Meles Zenawi as Monetary Architect: The Developmental State in Its Own Words

Taking him seriously

This series has, so far, treated the EPRDF’s monetary architecture as a problem to be diagnosed. That is the editorial position the series takes, and it will return to it. But before the diagnosis can be conclusive, the architecture has to be examined at its strongest — in the voice of the person who built and defended it, on the grounds he gave for building it. Meles Zenawi was, by some distance, the most intellectually serious head of government Ethiopia produced in the twentieth century, and the model he constructed was not an accident of expediency. It was a deliberate intellectual project. To dismiss it without engaging it is to argue against a strawman.

This article does what the rest of the series will not do: it states Meles’s case at his own preferred strength, in his own framing, with the evidence he would have cited. Only after that does it return to the critique. The reader who finishes the article unpersuaded by Meles is welcome to continue reading the series. The reader who finishes it persuaded is welcome to put it down and read Meles’s own writings instead. Either response is honest. What is not honest is to dismiss the developmental-state case without having met it.

The diagnosis

Meles’s case begins, in his 2006 essay African Development: Dead Ends and New Beginnings, with a diagnosis of African economic failure that is empirical rather than ideological.¹ Sub-Saharan Africa, by 2000, had experienced a quarter-century of decline relative to every other major region of the world. GDP per capita in many African countries was lower in real terms than it had been at independence. The neoliberal prescription that had been imposed on Africa through the IMF and World Bank structural adjustment programmes of the 1980s and 1990s had produced, on its own terms, failure. Growth rates were low. Industrialisation had been reversed in many countries. Capital had flowed out, not in. The promised benefits of market liberalisation had not materialised.

This is, importantly, an empirical observation rather than a theoretical claim. It does not depend on rejecting the theoretical apparatus of the Washington Consensus. It depends only on observing that, when the theoretical apparatus was applied, the predicted results did not appear. Meles’s move, characteristic of him, was to ask why — and to insist that the answer had to be sought in the specific characteristics of African economies, not in the universal claims of textbook economics.

The diagnosis he reached was that the Washington Consensus assumed institutional conditions that did not obtain in most African countries: deep capital markets that could finance long-horizon investment, competitive private sectors capable of replacing state-led activity, regulatory states capable of disciplining market actors, and political institutions stable enough to make the painful early-stage costs of reform politically sustainable. None of these conditions held in sub-Saharan Africa in the 1980s, and the Washington Consensus prescription, applied to economies without these foundations, produced what Meles called “deindustrialisation, agricultural stagnation, and the entrenchment of patron-client politics.”²

The diagnosis is not strawman. It is, in substantial part, factually correct. The structural adjustment era in Africa did not produce the growth its advocates promised; the institutional conditions for market-led growth were not in fact present in most African economies; and the social costs of the adjustment programmes were real and unevenly distributed. The argument that something else had to be tried because this had been tried and failed is, on the empirical record, defensible.

The prescription

What Meles proposed instead was the “democratic developmental state” — a state that consciously directed the economy toward industrialisation, infrastructure investment, and structural transformation, while seeking to maintain political legitimacy through electoral institutions and broad-based growth rather than through pure repression.³ The model he had in mind was East Asian: South Korea under Park Chung-hee, Taiwan under the KMT, Singapore under Lee Kuan Yew, and (with caveats) China under Deng Xiaoping. Each of these states had done what neoliberal theory said could not be done: industrialised rapidly, lifted hundreds of millions out of poverty, and transformed the structure of their economies — all under regimes of substantial state direction, deliberate market distortion, and (in most cases) authoritarian political control.

The Ethiopian application of this model, as Meles articulated it across multiple speeches and writings, had specific monetary components. Five are worth naming because they directly shape the policies this series has been critiquing.

First, directed credit. The state should ensure that credit flows to priority sectors — industry, infrastructure, export agriculture — even when market interest rates would direct it elsewhere. The mechanism was the state-owned banking system, supplemented by reserve requirements and forced bond purchases on private banks that would, in effect, redirect deposits toward state priorities.

Second, administered interest rates. Real interest rates should be low or negative to subsidise investment in priority sectors and to fund the state’s borrowing at affordable cost. This was a deliberate choice, defended as a transitional measure that would be relaxed once industrialisation had progressed sufficiently.

Third, capital controls. The capital account should remain closed to portfolio flows, with foreign direct investment permitted on a project-by-project basis. The reasoning was that open capital markets in a small developing economy generate volatile flows that disrupt long-horizon investment planning; the closed capital account was a deliberate restriction designed to preserve policy space.

Fourth, exchange-rate management. The exchange rate should be managed by the central bank, with the level chosen to support export competitiveness and the stability of the bilateral rate against major trading partners. Floating the currency would be premature; doing so before the export base was sufficiently diversified would expose the economy to unmanageable terms-of-trade shocks.

Fifth, monetary support for fiscal expansion. The central bank should be available to finance government deficits where those deficits funded productive infrastructure investment. The orthodox prohibition on monetary financing of deficits, Meles argued, assumed that all deficits were consumption-oriented; for an investment-oriented developmental state, the prohibition was misplaced.⁴

Taken together, these five components describe the EPRDF’s actual monetary architecture from 1995 through the end of Meles’s premiership in 2012 — and, with some modifications, well beyond. The architecture this series has been critiquing is, in this sense, the implementation of Meles’s intellectual project rather than its perversion.

The evidence he could cite

What is worth noting is that Meles had a body of empirical evidence on his side, by the standards of the GTP-era data, that supported the case for the developmental state. Between 2004 and 2012, Ethiopian official GDP grew at rates averaging 10 to 11 percent per year — among the fastest sustained growth rates in the world for a non-oil-producing developing country.⁵ Poverty headcount rates fell from roughly 45 percent in 2000 to roughly 30 percent by 2011.⁶ Primary-school enrolment rose from 30 percent in 1995 to over 80 percent by 2010.⁷ Infant and maternal mortality fell substantially. The road network doubled. Major infrastructure projects — the Gilgel Gibe hydropower complex, the early stages of the Grand Ethiopian Renaissance Dam, the Addis–Djibouti railway, the Ethiopian Airlines expansion — were planned and largely delivered during the Meles years.

These are not negligible accomplishments. By the metrics that any honest accounting would use for human development — life expectancy, education, infant mortality, access to electricity, access to safe water — Ethiopia made measurable, substantial progress under the Meles-era developmental state. The progress was real even if the official GDP figures were inflated, even if the income distribution was uneven, even if the political costs were high.

Meles could, and did, point to this record as vindication of his model. The Davos 2012 address, his last major international intervention before his death, included the line that has been quoted most often by his admirers: that the African neoliberal experiment had “failed” and that the developmental state was producing demonstrably better outcomes.⁸ The audience he was speaking to — Western corporate executives and political leaders attending the World Economic Forum — was the same audience that had championed structural adjustment for three decades. Meles was telling them, in their own conference centre, that their model had not worked and that his had.

The monetary case at full strength

Now state the monetary case at its strongest, in the framing Meles would have used.

The orthodox claim — Friedman’s claim, articulated at length in earlier articles of this series — is that monetary expansion above the rate of output growth produces inflation, and that the resulting inflation is, on net, harmful even when the cause it funds is meritorious. This is true in the steady state, Meles would say. But the steady state is not the relevant frame for a low-income economy attempting to industrialise. In the transition — the period during which the economy is being structurally transformed, during which new industries are being established, during which capital accumulation is required at rates faster than domestic savings can fund — some monetary expansion is necessary to bridge the gap.

The argument runs as follows. In a closed, low-saving economy, the constraint on investment is the available pool of credit. If the state expands credit through the central bank to finance infrastructure investment that will, in time, generate the productive capacity to validate the monetary expansion, the resulting inflation is a temporary cost paid for a permanent benefit. The infrastructure, once built, produces real output that re-anchors the price level. The argument is, in essence, that monetary policy should be conducted with reference to the long-run output potential rather than the short-run output level — that financing the long-run potential through temporary inflation is a defensible trade.

This is, in compressed form, the argument Meles would have made — and that his Ministry of Finance technocrats did, in fact, make in their internal documents and in their negotiations with the IMF.⁹ It is not a fringe view. It has serious advocates in the academic literature: Justin Lin’s “new structural economics” makes a related case; the Cambridge tradition associated with Ha-Joon Chang, Robert Wade, and Mushtaq Khan makes versions of it; the entire literature on state-led growth in East Asia is, at root, an empirical argument for this proposition.¹⁰

The reply this series gives, and will return to in articles 18 and 26, is that the empirical record of the trade — the cost in inflation against the benefit in infrastructure — does not, on the Ethiopian data, support the case. The infrastructure that was built during the GTP years generated real benefits, but it did not generate the export earnings or the productivity gains needed to validate the monetary expansion. The result was not temporary inflation paid for permanent benefit; it was permanent inflation paid for partial benefit, with the bill coming due in the 2023 Eurobond default and the 2024 birr collapse. The trade did not work for Ethiopia. Whether it could have worked in principle is a different question. Whether it worked here is the question this series asks, and the answer is no.

What Meles got right

Even from the critical position this series takes, several things in Meles’s framework are worth conceding.

He was right that the Washington Consensus, applied without adaptation to African conditions in the 1980s and 1990s, did not produce the predicted results. The empirical record is unambiguous on this. The Bolivias and Polands of the structural-adjustment record had institutional features that most African economies lacked, and the prescription did not travel as smoothly as its advocates expected.

He was right that the state had a role to play in African economic development that the Washington Consensus undervalued — particularly in infrastructure, in education, in public health, in the provision of credit to sectors where the market would not provide it on its own. The Ethiopian state’s accomplishments in these areas were real.

He was right that the East Asian experience demonstrated that economic transformation under state direction is possible — that the Korean and Taiwanese and Chinese trajectories existed and produced real results. The historical record cannot be denied.

He was right that political institutions matter, and that an inappropriate sequencing of reforms (full electoral democracy before institutional capacity is built) can produce worse outcomes than a sequenced approach. The Ethiopian experience under Abiy Ahmed after 2018, examined in articles 19 and 20, makes the point, in different ways than Meles meant it.

What he got wrong, the series will argue, is the specific monetary claim: that the inflation generated by central-bank financing of state-led industrial investment is a worthwhile cost. The evidence from Ethiopia is that the cost has been larger than the benefit, that the inflation has been more persistent than the temporary-transitional framing predicted, and that the institutional architecture set up to deliver the prescription has been more durable than its advocates expected — surviving Meles’s death and into Abiy’s reform-rhetoric era without serious modification until 2024.

What this article is not doing

This article is not endorsing the developmental-state critique. It is not, even in this sympathetic framing, agreeing that the trade was worth making. It is doing one specific thing: refusing to caricature the position, refusing to attribute it to stupidity or corruption, refusing to dismiss it without engagement.

The reader who comes to this series wanting Friedman-as-vindicated and Meles-as-villain will not find that here. The case for the Meles approach is real. The Ethiopian record under his leadership is mixed but, in important respects, positive. The intellectual project he led was serious, and the policies he championed were defensible on grounds that any honest critic must engage rather than dismiss.

The series’s argument, which the rest of Part III and beyond will develop, is narrower and more specific. It is that the monetary leg of the developmental-state package has, on the Ethiopian evidence, not delivered what its advocates expected, that the inflationary costs have outweighed the developmental benefits, and that the institutional architecture set up to deliver the monetary policy has — by surviving past the point where its theoretical advocates would have expected it to be relaxed — become an obstacle to the next stage of Ethiopian development rather than its enabler.

That is a narrower claim than “the developmental state is wrong.” It is also a more honest one. The next article turns to the specific mechanism — financial repression — by which the monetary architecture funded the state’s priorities, and at what cost to the rest of the economy.

---

Footnotes

1. Meles Zenawi, “African Development: Dead Ends and New Beginnings,” 2006, https://yale.app.box.com/s/me9rosgyk19wsv03u6spbngzdc5b8wpd; substantially revised version published as the introductory chapter in Good Growth and Governance in Africa: Rethinking Development Strategies, edited by Akbar Noman et al. (Oxford University Press, 2012). ↩

2. Meles, “African Development: Dead Ends and New Beginnings.” ↩

3. For Meles’s articulation of the “democratic developmental state,” see his speeches at the 2010 and 2012 World Economic Forum sessions in Africa, and the analysis in Sarah Vaughan, “Revolutionary Democratic State-Building: Party, State and People in the EPRDF’s Ethiopia,” Journal of Eastern African Studies 5, no. 4 (2011): 619–640. ↩

4. This articulation of Meles’s monetary views is reconstructed from his published writings, his Davos speeches, and the academic literature on his model; see particularly Christopher Clapham, “The Era of Haile Selassie,” and “The Era of Meles Zenawi,” in The Horn of Africa: State Formation and Decay (Hurst, 2017). ↩

5. World Bank, “Ethiopia Overview,” https://www.worldbank.org/en/country/ethiopia/overview, growth statistics for 2004–2012. ↩

6. World Bank, Ethiopia: Poverty Assessment (World Bank, 2015), reporting headcount rates over the relevant period. ↩

7. World Bank Education Data, Ethiopia primary enrolment rates, https://data.worldbank.org/topic/education?locations=ET. ↩

8. Meles Zenawi, speech at the World Economic Forum on Africa, 2012; for the context, see Léonce Ndikumana and James K. Boyce, “Africa’s Odious Debts,” in Africa’s Odious Debts: How Foreign Loans and Capital Flight Bled a Continent (Zed Books, 2011). ↩

9. For the technical articulation of the Ethiopian developmental-state economic framework, see Ministry of Finance and Economic Development, Growth and Transformation Plan, 2010/11–2014/15 (MoFED, 2010); and the post-GTP analyses in Tegegne Gebre Egziabher, ed., The Political Economy of Ethiopia’s Economic Reform (Forum for Social Studies, 2017). ↩

10. Justin Y. Lin, New Structural Economics: A Framework for Rethinking Development and Policy (World Bank, 2012); Ha-Joon Chang, Kicking Away the Ladder: Development Strategy in Historical Perspective (Anthem, 2002); Robert Wade, Governing the Market: Economic Theory and the Role of Government in East Asian Industrialization (Princeton University Press, 1990); Mushtaq Khan, “Beyond Good Governance,” in Towards New Developmentalism: Market as Means rather than Master, edited by Shahrukh Rafi Khan and Jens Christiansen (Routledge, 2011). ↩