War Finance Across Three Regimes: Paying for War with the Printing Press

A pattern that survives every regime change

Across the seventy-five years this series has covered — from the imperial monetary order of 1942 through the post-float architecture of 2026 — Ethiopia has fought, by my count, at least nine major military conflicts.¹ The Eritrean independence war (1961–91). The Ogaden war with Somalia (1977–78). The Tigrayan and Oromo insurgencies of the Derg years (1976–91). The Ethio-Eritrean border war (1998–2000). The 2006–09 intervention in Somalia. The various Ogaden insurgencies through the 2000s. The 2015–18 protest mobilisations that were, in some regions, militarised. The 2020–22 Tigray war. The 2023–present Amhara conflict and the persistent low-intensity insurgency in Oromia.

These wars had different causes, different durations, and different political settlements. What they share — across every regime that has fought them — is the method by which they were financed. Each war was paid for, in significant part, by central-bank advances to the government, by the suppression of import demand through forex rationing, by inflation that taxed the population in proportion to its holdings of birr, and by external borrowing whose terms reflected the wartime risk premium. The mechanism is constant. The wars are variable. The pattern is the pattern this whole series has been describing, applied to its most visible and morally consequential expression.

This article argues that the choice to monetise war debt is itself a policy choice, distinct from the choice to fight the war. The wars may, in some cases, have been unavoidable; the choice to print money to pay for them was not. Other countries facing comparable military pressures have, at various times, financed their wars through taxation, through formal debt issuance, through international cooperation, or through reductions in non-military expenditure. The Ethiopian choice has been, consistently, to print. The consequences belong on the ledger of each war’s costs alongside the casualties.

The mechanism

The mechanism by which war is financed through the central bank is worth setting out clearly because it is, in some respects, less obvious than it sounds.

A government at war spends more than its peacetime tax revenue can fund. The gap can be closed in four ways: by raising taxes; by issuing debt to domestic savers; by borrowing from abroad; or by borrowing from the central bank. The first option (raising taxes) is politically costly during war, because the population is already bearing the human cost. The second option (domestic debt issuance) requires a domestic capital market deep enough to absorb the issuance, which Ethiopia has not had. The third option (external borrowing) is constrained by the country’s external creditworthiness, which under wartime conditions is typically reduced rather than enhanced. The fourth option — central-bank borrowing — is therefore the residual, and it has been the option Ethiopian regimes have consistently chosen.

Central-bank borrowing for war has a specific monetary signature. The central bank credits the government’s account; the government spends the credit on military procurement, soldier pay, and emergency supplies; the credit becomes deposits in the recipient banks; the deposits become base money in the economy. The money supply expands by an amount roughly equal to the government’s wartime borrowing from the central bank. The inflation that results, with the standard lag described in article 6, taxes the population in proportion to its birr holdings.

This is not, in any meaningful sense, “free” financing. It is taxation through inflation, applied to a population that has not voted on it, distributed regressively across income classes (because lower-income households hold more of their wealth in cash and birr deposits than higher-income households do), and not visible on any budget that anyone can examine. It is the war tax that no one has to admit to imposing.

The Derg’s wars

The Derg’s wartime monetary record is the clearest case in Ethiopian history because the regime was at war almost continuously for its entire seventeen years in power, and the monetary consequences are documented in the data even where the official figures suppressed them.

Mengistu Haile Mariam’s 1984 announcement that 46 percent of GNP would be allocated to military spending was, even allowing for the conflation of GNP and government expenditure in the figure, a remarkable statement of priority.² In a country whose total tax revenue was 12–15 percent of GDP, allocating military expenditure at this level required financing that the budget could not provide. The financing came through central-bank advances on a scale that, by the late 1980s, had become the dominant source of base-money expansion in the economy.

The inflation that followed appeared in the data with the standard lag. The official inflation rate, suppressed by price controls, ran in the 15–20 percent range through the late 1980s. The true inflation rate, as measured by the parallel exchange rate and by open-market food prices, was probably in the 25–35 percent range. The wars in Eritrea and Tigray were, in the most precise causal sense, being paid for by the inflation that was destroying the purchasing power of Ethiopian wages, pensions, and savings.

This is the pattern in its purest form. The Derg fought wars it had partially chosen to fight (the Eritrean conflict could have been resolved diplomatically; the Tigrayan insurgency was a response to specific regime policies), it financed them through the central bank because it had foreclosed other financing options, and the population — including the populations the regime claimed to represent — paid the war tax through inflation that ate their real incomes year by year.

The EPRDF’s wars

The EPRDF inherited Mengistu’s wars unfinished but also fought new ones. The 1998–2000 border war with Eritrea (article 10) was, by some measures, more expensive in concentrated period than any single Derg conflict. Ethiopian military expenditure rose from approximately $95 million in 1998 to several hundred million per year through the war.³ The expenditure was financed through a combination of central-bank advances, emergency tax measures, and accelerated draw-down of external borrowing facilities.

The monetary consequences were visible in the late-1990s and early-2000s data. Inflation rose from the high single digits of the mid-1990s to double-digit rates during and after the war. The parallel exchange premium, which had narrowed sharply after the 1992 devaluation, widened again. The current-account deficit ballooned. The new regime’s earlier monetary achievements — the relative stability of 1992–97 — were partially undone by the wartime borrowing.

The EPRDF’s later interventions in Somalia (2006–09) and its sustained counter-insurgency operations in the Ogaden through the 2000s were smaller fiscal commitments than the Eritrean war, but they contributed to the persistent pattern of military spending that the developmental-state framework could not openly account for. Military expenditure under the EPRDF was, on the formal budget figures, modest by international comparisons. The off-budget elements — paramilitary operations through regional special forces, EFFORT-related security expenditures, intelligence-service budgets — added substantially to the actual total. The full fiscal picture is, even in retrospect, difficult to reconstruct.

The post-2018 wars

The Abiy Ahmed government’s wars are examined in article 20 (Tigray) and articles 19, 22 (Amhara and ongoing). The monetary mechanism, however, is the same.

The Tigray war’s cost has been variously estimated. Direct reconstruction needs are estimated at $20 billion — roughly one-fifth of Ethiopia’s pre-war GDP.⁴ The Oxford Initiative on Peace and Recovery in Ethiopia estimates cumulative economic losses (versus a no-conflict counterfactual) at approximately $125 billion by 2027.⁵ The fiscal cost of the war itself — military operations, displaced-population support, the partial collapse of tax collection in conflict-affected regions — was a major contributor to the deterioration in Ethiopia’s fiscal position through 2021–22.

The monetary consequences are visible in the inflation data. Ethiopian inflation, which had been running in the high teens before the war, accelerated to 30 percent and above by early 2022.⁶ The parallel exchange premium widened sharply through the same period, reaching over 100 percent by 2023. The 2024 float was, in significant part, an acknowledgment that the post-Tigray monetary position was unsustainable.

The 2023–present Amhara conflict and the persistent Oromia insurgency add to the ongoing cost, in ways the official figures do not fully report. Military expenditure in the most recent fiscal year, by IMF estimates, has continued at elevated levels even as the formal Tigray hostilities have ceased.⁷

The pattern across regimes

What is worth noting about the three-regime pattern is what does not vary across the changes of government.

Each regime has fought wars whose direct fiscal costs exceeded what the available non-monetary financing could fund. Each regime has filled the resulting gap, at least in part, through central-bank advances. Each regime has experienced the predictable monetary consequences: inflation, parallel-market premium widening, current-account deterioration. Each regime has, in its initial framing of the wars, presented them as defensive responses to external aggression and as costs the country had to bear. Each regime has, in subsequent assessment, faced the question of whether the wars were as unavoidable as their framing suggested.

This is not a claim that all the wars were unjust, or that the regimes could have or should have refused to fight them. The Eritrean conflict had real underlying causes; the Tigray war was triggered by a specific decision that the regime defended as necessary to preserve federal authority; the Amhara conflict has been driven by complex political and ethnic dynamics that have not had easy alternatives. The series’s argument is narrower.

The argument is that the financing method is a separable choice from the decision to fight. A regime that decides to fight a war can choose to finance it through tax increases (paid by the population in the same currency-of-the-day with which they pay other taxes), through formal debt issuance (paid back over time through future taxation), through international cooperation (which Ethiopia has had limited success obtaining for its more controversial conflicts), or through central-bank advances (paid for through inflation). The first three financing methods are visible. The fourth is hidden. The hiddenness is not an accidental feature; it is the political reason the method is chosen.

The Sowellian point is direct. A government that imposes a visible war tax has to defend the tax in front of the population that pays it. A government that imposes an invisible inflation tax does not. The choice between visible and invisible methods is, in essence, a choice about democratic accountability: do you let the population see, in their tax assessments, the cost they are bearing? Or do you let them experience it as “high prices” and “currency weakness,” which they may attribute to many things — foreign speculation, drought, the weather — but which are, in fact, the deferred bill for the wars their government has fought.

The multi-causal test, restated

A serious case must apply the multi-causal test honestly. Of the inflation Ethiopia has experienced across the period 1974–2026, how much is attributable to war finance and how much to the other factors this series has been tracking?

The answer is, honestly, that we cannot decompose this with full precision. The inflation generated by war finance is interwoven with the inflation generated by drought-related food-supply shocks, by global commodity-price movements, by domestic policy choices unrelated to war (financial repression, exchange-rate management), by external borrowing dynamics, and by population growth pressure on resources. Any decomposition would require counterfactual reasoning — what would inflation have been without the wars — that cannot be performed with empirical confidence.

What can be said is that the inflation episodes that coincide with the major military mobilisations are visibly larger than the episodes that occurred during peacetime. The 1985 inflation surge (Derg wars at full intensity), the 1998–2000 inflation rise (Ethio-Eritrean war), the 2008–11 inflation spikes (overlapping with the Somalia intervention and the early GTP investment surge), and the 2020–24 inflation episode (Tigray war and its aftermath) form a pattern. The peacetime years between them have lower inflation episodes attributable to the other factors but not the concentrated inflation that wartime monetary financing produces.

The pattern is not conclusive proof of causation. It is, however, consistent with the proposition that war finance is one of the major contributors to the inflation episodes, and that a country willing to fight wars without using the central bank to pay for them would have had measurably lower inflation across the period. The casualty count of each war, by this accounting, includes not only the soldiers killed in the fighting but also a fractional share of the welfare losses imposed by the inflation that funded the fighting. This is uncomfortable accounting. It is also accurate accounting.

The reform implication

What would a different monetary regime look like, in war finance terms?

The simplest form is a statutory prohibition on central-bank financing of government deficits. The 2025 reforms to the NBE charter, examined in articles 21 and 22, include provisions of this kind — the prohibition on monetary financing is one of the structural benchmarks under the IMF ECF programme.⁸ If the prohibition holds — and this is the open question — then future Ethiopian governments will be required to finance their wars through tax increases, formal debt issuance, or reductions in other expenditure. The cost will become visible. The democratic accountability of military decisions will, by this mechanism, be enhanced.

A more demanding form is operational independence of the central bank — the legal and institutional architecture that makes the central bank capable of refusing government requests for advances even when the government insists. The 2025 NBE reforms include some movement in this direction but, by the IMF’s own assessment, remain incomplete. The “remaining gaps with respect to governance and autonomy” flagged in the IMF’s January 2025 statement on the second ECF review are the gaps that would, in a stress scenario like a new war, determine whether the prohibition on monetary financing actually held.⁹

The series’s argument, developed in articles 22 and 27, is that the institutional reforms underway since 2024 are necessary but not sufficient, and that the test of their durability will come the next time the Ethiopian government faces a fiscal pressure it cannot meet through conventional means. The wars that have driven the war-finance pattern have not ended. The next test of the system will be the next war. Whether the system will hold is, in 2026, an open question.

The next three articles turn to the post-2018 period in detail: the reform rhetoric that maintained continuity, the Tigray war that exposed the underlying disorder, and the float that finally acknowledged it.

---

Footnotes

1. For a comprehensive listing, see Christopher Clapham, The Horn of Africa: State Formation and Decay (Hurst, 2017), and SIPRI’s Yearbook armed-conflict database. ↩

2. “1983–1985 famine in Ethiopia,” Wikipedia, citing Mengistu’s 1984 announcement. ↩

3. SIPRI Military Expenditure Database, Ethiopia entries 1996–2002. ↩

4. “Tigray war,” Wikipedia, citing the $20 billion reconstruction estimate; The New Humanitarian, “Weighed down by war and drought,” April 2024. ↩

5. Oxford Initiative on Peace and Recovery in Ethiopia, “The Costs of Ethiopia’s Internal Conflicts between 2020 and 2022,” https://www.ethiopiarecovery.ox.ac.uk/costsofconflicts/. ↩

6. World Bank, “Inflation, consumer prices (annual %) — Ethiopia,” 2022 series; Trading Economics, Ethiopia inflation history. ↩

7. IMF Country Report 25/188 (July 2025), discussion of continued elevated security expenditure. ↩

8. International Monetary Fund, “Ethiopia: First Review under the Extended Credit Facility Arrangement,” October 2024, on structural benchmarks for NBE-government financing. ↩

9. International Monetary Fund, “Ethiopia: Second Review under the Extended Credit Facility,” January 2025, on remaining gaps in NBE governance and autonomy. ↩