Public debt sustainability has returned to the centre of policy debates. Advanced economies emerged from the Covid-19 pandemic with historically high debt-to-GDP ratios, and geopolitical tensions have raised the prospect of sustained increases in defence spending. A key question for policymakers is how governments finance large spending shocks and which of their constituents ultimately bears their costs.

Recent policy discussions often emphasize the idea that low interest rates make government borrowing relatively cheap. For example, Blanchard (2019) argued that when the interest rate on government debt is persistently below economic growth, debt may be easier to sustain. Related work has emphasised the prolonged decline in safe interest rates and its implications for fiscal space (Furman and Summers 2020). Several VoxEU columns have explored similar themes (e.g. Choi et al. 2024, Auerbach and Gale 2025, Subacchi and van den Noord 2025, Acharya and Laarits 2026).

However, these debates focus primarily on peacetime dynamics. Our research  examines how government debt behaves during large, unexpected spending shocks such as major wars and the Covid-19 pandemic (Jiang et al. 2026). The historical evidence reveals a striking pattern: government bonds have repeatedly generated substantial real losses during these extreme episodes. They have even underperformed equities and real estates which are traditionally regarded as risky assets.

When government debt stops being safe

Government bonds issued by the US and the UK are widely viewed as the world’s benchmark safe assets. Their safety underpins financial regulation, monetary policy transmission, and global portfolio allocation. In economic downturns and financial crises, these bonds typically perform well, reflecting investors’ flight to safety.

Yet major wars and pandemic-scale emergencies present fundamentally different fiscal challenges. Over the past three centuries, wars have typically triggered large and sudden increases in government spending, averaging about 7% of GDP annually during the first four years of conflicts in our dataset. These shocks create enormous financing needs that governments rarely meet through tax increases alone.

Using long-run data on the full portfolios of US and UK government bonds, we document a consistent pattern. Figure 1 shows the time series of the return to the government bond portfolio relative to the GDP growth rate. Whenever there is a major war, we observe a sharp decline in the bond performance – wars are always disaster times for bondholders. Similarly, the bondholders also suffered large losses during the ‘war on Covid-19’.

Figure 1 Bond returns minus GDP growth

a) UK bonds

Note: Grey shaded areas indicate war periods. r = real bond return; g = real GDP growth.
War of Austrian Succession (1740−48), Seven Years’ War (1756−63), American Revolutionary War (1775−83), French−Napoleonic Wars (1792−1815), WWI (1914−18), WWII (1939−45), COVID−19 (2020−23).

b) US bonds

Note: Grey shaded areas indicate U.S. wars; light blue shading indicates American Revolutionary War (1775−83). r = real bond return; g = real GDP growth.
Wars shown: American Revolutionary War (1775−83), War of 1812 (1812−15), Mexican−American War (1846−48), Civil War (1861−65), Spanish−American War (1898), WWI (1914−18), WWII (1939−45), Korean War (1950−53), Iraq−Afghanistan War (2001−03), COVID−19 (2020−23).

We illustrate our formal analysis in Figure 2, which reports the average cumulative bond performance. Across wars and the Covid-19 pandemic, bondholders experienced average real losses of roughly 14% during the first four years of crises. These low real returns are large enough to substantially reduce the real value of government debt outstanding.

Figure 2 Bond cumulative performance during wars and pandemics

Note: The figure plots the average bond performance. The outcome variable is the log cumulative real return on the government bond portfolio in panel (a), the log cumulative nominal return on the government bond portfolio in panel (b), the log cumulative inflation in panel (c), the log cumulative return on the bond-minus-stock portfolio in panel (d), the log cumulative bond return minus GDP growth in panel (e), and the log cumulative return on the bond-minus-housing portfolio in panel (f). The horizontal axes denote the years since the start of the war. The orange bars indicate two-standard-error bands based on heteroskedasticity-robust standard errors.

Perhaps even more surprisingly, government bonds frequently underperformed riskier assets such as equities and real estate during wartime. As shown in panels (d) and (f) of Figure 2, the underperformance is substantial: after four years into the war, the cumulative bond return is more than 20% below the cumulative returns on equities and real estate. This sharply contrasts with their performance during financial crises or recessions, when government bonds typically outperform risky assets.

Inflation: The primary channel for debt reduction

A key mechanism behind these losses is surprise inflation. During wartime episodes, cumulative inflation averaged about 20% over the first four years of conflicts in the US and UK. Although nominal bond returns remained positive, inflation eroded their real value.

This pattern is consistent with longstanding fiscal theories of inflation, which emphasize how expectations of future deficits can influence price levels (Sargent and Wallace 1981, Leeper 1991, Sims 1994, Woodford 1995, Cochrane 1998). When governments face large unfunded spending increases, inflation can act as an implicit tax on holders of nominal government debt. In this sense, the governments shift risk onto debtholders
and spare the taxpayers and transfer recipients, consistent with the trade-off in our paper.

Importantly, our results show that inflation during wars was not simply a by-product of economic disruption. Instead, it often reflected deliberate policy choices that allowed governments to reduce their debt burdens without explicit default. For example, both the US and the UK repeatedly abandoned or suspended gold standard commitments during major wars, increasing monetary flexibility and allowing inflation to rise.

The Covid-19 pandemic displayed similar patterns. Despite initial expectations that safe-haven demand would push government bond prices up, inflation surged following the pandemic fiscal expansion, contributing to substantial declines in real bond returns.

Financial repression: Another key mechanism

A second important mechanism is financial repression – policies that reduce governments’ borrowing costs by influencing financial markets or investor behaviour. Historically, governments have employed a wide range of such policies. During WWII, for example, the US Federal Reserve implemented yield-curve control, capping Treasury bond yields and purchasing large quantities of government debt. The UK employed similar strategies, including regulatory measures that encouraged or required financial institutions to hold government securities.

These policies lowered nominal interest rates even as inflation increased, producing negative real returns for bondholders. In our data, nominal bond returns often rose slightly during wars, but not nearly enough to compensate for inflation. The combination of rising prices and suppressed yields generated substantial losses in real terms.

Recent research suggests that financial repression played a major role in reducing public debt after WWII (Reinhart et al. 2011, Acalin and Ball 2023). Our evidence shows that similar policies have been used repeatedly across centuries of fiscal crises.

Implications for debt sustainability

Our findings provide new perspective on the relationship between interest rates, growth, and fiscal sustainability. Much of the recent literature has emphasized the possibility that government borrowing may be relatively costless when interest rates remain below economic growth.

However, we show that average interest-growth differentials are strongly influenced by wartime and pandemic episodes. When these episodes are excluded, the US historically exhibits interest rates that exceed growth rates. Conversely, during major fiscal crises, real interest rates fall dramatically below growth largely because governments reduce the real value of debt through inflation and financial repression. Consequently, low interest-growth differentials during crises should not necessarily be interpreted as evidence that debt is easily sustainable. Instead, they often reflect implicit transfers from bondholders to taxpayers.

Rethinking the safety of government debt

Our results also challenge conventional views about safe assets. Government bonds provide insurance against many types of economic downturns, including financial crises and recessions. But they perform poorly precisely in states of the world associated with large fiscal shocks – such as wars or pandemics.

This finding highlights an important trade-off for policymakers. Protecting taxpayers from large spending shocks may require shifting part of the burden onto bondholders through inflation or financial repression. Economic theory suggests that such policies may be optimal when taxation is highly distortionary. However, they also reduce the safety of government debt and may raise borrowing costs over time if investors anticipate these risks.

Conclusion and lessons for today’s policy debates

Government bonds have long been considered the cornerstone of safe asset markets. Yet history shows that their safety is far from guaranteed during wars and pandemic-scale emergencies. Over three centuries of US and UK data, such episodes consistently produced large real losses for bondholders, primarily through inflation and financial repression. As advanced economies confront rising geopolitical risks and elevated public debt levels, these historical patterns carry several lessons for current fiscal and monetary policy debates, as well as for investors.

First, the apparent safety of government debt is conditional. Safe-asset status depends not only on credit default risk, but also on governments’ willingness to preserve the real value of debt during extreme fiscal shocks.

Second, inflation can play an important role in public debt dynamics, especially during large spending expansions. Policymakers confronting future crises may face pressure to tolerate higher inflation as a means of stabilising public finances.

Third, financial repression remains a relevant policy tool. While often less visible than explicit fiscal measures, regulatory and monetary interventions can significantly affect debt sustainability and investor returns.

Finally, the historical experience suggests that debt sustainability cannot be assessed using average interest-growth differentials alone. Policymakers must also consider how fiscal and monetary responses during crises affect the distribution of fiscal burdens across taxpayers and investors.

References

Acalin, J and L M Ball (2023), “Did the US really grow out of its World War II debt?”, NBER Working Paper No. w31577.

Acharya, V V and T Laarits (2026), “How the Tariff War Shock Affected the ‘Safe Asset’ Privilege of US Treasuries”, VoxEU.org.

Auerbach, A J and W Gale (2025), “Then and Now: A Look Back and Ahead at the Federal Budget”, VoxEU.org.

Blanchard, O (2019), “Public Debt and Low Interest Rates”, American Economic Review 109(4): 1197–1229.

Choi, J, D Dang, R Kirpalani, and D J Perez (2024), “Exorbitant Privilege and the Sustainability of US Public Debt”, VoxEU.org.

Cochrane, J H (1998), “A Frictionless View of US Inflation”, NBER Macroeconomics Annual 13: 323–384.

Furman, Jason, and Lawrence H. Summers. 2020), “A Reconsideration of Fiscal Policy in the Era of Low Interest Rates”, unpublished manuscript, Harvard University and Peterson Institute for International Economics.

Jiang, Zhengyang, Hanno Lustig, Stijn Van Nieuwerburgh, and Mindy Z. Xiaolan (2026), “Manufacturing risk-free government debt”, Journal of Financial Economics 176: 104203.

Leeper, E M (1991), “Equilibria under ‘Active’ and ‘Passive’ Monetary and Fiscal Policies”, Journal of Monetary Economics 27(1): 129–147.

Reinhart, C M and M B Sbrancia (2015), “The Liquidation of Government Debt”, Economic Policy 30(82): 291–333.

Reinhart, C M, J F Kirkegaard, and M B Sbrancia (2011), “Financial Repression and the Liquidation of Government Debt”, Finance & Development 48(2): 22–26.

Sargent, T J and N Wallace (1984), “Some Unpleasant Monetarist Arithmetic”, in Monetarism in the United Kingdom, Palgrave Macmillan.

Sims, C A (1994), “A Simple Model for Study of the Determination of the Price Level and the Interaction of Monetary and Fiscal Policy”, Economic Theory 4(3): 381–399.

Subacchi, P,and P van den Noord (2025), “International Trade Suppression and the Demand for US Treasuries”, VoxEU.org.

Woodford, M (1995), “Price-Level Determinacy without Control of a Monetary Aggregate”, in Carnegie-Rochester Conference Series on Public Policy, Vol. 43, North-Holland.