The Growth Story vs. the Inflation Ledger: Reading the GTP Years

Two ledgers, one decade

If you had only the GDP figures, the story of Ethiopia from 2010 to 2020 would read as one of the great development successes of the twenty-first century. The country grew at official rates between 8 and 11 percent per year for most of the decade, doubled its road network, built the largest hydropower dam in Africa, expanded its airline into the dominant continental carrier, opened a dozen industrial parks, electrified a railway to the sea, lifted millions out of poverty by every available headcount measure, and was held up by the World Bank and the African Development Bank as a model for the rest of the continent.¹

If you had only the inflation figures and the parallel exchange rate, the same decade would read as one of accumulating monetary disorder. Inflation breached 30 percent in 2008 and again in 2011, settled at 8–10 percent through the middle of the decade only by aggressive use of price controls, and resurged through 2017–18 as the underlying monetary pressure could no longer be contained. The parallel exchange premium, briefly compressed by the 2010 devaluation, widened again through the decade and reached over 50 percent by the time Abiy Ahmed took office.² The Eurobond issued in December 2014 — the country’s first sovereign borrowing in international capital markets — was, by 2018, already trading at distressed levels.³

These two ledgers describe the same decade. Both are honest reports of what the data show. The question this article addresses is: how can both be true at once, and what does the gap between them tell us about the model that was being run?

The argument is that the GTP years exhibited, in unusually pure form, the central trade-off of the developmental-state monetary model: real growth and infrastructure investment delivered in exchange for accumulating macroeconomic vulnerabilities that would, eventually, demand to be paid off. The growth was real. The vulnerabilities were also real. The post-2018 reckoning is, in significant part, the bill for the GTP years coming due. To understand the bill, you have to look at what the spending bought, and what it cost.

The plans

The Growth and Transformation Plan I (GTP I) ran from 2010/11 to 2014/15. Its successor, GTP II, ran from 2015/16 to 2019/20.⁴ The Plans were articulated by the Ministry of Finance and Economic Development under Meles Zenawi and, after his death in 2012, under his successor Hailemariam Desalegn. They were detailed, ambitious, and intellectually serious documents — each running to several hundred pages, each setting sector-by-sector targets across the entire economy, each grounded in the developmental-state framework Meles had articulated.

The headline targets were striking. GTP I committed to maintaining double-digit GDP growth, doubling agricultural production, expanding manufacturing output by a factor of 2.5, doubling the road network, doubling electricity generation, and reducing poverty headcount by half.⁵ GTP II raised the targets further. Meeting them required investment at scales the country had not previously attempted, financed at scales the domestic economy could not generate.

The financing strategy had three legs. The first was domestic resource mobilisation — the financial-repression machinery described in article 12, plus pension funds, plus state-enterprise retained earnings. The second was concessional external borrowing — World Bank IDA loans, African Development Bank facilities, bilateral concessional credits from China and a few other partners. The third, new for the period, was commercial external borrowing: the December 2014 Eurobond issuance, the Chinese commercial loans on harder terms than the early concessional deals, and the various supplier credits and Sinosure-backed facilities that financed specific projects. Article 17 examines this third leg in detail.

What is worth seeing about the financing strategy is that it required a particular set of macroeconomic conditions to remain coherent: GDP growth had to remain above the cost of debt service; the exchange rate had to remain stable enough that dollar-denominated debt did not become unmanageable in birr terms; the export base had to grow fast enough to generate the foreign-currency cash flows needed for debt service. The strategy was, in effect, a bet — a bet that the developmental investment would produce the growth and exports that would, in time, validate the borrowing.

The bet was not unreasonable when made. It was, however, a bet. And by 2018, the evidence was accumulating that the bet was not working out.

What the growth was

Before turning to the bet’s eventual cost, it is worth taking seriously what the growth actually was. Ethiopian official GDP grew at rates averaging approximately 10 percent per year through GTP I and at slower but still elevated rates through GTP II.⁶ Three observations about this growth are worth making.

First, the growth was real but probably overstated. The IMF’s Article IV consultations of the period repeatedly flagged data-quality concerns, particularly around the construction-sector deflator (which appeared to understate inflation, thereby overstating real growth in a sector that was, on the official figures, growing rapidly).⁷ Independent analyses by the African Development Bank and academic researchers have suggested that the true growth rate may have been 2–3 percentage points lower than the official figure across the decade.⁸ Even at the lower figure, however, the growth was substantial by any African or developing-country comparison.

Second, the growth was infrastructure-heavy. A large share of recorded growth came from public investment — roads, dams, railways, telecoms, urban construction — rather than from private-sector expansion. This is what the GTP framework intended: state-led capital formation that, in time, would generate productivity gains in the broader economy. The pattern is consistent with the East Asian developmental-state historical record.

Third, the growth was real-economy productive in important sectors and rentier in others. Agricultural productivity gains were real (though smaller than the official figures suggested); manufacturing expansion was real (though concentrated in a small number of export-oriented sectors); the rapid expansion of Ethiopian Airlines and the Ethiopian Shipping Lines was real and reflected genuine productive capacity. Other elements of the growth — the inflation of urban construction values, the expansion of state-enterprise loan books that funded each other, the round-tripping of credit between state institutions — were less unambiguously productive.

The honest assessment is that the GTP-era growth was meaningfully positive, meaningfully real, and meaningfully smaller than the headline figures suggested. The infrastructure assets that were built remain assets. The poverty reductions that occurred are not artefacts of accounting. But the model was operating closer to the limit of what its macroeconomic foundations could sustain than the official figures admitted.

What the inflation was

The other ledger tells the side of the story the growth figures suppressed.

The 2008 inflation episode is the most revealing. Headline CPI inflation in Ethiopia reached 64 percent year-on-year in mid-2008 — among the highest non-hyperinflationary rates in the world that year.⁹ The episode was, in the official narrative, attributed to global commodity price shocks: oil and food prices had risen sharply through 2007 and into early 2008. This was part of the story. It was not the whole story.

The whole story includes the fact that the broad money supply (M2) in Ethiopia had been growing at rates of 25–30 percent per year through 2007 and 2008, against nominal GDP growth in the 15–20 percent range.¹⁰ In Friedman’s framework, this excess money creation was producing inflation that was being held down, partially, by price controls — until the global commodity shock removed the ability of the controls to suppress the underlying pressure. The shock was real but it was triggering something the monetary stance had been building. A country running tighter monetary policy would have absorbed the same shock without the same inflation.

The 2011 episode followed the same pattern. Inflation reached 39 percent year-on-year in mid-2011, again attributed in official accounts to external shocks (drought, commodity prices, regional disruption) but again coinciding with a period of aggressive monetary expansion to fund GTP-era investment.¹¹ By 2012–14, inflation was brought down through a combination of tightened monetary policy (the introduction of the 27 percent rule and other measures described in article 12), price controls on a wider range of goods, and the introduction of subsidies on fuel and certain food items. The official rate fell into the high single digits by 2014.

But the underlying pressures did not go away; they were displaced. The parallel exchange rate widened. The current-account deficit grew. External debt accumulated. The fiscal accounts, on the consolidated definition that included state-enterprise losses, showed deficits that the headline figures concealed. The system was, in effect, holding the inflation lid down by sitting on it harder — by tightening capital controls, expanding price controls, deepening the financial-repression machinery. The mechanism worked, for a while. It was not durable.

The 2014 Eurobond as turning point

The December 2014 issuance of Ethiopia’s first sovereign Eurobond — $1 billion at 6.625 percent, ten-year maturity, due December 2024 — was the moment the GTP-era strategy reached its high-water mark and began its descent.¹² The bond was oversubscribed by international investors who, in the post-2008 yield-starved environment, were willing to take African frontier risk in exchange for the coupon. The Ethiopian Ministry of Finance pointed to the issuance as validation of the developmental model: international markets were prepared to lend the country money at reasonable rates, with no political conditions, to finance ambitious infrastructure plans.

What the Eurobond also did, in less-celebrated ways, was commit Ethiopia to a foreign-exchange obligation that could not be defaulted on without serious consequences. Concessional debt to the World Bank or to bilateral lenders can, in practice, be restructured with relatively limited reputational cost; sovereign Eurobonds, held by mutual funds and pension funds in the major financial centres, carry a different kind of cost when defaulted on. The 2014 issuance moved Ethiopia into a category of borrowing where the discipline of the market would, eventually, be tested.

By 2018, the test was beginning. The Eurobond was trading at significant discounts to par, reflecting investor concern about Ethiopia’s debt sustainability. The IMF Article IV of October 2018 explicitly raised concerns about debt distress.¹³ The Tigray conflict had not yet begun, but the macroeconomic vulnerabilities that the conflict would later exploit were already visible to anyone reading the data carefully.

The honest counter-reading

The strongest version of the GTP-era defence is that the model was, by any reasonable measure, working on its own terms through about 2015, and that what derailed it was a set of exogenous shocks — Meles’s death, the 2015–16 protest movements in Oromia and Amhara, the Tigray war that began in 2020, the COVID-19 pandemic — for which the developmental-state architecture cannot fairly be blamed. If those shocks had not occurred, the argument runs, the growth-and-debt strategy might have generated the export growth needed to validate the borrowing, and the bill that came due in 2023–24 might never have come due.

The argument has some force. Some of the shocks were genuinely exogenous, and a country that runs a high-investment growth model is inherently more exposed to shock than a country running a more conservative one. The counterfactual — what would the Ethiopian trajectory have looked like without Tigray, without COVID, without the political turbulence — is genuinely uncertain.

The series’s reply is twofold.

First, the political shocks were not as exogenous as the framing suggests. The Oromia and Amhara protests of 2015–18 were responses to specific policies of the EPRDF government — particularly the Addis Ababa Master Plan that triggered the initial Oromo protests, and the ethnic-federalist structure that made the protests politically channelable. The Tigray war was the result of a political decision by the post-EPRDF government, made by Abiy Ahmed and Isaias Afwerki, not an external shock visited on the country. Treating these as exogenous removes the political agency that produced them, which is dishonest accounting.

Second, even if the model had been allowed to run undisturbed by political shocks, the macroeconomic vulnerabilities — the accumulating external debt, the widening parallel premium, the persistent inflation, the over-extended state-bank loan books — were building toward an eventual reckoning regardless. The GTP-era growth strategy required continuously validating debt that was being accumulated faster than export earnings were growing. A model that requires perpetual continuation of favourable external conditions to remain sustainable is, in the relevant sense, not sustainable. The question of when the reckoning would come, not whether, was the open question.

The next decade

The GTP-era model collapsed in stages between 2018 and 2024. The reform-rhetoric phase (examined in article 19) attempted to manage the contradictions without fundamental change to the monetary architecture. The Tigray war (article 20) intensified the macroeconomic pressures. The 2023 Eurobond default (article 17) was the visible breaking point on the external-debt side. The 2024 float (article 21) was the visible breaking point on the exchange-rate side. The cumulative reckoning has consumed the post-Meles decade.

The series’s argument, restated: the GTP years were not a betrayed success but an unsustainable model that produced real but bounded benefits at real but accumulating costs. The growth ledger is true. The inflation ledger is true. Reading them together, in honest accounting, the model did not deliver what its advocates expected, and the cost has been borne disproportionately by the population least able to defend itself — the Ethiopian saver, the Ethiopian worker, the Ethiopian poor — through the channels articles 12, 13, and 23 examine in detail.

The next article opens Part IV with the symptom that was, across the entire EPRDF and post-EPRDF period, the most visible expression of the underlying disorder: the chronic dollar shortage.

---

Footnotes

1. World Bank, Ethiopia: Economic Update — Laying the Foundation for Achieving Middle Income Status (World Bank, 2015); African Development Bank, Ethiopia Country Strategy Paper 2016–2020 (AfDB, 2016). ↩

2. World Bank, “Inflation, consumer prices (annual %) — Ethiopia,” World Development Indicators. For the parallel premium across the decade, see Alemayehu Geda, “Ethiopia’s Economic Performance During the GTP Years,” Ethiopian Journal of Economics 27, no. 2 (2018). ↩

3. Trading Economics, Ethiopia government bond yield series, 2014–2018. ↩

4. Ministry of Finance and Economic Development, Growth and Transformation Plan I, 2010/11–2014/15 (MoFED, 2010); MoFED, Growth and Transformation Plan II, 2015/16–2019/20 (MoFED, 2015). ↩

5. MoFED, GTP I, executive summary. ↩

6. World Bank, “GDP growth (annual %) — Ethiopia,” World Development Indicators. ↩

7. International Monetary Fund, “Ethiopia: 2014 Article IV Consultation,” IMF Country Report 14/303 (October 2014); IMF Country Report 16/322 (October 2016). ↩

8. Alemayehu Geda, “Ethiopia’s Economic Performance During the GTP Years”; Stefan Dercon and Catherine Porter, “A Poor Life? Chronic Poverty and Downward Mobility in Rural Ethiopia,” in The Oxford Handbook of the Ethiopian Economy (Oxford University Press, 2019). ↩

9. World Bank, “Inflation, consumer prices (annual %) — Ethiopia,” 2008 figure. ↩

10. IMF Country Report 14/303, money supply growth data; National Bank of Ethiopia, Annual Reports 2007/08, 2008/09, 2009/10. ↩

11. International Monetary Fund, “Ethiopia: 2011 Article IV Consultation,” IMF Country Report 11/348 (December 2011). ↩

12. Federal Democratic Republic of Ethiopia, Eurobond Prospectus, December 2014. The 6.625 percent coupon and December 2024 maturity are documented in subsequent default-related coverage; see The EastAfrican, “Ethiopia bondholders’ $1bn debt restructuring talks collapse,” October 2025. ↩

13. International Monetary Fund, “Ethiopia: 2018 Article IV Consultation,” IMF Country Report 18/354 (December 2018). ↩