test that makes the state-led model work. Joe Studwell’s How Asia Works makes the most direct version of this argument: that what distinguished the successful East Asian developmental states (South Korea, Taiwan, Japan, China) from the unsuccessful ones (the Philippines, Indonesia for much of its history, many Latin American attempts) was that the successful ones imposed export discipline on the firms receiving state support.2 The firms had to compete in world markets and meet export targets; firms that failed were cut off from the subsidies. This export discipline provided the external standard that made the internal allocation of subsidies meaningful.
Fourth, land reform is the foundational step. Studwell again is most direct on this point: that successful late development has historically begun with substantial land reform that broke up the large estates of the traditional landlord class and distributed land to smallholders, with state support for productivity-enhancing investments. The East Asian successes all began this way. The Latin American failures, by contrast, retained the large-estate structure and tried to industrialise on top of it. The Ethiopian land reform of 1975 was, on this account, a correct step that places Ethiopia in the company of the successful East Asian cases on this dimension.
Fifth, the role of finance is to follow industrial policy rather than lead it. The developmental-state framework treats the financial sector as a tool of industrial policy rather than as an autonomous mechanism for allocating capital. The state-bank dominance, the directed lending, the financial-repression machinery — these were features the East Asian successes shared, and they were what allowed those states to fund their industrial transformation at the pace and scale required.
Each of these claims has genuine empirical support. The historical record supports the broad framework. The argument that purely market-led development is sufficient for late developers is, on the evidence, not credible. The developmental-state literature has, in this sense, won the broad argument about whether the state has a substantial role in late development.
Where the Ethiopian implementation differed
The series’s argument is that the Ethiopian implementation of the developmental-state framework was, in several specific dimensions, different from what the strongest developmental-state advocates would have prescribed. These differences matter for the evaluation.
The export-discipline element was missing. The state-favoured firms in Ethiopia — both state-owned enterprises and politically-connected private firms — received subsidised credit without the export-target requirement that the East Asian successes imposed. The DBE loans, the priority forex allocations, the various directed-lending instruments were not conditioned on the recipients’ export performance in a way that would have enforced market discipline. Studwell would, this article would suggest, regard this as a substantial deviation from the model he advocates. The Ethiopian developmental state was internally state-directed without the externally market-disciplined element that made East Asian state direction work.
The land reform did not include credit-extension to smallholders. The 1975 Ethiopian land reform broke up the large estates but, by maintaining state ownership and prohibiting land sale, did not extend the credit-enabling property rights to smallholders that the East Asian land reforms (in their later phases) did. South Korean and Taiwanese land reforms allowed smallholders to mortgage their land for productivity investments; Ethiopian land reform did not. The difference is examined in article 13. From the developmental-state literature’s own framework, this was a significant deviation.
The financial-repression machinery was not paired with developmental targeting. The state-bank dominance and directed-lending architecture that the East Asian developmental states used was, in their case, paired with explicit industrial-policy targeting that selected sectors based on credible analysis of comparative advantage and dynamic capability development. The Ethiopian system, particularly after the GTP era began, allocated subsidised credit based more on political-economic considerations than on industrial-policy analysis. The deviation from the East Asian pattern produced different results.
The monetary financing of fiscal deficits was, in important respects, not present in the East Asian success cases. South Korea, Taiwan, Singapore, and Japan all maintained relatively disciplined monetary policy throughout their developmental periods, with central banks that, while not formally independent in the modern sense, did not provide unlimited deficit financing. The Chinese case is more complex — substantial monetary expansion to fund state-enterprise investment did occur — but China’s central-bank lending was, for much of the relevant period, paired with binding capital controls that prevented the monetary expansion from producing the exchange-rate disorder Ethiopia experienced. The Ethiopian pattern of central-bank financing of deficits without the offsetting capital-control discipline produced the monetary disorder that the East Asian model would have avoided.
The article’s point is that the developmental-state model, properly implemented, does not require the specific monetary disorder Ethiopia experienced. The Ethiopian monetary disorder was, in significant part, the result of implementing the framework badly rather than of implementing the right framework. The strongest developmental-state advocates would, this article argues, recognise the Ethiopian implementation as a failure of their model rather than as a vindication of it.
What the rebuttal does not say
Three things this article is not claiming, because clarity about what is and is not being argued matters.
The article is not claiming that the developmental-state framework is wrong in principle. The framework has empirical support that no honest critic can dismiss. The argument that the state has a substantial role in late development is, on the evidence, broadly correct.
The article is not claiming that the Ethiopian development gains were illusory. The infrastructure was built, the poverty reductions occurred, the social services expanded. These gains were real and were achieved through the kind of state-led mechanisms the developmental-state framework prescribes.
The article is not claiming that pure market liberalisation would have produced better outcomes for Ethiopia. The 1980s African experience with structural-adjustment-style liberalisation is, on the evidence, not a compelling counterfactual to the GTP-era Ethiopian experience. The choice was never between Ethiopia’s developmental state and a successful market-led alternative.
What the article is claiming is narrower and more specific. The monetary architecture that Ethiopia operated was not the monetary architecture that the strongest developmental-state literature would prescribe. The deviations — missing export discipline, missing land-credit extension, missing capital-control discipline, missing industrial-policy targeting — produced specific costs that have, by 2024, accumulated to the point of the formal reckoning. A properly-implemented developmental-state model would have produced different monetary outcomes than the Ethiopian one did, and the difference matters.
The deeper question
The deeper question the article raises is whether the developmental-state model can, in principle, be implemented in conditions where the political-economy constraints article 25 examined are present. The East Asian successes were achieved by states that had, in specific ways, insulated their economic management from the patronage pressures the Ethiopian system has been subject to. The South Korean military government of Park Chung-hee had specific institutional arrangements that protected the Economic Planning Board from political-economic interference; the Taiwanese government under the KMT had similar protections; the Singaporean PAP government developed an administrative state with substantial autonomy from political patronage; the Chinese system under Deng Xiaoping and his successors developed analogous protections in specific sectors.
The Ethiopian political-economic system, as article 25 examined, has not developed these protections. The party-state-enterprise-banking complex that operates the developmental machinery is, in important respects, unable to insulate its industrial-policy decisions from the patronage logic that drives the broader political coalition. The monetary disorder is, in this sense, intrinsic to the way the developmental state has been operated in Ethiopia rather than incidental to it. A reform that wanted to preserve the developmental-state ambition while disciplining the monetary disorder would have to first solve the political-economy problem of how to insulate economic policy from patronage pressure. The article’s editorial position is that this is not impossible but is substantially harder than the standard developmental-state literature acknowledges.
The synthesis
What does this article, taken with article 11’s steelmanning of Meles’s own case, leave on the table?
It leaves an empirically-grounded acknowledgment that the developmental-state model has substantial intellectual merit and historical evidence behind it. It leaves an honest assessment that the Ethiopian implementation of the model was, in specific dimensions, a deviation from the strongest version of the prescription. It leaves a political-economy argument that the deviations were not accidental but were structurally produced by the Ethiopian political-economic configuration. And it leaves a path forward: the question is not whether to abandon the developmental-state ambition but whether to discipline it through monetary and institutional arrangements that the Ethiopian implementation, for political-economic reasons, has not provided.
This is, the series argues, the honest synthesis. It is not a refutation of the developmental-state literature. It is a critique of the specific Ethiopian implementation, made on terms the strongest developmental-state literature itself would recognise. The next article turns to what the discipline would look like — what sound money for Ethiopia would actually require, in concrete institutional terms.
Additional references engaged in this article: Robert Wade, Governing the Market: Economic Theory and the Role of Government in East Asian Industrialization (Princeton University Press, 1990); Justin Y. Lin, New Structural Economics: A Framework for Rethinking Development and Policy (World Bank, 2012); Dani Rodrik, One Economics, Many Recipes: Globalization, Institutions, and Economic Growth (Princeton University Press, 2007); Mushtaq Khan, “Beyond Good Governance,” in Towards New Developmentalism: Market as Means rather than Master, ed. Shahrukh Rafi Khan and Jens Christiansen (Routledge, 2011); Atul Kohli, State-Directed Development: Political Power and Industrialization in the Global Periphery (Cambridge University Press, 2004).
For the East Asian land-reform record and its connection to subsequent industrial development, see Studwell, How Asia Works, chapter 1; for the export-discipline argument, chapters 2 and 3.
For the comparative-development critique of the Ethiopian implementation, see Stefan Dercon, Gambling on Development: Why Some Countries Win and Others Lose (Hurst, 2022), chapter on Ethiopia.