1 Introduction

Rapid inflation has returned to the USA, suddenly and unexpectedly. In October 2019, the International Monetary Fund’s World Economic Outlook (IMF-WEO) forecast that inflation in the USA would be 2.4% in 2021, 2.3% in 2022, and 2.3% in 2023, continuing a downward trend that began in the mid-1980s. However, inflation was 4.7% in 2021, closed at 8.1% in 2022, and is now expected to be 3.5% in 2023 (and may well be higher). In this paper, we measure the effect of this unforeseen inflation on the value of fixed-rate dollar-denominated debt around the world. Our central argument is that the US inflation shock has caused a “Great Dilution” in the real value of dollar-denominated sovereign liabilities, and sparked a vast redistribution of wealth in the process.

Our analysis encompasses sovereigns around the world. Policymakers are rightly concerned about the macroeconomic fate of emerging markets as interest rates rise in the USA and elsewhere, as previous episodes have led to macroeconomic and debt crises. For instance, Pazarbasioglu and Reinhart (2022) argue that:

Tighter monetary policies in advanced economies are poised to push up international interest rates, which tends to put pressure on currencies and heighten the odds of default. ...Global financial conditions are set to deteriorate as central banks in advanced economies tighten policy to fight unexpectedly persistent inflation pressure.

Similarly, Acosta-Ormaechea et al. (2022) caution that:

With public debt-to-GDP ratios above pre-pandemic levels and borrowing costs rising amid higher local and global interest rates, countries will need to ensure the sustainability of public finances to help preserve credibility and rebuild fiscal space.


While these concerns are certainly warranted, we suggest they should take into account the reason for tightening financial conditions: the increase in US inflation, which in some cases cushions the effect of rising rates by reducing the real value of dollar-denominated sovereign liabilities issued by other countries.

Although we estimate the gains (and losses) to be substantial and policy-relevant, we note at the outset that our analysis focuses on the direct effects of surprise inflation on the burden of sovereign debt. We do not model the general equilibrium consequences on the cost of future borrowing, financial system stability, and the like.

The worldwide issuance of dollar-denominated debt has grown significantly in the recent era of global financial integration. Years of low inflation catalyzed the growth of financial assets issued at fixed interest rates and with long maturities. According to the Bank for International Settlements (BIS),Footnote 1 international issues of sovereign debt securities at fixed rates and with maturities longer than one year have represented more than 95% of all issues since 2013. Long-run fixed-rate dollar instruments are subject to larger valuation effects than if they had been issued at variable rates or issued at short maturities (as was the common practice prior to the “Great Moderation”).

In order to illustrate the distributive effects of the US inflation shock we conduct a few exercises to estimate the gains to sovereigns arising from the dilution of the value of long-term fixed-rate debt instruments due to US inflation.Footnote 2 Our analysis focuses on sovereign dollar-denominated debt, both because of data constraints and because sovereigns are the focus of the policy concerns. The amount of dollar-denominated debt issued in international markets is immense, totaling $11.1 trillion globally by the end of 2020, according to the Bank for International Settlements (Eren and Malamud 2022), of which $1.3 trillion corresponds to non-US long-term fixed-rate sovereign securities.Footnote 3 In addition, by the end of 2020, $20.7 trillion worth of long-term fixed-rate securities had been issued by the US government.Footnote 4

In order to estimate these effects we start with the USA, the biggest issuer of dollar-denominated debt. We first estimate the effect of surprise inflation on the $14.6 trillion worth of long-term treasury securities held by the public (including creditors abroad) and the Federal Reserve (Fed), to which we add the $2.1 trillion of cash notes. We show that the US government’s gain from unexpected inflation in 2021 and 2022 amounts to $1.4 trillion or 6.8% of GDP. Adding the gain on debt held by government agencies in pension programs and the like the number climbs to $1.9 trillion (9.2% of GDP). However, adding the Fed to the baseline scenario lowers the gain to $1.3 trillion or 6.3% of GDP.

Because we have good information about the foreign holdings of dollar-denominated assets (including cash), we can study how much of the $1.4 trillion gain for the treasury is paid by US non-residents. High US inflation generates a transfer to the US government from non-residents of about $542 billion (for comparison in 2020 federal government spending on defense was $777 billion and on Medicaid was $447 billion). About one-third of these gains come from Japan and China, two of the biggest holders of US treasuries. In all, about one-third of the gains accruing to the US Treasury are paid by non-residents.

However, these numbers can increase substantially under alternative scenarios that account for the possibility that high inflation persists in 2023 and beyond. In order to estimate the resulting gains to the US Treasury, we draw on information about the maturity structure of fixed-rate debt, as the gain accrues only on the debt that is not rolled over (once debt is rolled over we assume it fully internalizes future inflation). In the most extreme case where inflation returns to the 2% level in 2030, the gain to the US Treasury is anywhere between 16% and 20% of GDP depending on whether the holdings of securities by the Fed and treasuries by government agencies are included or not.

Of course, these gains to the sovereign may produce a variety of potentially undesirable results such as higher debt servicing costs in the future, greater interest payments on bank reserves, reduced international use of the dollar, or financial system instability. These considerations, albeit no doubt important, are largely outside the scope of our paper, which focuses specifically on the inflation-debt channel.Footnote 5

In the second part of the paper, we estimate the dilution of dollar-denominated long-term sovereign debt for other countries. This requires estimating how the GDP of other countries, measured in dollars, changes. We do this in two ways. First, we assume that US inflation transfers directly to local prices measured in dollars. If real exchange rates remain constant in the face of a US inflation shock, this relationship should be expected to occur. Moreover, we show this relationship is validated in the data. Second, we use actual observed increases in local prices in dollars to factor in the potential effect of recent US dollar appreciation and changes in the dollar real exchange rate.

Using the first method, we show that countries other than the US net a gain of over $104 billion from the unexpected US inflation shock of 2021–2022. Major winners in absolute dollar terms include middle-income countries like Turkey, Saudi Arabia, Argentina, Mexico, and Indonesia. Relative to the size of their economies, big winners include Oman and Qatar in the Middle East; Jamaica, Panama, and Uruguay in Latin America; as well as other countries such as Lebanon and Mongolia, all of which receive a one-time “transfer” larger than 2% of GDP. The second exercise does not modify the conclusions significantly. In fact, for the countries for which we have the data to perform this alternative computation, the gains actually increase from $100 billion for the years 2021 and 2022, to as much as $170 billion (because contrary to the widely held view, not all countries have experienced a real depreciation against the US dollar).

We then discuss how the inflation shock of 2021–2022 may carry over into the future and how this may compound the gains. As this exercise requires information about the maturity structure of fixed interest rate dollar-denominated debt, we conduct this exercise for a smaller group of countries. But the sample is quite representative of developed and emerging markets. Using the change in the IMF’s inflation forecast as a measure of the inflation shock, the gains do not increase significantly when introducing a forward-looking analysis. This is because the IMF expects inflation to return to steady-state value in the order of 2% as soon as 2024. In the most pessimistic case of inflation returning to its target in 2030, the countries excluding the USA receive a bonus that doubles our initial computation. These large gains, we believe, help to explain why the abrupt increase in interest rates in the USA has not produced the turmoil in emerging market sovereign debt associated with previous increases.Footnote 6 Here too, we note that the gains to these sovereigns could be balanced by various factors (for instance the short-run appreciation in the US dollar).

Our paper’s principal contribution is to the literature on the currency denomination of sovereign debt (Calvo 1988; Eichengreen and Hausmann 1999, 2005; Alfaro and Kanczuk 2018; Ontonello and Perez 2019; Ballard-Rosa et al. 2021; Sosa and Sturzenegger 2023). Although debt denominated in local currency provides a better hedge against negative domestic and external shocks, governments face the temptation to generate inflation and depreciate their currency to reduce the real value of their debt. In a classic study Calvo (1988) argued that the solution to this time-inconsistency problem was for countries to rely on debt that is denominated in foreign currency that cannot be diluted by inflation (though others (Frankel 2014) have pointed out that a reliance on dollar-denominated debt has contributed to severe contractionary balance sheet effects in currency crises). However, most of the prior research did not contemplate the impact of high inflation in the USA, i.e., in the currency that was intended to solve the credibility problem. This paper highlights how unanticipated nominal shocks in developed countries shape the fortunes of emerging country sovereigns in unexpected ways.

Our results also provide further evidence for the argument made by Reinhart et al. (2015) that rich countries lean far more heavily on heterodox measures such as surprise inflation to reduce their debt ratios (and less on running primary surpluses and other orthodox strategies) than is typically believed. Rigorous analyses by Hilscher et al. (2022) had concluded that the probability that US inflation would lower the real value of the debt was very low, both because the private sector holds relatively short maturity debt and because high inflation was believed at the time to be extremely unlikely. We show that the large inflation shock that has come to pass means that the USA is indeed inflating away a substantial fraction of its debt, as anticipated by Aizenman and Marion (2011). A contemporaneous paper by Pallotti (2022) studies winners and losers across economic sectors in the USA and comes to a similar conclusion to ours with regards to the gain to the treasury. Our paper complements this body of work by systematically documenting the impact of unanticipated inflation not only for the USA, but also for countries around the world. As such, we hope our paper will serve as a springboard to more fine-grained theoretical and empirical work that extends beyond our focus on dollar-denominated sovereign assets and liabilities.

2 The Impact on the US

Inflation in the USA reduces the real value of dollar-denominated government debt, generating a gain for the US Treasury. This effect is stronger when the debt has been issued at fixed rates and with long maturities. Even when there is some small dilution to short-term debt, to be conservative we only deal with debt longer than one year which we (and the US government) define as long-term debt.

By the end of 2020, long-term fixed-rate US securities totaled $20.7 trillion. This includes notes, bonds, and nonmarketable debt held by the public of $14.6 trillion,Footnote 7 plus $6.1 trillion of nonmarketable debt held by government agencies which we assume is long-term debt.Footnote 8 The amount of cash notes at the end of 2020 added an additional $2.1 trillion.Footnote 9

Because interest on debt compensates for expected inflation, it is the unexpected component of inflation that generates a transfer from creditors to debtors. Therefore, our starting point is an assumption about unexpected inflation. The October 2019 IMF-WEO (International Monetary Fund 2019) projected an inflation rate of 2.4% in 2021 and 2.3% in 2022 and 2.3% in 2024 for the USA (International Monetary Fund 2019, p.154), which we interpolate to an expected inflation rate of 2.3% for 2023. However, US inflation was 4.7% in 2021, 8% in 2022 and is currently projected to be 3.5% in 2023 according to the October 2022 IMF-WEO (International Monetary Fund 2022). The unanticipated inflation of 8% is simply equivalent to the sum of the actual inflation deviation from the forecast inflation over the two years (i.e., 4.7–2.4% + 8–2.3% = 8%).

We then apply this percentage of unexpected inflation (8%) to the total stock of long-term, fixed-rate dollar-denominated sovereign debt held by the public at the end of 2020 ($14.6 trillion). To this we add the full impact of inflation on US dollar cash notes. As prices in US dollars increase, the real value of these cash holdings diminishes. Unlike debt holders who receive interest, cash holders are not compensated for inflation and the government fully charges the “inflation tax” on them. Therefore we apply actual inflation in 2021 and 2022 (12.7%) to the stock of cash notes.

As might be expected, the gain for the USA is enormous. In dollar terms, the USA has shaved about $1.4 trillion in the purchasing power of its liabilities in 2021 and 2022. This is equivalent to a budget gain of 6.8% of GDP. Accounting for this “inflation tax” implies that the US government actually ran much smaller deficits over the years 2021 and 2022 than a straightforward examination of the nominal figures would suggest.

The inflation surprise may lead to higher interest rates on new debt issues in the future. Thus far, however, inflation projection-adjusted interest rates do not appear to have risen substantially in the USA (see appendix).

A significant portion of US debt is held by other government entities and the Federal Reserve,Footnote 10 so there is a discussion to be had on whether these holdings should be added to our calculation. Government agents’ $6.1 trillion holdings in short- and long-run treasuries could be added because most of these holdings are held in pension programs, social security trusts, and the like, whose beneficiaries, such as retirees, do not include the government.Footnote 11 Extending our baseline estimate to incorporate these government holdings, increases the gains to the US Treasury to $1.9 trillion or 9.2% of GDP. The evaluation of the Fed’s holdings is more complex so we analyze it separately in the subsection below.

2.1 The Fed

The Fed holds a large balance sheet of nominal assets and liabilities and therefore is likely to be strongly affected by an inflation surge. Roughly speaking we can split the main components of the Fed’s balance sheet as of December 2020 as in Table 1.

Table 1 Fed balance sheet Source: Federal Reserve Statistical Release. https://www.federalreserve.gov/releases/h41/20201231/
Full size table

On the asset side the Fed experiences a loss on its holdings of long-term US treasuries (those that are not inflation-protected) and mortgage-backed securities. This has to be compared to the gains obtained on the liability side: notes (which we already computed above) and the eventual gains on reserves.

The losses by the Fed on treasuries is significant and wipes out part of the gains estimated for the treasury above. Mortgage-backed securities losses should be estimated exactly as that of treasuries because it is the unexpected inflation component which provides a reduction in the value of the debt.

The Federal Reserve holdings of mortgage-backed securities totaled $2.1 trillion in December 2020; unanticipated inflation causes losses in the real value of these securities. If we apply our factor of 8% to this $2.1 trillion, we find these losses to be $165 billion; these losses on the Fed portfolio in turn reduce its net income, which would otherwise have been passed on to the treasury.Footnote 12

On the liability side cash notes should suffer the full effect of inflation, as discussed above. Reserves should incorporate the difference between inflation and the interest paid on reserves. The Federal Reserve began paying interest on these balances beginning in 2008; the rate was fixed at 0.15% till March 2022 and eventually rose to 4.40% in December 2022 (and 5.40% in July 2023).Footnote 13 Reserve balances totaled $3.14 trillion at the end of 2020.Footnote 14 Comparing cumulative inflation and the interest rate paid, the Fed gained $339 billion from reserve balances.Footnote 15

If we consider the combined gain from cash and reserves less the combined loss from treasuries and mortgage-backed securities, the Fed nets a gain of $141 billion or 0.7% of GDP.

In the previous section, we did not consider the assets and liabilities of the Fed as a whole. Because 0.7% of GDP is smaller than the 1.3% gain on cash estimated in the previous section, if we consider the overall effect of the Fed, the US gain actually falls from $1.4 trillion to $1.3 (from 6.8% of GDP to 6.3%).

2.2 Foreign Transfers to the USA from US Inflation

How much of the gains to the US Treasury from unexpected inflation are paid by the rest of the world? It is possible to estimate this transfer to the treasury because unlike the case for other countries, data are available on the individual country holdings for the two categories of US liabilities that are most exposed to inflation: long-term US treasuries and cash.

At the end of 2020, $7 trillion worth of US treasuries were held by non-residents of the USA, of which $5 trillion had long-term maturities, according to the Treasury Information Capital (TIC) System.Footnote 16 Figure 1 shows the share of treasuries held by foreigners over recent years.

Fig. 1
figure 1

Foreign holdings of long-term US securities. Source: Treasury Information Capital (TIC) System. The data correspond to foreign holdings of long-term US securities as fraction of the total long-term US securities

Full size image

Approximately, $947 billion worth of dollar bills were held abroad at the end of 2020, according to the US Federal Reserve (Bertaut et al. 2019).Footnote 17 Current estimates of cash holdings by country are not available. We follow prior research (United States Department of the Treasury 2006) that computed country-wise cash holdings based on fieldwork and cash shipments to each destination in 2006. To reach an allocation by country in 2020, we proportionally increase the individual 2006 country estimates by the increase in the holdings of cash abroad from 2006 to 2020 as reported by the Federal Reserve. This allows us to assign 55% of the total to individual countries. Figure 4 shows the evolution of cash holdings in the last decades estimated by Judson (2017). Notice that most of the cash holdings abroad are in $100 bills, which suggests that individuals use them as a store of value rather than for liquidity services (Fig. 2).

Fig. 2
figure 2

Total amount of US currency abroad. Source: Judson (2017)

Full size image

With these data we can then compute the losses accruing to countries from their holdings of US liabilities from the unexpected inflation shock of 2021 and 2022. Figure 3a, b plot the unexpected inflation effect from the holdings of long-term treasuries both in dollar values and as a percentage of GDP. The overall losses add up to $422 billion. Of these losses, 38% is accounted by Japan and China, which are major holders of US treasuries. The largest losses as a share of GDP correspond to East Timor (34% of GDP) and Luxembourg (17% of GDP); Hong Kong, Ireland, and Bahamas are also major losers.

When considering both the effect on treasuries and cash in Fig. 3c, d, the gains for the US Treasury (at the expense of non-residents) rises to $542 billion. Thus, fully one-third of the “inflation tax” in 2021 and 2022 is levied on non-residents abroad.

Countries that are major holders of dollar currency (relative to their GDP), such as Argentina, Cambodia, and Russia, emerge as significant losers of unexpected inflation in the USA. Argentina and Cambodia suffer a loss that is larger than 2% of GDP. As a share of GDP, East Timor, Luxembourg, Cambodia, Hong Kong, Argentina, Taiwan, Belgium, Singapore, and Ireland suffer the largest losses. As can be seen, the costs spread across rich and poor countries alike.

Fig. 3
figure 3

Redistribution to USA from other countries’ holdings of long-term US Treasury securities and cash. Sources: Treasury Information Capital System and Federal Reserve for Treasury holdings and Judson (2017), Bertaut et al. (2019) and US Federal Reserve for cash holdings. Panel b and d excludes two outliers: East Timor and Luxembourg

Full size image

2.3 The Effect of Future Inflation

Needless to say, our computation measures the reduction in the real value of liabilities arising from the inflation surprise in 2021 and 2022. But the real effect should consider additional inflation surprises beyond 2022. However, the estimate of the effect from future inflation depends on both how long we expect inflation to remain high and on the maturity structure of the debt.

Let us call \(\theta _0\) the intertemporal burden of debt at any time \(t=0\) for any country.Footnote 18 This intertemporal burden is defined by:

$$\begin{aligned} \theta _0=\sum _t \frac{q_t b_t}{ \epsilon _0 p_0 y_0}, \end{aligned}$$
(1)

where \(q_t\) and \(b_t\) are price (in USD) and quantity of debt due in period t. \(\epsilon _0\), \(p_0\) and \(y_0\) are the exchange rate (for the USA equal to 1), local prices and real GDP at the time of the computation \(t=0\).

The price of one unit of debt due in period t is

$$\begin{aligned} q_t= \frac{1}{(1+r_t)E_0\Pi _t^{US}}, \end{aligned}$$
(2)

where \(r_t\) and \(E_0\Pi _t^{US}\) represent the cumulative real interest rate and cumulative US inflation expected at time 0 until time t.

Given this definition it is easy to see that \(\Delta \theta\), the change in the debt burden when only the expectation of US inflation changes equals

$$\begin{aligned} \Delta \theta _0 = \frac{\sum _t \frac{b_t}{(1+r_t)E_0^*\Pi _t^{US}}}{\epsilon _0 p_0 y_0} - \frac{\sum _t \frac{b_t}{(1+r_t)E_0\Pi _t^{US}}}{\epsilon _0 p_0 y_0}, \end{aligned}$$
(3)

where \(E_0^*\Pi _t^{US}\) is the new expected inflation rate for the USA. If we assume the real interest rate to be constant and equal to zero, the expression can be simplified to

$$\begin{aligned} \Delta \theta _0 = \frac{1}{\epsilon _0 p_0 y_0} \left[ \sum _t b_t \left[ \frac{E_0\Pi _t^{US}}{E_0^*\Pi _t^{US}} -1 \right] \right] . \end{aligned}$$
(4)

In other words, the debt dilution corresponds to the effect of the higher inflation on the stock of nominal dollar debt, all expressed in terms of current GDP.

If we assume the structure of debt that is used typically in the sovereign debt literature (Hatchondo and Martinez 2009; Hatchondo et al. 2016) where debt is zero coupon and decays exponentially:

$$\begin{aligned} b_{t}= \left( \frac{1}{\eta } \right) ^{t} d_0, \end{aligned}$$
(5)

where the decay factor \(\eta\) is chosen to match the average maturity of the debt, we have a way of estimating (4). This is the equation we estimate.

In 2020, the average maturity of long-term debt (i.e., longer than one year maturity) in the USA was 8.3 years.Footnote 19 What remains for the computation is to obtain estimates of the change in inflation expectations, which we can take directly from the IMF World Economic Outlooks. In particular, we look at the change in expected US inflation between the October 2019 WEO and October 2022. The path for future inflation in both cases is presented in the first two rows of Table 2.Footnote 20

Table 2 Excess inflation by scenario\(^{20}\)
Full size table

Once we update our computation to include future periods with the inflation surprise of our Base Scenario (see first row of Panel B in Table 2), we find the gain for the USA is 7.2%. The results are impacted very little because the IMF expects inflation to return to steady state as soon as 2024, that is, that there will no further inflation surprises. (In fact, the increase comes from the persistent dilution on cash, as losses on bonds decrease due to the fact that we now allow debt to mature starting in 2021.)

This view, however, can be stressed by assuming that inflation declines to its steady-state value gradually over a longer period: 2026, 2028, and 2030. The second to fourth rows in panels A and B of Table 2 show these alternative disinflation hypotheses, and Table 3 shows the resulting gains to the USA. We assume in these computations that going forward higher interest rates fully compensate for inflation, including those for reserves, so that there are no further gains on debt that is rolled over, or on reserves. Still, if inflation takes longer time to decline, the gain for the USA scales up. They are 10.3% of GDP if inflation goes back to steady state in 2026, 13% if in 2028, and 15.7% if inflation goes back to steady state in 2030. By then, however, a third of the gain comes from the inflation tax on the money stock. The final rows show the estimates if we add holdings of treasuries by public agencies and, separately, the Fed’s balance sheet.Footnote 21

Table 3 Effect by inflation path scenarios (% GDP)
Full size table

The discussion on the potential of US inflation to dilute US debt has been discussed in the literature. Aizenman and Marion (2011) suggested this would be a mechanism to reduce debt and argued that a steady-state 4% inflation rate would reduce the real value of debt by about 20%. This was a result in line with Reinhart et al. (2015), who argue that heterodox solutions have been used extensively by developed country governments to deal with debt spikes. Given the maturity structure of US debt (heavily tilted toward the short term) and the low future likelihood of the type of inflation we have witnessed, Hilscher et al. (2022) had forecast that substantial debt dilution was unlikely, and argued that a debt reduction of more than 4% of GDP was all but impossible. The high inflation in the USA has led to a higher estimate (our 6.8% of GDP) in 2021 and 2022. However, if we allow inflation to persist at higher levels, our numbers approximate those of Aizenman and Marion (2011).

3 The Effect of US Inflation on Debt Burdens Around the Globe

We now turn to estimating the effect of US inflation on long-term dollar-denominated fixed-rate debt issued by other countries. As we show below, debt burdens will fall to the extent that US inflation increases the value of other countries’ GDP when measured in dollars.

3.1 Why US Inflation Matters Elsewhere

A helpful starting point to show how US inflation translates into dollar-denominated GDP of other countries is the simple purchasing power parity (PPP) relationship:

$$\begin{aligned} P_t = E P_t^*, \end{aligned}$$
(6)

where \(P_t\) is the price level in a specific country, and \(P_t^*\) is the price level in the USA. E is the exchange rate defined as the number of units of currency of that specific country per dollar.

This equation assumes all goods are tradable, or, alternatively, that there are no changes in the real exchange rate. Our relevant shock is an inflation shock in the USA, which should not change the real exchange rate.

PPP provides a simple exchange rate equation:

$$\begin{aligned} E=P_t/P_t^*, \end{aligned}$$
(7)

which states that the exchange rate will move according to the inflation differential. If the local country has higher inflation, its exchange rate will depreciate. But if the US inflation is higher the local country’s, then the currency will appreciate relative to the dollar. Now, if

$$\begin{aligned} GDP_t=P_t Q_t, \end{aligned}$$
(8)

where \(GDP_t\) is local currency nominal GDP, and \(Q_t\) is real GDP, then

$$\begin{aligned} \frac{GDP_t}{E_t}=GDP_t^{USD}=\frac{P_t}{E_t}Q_t=P_t^* Q_t, \end{aligned}$$
(9)

which shows that the local GDP measured in US dollars grows at the rate of US inflation.

The bottom line is that US inflation will increase the value of GDP in all countries, when measured in dollars, at the tune of the US inflation.Footnote 22 When payments are fixed in (nominal) US dollars, the real burden of these payments falls with US inflation.

Real exchange rates need not remain fixed. In the specific case of the inflation shock of 2021–2022, the sharp increase in interest rates in the USA led to a large real exchange rate appreciation of the US dollar (though this appreciation later reversed somewhat). A strong dollar appreciation may therefore lead to an increase in the debt burden in spite of higher US inflation.

What does the evidence have to say about how US inflation transfers to other countries? Using data from 1960 to the present, we can compute the dollar inflation in each country (we do this by dividing nominal GDP in dollars by real GDP, estimating a dollar deflator) and explore its relationship with US inflation for each country.Footnote 23 The results of this computation are shown in Table 4. Column 1 shows the relationship between yearly dollar inflation in each country and US inflation. The regression includes country fixed effects to control for country-specific drivers of the local real exchange rates. The coefficient is highly significant and larger than one. This indicates that an increase in US inflation is associated with a real dollar depreciation.

Table 4 Regression results
Full size table

Column 2 in Table 4 shows the same exercise with data collected in 5-year interval. Now, the coefficient reduces to 1. This means that US inflation will (sooner or later) imply a higher local price level in nominal dollars. Ceteris paribus, given trends in the real exchange rate US inflation translates to domestic dollar inflation basically one to one.

Of course, at least for 2020 and 2021 data, we can use actual dollar GDP numbers which are available. This should factor in the effect of the US appreciation, if any. We show below that our results hold regardless of which methodology we use.

3.2 The Effect of the 2021–2022 Inflation Shock

In estimating the effects of the inflation shock, we perform the analysis in two ways, first assuming constant real exchange rates, and second using actual real exchange rates.

3.2.1 Assuming Constant Real Exchange Rates

When real exchange rates are constant, dollar inflation in other countries equals that of the USA. This implies that we can apply the same methodology that we use for the US debt to estimate the (real) gain for the issuing sovereigns and the equivalent loss for the holders of that debt. In short, we can multiply the stock of long-term, fixed-rate dollar-denominated debt by the extent of unexpected US inflation to estimate the resulting gains to other countries.

Our data on the size, composition, and maturity of dollar-denominated sovereign debt comes from the BIS Debt Securities Statistics.Footnote 24 As can be seen in Fig. 4, almost all international sovereign debt is issued at fixed rates and with long maturities.Footnote 25

Fig. 4
figure 4

Share of long-term fixed-rate securities in total Sovereign International Issues. Source: BIS. The share is computed as the total long-term fixed-rate securities over the total securities (denominated in all currencies). All data refer to the outstanding issued in the fourth quarter of every year

Full size image

Although the amount of domestic currency debt issues has increased in recent decades,Footnote 26 the amount of total long-term fixed-rate dollar-denominated debt issued by countries other than the USA still totaled $1.3 trillion at the end of 2020.

As before, we apply the unexpected inflation (8%) to the total stock of long-term, fixed-rate dollar-denominated sovereign debt at the end of 2020. Figure 5a, b show our estimate of transfers to sovereigns. Figure 5a shows the absolute value in dollar terms, while Fig. 5b shows the values as a share of GDP for all countries except the USA (which is included in the bottom panel of Fig. 5). Data for each country is reported in Appendix.

Fig. 5
figure 5

Effect of unexpected inflation on value of Sovereign Liabilities. Source: BIS Table C3 “Debt securities issues and amounts outstanding, in billions of US dollars” and IMF-WEO (for forecast and actual inflation). The effect of unexpected inflation corresponds to the product between the stock of long-term fixed-rate securities denominated in US dollars and the 8% unexpected inflation. Figures for the USA include treasury securities and monetary base

Full size image

Although the issuance of dollar debt was originally thought to be a mechanism for tying the hands of sovereigns that otherwise faced the temptation to use inflation to reduce their debt obligations, US inflation now provides a means of debt dilution through, as it were, the back door. Argentina, Brazil, Indonesia, Mexico, Turkey, Saudi Arabia, the United Arab Emirates, Qatar, and Canada are some of the biggest beneficiaries of this debt dilution by absolute dollar value, with each country securing a windfall that exceeds $4 billion. Excluding the USA, the gains across all countries amounts to $104 billion, a number on the scale of total annual foreign aid flows.

If we focus on the impact as a percentage of GDP (in 2020), which provides a more appropriate measure of the effects, we see that the effect is larger for poorer countries. As share of GDP, the biggest beneficiary is Lebanon (9.8% of GDP, which is even higher than for the US!). Other major winners are countries such as Venezuela, Jamaica, Panama, Oman, Bahrain, and Qatar.

Real exchange rates may change as a result of interest rate hikes. In this first-cut analysis, by assuming PPP, i.e., that inflation differentials are exactly reflected in exchange rate changes, we abstracted from such realignments. However, in the short run, PPP may not hold. For example, recent interest rate hikes have led to significant short-run appreciation of the dollar, which actually decreases the US dollar-denominated GDP of other countries. This change is short run and likely to be transitory (as we showed above, also see (Taylor and Taylor 2004)). We can therefore interpret our results in this section as the eventual effect of unanticipated US inflation even if it does not reveal itself in the short run. Regardless, in the next subsection, way allow real exchange rates to vary.

3.2.2 Allowing Real Exchange Rates to Vary

The above estimation assumed real exchange rates to be constant. What if they are not? One way to relax the assumption is to compute the rate of price increases in dollars in each country and then compare this to the expected inflation rate. We do not need to speculate about this number, as it is readily and publicly available, at least for 2021 and 2022.

The exercise is to compute a country-wise dollar deflator, which can be computed by dividing nominal dollar-denominated GDP for each country by real GDP. We then subtract this number from the expected US inflation in 2019. To illustrate the methodology, Table 5 shows the nature of the exercise for two cases.

Table 5 Revised effect, selected cases
Full size table

Note that dollar prices in Canada and China increased faster than in the USA in 2021 but slower in 2022 (when the dollar appreciated). The last column shows the local dollar price inflation surprise, but subtracting US inflation. We extend this computation for all countries.

We should expect to see that countries for which the real exchange rate appreciates have higher price level increases when measured in dollars. Figure 6 shows that this is the case in our computation. As can be seen, a significant number of countries in our sample saw their currencies appreciate viz. the dollar.

Fig. 6
figure 6

Local US GDP deflator vs REER

Full size image

We then replicate our estimation of the reduction in debt burdens using actual dollar inflation. The country-by-country results are presented in Table 11 of Appendix. The exercise, somewhat surprisingly, shows that once the actual dollar GDP number is computed, the total gains to sovereigns are higher, though results differ by country. The results of this new inflation measure implies gains that are 1.7 times as large (from $100 billion for the countries with available data in 2021 and 2022, vs. $170 billion in the new formulation.) Figure 7 compares both methodologies. The correlation is positive.Footnote 27

Fig. 7
figure 7

Original vs. alternative effect (% of GDP of 2020)

Full size image

In short, correcting for changes in the real exchange rate strengthens our central finding: the current inflation spike has led to a substantial reduction in debt burdens for a significant number of countries.

3.2.3 Inflation Beyond 2023

The computations above measure the reduction in the real value of liabilities arising from the inflation surprise in the 2021–2022 period. Again, the real effects should consider the expected inflation increase in 2023 and beyond. As with the USA, the estimate depends on both the duration of the inflation shock and on the maturity structure of the debt.

To do this computation, we can use equation (4). While not as easily available, the average maturity of countries’ debt is available for a number of countries. Column 1 in Table 6 shows these maturities for selected countries for which maturity data is available. Maturities vary across countries.

What remains for the computation is to obtain estimates of the dollar inflation in each country. But from our analysis above, we can use US inflation, so again we use the change in expected US inflation, obtained from the WEO, as well as in alternative scenarios that we already showed in Table 2.

Table 6 Effect by inflation path scenarios (% GDP)
Full size table

The gains to other countries increase if the inflation shock persists. A more persistent inflation shock could mean, on average, the doubling of benefits to countries.

3.3 Caveats

Our results assume that there is inflation only in the USA, though inflation has risen in other parts of the world. Sovereigns that have issued debt in their own currencies will therefore gain from the reduction in the real value of their outstanding liabilities. Our estimates should be thought of as applying to the specific impact of US inflation rather than the consequences of global inflation in general. However, they bring attention to the declining value of sovereign debt stocks, an issue that has received insufficient attention.

In our computation, we have ignored holdings of US dollar-denominated assets by central banks. These holdings may imply a loss to sovereigns, thereby reducing the computed gains. We have ignored these holdings because it is difficult to find data on both the maturities and currency denomination of those assets. However, central banks frequently hold their assets as short-term instruments, so the omission may not be significant for our results (and indeed there is some evidence that holdings of dollar assets may be declining).

Our computations do not consider debts issued by multilateral financial institutions like the World Bank and IMF. The reason for this is that most multilateral debt is at variable rates (with a very small fraction of concessional loans at fixed rates).

We have focused on transfers arising from sovereign debt. However, governments account for only about one-quarter of the $4.2 trillion of dollar-denominated debt issued in emerging markets. The distributive impact of US inflation is thus more far-reaching than what we have estimated here. We leave it to future work to arrive at estimates of the scope of gains and losses for private creditors and debtors.

Finally, the large short-term gains to sovereigns may come with longer-term costs, such as higher interest rates on future debt issues and economic contractions associated with central banks’ efforts to curb inflation. However, at least for the USA, real interest rates on treasuries so far have not increased significantly (see appendix).

4 Conclusion

We have considered the distributive consequences of unanticipated inflation in the USA. The overall impact on the real value of sovereign liabilities is substantial and the largest beneficiary is the US Treasury. One-third of the inflation tax is levied abroad, particularly on large holders of US Treasury securities, including Japan and China, and countries whose residents hold significant stocks of dollar cash, such as Russia and Argentina. Only in 2021 and 2022, the USA has thus effectively received a transfer from the rest of the world of over $500 billion. But sovereigns other than the USA also secure substantial windfalls from the dilution of their dollar-denominated debt. The decline in the real value of non-US sovereign debt arising from unexpected inflation in 2021 and 2022 amounts to $104 billion, with a number of poor countries experiencing significant gains relative to their GDP. These gains come at the expense of private creditors and other sovereigns.

A key implication of our findings is that the widely anticipated turmoil in emerging market sovereign debt may be mitigated by the inflation windfall accruing to other countries. In addition, nominal US interest rates thus far have not risen as much as US inflation, so real interest rates remain negative. This distinguishes current policy from the 1980s, when real interest rates rose substantially, thereby precipitating the international debt crisis.Footnote 28 Seen from the perspective of sovereign debt issuers, the current international environment is therefore more benign than in the past. It is well-known that unanticipated inflation benefits debtors at the expense of creditors, but our work highlights the surprising set of winners and losers and the sheer size of the ongoing gains to sovereigns.