1. Sovereign Borrowing Outlook for OECD Countries

Abstract

Sovereign borrowing needs in the OECD area declined significantly in 2021 and 2022 from the record highs seen in the first year of the COVID-19 crisis as pandemic-related fiscal support policies were withdrawn. After declining for two consecutive years, sovereign borrowing needs are expected to increase in 2023 against the backdrop of weak growth prospects and rising interest rates.

This chapter examines public debt management challenges amid the current outlook of rising funding costs, high borrowing needs, heightened geopolitical uncertainties, and rapid monetary policy tightening. It mainly draws on OECD countries’ survey data to explore trends in sovereign borrowing needs, funding strategies, yield movements, interest rate risk, and the liquidity of government securities.

1.1. Introduction

This chapter analyses sovereign borrowing needs, redemptions, costs, and funding strategies for the period between 2007-22 and provides projections for 2023. The chapter first looks at trends in gross and net (of redemptions) borrowing needs, and then delves into funding strategies and costs, including the exposure of countries’ debt portfolios to interest rate hikes. Finally, it examines the implications of central banks’ (CBs) quantitative tightening (QT) programmes for sovereign issuances and the trends in trading liquidity of government securities. The analysis in this chapter is based primarily on data collected from OECD countries’ debt management offices (DMOs) covering realised figures for 2021, estimates for 2022 as of October 2022 and projections for 2023 (see Annex 1.A for details of the methodology used).

Key findings

OECD governments’ gross borrowing needs further declined to USD 12.2 trillion in 2022, down an estimated 15% from 2021, and 21% from the peak in 2020. This downward trend results from decreasing net borrowing needs, which have fallen substantially since 2020, as well as diminishing refinancing needs, which peaked in 2021 and decreased in 2022.
Gross borrowing needs are projected to increase by around 6% in 2023, to USD 12.9 trillion, due to the increase in both refinancing and net borrowing needs. While the COVID-19 crisis increased government net funding needs across the OECD area, this further increase is largely confined to those countries most impacted by Russia’s war of aggression against Ukraine.
Outstanding central government marketable debt for OECD governments remains elevated relative to pre-pandemic levels, but is falling as a percentage of GDP. The total debt stock in the OECD area increased by USD 10 trillion between 2019 and 2021 to USD 50 trillion and, is projected to reach almost USD 52 trillion in 2023. Outstanding central government debt declined as a share of GDP, from a peak of 90% in 2020 to 83% in 2022, and is projected to remain relatively stable in 2023 – though still about 10 percentage points above pre-pandemic levels.
The cost of new borrowing for OECD governments has more than doubled since 2021, driven predominantly by tighter monetary policy across most countries. The OECD average yield of sovereign bonds at issuance rose from 1.4% in 2021 to 3.3% in 2022. Importantly, the volume share of sovereign bonds issued with negative yields in the OECD area declined from 17% in 2021 to 8% in 2022, primarily comprising issuances made in the year’s first quarter.
Average term-to-maturity (ATM) of outstanding debt is at an all-time high, at eight years and two months in 2022, almost two years longer than the ATM in 2012, owing to sovereign issuers’ endeavours to minimise refinancing risks. Despite the extended maturities, debt redemption profiles are expected to be elevated and 47% of OECD marketable debt is anticipated to be redeemed or refixed under new interest rates until the end of 2025.
Quantitative tightening has implications for the investor base of sovereign debt, given that approximately one-quarter of the outstanding sovereign debt is held in the form of government securities by central banks, with half of this amount expected to mature by 2030
Sovereign debt managers perceive a decline in liquidity conditions in various markets, reflecting a confluence of factors, including macroeconomic uncertainty, monetary policy developments, geopolitical risks and deterioration of investors’ sentiment. In response, debt managers have taken measures to support market liquidity through benchmark bond programmes, communication strategies, buyback and switch operations, primary dealership systems and security lending facilities.

1.2. Trends in total gross and net borrowing by OECD sovereigns

1.2.1. Gross borrowing needs declined by approximately 15% in 2022, yet they remain 35% higher than the figures recorded in 2019, prior to the pandemic

Total gross borrowing by OECD governments, which soared to a record amount of USD 15.4 trillion in the first year of the COVID-19 pandemic, diminished in 2021 and is estimated to have declined further in 2022, to USD 12.2 trillion (Figure 1.1 Panel A), a larger decline than previously anticipated.1 Nonetheless, this downward trajectory is projected to reverse in 2023, with a 6% increase in borrowing requirements to USD 12.9 trillion, driven by slower economic growth prospects and amplified fiscal imbalances. Despite the decline in borrowing requirements from 2020 to 2022 (estimated at USD 3.3 trillion or 21%), governments continue to borrow roughly 45% more than in the previous decade, when borrowing amounts remained largely stable.

Figure 1.1. Borrowing, fiscal and debt outlook in OECD countries

In 2022, total sovereign debt issuance remained 35% above the figures recorded in 2019 prior to the COVID-19 pandemic. This is primarily due to a considerable increase in refinancing needs (i.e. issuances intended for rolling over debt) and, to a lesser degree, a modest rise in new borrowing requirements (i.e. gross borrowing net of debt redemption) (Figure 1.1 Panel B and C). More specifically, new borrowing requirements, which peaked in 2020 at USD 7.3 trillion, are estimated to have decreased to USD 2.0 trillion in 2022 before a modest projected increase in 2023, remaining 43% above the 2011-19 average. While net borrowing needs declined sharply after 2020, refinancing needs peaked in 2021 at USD 11.3 trillion and are anticipated to have dropped to USD 10.2 trillion in 2022.2 However, this is expected to reverse slightly in 2023, with refinancing needs rising to USD 10.6 trillion due to increased short-term securities issuances in 2022.

Figure 1.1 Panel B also depicts a noteworthy shift in the borrowing behaviour by governments relative to their fiscal needs over the past three years. Before 2020, net borrowing needs were below or very close to the government deficit, meaning that governments borrowed equally to or less than their fiscal deficits. In 2020, net borrowing requirements exceeded the government deficit by an unprecedented amount of USD 1.6 trillion. This implies that governments within the OECD region capitalised on favourable funding conditions to issue debt exceeding their fiscal requirements, in large part due to the uncertainty surrounding actual fiscal needs. Consequently, DMOs augmented their liquidity buffer in case of further upward revisions to funding requirements. In 2021 and 2022, this surplus cash balance was employed to finance a portion of the borrowing needs, resulting in government deficits surpassing the net change in marketable debt by USD 1.2 trillion and USD 0.2 trillion, respectively. This signifies that gross borrowing needs in 2020 were higher than expected and lower in 2021 and 2022 if the entirety of government deficits were financed by debt issuance within the respective period. The extent to which the surplus cash balances from 2020 issuances have been exhausted, and thus whether the reduction of the liquidity buffer will continue to smooth gross borrowing needs in the upcoming years, remains uncertain. In this Outlook’s forecasts for 2023, it is assumed that they will not.3

1.2.2. The OECD central government marketable debt stock decreased from 87% of GDP in 2021 to 83% in 2022 but remains 10 percentage points above pre-pandemic levels

Figure 1.1 Panel D shows that the outstanding amount of debt increased from approximately USD 40 trillion in 2019 to USD 49 trillion in 2020. Despite the phasing out of extensive pandemic-related fiscal support in 2021 and 2022, it rose to USD 50 trillion in 2021 and remained stable at that level in 2022. It is projected to reach approximately USD 52 trillion in 2023 on slower economic growth anticipated for the year, which could affect tax revenues, and the new expenditures arising from packages designed to mitigate the impact of Russia’s war of aggression against Ukraine, including measures to alleviate inflationary pressures on households.4 Additionally, as this analysis converts all values in local currencies to USD, foreign exchange rate movements and the conversion criteria can influence projections (Box 1.1).

While the outstanding debt level for the OECD area has increased in nominal terms, it has fallen as a share of GDP to an estimated 83% in 2022, down from its peak of 90% in 2020(Figure 1.1 Panel D). The ratio is projected to remain relatively stable in 2023, but is still about 10 percentage points above pre-pandemic levels.

An increasing outstanding amount of debt means that a higher amount of debt needs to be refinanced. In a period of rising interest rates, a larger volume of debt will need to be rolled over at higher rates, potentially adding strains to fiscal budgets. Furthermore, a combination of reduced market liquidity and a sizeable volume of debt to be refinanced could cause yields to be more volatile and sensitive to shocks.

Box 1.1. Definitions and conceptual approach: Revisions in the Survey template and methodology

The Sovereign Borrowing Outlook (SBO) publication draws mainly from the Central Government Marketable Debt and Borrowing Survey (BO survey), which collects gross borrowing requirements, redemption and outstanding debt amounts with a breakdown by type of instrument and currency. Originally developed in 2009 by the OECD’s Working Party on Debt Management (WPDM), the survey was revised in 2021.

The main revisions are the following:

i. Debt in foreign currency is now converted into national currency at the end of each year for years prior to the current year, and as of 30 June of the current year to estimate the current year and subsequent figures (previously, all figures collected through the survey were converted into USD using exchange rates from 1 December 2009);
ii. Inflation-linked debt: Information on inflation-linked debt (i.e. linkers) is now collected separately and includes accrued inflation when appropriate. Previously, these instruments were categorised under index-linked debt, which did not allow for a distinction between other index-linked and inflation-linked debt;
iii. New data items are now collected Sustainable bonds, DMO holdings and NextGenerationEU loans; and
iv. The survey now gathers comments on country-specific methodological aspects of the figures, which can be found in Annex A of this publication along with the revised definitions of variable rate notes, inflation-linked debt, and ATM.

As this version of the SBO incorporates revisions, the figures displayed differ from those in previous editions of the publication. One notable revision concerns the date for foreign currency conversions. The BO Survey was designed in 2009 after the Global Financial Crisis (GFC), and the foreign exchange rate (FX rate) used to convert local currencies to USD was fixed as of December 2009. During the GFC, the US dollar hit its lowest value against many OECD currencies, including the euro (EUR) (since its introduction) and the Japanese Yen (since 1995). Conversely, in the second half of 2022, strong monetary tightening in the US and a “flight-to-quality” movement from investors in response to elevated geopolitical and macroeconomic uncertainties helped the US dollar reach its highest value since 2002, 1990, and 1985 against the euro, Japanese yen, and British pound, respectively. These fluctuations in USD foreign exchange rates significantly impact the USD values used in the SBO for aggregating debt figures (as illustrated in Figure 1.2). For example, under the previous methodology based on 2009 FX rates, the outstanding amount of debt would have increased from USD 59.8 in 2021 to 62.5 trillion in 2022.

Figure 1.2. Central government marketable debt stock in the OECD under different exchange rate assumptions

1.3. The share of short-term instruments in the sovereign debt stock remained stable at 12% in 2022, slightly above the pre-pandemic levels

Figure 1.3 Panel A displays the composition of the debt stock in OECD countries, highlighting that the share of short-term instruments, after peaking at 16% in 2020, declined to 12% and remained stable at that level in 2022. The peak in 2020 was a consequence of DMOs’ use of short-term borrowings as a shock absorber to accommodate the considerable uncertainties and fiscal pressures caused by the pandemic. At the onset of the COVID-19 crisis, sovereign issuers in many OECD countries (e.g. France, Germany and the US) expanded their short-term borrowing programmes, which doubled the outstanding amount of short-term debt in 2020 compared to 2019 (Figure 1.3 Panel B).5 After rising rapidly in 2020, short-term instrument issuance has fallen notably over 2021 and 2022, as a result of the strategic choice of shifting from the money market to longer-term funding to both reduce interest rate sensitivity of the debt stock and rebuild contingency capacity if significant funding is needed again in short order. However, the share of short-term debt in total debt stock still remains slightly higher than pre-pandemic levels.

In nominal USD values, the outstanding amount of short-term instruments was estimated to be 45% larger by the end of 2022 compared to 2019 (USD 5.8 trillion compared to USD 4.0 trillion). The increase in the stock of fixed-rates, variable rates, and inflation-linked securities during the same period was much lower at 23%, 10%, and 5%, respectively. Although sovereign issuers reduced the share of short maturities in 2021, the pace of the reduction slowed in 2022, a year with high macroeconomic uncertainty and rising interest rates.

Figure 1.3. The composition of central government marketable debt by instrument type

1.4. Russia’s war of aggression against Ukraine affected the net borrowing of OECD countries, particularly those in proximity to the conflict area

Figure 1.4 illustrates the estimated net borrowing to GDP ratio fluctuations from 2021 to 2022 and the projected changes from 2022 to 2023 across OECD countries at the national level. Net borrowing requirements diminished, on average, by 1.3 percentage points (pp) of GDP in the OECD between 2021 and 2022, mainly due to declining pandemic-related government expenditures and stronger-than-anticipated tax revenues. At the same time, there was a wide variation between countries: in ten countries net borrowing requirements increased by over 0.5 percentage points of GDP in 2022 compared to 2021, whereas in 21 countries, they decreased by more than 0.5 percentage points. Net borrowing requirements are projected to increase between 2022 and 2023 by 0.3 percentage points of GDP on average, with a similar level of variation. A significant part of the rise in net borrowing requirements in 2023 is related to fiscal support measures. Notably, 10 countries among the 12 where net borrowing requirements are projected to grow by more than 0.5 percentage points of GDP from 2021 to 2023 are European. This reflects the impact of higher energy and food inflation in Europe and the fiscal support measures implemented to protect households and businesses from the consequences of elevated energy prices (OECD, 2022[1]). Five of the ten European countries share borders with Russia, Ukraine, or Belarus – namely Estonia, Finland, Latvia, Lithuania, and Poland. As a ratio to GDP, the 12 countries’ net borrowing requirements are projected to be, on average, two times greater than their 2019 pre-pandemic levels by 2023.

Figure 1.4. Net borrowing requirements as a ratio to GDP (%) by country in the OECD area

1.1. Funding strategies in times of high uncertainty

1.1.1. Sovereign issuers aim to minimise borrowing costs and risks by issuing in a regular, transparent, and predictable manner, regardless of the interest rate path

A funding strategy involves allocating financing needs across a variety of debt instruments over time. It requires two key elements: a forecast of gross borrowing and the composition of issuance by instrument. Its purpose is to meet governments’ financial needs while minimising borrowing costs and the level of risk. However, there is a trade-off between these two objectives. Short-term instruments typically carry lower costs but increase refinancing and interest rate risks. In contrast, instruments with longer maturities and fixed-rates minimise refinancing and interest rate risks but come at a higher cost. This is why sovereign issuers maintain a diverse range of nominal and price-indexed instruments to minimise costs while controlling risks across the maturity spectrum. The precise choice of instruments depends on various factors, such as instrument liquidity, investor demand and base, the level of financial market development, the outstanding amount of debt maturity and macroeconomic conditions.

There are compelling theoretical arguments for sovereign issuers to adopt a regular and predictable funding strategy regardless of expected interest rate movements. These include improving market confidence, reducing the underwriting burden on Primary Dealers (financial firms that purchase sovereign bonds directly from governments to resell them to investors), facilitating investor planning, and minimising the risk of insufficient offering demand. DMOs achieve these objectives by publishing their borrowing programmes and goals in advance, issuing press releases, and regularly communicating with the market. Even during times of high volatility, sovereign issuers aim to maintain a predictable and regular issuance strategy. Attempting to predict market movements and adapting the strategy accordingly may yield short-term benefits for the sovereign issuer. However, over the medium- to long-term, investors would adjust, increasing market volatility and negatively affecting sovereign costs and risks.

1.1.2. Market volatility and unpredictability of funding needs impacted borrowing plans, which were revised in 2022 in varying ways across the OECD

The recent survey on Primary Market Developments indicates that sovereign DMOs have been experiencing several challenges in their operations, such as high market volatility (reported by 88% of respondents), uncertain cash flow (70%), and unpredictable funding needs (68%). These challenges made it difficult to implement borrowing plans as initially intended.6 The COVID-19 crisis led to increased uncertainty surrounding financing needs and investor demand for certain security instruments. Russia’s war of aggression against Ukraine further exacerbated the situation by causing a sharp rise in food and energy prices, adding to global inflation at a time when prices were already rising rapidly. Global financial conditions tightened sharply as central banks increased interest rates in an effort to lower inflation (OECD, 2022[1]). This tightening in financial conditions raised the perception of a possible recession. Borrowing needs became less predictable as tax revenues tend to grow with inflation, as do expenditures where countries have pursued relief packages aimed at mitigating the effects of food and energy inflation on households and businesses.

Sovereign issuers have adjusted their borrowing strategies to the current economic outlook, adopting more flexible yet transparent approaches by modifying the volume, number, and instrument composition of planned issuances over the year (Figure 1.5). The most common changes in borrowing plans were related to the number of issuances across the yield curve, the composition of issuances and the introduction of new maturity lines. Other frequent changes were related to the frequency of auctions, the number of new types of securities, the use of syndication, and the use of buyback operations.

France, Italy and Japan offer examples of countries that revised funding plans or borrowing needs in different ways in 2022. Japan revised up its funding needs in May and November of 2022 by JPY 2.7 trillion and 9.7 trillion respectively (approximately USD 92 billion combined), both reflecting increasing fiscal needs (Ministry of Finance, 2022[6]; Ministry of Finance, 2022[7]); with adjustments between fiscal years, market issuance by periodic auctions remained unchanged in May, and increased by JPY 4.5 trillion in November. France revised its borrowing requirement upwards by EUR 13.4 billion in July 2022 to accommodate the impact of a supplementary budget bill that reflected measures to mitigate the decrease in purchasing power due to inflation. Nevertheless, this did not affect its funding strategy given that this additional expenditure was funded by the Treasury’s account, whose cash availability rose in 2020 and 2021 (Agence France Trésor, 2022[3]). Italy cancelled bond auctions scheduled on 11 and 25 August 2022, and on 28 and 29 December 2022, the first due to the large cash availability of the Treasury and the second because the funding target of the year was already reached (Ministero dell’Economia e delle Finanze, 2022[4]; Ministero dell'Economia e delle Finanze, 2022[5]).

Figure 1.5. Changes in borrowing plans in 2022 (number of countries)

1.1.3. Long-term securities constituted 56.3% of 2022 issuances, less than in 2021

The shares of short-term and inflation-linked securities in total issuance are estimated to have grown moderately in 2022 compared to 2021 (Table 1.1): from 37.3% to 43.7% for short-term securities, and from 2.4% to 2.8% for inflation-linked securities. The fixed-rate share is estimated to have decreased from 57.3% to 50.4% between 2021 and 2022. From the demand side, this might capture the fact that investors are protecting themselves from inflation and interest rate risk by preferring inflation-linked and short-term securities. On the supply side, sovereign issuers may also be utilising short-term instruments to maintain a more flexible borrowing plan, as they can help to smooth (historically high) financing needs, and to avoid the risks associated with the issuance of long maturities when rates and volatility are high. These two forces contribute to the widespread reduction in the term spread (refer to Figure 1.8).

Importantly, these trends varied across countries – the rise in the issuance of short-term securities as a proportion of total issuances was especially observed in the United Kingdom (from 17% in 2021 to 29% in 2022), the United States (from 42% to 49%), Canada (from 41% to 66%), and Switzerland (from 57% to 82%). Conversely, this share decreased in the European Union (EU) area (from 28% to 25%) during the same period and remained relatively stable in Japan. For 2023, it is projected that countries will experience a gradual convergence to the composition observed in the pre-pandemic era, with the relative share of short-term issuances decreasing by 0.2 percentage points. Box 1.2 delves into the persistence of short-term instruments on countries’ debt portfolios since the pandemic.

Table 1.1. Funding strategy based on gross marketable borrowing needs in the OECD area
OECD Countries
	2017-19 Average	2020	2021	2022e	2023p
Short term	3.2 \| 36.8%	6.7 \| 43.7%	5.4 \| 37.3%	5.3 \| 43.7%	5.6 \| 43.5%
Long term	5.5 \| 63.2%	8.7 \| 56.3%	9 \| 62.7%	6.8 \| 56.3%	7.3 \| 56.5%
Fixed-rates	4.9 \| 56.5%	8 \| 51.9%	8.3 \| 57.3%	6.1 \| 50.4%	6.5 \| 50.3%
Inflation-linked	0.3 \| 3.6%	0.3 \| 2.2%	0.3 \| 2.4%	0.3 \| 2.8%	0.4 \| 2.8%
Variable rate	0.3 \| 3%	0.3 \| 2.2%	0.4 \| 3%	0.4 \| 3%	0.4 \| 3.4%
Trillion USD \| as a share of total borrowing
Note: Standardised gross borrowing requirements are shown. Values in local currencies are converted to USD using the exchange rate on the last day of the year. Cash management related issuances are excluded. For 2023 exchanges rates are fixed.
Source: 2022 Survey on Central Government Marketable Debt and Borrowing; OECD Economic Outlook 112 (November 2022); IMF (2022[2]), World Economic Outlook, https://www.imf.org/en/Publications/WEO/Issues/2022/10/11/world-economic-outlook-october-2022; Refinitiv, national authorities’ websites and OECD calculations.

Box 1.2. Managing public debt: A stock-and-flow perspective

Sovereign issuers aim to minimise borrowing costs subject to a prudent level of risk and may do so by pursuing a particular debt structure. Debt portfolios vary over time depending on the dynamics between gross financing needs (i.e. refinancing needs plus net borrowing needs) and bond issuances. In this respect, the composition of gross financing needs influences borrowing strategies through two main channels. Firstly, on the bond investors’ side, as they typically pursue a consistent structure for their investment portfolio over time, they will need to reinvest the proceeds of maturing issues into similar securities from the same issuer to maintain their portfolio composition, thereby supporting demand for refinancings. Net borrowing needs, which are associated with fiscal needs (but not identical as fiscal deficits can be funded through means other than debt), on the other hand, represent additional exposures in the market for investors. Secondly, on the sovereign issuers’ side, the evolution of the debt stock composition is a function of both maturing securities and the issuances made to refinance them. Consequently, the rate at which DMOs can converge their debt portfolios to a benchmark debt portfolio is directly related to issuance and refinancing needs.

Figure 1.6. displays the borrowing needs and strategies in 2022 and 2021 in the OECD area. In 2021, refinancing needs from short-term instruments reached USD 6.7 trillion, while countries issued only USD 5.4 trillion in short-term instruments, reducing their share in debt portfolios (refer to Figure 1.3). Conversely, fixed-rate instruments’ refinancing needs were USD 4.1 trillion in the same year, while their issuances reached USD 8.3 trillion. As the net issuance of fixed-rates bonds (USD 4.2 trillion) was greater than net borrowing needs (USD 3.2 trillion), the exceeding amount was used to fund short-term redemptions. This has resulted in an increase in the share of the fixed-rates in debt portfolios. In 2022, the proceeds raised through the issuance of fixed-rates were only sufficient to cover refunding needs from fixed-rates and new borrowing needs. Consequently, the relatively large outstanding amount of bills, inherited from 2021, was left virtually unchanged in 2022.

Figure 1.6. Borrowing needs and funding strategies in the OECD area (USD trillion)

Figure 1.6 also illustrates recent changes in inflation-linked debt stock in the OECD area. For example, the share of inflation-linked instruments in total issuances is estimated to have grown by approximately 27%, from 2.2% in 2020 to 2.8% in 2022 (Table 1.1). Despite the increased emphasis on inflation-linked instruments in borrowing strategies, their outstanding amount is estimated to have decreased in both nominal (from USD 3.40 to 3.35 trillion) and relative terms (from 6.9% to 6.8% of the debt stock) between 2020 and 2022. This was mainly the result of large redemptions of inflation-linked instruments in 2022 (increasing from USD 213 to 0.5 trillion from 2021 to 2022), that significantly exceeded the issuance amount, leading to a decrease in the direct exposure of countries to the changes in inflation.1

← 1. Carrying short-term and variable-rate debt instruments also increases the exposure of sovereign issuers to inflation. That is because interest rates are correlated with inflation (and, thus, it affects the costs of variable rate instruments) and with the refinancing costs of short-term debt.

1.2. Rising borrowing costs in the OECD area in 2022

1.2.1. Inflationary pressures have prompted rapid monetary tightening, moving the OECD average policy rates from 0.8% to 5.3%

The secular declining trend of nominal and real interest rates was among the most consequential macroeconomic and financial developments in recent history and was particularly pronounced in the aftermath of the GFC. This decline has been linked to structural drivers including demographic trends (Lunsford and West, 2019[8]), changes in saving and investment propensities (Rachel and Summers, 2019[9]) and regulatory reforms (Ranaldo, Schaffner and Vasios, 2021[10]). Although none of these drivers reversed in 2022, inflationary pressures have prompted central banks to raise policy rates, halting a decades-long trend.

Inflation reached a nearly four-decade high across many OECD countries in 2022. Russia’s war of aggression against Ukraine aggravated supply disruptions caused by measures to contain the COVID-19 outbreak, driving up energy and food prices as demand and employment in OECD economies rebounded from a severe recession (OECD, 2022[1]). Once inflation becomes entrenched, the self-reinforcing feedback between prices and wages makes the costs of transitioning back to a low-inflation regime high – requiring a stronger monetary response and raising the chances of a recession (BIS, 2022[11]). Central banks, by and large, have embraced rapid monetary tightening to avoid persistent inflation (Figure 1.7).

Figure 1.7 depicts the increase in policy rates across the OECD. The time series of policy rates in Panel A demonstrates that for the United States, the euro area, and the United Kingdom, policy rates reached their highest levels since 2008. A similar pattern can be observed for other OECD countries, as represented by the shaded area in the chart. Panel B indicates that, on average between 2021 and 2022, policy rates shifted from 0.8% to 5.3%. Emerging OECD countries and European countries outside the euro area make up the most significant increases in this period (average growth of roughly eight percentage points). In contrast, the smallest increases were in the euro area, Korea, Nordic countries and Switzerland. The Bank of Japan (BoJ) did not raise interest rates, and the Turkish Central Bank reduced policy rates during the period. Central banks’ responses also varied in terms of the timing of the introduction of monetary tightening policies (Figure 1.7 Panel C), with four central banks rising policy rate hikes in the first half of 2021, seven in the second half of 2021, and eight until July 2022, including the Federal Reserve (Fed) and the European Central Bank (ECB).

Figure 1.7. Central banks’ policy rates in the OECD area (selected OECD countries)

1.2.2. Yield curves shifted up, with the term spread shrinking in most OECD countries, eventually flattening or inverting curves

With policy rates moving upward, bond yields tend to follow suit.7 Figure 1.8 shows the changes in two- and ten-year benchmark yields for selected OECD countries between December 2021 and December 2022. More specifically, the OECD average for the 2-year benchmark yield increased from 0.5% to 3.3% while the ten-year one rose from 1.1% also to 3.3%, which implies that the shape of the OECD “average curve” changed from upward-sloping to flat. This pattern is also visible in Figure 1.8 Panel C, displaying a contraction in the term spread (difference between 10- and 2- year benchmark yields) across the OECD area, a common occurrence during periods of monetary tightening as short-term yields are more affected by the overnight policy rates while longer-term yields reflect to a large extent market expectations for the steady-state neutral rate.8 The average reduction in the term spread was 0.6 percentage points, which precisely matches the average term spread in 2021, signifying that the average term spread reached zero in December 2022. Excluding Greece and Japan, where it increased by 0.3 percentage points and 0.2 percentage points, respectively, and Italy and Türkiye, where it remained constant, the term spread declined in all other 24 countries in the figure. The largest decreases took place in Chile (2.8 percentage points), the United States (1.5 percentage points, where the yield curve was inverted as of December 2022, as shown in Figure 1.8 Panel D), Canada (1.3 percentage points), and Israel (1.1 percentage point), demonstrating that this movement affected both large and smaller issuers.9

Figure 1.8. Change in benchmark yields between December 2021 and December 2022

The shape of yield curves is, nevertheless, relevant for sovereign issuers’ borrowing strategies given its implication in the classic trade-off between borrowing costs and risks (OECD, 2012[12]). In the more common scenario of upward-sloping curves, long maturities tend to have lower risks as they reduce refinancing risks and lock in a coupon for a longer period, protecting it from changes in short-term interest rates. For opposite reasons, issuances of short maturities tend to elevate refinancing and interest rate risks while minimising borrowing costs in the short run. When curves are inverted, however, although long maturities can be cost-minimising, there is also the potential for regret risk associated with the issuance of longer securities – issuers are locking their yields in for the long term during a period of higher yields. This can also partially explain why some sovereign issuers opted to increase short-term borrowing in 2022 (refer to Table 1.1). There are also liquidity considerations that affect borrowing strategies. When a government issues a relatively high volume of long maturities, the market buys the duration risks (i.e. exposure to future interest rate hikes). If there is insufficient demand for duration risk, the premium on long maturities will increase, affecting long-term yields. Against this background, a relatively balanced issuance split across the maturity spectrum is often targeted even when curves are inverted.

1.2.3. Sovereign bond yields increased at the fastest and strongest rate in the past three decades coupled with widening spreads between countries

Another important development has been the widening of spreads between countries, particularly at the longer end of the yield curve. For instance, in the euro area, where countries are subject to the same monetary policy, the spreads between Germany’s and Southern European countries’ yields have broadened from December 2021 to December 2022 for the 10-year benchmark, more precisely by 1.1, 0.7, 0.5 and 0.4 percentage points for Greece, Italy, Portugal, and Spain, respectively. The spreads of the 2-year benchmark for these countries remained largely stable in the same period – in fact, it decreased for Greece by 0.5 percentage points and did not vary by more than 0.1 percentage point in the other three countries.

Regarding the relative speed and intensity of the surge in bond yields, Figure 1.9 demonstrates that the increase in yields, which commenced in 2021, ranks among the most rapid relative to other recent periods of sustained and significant yield increases – 1994-95 (the 1994 Bond Market Crisis), 1999-2000 (the Dot-com Bubble), 2004-07 (the Housing Market Boom), and 2016-18 (the attempt to normalise monetary policy post-GFC). Figure 1.9 Panel A reveals that the increase in the 2-year bond yield benchmark between September 2021 and December 2022 was the sharpest among the last five periods of sustained rising yields on average for the G7 countries. The yield on the 2-year bond benchmark climbed quickly starting from September 2021 in anticipation of monetary policy tightening. Compared to other episodes shown in the figure, this was the strongest rise. Similarly, the 10-year bond yield, which began to rise slowly the second half of 2021 when inflation started to escalate in the OECD area, picked up during 2022. The recent episode of increase in the 10-year benchmark yield exceeded all those from other periods of sustained growth in yields except from the tightening cycle in 1994-95.

Figure 1.9. Changes in G7 benchmark bond yields following the start of the central banks tightening cycle

1.2.4. The increases in nominal yields were largely driven by real yields in the US and UK, and inflation expectations in the euro area

Breakeven inflation captures inflation expectations and inflation risk premiums, with the latter referring to the risk that investors take to lock in their returns to the current inflation expectation (and, thus, correlated with uncertainty about future inflation). Real yields are associated with growth and growth volatility expectations, as they serve as a proxy for the rate of return on investments. An increase in real yields can signal improved growth prospects or heightened economic uncertainty. Given that growth prospects for the OECD area have been revised downward since 2021, this uptick in real yields can be interpreted, as a consequence, of increased macroeconomic and geopolitical uncertainty.

Figure 1.10 illustrates the variation in real yields across selected countries between the last quarter of 2021 and 2022. The US real yield curve shifted from negative and upward-sloping to downward-sloping and entirely positive. Germany, France and the UK also experienced upward shifts in their yield curves, however, with France and the UK surpassing zero thresholds only for longer maturities and Germany’s curve remaining negative but nearing zero. Real yields rose equally or more than nominal yields in the UK and the US, and less than nominal yields in France and Germany. Thus, the growth in yields was primarily driven by real rates in the United States and the United Kingdom, while in France and Germany, it was influenced by a combination of both an increase in breakeven inflation and real yields (Figure 1.7).

Figure 1.10. Real yield curves for selected sovereign issuers in the OECD area

1.2.5. Positive real yields can put more constraints on governments’ fiscal policy

Prior to 2022, negative real yields were prevalent across developed economies, allowing governments, in principle, to sustain higher fiscal deficits without jeopardising their debt sustainability (Blanchard, 2023[13]). However, given that interest rates are endogenous to fiscal instances (i.e. they depend on how market participants perceive fiscal policy to be sustainable), revising down fiscal targets can lead to increasing rates. When real rates rise, the fiscal balance further deteriorates due to escalating borrowing costs, creating a feedback loop that, in extreme scenarios, can result in a debt default. Although the impact of fiscal policy on borrowing costs is more pronounced in emerging markets due to their heightened vulnerability to debt repayment, developed countries may also face fiscal constraints, even when real yields are negative. A notable example of this was the severe market stress in the UK bond market during September 2022, where a fiscal announcement in the United Kingdom immediately affected bond yields, imposing high costs on the government, which was subsequently compelled to revise the recently announced fiscal plan within a month (Box 1.3).

Box 1.3. The September 2022 UK bond market: from a public debt management perspective

On 23 September 2022, the UK Government published a Growth Plan presented in a fiscal statement, which included unfunded tax cuts. This extension of the unfunded tax cuts and the absence of independent scrutiny by the Office for Budget Responsibility came as a surprise to the UK Government bond (gilt) market and immediately resulted in a sharp increase in long-term bond (gilt) yields. This sudden and sharp rise in yields also led to urgent margin calls (i.e. demands for additional capital to cover leveraged positions due to the depreciation of the value of collateral) for investment vehicles (particularly pension funds) engaged in liability-driven investments (LDI) (BIS, 2022[14]). To raise cash in order to meet demands for collateral, or to reduce exposure to collateral calls, some LDI funds sold gilts, which further reduced particularly long-dated gilt prices, increasing yields and forcing those funds to post more collateral, thereby creating a negative spiral of rising yields and falling prices in the gilt market.

To stabilise markets, and assist pension funds in managing their liquidity and in closing out their positions, the Bank of England (BoE) announced on 28 September 2022 temporary and targeted purchases of long maturity gilts (Bank of England, 2022[15]). Although the BoE successfully helped stabilise the market, yields remained notably higher than before the announcement of the Government’s Growth Plan (Figure 1.11).

Yields only returned to pre-announcement levels in late October 2022 when a reversal of most of the proposed tax cuts and a reduction in the scope of the relief package set out in the Growth Plan were announced (HM Treasury, 2022[16]). During the period between 23 September and 20 October 2022, the average yield to maturity of conventional gilts rose to 4.2%, higher than both the period prior to the announcement of the Growth Plan and the subsequent months. During this period, yields on government bonds in other major markets were also increasing. However, compared to other large issuers in the OECD area, the increase in the UK’s 10-year benchmark yield in this period was more than 50% higher than the average increase for Canada, France, Germany, Italy and the US in the same period.

From a debt management perspective, the new fiscal plan necessitated a significant increase in the UK Government’s borrowing requirement (HM Treasury, 2022[17]). As usual on the day of a fiscal event, the UK’s Debt Management Office (UK DMO) published on 23 September 2022 a revised financing remit for the financial year 2022-23. The remit for each financial year is first published shortly before the start of the relevant financial year, with subsequent revisions typically occurring twice a year (a technical revision in April, and a further revision during the autumn alongside a fiscal event) (UK DMO, 2023[18]). The UK DMO’s net financing requirement was revised up by GBP 72.4 billion to GBP 234.1 billion for the financial year 2022-23, representing an increase of approximately 45% since the previous remit revision that was published on 26 April 2022. The planned composition of gilt issuance also changed, with the proportion of short maturities rising from 32% to 38%, alongside a decrease in the share of long maturities and index-linked securities. Throughout this challenging period between 23 September and 17 November, the DMO remained committed to its already-announced issuance programme and successfully raised more than GBP 31 billion via 11 different gilt operations, notwithstanding the volatility of the market; and another GBP 31 billion in Treasury Bill sales.

On 17 November 2022, following the announcement by a new government of a reversal of most of the policies set out in the Growth Plan (UK DMO, 2022[19]) and alongside the publication of the Economic and Fiscal Outlook by the independent Office for Budget Responsibility (OBR), the UK DMO published a further revision to its borrowing remit. This time, the net financing requirement was revised down by GBP 31.4 billion to GBP 202.7 billion for the fiscal year.

Figure 1.11. The amount issued and auction yields during the period of market stress

Overall, the stress episode in the gilt market offers insights into the impact of a large unanticipated fiscal loosening on the cost of debt issuance. It is important to note that the increase in borrowing costs during this period, part of which may have been attributable to the market’s reaction to the announcement, will persist until the gilts issued during this period mature. Considering that the nominal yield curve for UK Government bonds was approximately flat in late 2022 (refer to Figure 1.8), this difference in yields cannot be explained solely by the maturities of the debt issued. It is also likely that it is attributable to the reaction that long-end investors had to the 23 September 2022 announcement. If the spread between auction yields and the BoE’s policy rate had remained constant between 23 September and 20 October 2022 (at 1.0pp from July to mid- September 2022, instead of 1.46 percentage points from 23 September to 20 October 2022), the issuance of GBP 17.8 billion of gilts at fixed rates would have cost approximately GBP 2.0 billion less for the UK Government until 2061 (the maturity year of the longest gilt issued over the period).

These events highlight that, even in the relatively liquid gilt market, specific sectors of the investor base can be detrimentally impacted by unexpected events. From a debt management perspective, it also demonstrates that interventions by the authorities aimed at stabilising markets can be successful and underlines the importance of maintaining a credible institutional framework for debt management and fiscal policy.

1.2.6. Average yield to maturity at issuance rose less than secondary market yields, with the volume share of securities with sub-zero yields gradually falling to zero

Yield to maturity (YTM) at issuance reflects the actual borrowing costs borne by sovereign issuers, as opposed to benchmark yields, which represent yields from secondary market transactions for specific representative securities. The OECD average YTM increased from 1.4% in 2021 to 3.3% in 2022 (Figure 1.12, Panel A), a slightly smaller rise compared to the movements observed in benchmark yields (see Figure 1.8). In general, emerging economies and countries near the conflict in Ukraine experienced the highest increases in the average YTM, such as Colombia, the Czech Republic, Hungary, Latvia, Lithuania, Mexico, Poland, the Slovak Republic, and Türkiye. Other EU countries and Japan experienced the lowest increases in the average YTM. The UK and the US both experienced a relatively high YTM increase compared to the countries in the EU area, likely explained by the proportionally higher increase in their policy rates (refer to Figure 1.7). It is also noteworthy that negative average YTMs in 2021 (found in Austria, Denmark, France, Germany, the Netherlands, and Switzerland) vanished in 2022.

Figure 1.12. Yield to maturity at issuance and volume issued in 2022

Figure 1.12 Panel B displays the volume share of fixed-rate bond issuances by yield category in 2022. The volume share of negative yield issuances has diminished significantly between 2021 and 2022, notably from 51% to 34% in Japan and 50% to 7% in the euro area. Conversely, the share of issuances with YTM above 2% rose in the same period, more specifically from roughly 2% in the euro area to 36%, from 9% to 32% in the United States, from 0% to 67% in the United Kingdom and from 5% to 78% in Canada.

Breaking down yield to maturity at issuance by month in 2022 shows that the share of primary issuance at negative yields was largest in the first quarter and decreased gradually through the year (Figure 1.12 Panel C). More precisely, until April, a portion of the boxplots' bottom whiskers, which approximately represents 25% of the volume of the issuances, had a yield below zero.10 This share declined gradually, reflecting that the volume share of the issuances with a yield below zero decreased from 18.5% in January to 5.4% in April and finally to zero in December.

Figure 1.12 Panel D presents the percentage of the amount issued across 2022 in the OECD area, showing that funding activities were more concentrated until August when borrowing conditions were relatively more favourable. By the end of June, DMOs had already borrowed 57% of the year’s borrowing needs and 73% by the end of August. This concentration, to some extent, clarifies why the average yield at issuance was generally below benchmark yields: many issuances occurred when policy rates had not yet increased and, thus, benefited from relatively favourable funding conditions.

1.3. Interest rate sensitivity in times of rising interest rates

1.3.1. The average term-to-maturity for OECD countries remained at a record high

The effect of rising government bond yields on borrowing costs is highly dependent on issuers’ debt portfolio features, such as maturity and instrument composition.11 Figure 1.13 Panel A demonstrates that the average remaining time-to-maturity for all securities composing a debt portfolio (ATM) reached a record high level in 2021 at approximately eight years and two months and remained unchanged in 2022. This can be explained by the fact that, on aggregate, in 2022 countries issued securities with roughly the same ATM as their maturing debt (see Box 1.2 for more details). In addition, the ATM of long maturities issued in 2022 moved from nine years and two months in 2021 (a record high) to nine years in 2022, both above the ATM of long maturities maturing in 2021 and 2022, respectively of six years and four months, and five years and eight months. Regarding new lines, eight OECD countries issued their longest instrument in 2022.12

Figure 1.13. The average term-to-maturity for OECD central governments’ marketable debt

OECD averages hide considerable variation across countries. Country-specific data indicates that the average ATM decreased in 11 countries in 2022. (Figure 1.13 Panel B). Among these, nine are in Europe, which could result from the uncertainties stemming from Russia’s war of aggression against Ukraine. In seven countries, the ATM remained roughly unchanged, while in 17 countries, it increased by an average of three months. Large issuers predominantly featured in countries with a rising ATM (e.g. Canada, France, Germany, Japan, and the United States), except the United Kingdom (approximately unchanged ATM) and Italy, where a modest decrease of less than one month was observed.

1.3.2. The average time-to-refixing figures indicate larger exposure to interest rate changes than the average term-to-maturity implies

Figure 1.14 Panel A contrasts the 2022 ATM with the average time-to-refixing (ATR, i.e. the weighted average of time to maturity for fixed-rates and time to refixing (adjusting) for variable rates and inflation-linked instruments within a debt portfolio). ATR is a superior indicator of borrowing cost sensitivity to interest rates, as it also accounts for the refixing effects of floating rates and index-linked instruments. On average, ATR is approximately one year shorter than ATM.

Figure 1.14. Marketable debt’s average maturity and refixing

The United Kingdom has the most significant difference between ATM and ATR, primarily because 23% of its debt stock comprises inflation-linked securities, some with very long maturities of up to 50 years. In the ATR, these securities are weighted by their refixing term, rather than their maturity in the ATM. Despite this, the UK’s ATR remains relatively long, exceeding 10 years, as it has the highest ATM in the OECD. Other countries with a moderate share of inflation-linked or variable rate instruments (more than 20%) and, consequently, a large difference between their ATM and ATR are Costa Rica, Iceland, Israel, and Poland. It is worth noting that there is a premium paid by investors to protect them from future hikes in inflation and interest rates and, therefore, although floating rates and index-linked securities do increase sovereign issuers’ exposure to interest rates, they can also be cost-minimising (see Box 1.4).

ATM and ATR are widely utilised indicators of market-risk exposure by debt management offices. However, since they do not capture the short-term and medium-term exposure to interest rate increases, countries also monitor the proportion of debt maturing or refixing in the upcoming years.13 Figure 1.14 Panel B shows that 29%, 39%, and 47% of the outstanding debt from OECD countries is due to mature or be refixed by 2023, 2024, and 2025, respectively. Thus, by the end of 2025, 47% of the total OECD debt stock will be refunded or refixed under new interest rate conditions.

A country with an 83% central government marketable debt-to-GDP ratio (the 2022 OECD average), and an average yield at issuance of 1.1% (the December 2021 OECD average for the 10-year benchmark), will see an increase in interest payments from 0.9% to 1.6% of GDP when refunding 47% of its debt stock at a 3.3% yield (the December 2022 OECD average for the same benchmark). This represents an 80% increase compared to the average OECD country in 2021. The 0.7 percentage point GDP increase, for example, is more than what OECD governments spent on environmental protection or housing and community amenities.14 It is important to note that this exposure varies considerably among countries. Figure 1.14 Panel B shows that the percentage of debt stock maturing or refixing within the next three years ranges from 22% in Luxembourg to 85% in Israel, with relatively high levels in the United States (56%) and Canada (54%), and lower levels in the EU area (37%).

Box 1.4. The costs and benefits of inflation-linked instruments (linkers)

Inflation-linked sovereign bonds have been part of the funding strategies of sovereign borrowers in the OECD for decades, originating in 1981 in the UK and later in many other OECD countries, for instance, in 1985 in Australia, 1991 in Canada, 1994 in Sweden, 1997 in the US, 1998 in France, 2003 in Italy and 2004 in Japan, 2006 in Germany, 2015 Belgium, and 2017 Ireland (OECD, 2017[22]). As of 2022, half of OECD member countries, or 19 nations, incorporated these instruments into their borrowing strategies. They are also accounting for a large share of borrowing strategies and debt stock – linkers’ share of the debt stock rose from 6.4% in 2007 to 8.1% in 2019, before the COVID-19 outbreak, especially pushed by strong and sustainable demand from pension funds and insurance companies. Then it fell to 7.0% in 2022, as the share of short-term instruments grew in response to the pandemic (see Figure 1.3).

Linkers present several advantages to sovereign borrowers, including i) Facilitating portfolio lengthening and smoothing of payments, as they protect against inflation, encouraging investors to accept more duration risk (in the OECD area, the average maturity of inflation-linked securities is 4.7 years longer than the rest of the debt portfolio); ii) Expanding and diversifying the investor base, which plays a crucial role in building funding strategy resilience, improving market liquidity, and managing funding risk, particularly with insurance companies and pension funds, whose liabilities are also inflation-linked; and iii) Enabling better debt portfolio positioning regarding cost and risk trade-offs, particularly when borrowing costs can be reduced due to a premium that investors are willing to pay for inflation protection.

The exact cost of a linker can only be determined after it matures and is compared to its counterfactual. If the realised inflation rate over the life of the bond is, on average, below the breakeven rate at issuance, linkers would be cost-efficient, and vice versa. Figure 1.15 compares breakeven inflation and realised inflation for selected OECD countries with more liquid linkers’ markets in the same period. It demonstrates that the costs of linkers, relative to their fixed-rate counterparts, have varied significantly across countries. In the United Kingdom, aside from 2022, breakeven inflation remained above the realised inflation for the same year, implying that investors paid a premium for inflation protection. The country benefited from reduced borrowing costs compared to issuing fixed-rate instruments with a similar term. This was also the case in France, Italy and the United States, but only before 2016. In Canada and Germany, realised inflation tended to align with breakeven inflation before 2016 but was slightly higher afterwards. Notably, in all of these countries, inflation exceeded its breakeven in 2022, highlighting that the magnitude of inflationary pressures took the markets by surprise.

Despite the recent increase in the relative costs of linkers compared to fixed-rates, it may be premature to conclude that they will be more expensive than their fixed-rate counterparts. Linkers typically have long maturities, and their costs depend heavily on future inflation. Surpassing the breakeven inflation for several years, even substantially, does not imply that linkers will be costlier than their fixed-rate counterparts. Decisions regarding the use of linkers are multifaceted and contingent upon the demand and liquidity of each sovereign bond market, which debt management offices carefully scrutinise. For example, Canada discontinued its Real Return Bonds programme due to insufficient demand in the last quarter of 2022 (Government of Canada, 2022[23]). New linkers’ lines were also issued in 2022 in Denmark, France, Iceland, Italy, Japan, Korea, Türkiye and the United States.

Figure 1.15. Realised headline inflation (line) versus breakeven inflation (shaded area)

1.3.3. Sovereign debt becomes more sensitive to interest rates when considering the effect of central banks’ holdings

A significant consequence of large-scale central bank purchases of government debt, known as quantitative easing (QE), is the heightened exposure of the public sector to interest rate fluctuations. Central banks acquire sovereign bonds, funding these purchases by issuing bank reserves. The remuneration of these reserves, accessible on demand for the banking sector, is closely linked to the policy rate. Consequently, sovereign instruments on central banks’ balance sheets function like floating rate notes. Figure 1.16 illustrates the adjusted ATR for the Euro area and G7 countries when central bank holdings are considered to have zero maturity, as they are funded at an overnight rate. Given that central banks hold a substantial proportion of their respective countries’ debt (see Figure 1.17) and that the ATM of their holdings is lengthy, treating them as floating rate notes results in a significant reduction in ATR. The average ATM for selected countries in Figure 1.16 of 8.6 years and ATR of 8.0 years decrease to 5.6 years when adjusted for CB holdings.

Thus, in addition to the two already explored drivers of short-term and medium-term sensitivity to interest rates (i.e. maturing bonds that are expected to be refunded under new interest rates, and floating rates and index-linked securities that are refixed under new rates or indexes), there is a third one that refers to central banks’ government security holdings, which are refixed overnight. Under this enlarged public sector perspective that also covers central banks’ balance sheet, 80%, 71%, 71%, and 67% of Japan, the United States, Canada, and the United Kingdom’s debt matures or re-fixes (because they are floating rates or index-linked instruments, or fixed-rates held by central banks) by 2025, compared to 38%, 56%, 55%, and 36% when the analysis is limited to the central government, respectively.

Figure 1.16. Adjusted ATR to consider the effect of central banks’ holdings

1.4. Implications of monetary policy normalisation and quantitative tightening for sovereign debt markets

1.4.1. Some central banks shifted from net buyers to net sellers of sovereign bonds

As shown in the previous section, sovereign issuers have significantly lengthened the maturity of their debt portfolios over the past decade in a favourable funding environment facilitated by the asset purchase programmes implemented by many central banks in the OECD area (i.e. quantitative easing – QE). These programmes aimed to lower long-term funding costs and were generally successful in achieving this objective, with bond acquisitions of 1% of GDP having been estimated to decrease long-term yields by approximately 5 to 10bps on average (OECD, 2022[1]). In response to increasing inflationary pressures, central banks have begun to downsize their substantial bond holdings (i.e. quantitative tightening – QT), which is likely to contribute to tighter funding conditions. The downsizing strategy differs considerably across OECD countries, with some central banks actively selling bonds, while others are not (fully or partially) reinvesting the proceeds from maturing bonds.15

In 2022, the government security holdings of the Bank of Canada (BoC), the BoE, and the Fed declined, while the ECB maintained a broadly stable holding, and the BoJ saw continued growth (Figure 1.17. G7 Central banks’ government security holdings Panels A and B).16 The monthly changes in the holdings of the first three central banks transitioned from an average increase of USD 100 billion in 2021 to a decrease of USD 25 billion in 2022. This shift resulted in a reduction of USD 181 billion, USD 60 billion, and USD 53 billion in the holdings of the Fed, BoC, and BoE, respectively. These figures equate to approximately half, one-third, and 5% of the 2022’s issuances of long-term instruments by their respective governments. In contrast, the ECB increased its government security holdings by USD 113 billion in the first half of 2022 but kept it broadly stable in the second half, indicating a shift in the ECB's policy. Meanwhile, the BoJ expanded its government security holdings in both semesters of 2022, with a growth of USD 166 billion in the first half and USD 213 billion in the second half.

Figure 1.17. G7 Central banks’ government security holdings

G7 central banks’ balance sheets reached a record high in July 2022 at USD 13.7 trillion, as the pace at which the government security holdings of the BoJ and the ECB rose was faster than the decline in the holdings from the BoC, BoE and the Fed (Figure 1.17. G7 Central banks’ government security holdings Panel C). The government security holdings of the three G7 central banks that started to downsize their balance sheets in 2022 (BoC, BoE and the Fed) peaked in February 2022, reaching USD 7.0 trillion, and declined by 5% to USD 6.7 trillion. Since the start of the downsizing, the BoC, BoE and Fed decreased their balance sheets by 19%, 2%, and 5%, respectively.

1.4.2. Over 50% of central banks’ government bonds holdings mature by 2030

Quantitative tightening impacts DMOs’ operations in three ways: through the portfolio balance channel, which may lead to higher yields as investors absorb increased net issuance; through a signalling effect, where market participants may interpret balance sheet reduction as a sign of future policy rate changes; and by increasing primary dealers’ search costs and risk premia due to reduced activity of central banks (OECD, 2022[24]). In all of these cases, the pressure put by QT programmes will vary with their pace and magnitude, both of which depend on the maturity profile and size of their holdings.

Figure 1.18 Panel A displays the maturity profile of three selected G7 central banks,17 revealing that over 50% of the government security holdings on the Fed’s, BoE’s, and BoC’s balance sheets will mature by 2030. Among these, only the BoE has more than 10% of its government security holdings expected to mature after 2030, which also reflects the fact that the United Kingdom has longer lines than these two sovereign issuers (i.e. the longest line is 2073 for the UK, 2052 for the US, and 2064 for Canada). Panel B shows the relative size of CBs’ government security holdings in the OECD area, revealing that on average 25% of the outstanding amount of government debt is held by central banks. This figure ranges from zero in Chile, the Czech Republic, Denmark, and Norway to 50% in Portugal.

Figure 1.18. The impact of QT overtime: the maturity profile and size of holdings

Note: In Panel B the following countries are not displayed as their central banks hold no government security: CHL, CZE, DEN and NOR.

Source: OECD calculations based on information from the Bank of Canada, Bank of England, Bank of Japan, European Central Bank, Federal Reserve, 2022 Survey on Central Government Marketable Debt and Borrowing, OECD (2022[1]), OECD Economic Outlook, Volume 2022 Issue 2, https://doi.org/10.1787/f6da2159-en; IMF (2022[2]), World Economic Outlook, https://www.imf.org/en/Publications/WEO/Issues/2022/10/11/world-economic-outlook-october-2022; Refinitiv, and national authorities’ websites.

1.5. Deteriorating liquidity conditions in sovereign bonds markets

1.5.1. Liquidity deteriorated in 2022, with the domestic sovereign bond markets being more affected than the repo, derivatives and foreign bond markets

Market liquidity (i.e. the degree to which an asset can be traded without impacting its value) is paramount for markets to function efficiently and, therefore, is among the most important factors monitored by DMOs. With low liquidity, investors must bear greater costs to enter or exit a position due to price movements. The extent to which prices move during transactions depends heavily on the volume being traded, the asset being exchanged, and the specific market dynamics in which the trade occurs. Sovereign bond markets are among the most liquid ones due to several reasons, including the availability of high-quality and voluminous securities with various maturities and the high demand of sovereign bonds by a variety of investors. A highly liquid sovereign bond market enables DMOs to adapt more easily to changes in borrowing needs, as the market will absorb the increase in volume without affecting prices (much), thus facilitating the pursuit of their funding strategies, which minimise costs while keeping risks controlled.

A recent survey on Liquidity in Secondary Government Bond Markets conducted among OECD member and accession countries, indicates that sovereign issuers observed a deterioration in liquidity metrics across asset classes, including in the highly liquid sovereign bond market of OECD countries. DMOs monitor a myriad of indicators on sovereign bonds’ market liquidity, including bid-ask spreads, turnover, volatility metrics, free float, and the average bid-to-cover ratio among others.18 Figure 1.19 Panel A illustrates that 2022 was the first time since 2017 that more than half of the DMOs from OECD and accession countries reported a decline in liquidity conditions of domestic sovereign securities.19 Exceptions are Brazil, Bulgaria, the Czech Republic, Mexico and Türkiye, where liquidity conditions were reported to improve (Figure 1.19 Panel B), particularly in their foreign bonds market due to rising demand for investments in hard currency (i.e. flight-to-quality phenomenon).

Figure 1.19. Trends in liquidity conditions of domestic sovereign securities

Compared to the derivative, repo and foreign bond markets, the domestic bond market was the one whose liquidity conditions deteriorated more frequently across OECD and accession countries (Figure 1.18 Panel C). One reason is that the repo and derivative markets benefited more directly from the support of security lending facilities, operated either by the monetary authority or the DMOs themselves.

1.5.2. Macroeconomic uncertainty, geopolitical risks, declining investor sentiment, and a reduction in free float contributed to a deterioration in liquidity

The primary factors affecting market liquidity include macroeconomic uncertainty, monetary policy developments, geopolitical risks, and deterioration of investor sentiment (Figure 1.20 Panel A). A monetary policy direction reversal following a decade of historically low yields and abundant liquidity has generated considerable macroeconomic uncertainty. The rapid withdrawal of supportive monetary policy implemented since the GFC resulted in significant asset depreciation and rising volatility. Typically, a one-percentage-point change in interest rates implies a change in a bond’s price equal to its duration (i.e. the weighted average time to receive all the bond’s cash flows) (OECD, 2017[22]), indicating that the average 4.5 percentage points rise in policy rates across the OECD could lead to a nearly 45% loss for a bond with a 10-year duration. In addition to these conjunctural factors affecting market liquidity, DMOs have reported that digitalisation and algorithmic trading may heighten the risks of a market squeeze (Figure 1.20, Panel B). Specifically, as the share of liquidity provided by algorithmic providers has increased over the past decade, liquidity has become less resilient, often declining during periods of extreme stress.

Another crucial development impacting market liquidity pertains to changes in the sovereign debt investor base, leading to a decline in free float (i.e. the share of debt that can be exchanged in markets). Two characteristics of the investor base are particularly relevant for market liquidity: diversity; and willingness to trade. Firstly, a more diverse investor base increases the likelihood of having investors on both sides of a transaction (buy and sell) – if the investor base were homogeneous, all investors would want to buy or sell simultaneously, hindering transactions and increasing market volatility. Secondly, if investors hold their assets until maturity, they will not exchange them, thus not providing liquidity to the markets. Consequently, markets are more likely to be liquid when their investor base is diverse and actively trading their holdings.

Figure 1.20. Factors affecting market liquidity

Since the GFC, investor bases have significantly changed in many OECD countries, with central banks becoming one of the primary investors. Figure 1.21 Panel A indicates that since 2010, the proportion of sovereign debt held by domestic central banks has increased from 3% to 20%, reducing the portion held by domestic investors (banks and non-banks – by 9 percentage points and 5 percentage points, respectively) and foreign investors (banks and non-banks – by 3 percentage points and 4 percentage points, respectively). Another increasingly relevant investor type is foreign official institutions, which held, on average, 16% of OECD countries’ sovereign debt in 2021, up from 8% in 2005. These two investor types tend to hold these securities long term – they do not actively engage in sell-side operations. As a result, they tend to reduce the free float of the debt and, consequently, market liquidity. Specifically, the free float decreased for all countries displayed in Figure 1.21 Panel B, with the average decrease reaching 26 percentage points. Additionally, some countries reported that due to deteriorating market conditions, some Primary Dealers became less active (note the significant decrease in the domestic banking sector in Figure 1.20 Panel A), while other market participants, such as hedge funds, filled the gap left by them, with some of these participants ultimately acting as “shadow dealers”.

Figure 1.21. Sovereign debt investor base for selected OECD countries

1.5.3. Sovereign issuers have been supporting liquidity through various means, including improved communication, tapping existing securities and buybacks

Considering the influence of liquidity on borrowing costs and market resilience, DMOs strive to actively enhance the liquidity conditions of sovereign bond markets. DMOs have improved communication with market participants (e.g. through more frequent communication), tapped existing securities (particularly off-the-run bonds), conducted buyback and switch operations, provided security lending facilities,20 and implemented measures to enhance the market capacity of Primary Dealers 21 (e.g. reducing the minimum quotation time and coverage, broadening requirements for bid/ask spreads) (Figure 1.22). These measures supplement the general strategies followed by DMOs to support liquidity, through which they aim to establish large benchmarks with transparent and predictable issuances. For smaller issuers, auctions serve as the primary liquidity points; thus, they may increase the frequency of auctions to improve liquidity.

Primary Dealers (PDs), who are among the most crucial providers of liquidity in the sovereign bond market, face elevated risks during periods of high volatility, as their holdings may depreciate while they fulfil their role as intermediating parties. QT can also impact operations by reducing central banks’ role as investors during periods of elevated borrowing needs, leading to increased search costs for PDs and potentially extending the time securities are held in their portfolios. Additionally, they have been affected by new capital and liquidity regulations put in place during a time of abundant liquidity and low volatility in the early 2010s as a response to the GFC (OECD, 2013[25]; 2014[26]). For these reasons, despite the observed deterioration in market liquidity in numerous OECD countries, DMOs tended not to increase the burden on PDs in their liquidity provision role – specifically, more than 80% of DMOs reported that they did not impose new requirements on market-makers. However, existing requirements were largely maintained, such as those related to bid/ask spreads, the size of offers, and secondary market performance, among others.22

Figure 1.22. Measures implemented by sovereign issuers to improve liquidity conditions

Sovereign bond markets have been shifting away from an environment characterised by low interest rates and abundant liquidity. While this is likely negatively impacting liquidity, it is not clear whether current conditions are abnormally illiquid or whether liquidity was abnormally high during the period of quantitative easing implemented since the GFC. Markets may be experiencing a temporary liquidity squeeze – in this scenario, when the global macroeconomic and geopolitical outlook stabilises, liquidity could improve and potentially return to levels seen in the previous decade (the reduction would be merely cyclical). Alternatively, the diminished role of central banks in sovereign bond markets, coupled with a decline in the role of traditional banks and the broader adoption of algorithmic trading, may structurally impact liquidity; consequently, even when the outlook becomes less uncertain, liquidity might not recover to pre-2022 levels. Since poor liquidity in bond markets could sharply amplify asset price moves and shocks, policy makers should remain vigilant and address potential vulnerabilities to prevent any systemic event that may adversely affect market confidence.

References

[3] Agence France Trésor (2022), Monthly Bulletin, https://www.aft.gouv.fr/files/medias-aft/7_Publications/7.2_BM/386_Monthly%20bulletin%20July%202022.pdf (accessed on 2023 14 April).

[29] Bank of Canada (2022), Government of Canada Bond Purchase Program, https://www.bankofcanada.ca/markets/market-operations-liquidity-provision/market-operations-programs-and-facilities/government-canada-bond-purchase-program/.

[34] Bank of England (2022), Asset Purchase Facility: Gilt Sales – Market Notice 20 October 2022, https://www.bankofengland.co.uk/markets/market-notices/2022/october/asset-purchase-facility-gilt-sales-market-notice-20-october-2022.

[20] Bank of England (2022), Bank Rate history and data, https://www.bankofengland.co.uk/boeapps/database/Bank-Rate.asp (accessed on 3 March 2023).

[15] Bank of England (2022), Gilt Market Operations – Market Notice 28 September 2022, https://www.bankofengland.co.uk/markets/market-notices/2022/september/market-notice-28-september-2022-gilt-market-operations (accessed on 30 March 2023).

[11] BIS (2022), Annual Economic Report: June 2022, https://www.bis.org/publ/arpdf/ar2022e.pdf.

[14] BIS (2022), BIS Quarterly Review: December 2022 International banking and financial market developments, http://www.bis.org/publ/qtrpdf/r_qt2212.htm.

[13] Blanchard, O. (2023), Fiscal Policy Under Low Interest Rates, MIT Press, https://doi.org/10.7551/mitpress/14858.001.0001.

[36] Board of Governors of the Federal Reserve System (2022), Plans for Reducing the Sise of the Federal Reserve’s Balance Sheet, https://www.federalreserve.gov/newsevents/pressreleases/monetary20220504b.htm.

[35] Debrun, X., J. Ostry and C. Wyplosz (2020), Debt Sustainability, Oxford University.

[28] Engstrom, E. and S. Sharpe (2022), “(Don’t Fear) The Yield Curve, Reprise”, FEDS Notes, Vol. Washington: Board of Governors of the Federal Reserve System.

[31] European Central Bank (2022),, Monetary Policy Statement – Press Conference, https://www.ecb.europa.eu/press/pressconf/2022/html/ecb.is221215~197ac630ae.en.html.

[37] European Central Bank (2022), Securities lending of holdings under the asset purchase programme (APP) and pandemic emergency purchase programme (PEPP), https://www.ecb.europa.eu/mopo/implement/app/lending/html/index.en.html (accessed on 6 March 2022).

[33] European Central Bank (2022), “The Transmission Protection Instrument”, Press Release, https://www.ecb.europa.eu/press/pr/date/2022/html/ecb.pr220721~973e6e7273.en.html.

[38] Federal Reserve (2022), “Repo and Reverse Repo Agreements”, Federal Reserve Bank of New York, https://www.newyorkfed.org/markets/domestic-market-operations/monetary-policy-implementation/repo-reverse-repo-agreements (accessed on 3 March 2022).

[23] Government of Canada (2022), Update on the 2022-23 Debt Management Strategy, https://www.budget.canada.ca/fes-eea/2022/report-rapport/anx2-en.html.

[16] HM Treasury (2022), Chancellor brings forward further Medium-Term Fiscal Plan measures, https://www.gov.uk/government/news/chancellor-brings-forward-further-medium-term-fiscal-plan-measures (accessed on 3 March 2023).

[17] HM Treasury (2022), The Growth Plan 2022, https://www.gov.uk/government/publications/the-growth-plan-2022-documents/the-growth-plan-2022-html#:~:text=The%20Growth%20Plan%202022%20makes,sustainable%20funding%20for%20public%20services (accessed on 30 March 2023).

[2] IMF (2022), World Economic Outlook, https://www.imf.org/en/Publications/WEO/Issues/2022/10/11/world-economic-outlook-october-2022.

[27] Lane, P. (2019), “The Yield Curve and Monetary Policy”, Speech in the Centre for Finance and the Department of Economics at University College London.

[8] Lunsford, K. and K. West (2019), “Some Evidence on Secular Drivers of US Safe Real Rates”, American Economic Journal: Macroeconomics, Vol. 11, pp. 113-139.

[4] Ministero dell’Economia e delle Finanze (2022), Modifications to the government bond auction calendar for August 2022, https://www.dt.mef.gov.it/export/sites/sitodt/modules/documenti_en/debito_pubblico/modifiche_calendari/Modifications-to-the-Government-bond-auction-calendar-August-2022.pdf (accessed on 14 April 2023).

[5] Ministero dell’Economia e delle Finanze (2022), Changes to December 2022 auction calendar, https://www.dt.mef.gov.it/export/sites/sitodt/modules/documenti_en/debito_pubblico/modifiche_calendari/Modifications-to-the-Government-bond-auction-calendar-December-2022.pdf (accessed on 14 April 2023).

[6] Ministry of Finance (2022), JGB Monthly Newsletter, https://www.mof.go.jp/english/policy/jgbs/publication/newsletter/jgb2022_05e.pdf (accessed on 4 April 2023).

[41] Ministry of Finance (2022), JGB Monthly Newsletter April 2022, https://www.mof.go.jp/english/policy/jgbs/publication/newsletter/jgb2022_04e.pdf (accessed on 14 April 2023).

[7] Ministry of Finance (2022), JGB Monthly Newsletter November 2022, https://www.mof.go.jp/english/policy/jgbs/publication/newsletter/jgb2022_11e.pdf (accessed on 14 April 2023).

[40] Ministry of Finance (2022), JGB Monthly Newsletter November 2022, https://www.mof.go.jp/english/policy/jgbs/publication/newsletter/jgb2022_11e.pdf (accessed on 4 April 2023).

[1] OECD (2022), OECD Economic Outlook, Volume 2022 Issue 2, OECD Publishing, Paris, https://doi.org/10.1787/f6da2159-en.

[24] OECD (2022), OECD Sovereign Borrowing Outlook 2022, OECD Publishing, Paris, https://doi.org/10.1787/b2d85ea7-en.

[22] OECD (2017), OECD Sovereign Borrowing Outlook 2017, OECD Publishing, https://www.oecd-ilibrary.org/governance/oecd-sovereign-borrowing-outlook-2017_sov_b_outlk-2017-en.

[26] OECD (2014), OECD Sovereign Borrowing Outlook 2014, OECD Publishing, https://www.oecd-ilibrary.org/governance/oecd-sovereign-borrowing-outlook-2014_sov_b_outlk-2014-en.

[25] OECD (2013), OECD Sovereign Borrowing Outlook 2013, OECD Publishing, https://www.oecd-ilibrary.org/governance/oecd-sovereign-borrowing-outlook-2013_sov_b_outlk-2013-en.

[12] OECD (2012), OECD Sovereign Borrowing Outlook 2012.

[42] OECD Publishing, P. (ed.) (2023), OECD Interim Economic Outlook of March 2023: A fragile recovery, https://doi.org/10.1787/d14d49eb-en.

[9] Rachel, L. and L. Summers (2019), “On Falling Neutral Real Rates, Fiscal Policy, and the Risk of Secular Stagnation”, BPEA Conference of 7 March-8.

[10] Ranaldo, A., P. Schaffner and M. Vasios (2021), “Regulatory effects on short-term interest rates”, Journal of Financial Economics, Vol. 141/2, pp. 750-770, https://doi.org/10.1016/j.jfineco.2021.04.016.

[39] Reserve Bank of Australia (2022), From QE to QT – The next phase in the Reserve Bank’s Bond Purchase Program, https://www.rba.gov.au/speeches/2022/sp-ag-2022-05-23.html.

[30] Reserve Bank of New Zealand (2022), Reserve Bank details planned sales of New Zealand Government Bonds, https://www.rbnz.govt.nz/hub/domestic-markets-media-releases/reserve-bank-details-planned-sales-of-new-zealand-government-bonds.

[32] Sveriges Riksbank (2022), “Policy rate raised by 1 percentage point”, Press Releases, https://www.riksbank.se/en-gb/press-and-published/notices-and-press-releases/press-releases/2022/policy-rate-raised-by-1-percentage-point/.

[18] UK DMO (2023), Revision to the DMO’s financing remit 2022-23: September 2022, https://www.dmo.gov.uk/media/pddhdwib/pr230922_2.pdf (accessed on 3 March 2023).

[21] UK DMO (2022), Gilt Market: Data, https://www.dmo.gov.uk/data/gilt-market/ (accessed on 3 March 2032).

[19] UK DMO (2022), Revision to the DMO’s financing remit 2022-23: Autumn Statement 2022., https://www.dmo.gov.uk/media/jhebewha/sa171122.pdf (accessed on 30 March 2023).

Annex 1.A. Methods and sources

Definitions and concepts used in the Sovereign Borrowing Outlook Survey

The Borrowing Outlook survey collects gross borrowing requirements, redemption and outstanding debt amounts with breakdown of these items by maturity, currency, interest rate types and ESG-labelling. It also collects data on DMOs’ holdings, NextGenerationEU loans and country-specific methodological aspects. It uses core definition of sovereign debt, so-called central government marketable debt, mainly due to its comparability and collectability. This measure, directly linked to the central government budget financing, enabled the OECD to collect not only for realisations but also for estimates of government borrowing requirements, funding strategies as well as outstanding debt with instruments, maturity and currency types.

Coverage of institutions: Central Government

The coverage of institutions by debt statistics varies from public sector to central government. Public sector stands as broadest institutional coverage, as it includes local governments, state funds financial and non-financial public corporations as well as central government debt. General government definition, which is used for example by the OECD System of National Accounts (SNA), consists of central government, state and local governments and social security funds controlled by these units. Central government covers all departments, offices, establishments and other bodies classified under general government, which are agencies or instrument of the central authority of a country, except separately organised social security funds or extra-budgetary funds. In terms of layers of coverage of institutions, central government stands out as the core definition. Debt of the central government is raised, managed and retired by the national DMOs on behalf of the central government. Hence, the advantage of this relatively narrow definition of debt is that it enables countries to provide comparable figures, in particular for the estimations.

Coverage of types of debt: Marketable debt

In terms of instruments, liabilities can be in the form of debt securities, loans, insurance, pensions and standardised guarantee schemes, currency and deposits, and other accounts payable. Debt items can be classified as marketable and non-marketable debt. While marketable debt is defined as financial securities and instruments that can be bought and sold in the secondary market, non-marketable debt is not transferable. For example, bonds and bills issued in capital markets are marketable debt; multilateral and bilateral loans from the official sector are non-marketable debt.

The Borrowing Outlook survey focuses on marketable debt instruments, while most government debt statistics (e.g. OECD SNA, EU Maastricht debt, and IMF Public Sector Debt Statistics) cover both marketable and non-marketable debt items. OECD governments are financed predominantly by marketable debt instruments. This is a central definition for every analysis concerning various issues around debt management including borrowing conditions, portfolio composition, investor preferences and market liquidity. An advantage of using this definition is to indicate to investors which instruments are available for trade in the secondary markets, and which are not. Another reason is for the issuer to calculate different characteristics of the debt, such as duration or time to maturity, which in the case of non-marketable debt would present a difficult issue.

Terminology

Standardised gross borrowing requirement (GBR) for a year is equal to the net borrowing requirement during that year plus the redemptions of long-term instruments of the same year and the redemptions of short-term instruments issued in the previous year. Therefore, this indicator captures the issuances of all securities excluding those that were issued and redeemed in the same calendar year. In other words, the size of GBR in calendar year amounts to how much the DMO needs to issue in nominal terms so as to fully pay back maturing debt issued in previous years plus the net cash borrowing requirement through any issuance mechanism.
Net borrowing requirement (NBR) is the amount to be raised for the current budget deficit. While refinancing of redemptions is a matter of rolling over the same exposure as before, NBR refers to new exposure in the market.
The funding strategy involves the choice of i) money market instruments for financing short-term GBR and ii) capital market instruments for funding long-term GBR. The strategy entails information on how borrowing needs are going to be financed using different instruments such as long-term, short-term, nominal, variable-rate, indexed bonds and FX-denominated debt.
Gross debt, or debt stock, corresponds to the outstanding debt issuance at the end of calendar years. This measure does not take the valuation effects from inflation and exchange rate movements; thus it is equal to the total nominal amount that needs to be paid back to the holders of the debt.
Redemptions refer to the total amount of the principal repayments of the corresponding debt including the principal payments paid through buy-back operations in a calendar year.

Regional aggregates

Total OECD area denotes the following 38 countries: Australia, Austria, Belgium, Canada, Chile, Colombia, Costa Rica, the Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Israel, Italy, Japan, Korea, Latvia, Lithuania, Luxembourg, Mexico, the Netherlands, New Zealand, Norway, Poland, Portugal, the Slovak Republic, Slovenia, Spain, Sweden, Switzerland, Türkiye, the United Kingdom and the United States.
OECD accession countries include Bulgaria, Brazil, Croatia, Peru and Romania.
The G7 includes seven countries: Canada, France, Germany, Italy, Japan, the United Kingdom and the United States.
The OECD Euro area includes 17 Member countries: Austria, Belgium, Estonia, Finland, France, Germany, Greece, Ireland, Italy, Latvia, Lithuania, Luxembourg, the Netherlands, Portugal, the Slovak Republic, Slovenia and Spain.
In this publication, from a public debt management perspective, the Emerging OECD group (i.e. OECD emerging-market economies) is defined as including seven countries: Chile, Colombia, Costa Rica, Hungary, Mexico, Poland and Türkiye.
The euro (EUR) is the official currency of 20 out of 28 EU Member countries. These countries are collectively known as the Euro area. The Euro area countries are Austria, Belgium, Cyprus, Estonia, Finland, France, Germany, Greece, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, the Netherlands, Portugal, the Slovak Republic, Slovenia and Spain. In this report, the Euro area covers only the countries that are simultaneously in the Euro area and in the OECD.

Calculations and data sources

Estimates that are presented as a percentage of GDP, for consistency reasons, use GDP estimates from the last OECD Economic Outlook in the previous year (so November 2022 for this publication) and are calculated using nominal GDP data.
Debt is measured as the face value of current outstanding central government debt. Face value, the undiscounted amount of principal to be repaid, does not change except when there is a new issue of an existing instrument. This coincides with the original promise (and therefore contractual obligation) of the issuer. DMOs often use face value when they report how much nominal debt will mature in future periods. One important reason for using face value is that it is the standard market practice for quoting and trading specific volumes of a particular instrument.
Currencies are converted into USD using flexible exchange rates using data sourced from Refinitiv. The effects of using alternative exchange rate assumptions (in particular, fixing the exchange rate versus using flexible exchange rates) are illustrated in Box 1.1.
All figures refer to calendar years unless specified otherwise.
Aggregate figures for gross borrowing requirements (GBR), net borrowing requirements (NBR), central government marketable debt, redemptions, and debt maturing are compiled from answers to the Borrowing Survey. The OECD Secretariat inserted its own estimates/projections in cases of missing information for 2022 and 2023, using publicly available official information on redemptions and central government budget balances. Where government plans have been announced, but not incorporated into financing plans as of the end of December 2022, they are not included in the projections presented in this publication. Also, the latest estimates of government net lending in the OECD Economic Outlook database are used in estimating some missing data.
Both the 2022 OECD Survey on Primary Market Developments and the 2022 OECD Survey on Liquidity in Secondary Government Bond Markets were carried out in October 2022.
Yield group debt calculations in Figure 1.12 Panel B are based on all issuances and re-openings of fixed-rate bonds (i.e. data excludes: short-term instruments, indexed linked, floating rate instruments and strips) and for comparability reasons only bonds issued in USD, EUR), JPY, GPB and Canadian dollars (CAD) were chosen. Data is sourced from Refinitiv.

Revised definitions

Inflation-linked securities are instruments with coupon and/or principal payments which are linked to an inflation index. Includes accrued inflation for all years up to and including the current year of the survey as of the reporting date.
Variable rate notes have a floating or variable interest rate or coupon rate. It is a long-dated debt security whose coupon is refixed periodically on a “refix date” by reference to an independent interest rate index such as LIBOR or Euribor. For example, medium and long-term floating rate notes (FRNs, or colloquially as floaters) are debt obligations with variable interest rates that are adjusted periodically (typically every one, three, or six months). The interest rate is usually fixed at a specified spread over one of the interest rate indices. For projections of variable rate debt, the rate at the level of the last settled coupon is used.
Average term-to-maturity figures follow the same coverage described at the beginning of this Annex.

Country-specific methodological aspects

Annex Table 1.A.1. Average term-to-maturity: country comments
Country	Note
Canada	The debt stock includes both domestic and foreign marketable securities. The ATM is reported as at the end of each calendar year from 2007 to 2022.
Colombia	All marketable debt in Colombia corresponds to domestic bonds (TES) and foreign bonds.
Costa Rica	Maturity of domestic debt and Eurobonds of external debt.
Estonia	The figures of Estonia take into account the average life of total debt including IFI loans, T-bills and Eurobonds.
France	Excludes swap effects.
Germany	Calculation excluding holdings in own stock. Inflation-linked securities weighted with 0.75.
Hungary	Data excludes retail securities, locally issued FX bonds, loans and since 2020 also excludes the non-marketable bonds issued to municipalities. Data includes cross-currency swaps.
Italy	No security has been excluded; swap effects, buybacks and repos are excluded.
Japan	MOF announces ATM, based on Fiscal Year, not Calendar Year. Figures from 2007 to 2021 exclude saving bonds. Figures of 2022 are estimated and include saving bonds.
Netherlands	The information is based on the data of T-bill and Bonds.
New Zealand	The calculation is marketable securities only and, therefore, excludes the non-market securities held by the Reserve Bank of New Zealand and the Earthquake Commission. However, it includes securities held by the Reserve Bank of New Zealand that were purchased under their Large-Scale Asset Purchase programme and Government Bond repurchases.
Sweden	End of year. Government bonds; Inflation-linked bonds; Public bonds, foreign currencies; Green bonds; T-bills.
The United Kingdom	ATM is weighted by the nominal amounts outstanding of gilts and T-bills issued for debt management purposes, as of the reporting date. Nominal amounts of gilts include government holdings; nominal values of index-linked gilts also include accrued inflation as of the reporting date.

Annex Table 1.A.2. Borrowing Outlook survey: country comments
Country	Variable	Comment
Austria	ESG-labelled of which: green bonds only	Includes also Green Treasury Bills with a Maturity less than 1 year (debt stock: 1 billion in 2022 and 2 billions in 2023)
Belgium	General comments on Methodology and concepts	Central Government Gross Borrowing Requirement reporting on final debt positions (i.e. after swap).
	T-bills	Treasury Certificates (TC)
	Other (cash etc..)	ECP
	Long-term instruments	OLOs, State bonds and EMTN
	Fixed Rate of which: zero (or negative) coupon bonds	OLOs and EMTN (after swap) with nominal coupon <= 0%
	Long-term marketable debt (from row 15)	Long-term final debt positions (after swap)
	Long-term marketable debt of which: issued in national currency	Long-term final debt positions (after swap) in euro
	Long-term marketable debt of which: issued in foreign currency	No long-term final debt positions (after swap) in foreign currency
	ESG-labelled of which: green bonds only	Green OLOs
	Total of DMO / central government holdings	Securities held in portfolio
	Short-term DMO / central government holdings	OLOs held in portfolio
	Long-term DMO / central government holdings	TCs held in portfolio
	Loans from other financial facilities (e.g. ESM, EFSF, SURE, NextGenerationEU etc.)	SURE (central government only)
Canada	T-bills	For current calendar year 2022, we have provided all numbers as of August 2022, as year-end estimates are not public and, in some cases, not projected.
	Other (cash etc..)
	Fixed rate
	Inflation linked
	Marketable debt held by non-residents	Non-resident holdings for 2022 are reported as of the latest published numbers for July 2022, and there are no forward-looking estimates.
	Average maturity of marketable debt	Because of data constraints, ATM for years 2006-10 are net of assets and as at fiscal year-end (March). ATM for 2022 is as of August 2022, as we do not project it for remaining months.
Denmark	Short term instruments	Please note that the numbers for borrowing in short-term instruments are net borrowing numbers.
Denmark	Inflation linked	Issuance in the inflation-linked bond includes the index revaluation. Outstanding amount in the inflation-linked bond is measured at indexed nominal value year-end.
France	Long-term instruments	For 2023 the amount of gross issuance does not include potential buybacks.
	Fixed rate	Redemptions integrate buybacks. The 2023 redemptions do not include potential buybacks.
	Inflation linked	For 2023, AFT will maintain its inflation-linked bonds issuance commitment around 10%
	Debt Service – Principal	Principal integrates the accumulated inflation for linkers
Germany	Marketable debt held by non-residents	Estimate. Short- and long-term. Rounded.
	Average maturity of marketable debt	Calculation excluding holdings in own stock. Inflation-linked securities weighted with 0.75.
	Debt Service – Principal	Money market instruments included at cash value.
Hungary	General comments on Methodology and concepts	We converted all foreign currency debt into national currency rate as 31 December of each year (realisations, officials FX rates of NBH), and we used 404 EUR/HUF rate for current year (2022, planning rate).
Ireland	General comments on Methodology and concepts	Ireland currently has two 0% coupon Government bonds outstanding, one of which matures on 18 October 2022. Ireland also has one green bond with an o/s balance of EUR 6.85bn. The 0% coupon bonds and green bond are included in the long-term instruments categories 1.2 and 1.M.2. Ireland also has three inflation linked bonds with a total balance o/s of c. EUR 1.1bn. As these bonds were all issued as private placements they are included within the non-marketable debt category in this survey.
	Marketable debt held by non-residents	We do not have data on non-resident holdings – the Central Bank of Ireland publishes some information on Holdings of Irish Long-Term Government Bonds
	Total of DMO / central government holdings	Slightly unclear what is being requested here – is it Irish sovereign marketable debt held by Government/public sector bodies?
Israel	General comments on Methodology and concepts	No applicable data earlier than 2011
Italy	Average maturity of marketable debt	With regard to marketable debt issued by central government and held by the same sector, please see data published in the “Financial accounts” (table 21) on the Bank of Italy website: “https://www.bancaditalia.it/pubblicazioni/conti-finanziari/index.html?com.dotmarketing.htmlpage.language=1” The Italian Treasury could only provide data about its bond portfolio used for Repo activity, which is not accounted for in the “Total Marketable Debt” above
Japan	T-bills	T-bills for cash management purposes are excluded. They are included in “Short-term Other (1.1.2)”.
Japan	Average maturity of marketable debt	MOF announces ATM, based on Fiscal Year, not Calendar Year. Figures from 2006 to 2021 exclude JGBs for retail investors.
Spain	General comments on Methodology and concepts	In the interest of not delaying even further the delivery of the data, we haven’t been able to provide information on Accrued Inflation. Although we do have this information, we don’t have it readily available and prepared for this format. We hope to provide it shortly or include it in next year’s outlook.
Sweden	Central Government Gross Borrowing requirement	The total marketable debt and the non-marketable debt adds up to the official measure of government debt and is defined based on principles laid down at EU level. It accounts for the Central Government gross debt, without regard to the assets. For the years in the forecast, it is nominal (face value) amount, incl. assets under management and is the sum of the amounts that the Debt Office is committed to paying when a debt instrument matures and receives at maturity if it is an asset. The amount is reported in SEK at the exchange rate at the time of borrowing.
	T-bills	Borrowing in the money market corresponds to outstanding stock at year-end. Redemptions: Initial stock maturing within 12 months.
	Fixed rate	Government bonds; Inflation-linked bonds; Public bonds, foreign currencies; Green bonds;
	Average maturity of marketable debt	End of year. Government bonds; Inflation-linked bonds; Public bonds, foreign currencies; Green bonds; T-bills
	Other non-marketable (e.g. government bonds for retail investors or retails bonds, saving bonds, etc.)	Other capital market debt: Assigned bonds, Private placements foreign currencies Liquidity management: Deposits; accounts; Securities holdings; Commercial paper, foreign currencies; Repos, other securities; Liquidity bills; Repos, Swedish Government papers; Collateral. Borrowing liquidity management corresponds to outstanding stock at year-end. Redemptions liquidity management: Initial stock maturing within 12 months Lottery bonds, National debt savings Derivatives market
	Debt Service	Government bonds; Inflation-linked bonds; Public bonds, foreign currencies; Green bonds; T-bills; On-lending
Türkiye	Inflation Linked of which: accrued inflation	With respect to the Central Government Accounting Regulation in Turkey, accrued inflation is reported as interest. Therefore, it is not included in central government debt stock statistics. In this regard, we will continue not to include accrued inflation as memo item.
Türkiye	Total of DMO / central government holdings	Total of central government holdings is shared since the data could not been broken down in terms of term.
United Kingdom	General comments on Methodology and concepts	* Government holdings: 1. are included in Debt stock data (rows 15 to 19 and in ATM). Government holdings of index-linked gilts include inflation uplift as of the reporting date. 2. are excluded from S2 debt service data as well as historical redemptions and gross borrowing data. ATM is weighted by the nominal amounts outstanding of gilts and T-bills issued for debt management purposes, as of the reporting date. Nominal amounts of gilts include government holdings; nominal values of index-linked gilts also include accrued inflation as of the reporting date. Gross borrowing data are cash amounts rather than nominal. Debt stock and redemptions data are nominal amounts; however holdings from overseas investors are market value data (they are sourced from the Office for National Statistics). T-bill data include only T-bills issued for debt management purposes i.e. they exclude T-bill issued for cash management purposes.
	Debt Service – Principal	Data exclude DMO holdings of gilts as well as future inflation uplift.
	Debt Service – Interest
United States	Fixed Rate of which: zero (or negative) coupon bonds	For this section, the U.S. Treasury has listed the volume of principal-STRIPS outstanding. (https://www.treasurydirect.gov/marketable-securities/strips/)

Notes

← 1. This analysis considers the standardised method to compute the gross borrowing needs. For details, see Annex 1.A.

← 2. Refinancing requirements were assessed utilising the standardised method. In other words, for a specific calendar year, the method takes into account all the long-term maturities that were due for redemption within that year, as well as the stock of short-term debt from the previous year which, by definition, will be redeemed in the following year. Consequently, the method excludes the redemption of all securities issued and redeemed within the same year.

← 3. It is important to recognise that three primary sources contribute to the discrepancies between the government deficit and net borrowing requirements, complicating the determination of the precise proportion attributable to cash balance management. Firstly, there is a distinction in scope: while the latter pertains to the central government, the former encompasses the general government, which includes state and local governments as well as social security funds. Secondly, the stock of marketable debt may not increase in parallel with the government deficit if the government employs its financial assets to meet fiscal needs beyond cash balances, such as through the sale of real estate and state-owned enterprises. Lastly, governments may augment their net liabilities via alternative means distinct from marketable debt, such as loans or arrears encompassing other accounts payable.

← 4. Global GDP growth is estimated to reach 3.2% in 2022, approximately half the rate observed in 2021, and is projected to decelerate further to 2.6% in 2023, significantly below the rate anticipated prior to the war in Ukraine (OECD, 2023[42]). It is worth noting that there are moderate downside risks to this projection, particularly the potential for even slower economic growth due to the tense geopolitical situation (including the escalation of the War in Ukraine), the uncertain magnitude and duration of monetary tightening, duration risks in the business models of financial institutions (as exemplified by the failure of the US Silicon Valley Bank in March 2023), scarcity of energy supplies, and global trade-related tensions.

← 5. In normal times, short-term debt such as T-Bills act as a cash management instrument, whose issuance volume is determined by the timing and size of government receipts and outlays. During crisis periods, such as the 2008 financial crisis and the recent COVID-19 crisis, T-Bills play an important role as ‘shock absorbers’ in sovereign financing. When there are unexpected changes in financial requirements, sovereign debt managers can delay adjusting debt issuance rates during the year, typically through changes in the issuance of T-bills (OECD, 2022[24]).

← 6. A borrowing plan, based on funding strategies, often sets out an issuance calendar with information on issuance methods and the use of debt management operations and tools.

← 7. Monetary policy influences the yield curve (i.e. yields for different maturities) through at least two channels: 1) forward guidance (i.e. changes in the expectations of future levels of interest rates) and 2) term spread (i.e. the excess return required by investors to hold a bond to maturity net of the expected return from continually reinvesting at the short-term rate over that same time horizon) (Lane, 2019[27]). More specifically, an increase in policy rates under a scenario of unchanged expectations of future levels of interest rates would affect the yield of short-term investments, while its effect on long maturities tends to be less pronounced as it is “weighted” by the (held constant) expectations of future levels of interest rates. In other words, the increase affects short-term maturities more than their long-term counterparts. However, if central banks communicate about future hikes, they will influence the intermediate and long maturities accordingly. Regarding the term spread, as there is virtually no credit risk in sovereign securities of OECD countries, it captures mostly the duration risk – the impact on the securities’ price that comes from changes in the yield curve. As future interest rates are affected by future decisions of central banks, which are navigating a trade-off between controlling inflation while avoiding causing unnecessary damage to the economy (BIS, 2022[11]), the shape of the yield curve is affected by expectations on inflation, growth and employment outlooks.

← 8. It is important to note that the phenomenon of short-term yields rising more significantly than their long-term counterparts is commonly observed during periods of monetary tightening. This occurrence, known as bearish flattening, leads to a flatter yield curve in a bearish bond market, characterised by falling prices due to rising interest rates. This dynamic arises from market expectations concerning future interest rates, growth, and inflation. A flattening yield curve indicates slowing economic growth, which results from higher funding costs and, consequently, reduced long-term real returns. Investors may opt to purchase long-dated maturities as a protective measure against this downturn. In this scenario, investors anticipate that central banks will accommodate their current monetary policy stance, and the potential impact of future policy rates will be factored into long-term yields (i.e. due to the arbitrage principle, long-term yields are just the weighted average of the future policy rates).

← 9. Although curve inversions have been associated with recessions in the past, they are merely indicators reflecting market expectations about the path of inflation and monetary policy and should not necessarily be regarded as a harbinger of economic downturns (Engstrom and Sharpe, 2022[28])..

← 10. The bottom whisker extends from the lower edge of the box (which represents the 1^st quartile, or 25^th percentile) to the smallest data point that is not considered an outlier. Outliers are considered as data points that are below the 1^st quartile minus 1.5 times the interquartile range (i.e., the difference between the 3^rd and 1^st quartiles).

← 11. Firstly, the maturity of the current debt portfolio determines the outstanding amount of debt to be refinanced during the period of high interest rates. Suppose a fixed-rate bond matures only after this period. In that case, the yield at issuance of a debt portfolio (i.e. the one that proxies the borrowing cost to the issuer) will be unaffected by the current interest rate hike, and debt holders will only experience losses. Secondly, the outstanding amount of floating-rate and index-linked instruments, as the principal and (or) the coupons of these instruments are refixed, usually within a period of fewer than six months, based on the new index or rate of reference. Thirdly, the share of debt denominated in foreign currency, as borrowing costs in local currency depend on foreign exchange rate movements, which are linked to differences in the expectations on interest rates and inflation of the issuing country and the country whose currency the debt is denominated in. Particularly during times of uncertainty, investors tend to prefer holding the debt of developed economies, whose macroeconomic fundamentals are typically more robust, increasing the cost in the domestic currency of the debt denominated in hard currencies.

← 12. These include Estonia (2032, previously 2030), Germany (2053, previously 2052), Iceland (2042, previously 2033), Japan (2062, previously 2061), the Netherlands (2054, previously 2052), Norway (2042, previously 2023), and the United States (2052, previously 2051).

← 13. The answers from the 2022 Survey on Central Government Marketable Debt and Borrowing revealed that virtually all OECD countries compute the ATM and more than 60% the ATR, while roughly 90% computes the share of debt maturing in the next few years. Theoretically, two countries may have identical ATM and ATR values but experience significantly different exposures. For example, consider a country with a 10-year ATR and a debt portfolio consisting solely of zero-coupon fixed rates maturing in 10 years. Its exposure to interest rate fluctuations differs greatly from a country with the same ATR, with half of its debt portfolio maturing in the next year and the other half in 20 years. In the latter scenario, half of the country’s debt portfolio would be refinanced under new interest rates, while in the former case, the issuer remains unaffected. It is worth noting that this is a (very) stylised example with the sole purpose of showing how the ATM/ATR can be misleading. No OECD country has such a small diversity of bond types across the maturity spectrum.

← 14. According to the OECD Government at Glance 2021 for 2019. The amount spent on these government functions was 0.5% and 0.6% of GDP, respectively.

← 15. Selected examples are: 1) Australia: Since February 2022, the Reserve Bank of Australia ceased bond purchases while ruling out bond sales to minimise market impact, gradually decreasing bond holdings (Reserve Bank of Australia, 2022[39]); 2) New Zealand: On 23 February 2022, the Reserve Bank of New Zealand agreed to initiate a gradual reduction of bond holdings by selling a limited number of securities in order of maturity date, starting with the longest maturity, while allowing shorter-maturity bonds to mature without reinvesting the proceeds (Reserve Bank of New Zealand, 2022[30]); 3) Canada: Since April 2022, the Bank of Canada (BoC) discontinued its bond purchase programme (Bank of Canada, 2022[29]); 4) The United States: On 4 May 2022, the Federal Reserve Board communicated that reductions of the System Open Market Account’s (SOMA) balance sheet would be gradual, predictable, and achieved by capping the number of reinvestments from the proceeds of matured bonds (Board of Governors of the Federal Reserve System, 2022[36]); 5) The United Kingdom: In September 2022, the Bank of England’s (BoE) Monetary Policy Committee decided (and implemented the decision) to commence selling UK Government bonds (gilts) from November onwards, with a quarterly schedule targeting short- to medium-term maturity sectors (3 to 20 years) (Bank of England, 2022[34]); 6) Sweden: On 20 September 2022, Sweden’s central bank (Sveriges Riksbank) announced that the asset purchase programme was expected to cease at year-end, leading to a gradual decline in its balance sheet across all maturities (Sveriges Riksbank, 2022[32]); and 7) The European Union: On 15 December 2022, the ECB president, Christine Lagarde, announced in a press conference that from March 2023, the ECB’s holdings would decline at a measured and predictable pace, as it would not reinvest all proceeds from matured securities (European Central Bank, 2022[31]) The ECB may still purchase securities through the Transmission Protection Instrument (TPI) to smooth monetary policy stance across Euro area countries via targeted bond purchases for countries with sound fiscal and macroeconomic policies (European Central Bank, 2022[33]).

← 16. It is worth noting that the movements in central banks’ government holdings differ from the movements of their total holdings. The BoJ, for instance, decreased its total holdings by 2.6% between the first quarter of 2022 and the first quarter of 2023.

← 17. These central banks were selected given the data available on the maturity of their government security holdings.

← 18. Based on the discussions and background documents from the 2022 meeting of the Working Party of Debt Management. Bid-ask spreads refer to the difference between the bid and ask prices; turnover to the traded volume as a share of the outstanding amount, free float to the share of the outstanding amount that can be freely traded on the market, and the average bid-to-cover ratio to the dollar amount of bids received in a treasury security auction versus the amount sold.

← 19. The OECD Council adopted Accession Roadmaps for Brazil, Bulgaria, Croatia, Peru, and Romania in June 2022.

← 20. Several DMOs and CBs across the OECD have bolstered market liquidity through security lending facilities. These facilities function in the repurchase agreement market (Repo market) – a market in which a counterparty sells a security with a simultaneous agreement by the seller to repurchase the same security from the same buyer at a pre-agreed price. Through this operation, the buyer acquires a security that can be utilised for transactions in the market, thereby enhancing liquidity. CBs have become active in this market, given their extensive range of securities purchased during quantitative easing. For instance, the New York Fed’s Open Market Trading Desk is authorised to conduct repo and reverse repo transactions to support effective monetary policy implementation and facilitate smooth market functioning (Federal Reserve, 2022[38]). The ECB has permitted repo transactions using securities acquired under the public sector purchase programme (PSPP) since 2015 (European Central Bank, 2022[37]). Given the importance of market liquidity to sovereign issuers, some DMOs have also started operating repo facilities. Repo facilities managed at the DMO level can provide liquidity even for off-the-run securities and those not purchased by the central bank during liquidity squeeze, thereby rendering markets more resilient to shocks. DMOs can also utilise the Repo facility to conduct reverse Repos, which can offer them a more favourable yield compared to central bank deposits, thus improving their cash management. As of 2022, 26 OECD and accession countries operate a lending facility, 11 are considering implementing one, and only two countries (Korea and Estonia) neither possess nor contemplate having such a facility (calculations based on the OECD member and accession countries’ responses to the 2022 OECD Survey on Liquidity in Secondary Government Bond Markets).

← 21. Primary dealers (PDs) are financial institutions (i.e. banks or securities firms) that are entitled to buy government securities in primary markets to resell them to others, thus acting as market makers of government securities.

← 22. These requirements are, in many cases, associated with benefits such as enhanced communication with the DMO, direct access to auctions, and access to security lending facilities, among others.

╳

Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

https://doi.org/10.1787/09b4cfba-en

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at https://www.oecd.org/termsandconditions.

Table 1.1. Funding strategy based on gross marketable borrowing needs in the OECD area
OECD Countries
	2017-19 Average	2020	2021	2022e	2023p
Short term	3.2 \| 36.8%	6.7 \| 43.7%	5.4 \| 37.3%	5.3 \| 43.7%	5.6 \| 43.5%
Long term	5.5 \| 63.2%	8.7 \| 56.3%	9 \| 62.7%	6.8 \| 56.3%	7.3 \| 56.5%
Fixed-rates	4.9 \| 56.5%	8 \| 51.9%	8.3 \| 57.3%	6.1 \| 50.4%	6.5 \| 50.3%
Inflation-linked	0.3 \| 3.6%	0.3 \| 2.2%	0.3 \| 2.4%	0.3 \| 2.8%	0.4 \| 2.8%
Variable rate	0.3 \| 3%	0.3 \| 2.2%	0.4 \| 3%	0.4 \| 3%	0.4 \| 3.4%
Trillion USD \| as a share of total borrowing
Note: Standardised gross borrowing requirements are shown. Values in local currencies are converted to USD using the exchange rate on the last day of the year. Cash management related issuances are excluded. For 2023 exchanges rates are fixed.
Source: 2022 Survey on Central Government Marketable Debt and Borrowing; OECD Economic Outlook 112 (November 2022); IMF (2022[2]), World Economic Outlook, https://www.imf.org/en/Publications/WEO/Issues/2022/10/11/world-economic-outlook-october-2022; Refinitiv, national authorities’ websites and OECD calculations.