1. Sovereign borrowing outlook

This chapter analyses sovereign debt developments for the period of 2008-24 with a focus on the OECD area. It explores borrowing requirements and funding strategies; debt-to-GDP ratio dynamics; borrowing costs and yield curves; interest payments and refinancing risks; and the effects of central banks’ balance sheet reduction on the investor base and market liquidity. This chapter draws mainly on responses received to an annual survey on the marketable debt of central governments. It provides data up to 2022, with 2023 estimates and 2024 projections. Details on the methodology can be found in ‎Annex 1.A.

The decline in gross borrowing by OECD governments in 2021 and 2022 halted in 2023. After reaching a high of USD 15.4 trillion in 2020, it reduced to USD 12.1 trillion in 2022, with estimations indicating a rise to USD 14.1 trillion in 2023 (Figure ‎1.1, Panel A). The 2024 projection of USD 15.8 trillion would surpass the peak reached during the pandemic, and mark a nearly 45% and 25% rise in real terms from the 2019 and 2022 levels, respectively. Box ‎1.1 explores these trends for Emerging Market and Developing Economies (EMDEs).

Gross borrowing needs consist of net borrowing and refinancing requirements, with the former driving the increase in the gross figure in 2023. Net borrowing peaked in 2020 at USD 7.3 trillion but declined to roughly USD 2.1 trillion in 2022, close to pre-pandemic levels. However, rising spending, notably on social security and cost-of-living-related measures, has reversed this two-year downward trend (OECD, 2023[1]) with net borrowing estimated to rise to around USD 3.6 trillion in 2023 then decline slightly to USD 3.2 trillion in 2024. That net borrowing levels are now more closely aligning with fiscal deficits suggests cash balances have normalised somewhat, after surging during the pandemic.1

Refinancing requirements are anticipated to be the primary driver of higher gross borrowing in 2024. These amount to approximately USD 7 trillion before the pandemic but have been above USD 10 trillion since 2021, with a projected peak of USD 12.6 trillion in 2024. This projected record high is attributed to the rise in short-term borrowing, which needs to be refinanced within a year, during 2023-24.

Driven by higher net borrowing, the OECD debt stock is expected to reach USD 54 trillion in 2023 and is projected to climb further to USD 56 trillion in 2024 (Figure ‎1.1, Panel B). In terms of issuers, the United States' portion of the OECD debt stock has risen notably in recent years, constituting nearly half of the total in 2023. The EU’s share is 20%, Japan’s 16%, the United Kingdom’s 6%, with the remaining OECD countries combined accounting for 9%. Significant shifts have occurred over time, with the United States’ share increasing by 7 percentage points (pp) since 2019 and doubling since 2008. This increase is attributed not just to the growth in US debt but also to the appreciation of the US dollar.2

The rise in gross borrowing in the OECD was mainly driven by a few large issuers, with a decrease in the gross borrowing to GDP ratio estimated in more countries than an increase. In only 10 OECD countries, there is an increase exceeding 1pp estimated for 2023; this includes large issuers such as Italy, the United Kingdom and the United States. Meanwhile, a decline exceeding 1pp is estimated for 12 countries, including France, Germany, and Japan (Figure ‎1.2, Panel A). Compared to their 2015-19 averages, increases above 5 pp in this ratio are estimated for Austria, Finland, Germany, and the United States.

Net borrowing levels have varied widely across countries, with the OECD average remaining stable at around 2.5% of GDP in 2023, around 1 pp higher than the 2015-19 averages (Figure ‎1.2, Panel B). Notable increases in net borrowing to GDP ratios in 2023 are estimated for the United Kingdom, where a 5pp rise will be largely due to the funding of cost-of-living-related measures; Israel, with a 4pp increase driven in part by the conflict in the Middle East; and Italy, with a 3pp rise driven mainly by tax cuts and new expenditures. Fiscal consolidation led to the steepest falls in Colombia and Hungary (OECD, 2023[1]).

While sustained increases in net borrowing pose fiscal challenges, low and persistent fiscal surpluses may raise certain concerns such as government access to debt funding and the long-term viability of the government securities market. In recent years, countries such as Denmark, Ireland, Norway, and Sweden have run persistent fiscal surpluses (Figure ‎1.2, Panel B), which may pose challenges in developing credit risk-free yield curves and supporting liquidity in government bond markets. When the supply of government bonds is limited due to diminishing government debt, they could become less liquid and more expensive. A deterioration in the liquidity and efficiency of government securities implies a less useful benchmark for the risk-free interest rate and for pricing and hedging positions in other markets. Additionally, the government’s funding needs, which might temporarily disappear, may resurface, leading to market access challenges in periods of heightened borrowing needs (Bjellerup and Rådahl, 2023[3]). Finally, lower market volumes can affect the viability of the primary dealer business models, which can be crucial for market access.

Refinancing requirements in the OECD area overall have remained largely stable in 2023 but have risen significantly in some countries since the pandemic (Figure ‎1.2, Panel C). Unlike new borrowing, refinancing requirements are not subject to immediate fiscal fluctuations but rather linked to the size and composition of debt portfolios. In 2023, OECD refinancing to GDP ratios remained broadly in line with 2015-19 averages. The United States and Japan individually stand out and are projected to contribute to a USD 5.0 trillion surge in OECD refinancing requirements from 2019 to 2024.

Short-term debt (securities issued with less than a one-year maturity) accounted for 50% of gross borrowing in 2023 and is estimated to account for 51% in 2024, surpassing the previous peak in 2020 of 43% (Figure ‎1.5, Panel A). This shift corresponds to a decrease in the issuance share of fixed rate debt, while the shares of inflation-linked and variable-rate instruments are projected to remain stable at 3%, despite large fluctuations in inflation levels and policy rates during 2021-23. Although there has been a rise in short-term borrowing, the ATM for OECD debt has remained largely stable at around 8 years (see Figure ‎1.20).

The historically high share of short-term borrowing stems largely from macroeconomic and geopolitical uncertainties combined with quantitative tightening (QT) and higher levels of gross borrowing,3 which made it difficult for some sovereign issuers to issue more fixed rate debt. As a result, short-term debt issuance accounted for more than half of the new borrowing (USD 2.0 trillion out of USD 3.6 trillion) in 2023, further increasing the share of short-term debt in the outstanding debt stock (Figure ‎1.5, Panel B).

Issuing short-term instruments provides significant operational flexibility to debt managers, aiding in managing uncertainties and temporary shocks by allowing issuers to quickly raise large levels of funding. During periods of heightened uncertainty and/or when gross borrowing is higher, investor demand typically shifts towards the safest, most liquid assets (often short-term instruments). On the supply side, crises introduce uncertainties about the extent and duration of revenue shortfalls and costs associated with government support measures; short-term instruments allow for better matching of inflows with these outflows, and faster repayment if the revenue shortfalls subside. For these reasons, short-term instruments were used as shock absorbers at the onset of the COVID-19 pandemic (OECD, 2022[4]).

The increased share of short-term borrowing in the OECD in 2023 was predominantly driven by a small group of large issuers. While the share of fixed-rate borrowing in the OECD area declined overall (Figure ‎1.6, Panel A), it actually increased in 19 individual countries, declined in 14 and remained stable in 5. Increases in the share of short-term borrowing by large issuers, such as Canada, France, Germany and especially the United States (where it jumped from 49% of total issuance in 2022 to 62% in 2023) outweighed the increase in the share of fixed-rate issuance in half of OECD countries.

Short-term borrowing in the United States has risen especially since June 2023, mainly to increase the balance in the Treasury General Account. This declined rapidly in the months prior to the suspension of the debt ceiling rule, leading the US Treasury to, in part, finance the federal government by drawing down this balance. Market conditions in the second half of 2023 were shaped by this increased supply of US treasuries in conjunction with QT, and higher long-term yields, indicative of possible changes in long-run neutral interest rates and term premia. The Treasury Borrowing Advisory Committee (TBAC), in November 2023, suggested that this trend of greater short-term borrowing in the United States might persist until Q2 2025 (TBAC, 2023[5]; TBAC, 2023[6]). Likewise, the United Kingdom reduced the issuance share of long-dated (and inflation-linked) bonds in 2023-24 compared to 2022-23 as pension funds’ appetite for such debt wanes as part of a structural change to demand (UK DMO, 2024[7]).

Twenty OECD countries issue inflation-linked securities (linkers) offering a typically much-demanded risk-free product which provides an inflation hedge for investors. This can help to broaden the investor base for government debt and can also represent cost-effective financing for issuers. With inflation higher in 2023, 14 of these countries reduced the share of linkers they issued (Figure ‎1.6, Panel C). This trend emerged despite stable or increasing market demand for linkers according to survey results, indicating a strategic shift in debt portfolio management.4 Colombia recorded the highest borrowing share in linkers at 53%, whilst the average amongst all 20 issuers was 9%. In 2023, only Colombia raised its linker issuance share by over 1pp, while the issuance share in countries such as Italy and the United Kingdom declined by more than 1pp. Notably, Germany ceased its linker programme in 2023, as Canada did in 2022 (Government of Canada, 2022[8]).

Owing to higher gross borrowing, and the loss of central banks as net buyers of sovereign debt, several sovereign issuers have placed a greater focus on capturing demand from retail investors. Survey findings show that 13 OECD countries are giving more consideration to the retail sector in 2023-24. Given the higher yields, government bonds are an increasingly attractive investment for typically more price sensitive household investors. Thus, sovereign issuers are attempting to capture this demand through retail products and/or programmes, with retail purchases reaching unprecedented levels in 2023 (Box ‎1.2).

Sovereign Debt Management Offices (DMOs) predominantly operate as price takers, acting as intermediaries to align gross borrowing with market funding. Both the borrowing levels and market conditions are largely outside the control of DMOs. Consequently, their primary role is to decide the instrument and maturity composition under the constraints of the level of borrowing and the timing of government cash flows and market access. The decision involves selecting from various debt instruments, each with its own investor demand profile, limiting the operational deliverability of certain debt portfolios.

Sovereign issuers select an issuance composition with the purpose of lowering borrowing costs over the long-term, while maintaining manageable risk levels. This decision requires navigating a trade-off between cost and risk: short-term instruments, while less costly, carry higher refinancing and interest rate risk, whereas long-term instruments typically help mitigate against these risks but often at a higher cost to the issuer. Thus, to meet their long-term mandate, sovereign issuers typically opt to issue at a range of maturities, with programmatic issuance patterns.

The issuance strategy will always be somewhat constrained by uneven demand across maturities and the need to support market liquidity. Long-dated securities, typically bought by pension funds and insurance companies, face lower overall demand, limiting sovereign issuers’ strategic flexibility. To enhance market liquidity, DMOs often re-open individual lines multiple times, building them up to a benchmark size, but this can also be costly. This is particularly the case with long-dated securities, where the greater duration reduces the bond market value more significantly during periods of higher rates, resulting in relatively lower cash proceeds raised for the same amount of debt issued.

In 2023, most DMOs in the OECD identified market volatility, rising yields and cash flow forecasting as primary concerns (Figure ‎1.8, Panel A). Inflation, though on a downward trend, remained above target, with higher services price inflation proving particularly sticky. Economic growth has been dampened by tighter monetary policies and subdued global trade (OECD, 2023[1]; OECD, 2024[16]). Geopolitical tensions, including the war in Ukraine and the conflict in the Middle East, are contributing to market volatility. Despite improved cash flow stability compared to the pandemic period, funding needs still show higher-than-average fluctuations, with public finances particularly under strain due to structural factors such as demographic shifts and decarbonisation efforts, and as a result of the current environment of higher interest costs and slower output growth.

Macro-financial conditions significantly impact the costs of risk minimisation. In uncertain economic times, investors' reluctance to take longer-term risks can increase the cost of issuing at longer maturities. Consequently, low risk portfolio costs rise in such scenarios. Furthermore, in the current environment of higher interest rates, committing to long-term maturities might risk locking in these higher rates over an extended period (regret risk), underscoring that cost and risk considerations extend beyond the current shape of yield curves. In 2022-23, cost-efficient risk curves took a parallel shift higher.

Issuance strategies remain flexible across all OECD countries in an uncertain environment. All DMOs use flexible strategies to handle risks, with 81% using liquidity buffers and 73% adapting their communication strategies (Figure ‎1.8, Panel B and C). The value of liquidity buffers lies in ensuring the availability of cash during unpredictable market conditions, without the need to match issuances so closely to outflows. This supports a more flexible issuance approach. Communication strategies are also crucial to finding the right balance between flexibility and predictability.

Flexibility has limits, with large issuers often favouring a programmatic approach to issuance. Sudden shifts in issuance strategies by major issuers can immediately impact the yield curve. This can increase volatility and complicate their ability to meet their mandates and ensure market access at all times. When strategic alterations are necessary, issuers will look to provide as much advance notice as possible. For instance, in 2023, the United States was the sole large issuer to alter the instrument composition of its issuance plan (towards shorter-dated instrument), and they provided a timeline with this shift (US Treasury, 2023[17]).

Smaller issuers typically follow a more flexible approach to debt management, given that strategy changes typically have less impact on their markets. This explains why large-scale annual changes in issuance strategies are more frequently undertaken by smaller issuers (Figure ‎1.6). It also accounts for why, aside from the United States, all twelve countries that modified the instrument composition of their initial issuance strategy are relatively small issuers (Figure ‎1.8, Panel C).5

Flexibility also includes adaptability of issuance techniques and changes to the auction calendar and announcements. In 2023, this included changes in the frequency of auctions, use of switches, use of syndications, private placements, buybacks and post-auction facilities (Figure ‎1.8, Panel D). Other examples include the use of dual/triple-tranche transactions, allowing an issuer to offer multiple debt securities with varied terms within a single issuance, catering to diverse investor preferences and market conditions; and announcing in advance the total issuance volume but delaying the instrument composition announcement until closer to the operation. These methods allow the DMO to better match volatile funding needs with demand for certain securities.

These changes are balanced with practices that boost predictability. Examples include having a stated minimum target size for the outstanding amount of benchmark securities; being transparent in market consultations, feedback and adjustments to avoid unexpected supply movements; keeping an issuance calendar with weekly issuances of various securities; and opting wherever possible for incremental adjustments to the pre-scheduled auction sizes and calendar, allowing for some fine-tuning to the very latest market conditions.

The aggregate debt-to-GDP ratio in the OECD, measured using central government marketable debt, is estimated to remain largely stable through 2022-23 at 83% and is projected to increase slightly to 84% by 2024 (Figure ‎1.9, Panel A). This figure masks the high variability in debt-to-GDP ratios across countries (Figure ‎1.9, Panel D). An increase of over 1pp is projected in nine countries, led by the United States, where it is projected to increase by 3pp. In contrast, 12 countries' debt-to-GDP ratios are projected to fall by over 1pp, with Japan, Portugal, and Spain, where it is projected to fall by over 5pp.

Debt-to-GDP ratios in 2023 surpassed pre-pandemic levels by roughly 5pp on average and 10pp on aggregate in the OECD area.6 The ratio is higher in 24 countries and has increased by over 8pp in all the G7 countries. Japan and US debt-to-GDP ratios have risen by over 24 and 22pp, respectively, since 2019. Over the same period, debt-to-GDP ratios decreased in 14 countries, falling by over 10pp in Ireland, Poland, Portugal and Sweden. It is worth noting that debt-to-GDP ratios and movements vary greatly depending on debt assessment methodologies, with the OECD’s SBO, National Accounts (NA) and Economic Outlook (EO) using different approaches (see Box ‎1.3).

The share of short-term debt in the OECD debt stock has grown and is projected to reach 16% by 2024 (Figure ‎1.9, Panel B), the highest value in 15 years aside from 2020. Concurrently, the proportion of fixed-rate instruments is estimated to have fallen from 79% in 2022 to 76% in 2023 – aside from 2020, this is the lowest figure since 2010, a period during which central banks' bond holdings were negligible.

Despite this decline, fixed rate debt continues to constitute over half of the outstanding debt in all OECD countries in 2023. The share of short-term debt ranges from 0% in 9 OECD countries to 22% in the United States, while the share of linkers ranges from 0% in 17 countries to 32% in Chile (Figure ‎1.9, Panel C).

The decrease in debt-to-GDP ratios between 2021 and 2023 is driven mostly by higher inflation, which boosted nominal GDP growth, and partly by real GDP growth (Figure ‎1.11). Since 2021, inflation contributed to a decrease of 13pp in these ratios while real GDP growth accounted for an 8pp decrease.7 The real yield, which combines the impact of nominal rates and inflation, was negative in nearly all OECD countries.8 Most of the OECD debt was issued when inflation and interest rates were lower, carrying lower costs (Figure ‎1.19). Thus, the current level of nominal GDP growth exceeded the effective interest rates (i.e. the ratio of interest expenses to the debt stock), helping to reduce debt-to-GDP ratios.

With falling inflation and an increasing amount of debt issued under higher rates, the declining trend in debt-to-GDP ratios in 2021-23 is set to reverse in 2024. While inflation’s effect on the nominal GDP is immediate and short-lived, the effects of higher borrowing costs on interest expenditure are gradual and long-lasting. Inflation has the initial effect of lowering debt-to-GDP ratios but, if persistent, its impact in the medium to long run is to increase the inflation premium and debt financing costs.

Inflation does little to improve debt sustainability, as debt-to-GDP trajectories are mainly driven by fiscal balances in the long-term, with unexpected inflation providing only a short-term impact. Historical data indicates the crucial role of primary balances in shaping these trajectories and suggests caution in overinterpreting recent declines (Rawdanowicz et al., 2021[19]; Acalin and Ball, 2023[20]). To avoid an adverse feedback loop of rising interest rates, slow growth and expanding deficits, prudent fiscal policies supported by credible medium-term budget frameworks are essential. Due to these potential feedback loops, for example, the EU is considering incorporating debt sustainability analysis in its new fiscal rules framework.9

Borrowing costs for new debt, measured by the yield to maturity at issuance (YTM), rose from approximately 3% in 2022 to 4% in 2023 on average across OECD countries. This increase is about half of the almost 2pp rise seen in 2022, taking borrowing costs in 2023 to nearly twice their 2015-19 average (Figure ‎1.12). Moreover, unlike in 2022, when all countries experienced an increase, five countries saw a reduction in their average YTM in 2023, with some central banks lowering rates last year (e.g. Chile, Czechia and Poland). This suggests that the rise in YTM is slowing, with further increases unlikely as policy rates are expected to fall in most countries in 2024 (OECD, 2024[16]).

Negative nominal rates, common in many European countries in 2020 and 2021, were only present in Japan in 2023 (Figure ‎1.12, Panel B). Additionally, the percentage of debt issued at borrowing costs below 2% fell sharply from nearly 40% in 2022 to 15% in 2023, whereas the proportion of debt issued with a yield exceeding 4% more than tripled, from 11% in 2022 to 37% in 2023.

The rise in average YTM during 2023 was more pronounced in European countries. Of the 20 countries experiencing the highest increases in YTM at issuance, 13 belonged to the euro area and 5 were non-Euro European nations. This pattern reflects the ECB’s intensified monetary tightening in 2023, a shift from 2022 when euro area nations had some of the smallest yield and policy rate increases (Figure ‎1.13, Panel A). Concurrently, 2023 saw a narrowing of the spread over German bonds in 12 European countries, while only 3 witnessed a widening, possibly reflecting a perception of lower risk, and the mix of the ECB's reinvestments under their asset purchase programmes.

The key driver of the rise in yields is expectations for central bank policy rates, with the average policy rate in the OECD rising from 5% in 2022 to 6% in 2023 (Figure ‎1.13, Panel A). In 2023, 15 out of 20 OECD central banks raised their policy rates. Notably, the ECB raised policy rates by 2.0pp, the second largest increase in the OECD behind the Icelandic central bank. The BoJ has kept policy rates near zero since 2016 while central banks in Chile, Czechia, Hungary, and Poland began to lower rates in 2023.

Nominal rates are typically determined by three factors: real interest rates, inflation expectations and a risk premium. Real interest rates refer to the real returns that consumers and savers demand in exchange for postponing consumption, being an estimation of the real returns on investments and largely determined by real economic growth expectations and volatility. Inflation expectations are incorporated into nominal interest rates to offset savers' anticipations of changing price levels. The risk premium compensates investors for the uncertainty around future inflation levels. This premium tends to be higher for bonds with longer maturities due to the increasing uncertainty of longer-term expectations, whereas short-term bonds are more directly affected by prevailing policy rates and more immediate inflation expectations.

In 2023, there was a consistent rise in the average yields across the OECD, coinciding with a decline in breakeven inflation rates and an increase in real yields (Figure ‎1.13, Panel B, C and D).10 According to the no-arbitrage principle, which posits that securities with the same cash flows should have identical prices, the spot rate of a bond is equivalent to the spot rate of a shorter-term bond and the series of forthcoming short-term forward rates until the bond's maturity. This implies that market pricing reflects expectations that future short-term forward rates will stay high, suggesting policy rates will not return to the near zero level of much of the last decade in the medium-term – the implication being that rates will be kept higher for longer.11

The term spread (i.e. the yield difference between the secondary market yield of the 10- and 2-year benchmark bonds) is a key indicator for sovereign issuers to consider when balancing cost and risk. This typically positive spread arises from the term premium, the extra return demanded by investors for the risk of buying longer-term bonds. The term spread varies, influenced by factors such as the supply and demand of bonds across the curve, inflation and interest rate expectations, the business cycle and output growth. DMOs can use this spread to help assess the trade-off between short-term debt, typically with lower cost and higher liquidity and refinancing risk, and long-term debt, typically with lower refinancing risk and liquidity but higher cost.

The average term spread in the OECD reached historical lows, approaching zero in 2023 (Figure ‎1.14, Panel A). This figure has been trending downwards since 2009, largely driven by increased demand for longer maturities in search of positive yields due to central banks taking interest rates to at or near zero and QE programmes; expectations for lower levels of future inflation; and subdued growth, which diminishes expected returns on future investments (Cohen, P. and Xia, 2018[22]). This downward trend halted in 2021 amid heightened uncertainty about the trajectory of policy rates and inflation but has continued over 2022-23 with inflation easing and markets anticipating declining policy rates in 2024.

Throughout 2023, the term spread moved in different directions, mirroring changing conditions throughout the year (Figure ‎1.14, Panel B). It remained stable from January through May before dipping in June and July, influenced by easing inflation in the United States and Europe and growing recession concerns. From July to December, there was a reversal, with the spread increasing once again due to supply-demand imbalances triggered by higher borrowing needs, especially in the United States, and QT.

The term spread declined overall in more than two-thirds of countries in 2023, but the dynamics varied between euro area and non-euro area economies (Figure ‎1.14, Panel C). In the euro area, all countries experienced a decline in term spreads, by an average of 0.5pp, mirroring the widespread decrease in 10-year yields in the region. Outside the euro area, term spreads moved in both directions, with half of the countries experiencing a rise in 10-year yields and the other half a decline. Compared to 2015-19 levels, term spreads rose only in Japan, reflecting the increased flexibility of the BoJ’s yield curve control (BoJ, 2023[23]).

Nominal yield curves in G7 countries remained largely unchanged and inverted in 2023, contrasting the pre-pandemic period when all G7 countries had mostly upward-sloping curves (Figure ‎1.15). Canada, Japan and the US nominal yield curve remained largely unchanged in 2023. In the same period in the United Kingdom, nominal yields increased at the short end but slightly declined in the 5, 7 and 10-year benchmarks, leading to further inversion. Short-end yields also rose in France, Germany and Italy in 2023, while their nominal yield curves displayed varying degrees of inversion.

Yield curves typically slope upwards, reflecting risk premiums that align with bond duration, whereas downward slopes suggest anticipated reductions in policy rates. The effect of current monetary policy on long-term rates is often limited, given that the market always expects policy rates to converge to the neutral rate in the long run. Given the cyclical nature of high short-term rates during periods of high inflation, yield curves often invert in these periods.12

Inverted yield curves can impact the cost-risk trade-off between alternative borrowing strategies. They imply that longer maturities bear lower costs and carry fewer risks compared to shorter maturities, as rates are fixed for a longer period at a lower rate. However, as inverted curves anticipate falling interest rates, locking into a certain yield level in a longer term when interest rates may be about to fall can lead to regret risk. Thus, some DMOs are opting to issue more at shorter maturities. Additionally, issuance strategies should factor in demand patterns across the maturity spectrum, with certain investors persistently favouring specific parts of the curve, such as pension funds' preference for longer-term investments and money market funds' (MMF) preference for shorter-dated instruments. Thus, a well-balanced issuance strategy, with regular supply at key benchmark maturities across the curve, is often the norm regardless of the shape of the curve.

Similar to nominal yield curves, real yield curves in G7 countries remained largely unchanged in 2023, with all G7 nations, aside from Japan, having displayed largely positive real yield curves. Real yield curves finished 2023 inverted at the short end in all G7 countries aside from Canada and Japan.

The main implication of positive real yields is the pressure it puts on primary balances, raising the risk of a vicious cycle of increasing deficits through higher interest payments and debt. The disparity between real interest rates and GDP growth is a key factor in debt-to-GDP dynamics. Most government debt in the OECD was accumulated when GDP growth exceeded real rates, allowing governments to run primary fiscal deficits and still have falling debt-to-GDP ratios. Now, with OECD debt at record levels and being refinanced with positive real interest rates, most governments must achieve primary balance surpluses to stabilise debt-to-GDP ratios. However, achieving primary balance surpluses in the current context of heightened spending pressures may be harder than in the past (OECD, 2023[1]). DMOs, acting as intermediaries between governments and markets, can only determine the timing and terms under which the debt matures, without significantly impacting debt-to-GDP trajectories.

Interest expenditures in the OECD are expected to have risen gradually from 2.3% in 2021 to 2.9% of GDP in 2023 (Figure ‎1.16, Panel A). Estimates show an increase in this ratio in 20 countries, a negligible change in 9, and a decline in 4.13 Notably, this ratio is expected to rise by more than 0.5pp of GDP in 2023 in Czechia, Greece, Hungary, New Zealand and the United States. Against this overall upward trend, Italy and the United Kingdom are expected to see interest payment-to-GDP ratios reduce in 2023, compared to 2022, due to declining inflation. This will lead to lower expenses on their stock of linkers, which comprise around 10% and 25% of their respective debt portfolios.

The expected interest expenditure-to-GDP ratio in the OECD is just 0.3pp higher in 2023 than 2015-19 levels (Figure ‎1.16, Panel B). Despite the higher yield environment, only 15 countries are expected to have higher interest payments as a percentage of GDP in 2023 compared to pre-pandemic levels. In this period, the United Kingdom has experienced the largest increase, at 1.8pp of GDP, nearly twice the second largest, which is Hungary at 1.0pp, and three times that of France and the United States, both at 0.6pp. This increase amounts to almost two-thirds of the UK’s annual investment expenditure (OECD, 2023[18]).

Euro area countries have experienced relatively lower interest payment levels and less movement in these levels. In the euro area, the share of interest expenditure to GDP typically remains lower than in countries with comparable debt-to-GDP ratios, with the exception of Italy and Greece, reflecting the ECB’s lower policy rates, and expectations for lower inflation in the future (Figure ‎1.16, Panel C). Additionally, euro area countries have seen a comparatively smaller rise in interest expenditure due to their debt profiles (Figure ‎1.16, Panel D).

The increase in the interest expenditure-to-GDP ratio in 2023 is smaller than the jump in the average cost of issuance of 1pp (Figure ‎1.12, Panel A), as changes in the interest expenditure of OECD countries are typically gradual, largely due to the predominance of fixed-rate debt in their portfolios. Under accrual accounting, the predominant method in the OECD area, interest payments accrue every year proportionally to their yield at issuance and outstanding amounts.14 As it is rare for an OECD country to refinance more than a quarter of its debt annually, shifts in interest payments tend to be incremental (Figure ‎1.16, Panel B). Linkers and floaters differ as their costs follow the prevailing inflation and interest rates, respectively.

Interest payments on inflation-linked bonds were significantly impacted by the sudden increase in inflation in 2022-23 but are expected to fall going forward. In 2022, these bonds made up about 8% of the OECD's total debt stock but accounted for around 30% of the interest payments (Figure ‎1.17). In 2023, even though the share of inflation-linked debt in the total stock stayed relatively unchanged, the share of interest payments it accounts for has halved. However, this is still roughly twice the proportion of these instruments in the debt stock.

Interest payments on floating-rate instruments are anticipated to decline along with policy rates in 2024. As they only represent roughly 2% of the OECD debt stock and 4% of the interest payments in 2022-23, the reduction in interest payments from floaters will be minor compared to those from linkers.

Short-term instruments exhibit a similar pattern to floaters, although there is a time lag, and they constitute a far greater share of the debt. While the borrowing costs of floating rates quickly align with policy rates, short-term instruments are refinanced at the prevailing rates when they mature. This meant that, in 2022, interest payments from these instruments were relatively modest, with many of them being issued in late 2021 or early 2022 at lower rates. However, in 2023, they accounted for 15% of the debt portfolio but 23% of interest payments. Due to the time lag of up to one year, a decrease in the associated interest payments is expected only by 2025, assuming that policy rates fall in 2024 (OECD, 2024[16]).

Fixed rate instruments, which account for nearly 80% of the OECD debt portfolio, are expected to lead to higher interest payments in the short- and medium-term. In 2009-20, the share of interest expenditure for fixed rates was, on average, nearly 10% higher than its share of the debt stock. This disparity arose because much of the fixed rate debt outstanding was issued in the decade prior, when rates were higher than in the period 2009-20. Moving forward, a similar dynamic is anticipated, with recently issued fixed rates launched at higher rates, contributing to higher interest payments even as policy rates fall.

In theory, fixed-rate debt is associated with higher issuance costs due to duration risk, while inflation-linked debt is typically less costly for the issuer, as investors are willing to pay a premium for inflation protection. However, in 2022-23, with inflation way significantly above target, linkers were more expensive to issue than fixed rate debt. It is important to note that the true cost-effectiveness of an inflation-linked bond can only be assessed at maturity by comparing it with a fixed-rate counterpart. If the average inflation rate over the bond's lifetime is lower than the breakeven rate at issuance, the linker was cost effective, and vice versa if it is higher.

An analysis of matched pairs of inflation-linked and fixed-rate bonds shows that linkers are typically cost-effective. In 66% of the matched pairs maturing between 2011-21, linkers incurred lower costs than their fixed-rate counterparts (Figure ‎1.18, Panel A). Conversely, in 85% of the matched pairs that matured between 2022-23, fixed rates were more cost-effective. When analysing bonds based on their issuance year, linkers issued between 2011-15 tended to be cost-efficient but the opposite can be said for those issued between 2016-18 (Figure ‎1.18, Panel B). Looking forward, as inflation falls back towards target and linker payments decline, fixed rates are expected to be the main driver behind higher interest payments.

Beyond potential cost efficiencies, inflation-linked bonds can provide several advantages to DMOs (OECD, 2023[24]). From a fiscal perspective the correlation between the interest payments on linkers and government revenues can stabilise the ratio between these variables.15 Also, the market for linkers differs from that of conventional bonds, enabling sovereign issuers to diversify their investor base and to extend the maturity of their debt portfolio, as buyers of linkers often prefer longer maturities.16 Finally, they can contribute to price stability by signalling the government's commitment to controlling inflation.17

Inflation-linked instruments can also come with other considerations for issuers. These considerations include the potential for high volatility in the level of interest payments, since the benefits of lower costs associated with linkers accrue gradually, in contrast to the sharp increases in payments during periods of higher inflation. This dynamic can result in asymmetrical reputational risk for the product, with the cost-effectiveness of linkers being realised incrementally over many years, while the cost of financing them surges and is very visible when inflation spikes.

About 60% of the outstanding OECD debt stock has coupons fixed at pre-2022 yield levels and is likely to be refinanced at higher rates in the future (Figure ‎1.19, Panel A). Roughly 30% of OECD debt was issued in 2022-23, when rates had already begun increasing, and a further 10% of the debt stock consists of linkers and floaters, for which interest has already been adjusted to higher rates and inflation. Thus, in the short- to medium-term, higher interest payments are likely to result from the refinancing of fixed rate issuances and new borrowing.

The current tightening cycle has had a significant impact on borrowing costs, with 30% of OECD debt issued in 2022-23 accounting for around 40% of interest expenses. This marks the first occasion since the data series began in 2009 that the proportion of interest payments from debt issued in the current and preceding year exceeds the share of the debt stock issued in that same period (Figure ‎1.19, Panel B).

Concurrently, debt issued during the peak of the pandemic in 2020-21 amounts to nearly the same 30% of the current OECD debt stock, yet it represents less than 15% of the interest payments. In addition, around 7% of the debt stock that was issued before 2012 accounts for 13% of the interest payments, due to the relatively higher yields prevalent during and right after the GFC.

Looking forward, fixed-rate instruments worth nearly a quarter of OECD GDP are set to mature by 2026 (Figure ‎1.20, Panel A). In five OECD countries, fixed rates exceeding 20% of GDP will mature in this period, including Japan (52%), Italy (33%), the United States (27%), Spain (27%) and France (20%).18 These countries face heightened refinancing risks if higher interest rates persist for much of this period.

The lower the YTM at issuance on maturing debt, the greater the fiscal impact of refinancing that debt at higher yields. The OECD average YTM of fixed-rate debt due by 2026 ranges from 2% to 3%, compared to fixed-rate issues in 2023 which on average were yielding around 4% (Figure ‎1.20, Panel B). Costa Rica and Slovenia, which issued fixed rates in 2023 with a YTM below that of the debt that matures between 2024 and 2026 might see reduced interest payment pressures upon refinancing in the next three years. Conversely, 19 OECD countries, including all G7 nations except Japan, issued fixed rate debt in 2023 at a YTM that was more than 1pp higher than their fixed rate debt that matures by 2026.

Reflecting these disparities between the YTM of the maturing fixed-rate debt and that of new issuances, OECD government interest expenses are estimated to climb by 0.5pp of GDP by 2026, solely due to fixed-rate refinancing (Figure ‎1.20, Panel D).19 This projected increase is the same as the average OECD government's annual expenditure on environmental protection (OECD, 2023[18]) and is relatively higher for a few large issuers including the United States, Italy, France and the United Kingdom.

Compared to the increase in yields, the projected growth in interest payments due to debt refinancing is relatively small, thanks to the average maturity in OECD member's debt portfolios being approximately 8 years as of 2023 (Figure ‎1.20, Panel C). Over the last two decades, sovereign issuers have generally lengthened the maturity profile of their debt portfolios (Figure ‎1.20, Panel E), largely to help reduce refinancing risk. Thus, more than half of the current OECD debt matures or refixes beyond 2027, partially insulating countries from the impact of the current tightening cycle on interest payments, revealing the benefit of longer maturity profiles.

The combination of central banks' large sovereign bond holdings and monetary tightening significantly impacts government security markets and sovereign debt portfolios. The large-scale purchases of government securities by central banks were executed through the issuance of bank reserves in exchange for securities. When this is considered in the context of the entire public sector balance sheet, central banks' bond holdings essentially act as floating rates with private creditors. This transforms about one third of the government debt into floating rates (Figure ‎1.20, Panel A), reducing the average ATR for OECD countries with QE programmes from 8 to 5 years (Figure ‎1.20, Panel C). This shift obscures not only the interest rate risks that governments face but also the interest costs; instead of higher interest payments, increased costs manifest as reduced profit remittances from central banks or demands for indemnification.

Higher interest rates diminish profits or cause losses on some central bank bond holdings (ECB, 2023[25]; OECD, 2023[26]). This happens in two ways. First, they raise the cost of liabilities linked to policy and deposit rates. This has a higher impact where central banks engaged in QE, as they have more liabilities from issuing interest-bearing bank deposits to buy bonds. Second, where central banks conducted a programme of QE and either use mark-to-market accounting or are actively selling bonds, higher rates reduce the value of their holdings, as higher discount rates are applied to their fixed cash flows.20

Central bank losses can impact government's gross borrowing needs as they halt profit remittances or trigger indemnification. Central banks usually profit from seigniorage revenue, derived from issuing non-interest-bearing currencies to acquire interest-bearing assets. These profits vary from 0.1 to 0.9% of GDP (Reis, 2013[27]). Losses can lead to suspended remittances and, depending on the country, may even lead to recapitalisation or indemnification by the central government (Bell et al., 2023[28]).21

Central bank losses in the United Kingdom have resulted in direct indemnification by the government whilst there has been a halt in the remittances in the United States and in some EU countries. The Bank of England (BoE), which uses mark-to-market accounting, and whose bond holdings have the highest average maturity of any OECD country (Figure ‎1.20, Panels C and E), faces substantial losses. Since October 2022, the UK government has allocated about 1% of GDP to indemnification, with losses anticipated to rise to 5% of GDP (UK Office for Budget Responsibility, 2023[29]). In the euro area where central banks were previously transferring around 0.1% to 0.2% of euro area GDP annually to member countries, remittances are projected to stop altogether for a total of 11 years in Germany, 4 years in the Netherlands, 3 years in France and 2 years in Spain (Belhocine, Bhatia and Frie, 2023[30]). The US Federal Reserve, after a decade of annual remittances paid to the US federal government which rose to as high as 0.5% of GDP in 2020, ceased to make a profit on its holdings in 2022 Q3 and is not expected to resume paying remittances until 2027 (Castro and Jordan-Wood, 2023[31]).

In 2023, central banks’ average government bond holdings as a percentage of the national government debt decreased from 37% to 33% in 21 selected OECD countries (Figure ‎1.21, Panel A).22 Notable declines occurred in Canada, Finland, and Baltic countries.

The four largest central banks that engage in QT – the Federal Reserve, ECB, BoE, and Bank of Canada (BoC) – reduced their government security holdings from USD 11.5 trillion in 2021 to USD 10.0 trillion in 2023 (Figure ‎1.21, Panel B). An additional decrease of USD 1.0 trillion is projected for 2024. This corresponds to an annual decrease of around 7% in their government security holdings over 2021-24.

QT can be achieved by central banks not or only partially reinvesting the proceeds from their maturing bond holdings, known as passive QT, or through a combination of not reinvesting and directly selling some of their outstanding bond holdings, known as active QT. The ECB and central banks of Canada and the United States are adopting a passive QT approach (BoC, 2022[32]; ECB, 2023[33]; OECD, 2023[34]; BIS, 2023[35]; Fed, 2022[36]). Meanwhile the Bank of England is adopting an active QT approach. (BoE, 2022[37]). Conversely, the BoJ has an ongoing QE programme due to Japan’s low inflation levels (BIS, 2023[35]).23 Among these central banks, the BoC is forecasted to see the largest decline in government security holdings over 2021-24, of 50%. This compares to a roughly 25% reduction for the Federal Reserve, 40% for the BoE, and 7% in the ECB. The ECB, adopting a slower pace of balance sheet reduction, is projected to surpass the Federal Reserve in having the largest holdings of government securities during 2024.

The pace of QT programmes will tend to align with the maturity profile of central banks' bond holdings (Figure ‎1.21, Panel C). Nearly 50% of the BoC's government holdings mature by 2026, compared to 40% for the Fed and 30% for the BoE. In contrast, about 50% of the BoE's government security holdings mature beyond 2031, whereas for the Fed this figure is 40%, and for the BoC, it is 30%.

As central banks’ government bond holdings are predominantly fixed-rate, QT has significantly increased the supply of these securities to the market. New borrowing in fixed rates as a share of their respective outstanding amounts in 2023 was 1% for Canada, 2% for the United States and 6% for the euro area and the United Kingdom. Due to QT, the extra supply to the market rose to 9% of this figure in Canada, 6% in the United States, 7% in the euro area and 9% in the United Kingdom (Figure ‎1.21, Panel D).

QT implies more debt for the market to absorb, and unless debt managers alter their issuance strategy, there is also more net duration for the market to absorb.24 The ATM of central bank holdings is 1 year longer than the ATM of the bonds in the market in Canada and the United States (Figure ‎1.21, Panel E). The United Kingdom faces additional pressure in terms of duration supply to the market, as the BoE's active QT has led to the sale of bonds with an ATM of close to 20 years.25

In addition to defining the instrument and maturity composition of issuances, another crucial task of public debt management (PDM) is developing a diverse investor base (OECD, 2019[38]). This diversity is key, due to the varying behaviour and investment needs of different investors. It lessens the probability of uniform buying or selling movements, contributing to deeper and more liquid markets. It also enables sovereign issuers to better balance cost and risk trade-offs, with some investors seeking high returns through risk-taking and others preferring more conservative and programmatic strategies. Geographical diversification can further support demand, leaving an issuer less reliant on the macro-financial conditions of specific regions.

Sovereign issuers engage in many activities to develop a stable investor base, including enhancing transparency in debt operations transparency, frequently engaging with investors, diversifying their product offerings, and taking measures to boost market liquidity. A consistent and predictable issuance strategy can play a crucial role in building a stable investor base. Additionally, sovereign issuers from more than 50 countries, of which nearly 30 are in the OECD area, have issued sustainable bonds, with one of their main objectives for doing so being the diversification of the investor base (as discussed further in Chapter 3).

In the past 15 years, the investor base for OECD general government debt has shifted significantly as a result of the greater role of domestic central banks in these markets. Their share of sovereign bond holdings jumped from 5% in 2008 to 27% in 2022 (Figure ‎1.22, Panel A), reflecting their increased purchases of new and existing government bonds against a backdrop of growing supply. Simultaneously, other domestic investors' shares were reduced, with domestic banks' holdings dropping from 46% to 33% and the holdings by the non-banking sector declining from 20% to 14%.

Foreign investors' shares fluctuated between 28% and 31% during 2008-19, but fell in 2020 as central banks accelerated or re-introduced QE programmes in response to the pandemic. Foreign official investors' shares dropped from a high of 17% in 2013 to 12% in 2022. This coincides with diverging growth rates between the OECD government debt stock and global foreign exchange (FX) reserves (Figure ‎1.22, Panel D); Although both grew at a similar rate between 2008-15, FX reserves rose by 9% in 2015-23, while the OECD debt stock surged by 57% in the same period (Figure ‎1.22, Panel C).26

Compared to the period 2008-22, survey responses indicate that in 2023 there was a partial reversal in the investor base composition trends, with a decrease in the share of domestic central banks and a rise in the non-bank financial sector's share (Figure ‎1.22, Panel B). In 2023, the share of domestic central banks’ sovereign debt decreased in 16 countries, while it increased for domestic asset managers and hedge funds in 11 countries, for institutional investors increased in 9, and for of other investors increased in 10. This indicates a growing absorption of government bond supply by the non-banking financial sector in times of monetary tightening and rising yields (Figure ‎1.22, Panel D).

As QT progresses, it remains unclear which investors will absorb the additional supply of government bonds and how maturity structures and yields will evolve. This will affect the cost and risk profiles of sovereign debt portfolios. The change will not simply be a shift back to the investor base before the major QE programmes which began in 2008. Currently, the OECD central government marketable debt stock, as a share of GDP, stands at around 83%, compared to 53% in 2008, with the 30pp increase largely accounted for by central banks’ bond holdings. As a result of QT and continued high borrowing, the market will have to absorb a record level of net supply, with potential consequences for borrowing costs and maturity profiles.

The composition of sovereign debt investors is expected to increasingly include more price sensitive investors, which could put upward pressure on yields. This contrasts with central banks, whose acquisition of government bonds is largely detached from the current bond prices as they are guided by mandates for price stability. Likewise, under QT, central banks tend to be price insensitive sellers, ready to sell if required by their mandates, irrespective of market prices.27

Foreign investors' medium-term demand may rise as the current high FX hedging costs decline, but their demand for hard currency could decrease if economies perform more robustly and avoid a downturn. Tightening cycles typically elevate FX hedging costs, as investors need to hedge against a stronger currency amid growing interest rate differentials and heightened FX volatility. Furthermore, if a soft economic landing materialises, it might boost the appeal of government securities in weaker currencies compared to hard currencies, reducing the demand for the sovereign debt of most OECD countries. Conversely, a downturn would likely result in “safe havens”, such as Germany and the United States, benefitting from a “flight to safety”, which can translate into lower borrowing costs.

Domestic banks’ demand for government securities is influenced by the relevant regulations. Post-GFC Basel III’s enhanced liquidity requirements caused a notable rise in banks' demand for these securities in the last decade but also reduced the likelihood of new regulations, given that the banking sector has largely remained resilient during the ongoing monetary tightening cycle (FSB, 2023[41]). Yet, following the banking turmoil in the first semester of 2023, the United States is contemplating new capital treatment rules, which could elevate the demand for government securities by US-based institutions (CRS, 2023[42]).

Banks often meet liquidity requirements by holding government securities or central bank reserves, basing their choice on yield levels (Eren, Schrimpf and Xia, 2023[43]), duration risk, and currency considerations. During monetary tightening, duration risks prompt banks to opt for bank reserves over bonds, although they revert to bonds under improved funding conditions. Thus, banks are not expected to be major buyers of government bonds under QT, as evidenced by the decrease in their government bond holdings in 2022. However, in the medium-term, with a potential shift towards monetary easing, banks’ demand for government securities might rise, provided that the yield differential relative to central bank reserves offsets the higher duration risk.

As foreign investors and domestic banks reduce their holdings of government securities during the tightening cycle, the domestic non-bank sector is expected to emerge as the primary marginal buyer of new government securities. Thus, the trend witnessed in 2022-23 is expected to persist with the tightening cycle. This varied group encompasses the most price sensitive investors (Fang, Hard and Lewis, 2023[44]) such as investment funds, hedge funds, and households. The latter bought a record amount of government bonds through retail programs in 2022-23 (Box ‎1.2).

Pension funds and insurance companies' demand for government bonds will probably align with long-term yield trends. These investors seek higher yields and longer maturities to match their liabilities. Yet, uncertain macroeconomic conditions in 2023 impacted demand for long-duration bonds in some countries.28 Regulatory mandates for pension funds to maintain minimum government bond holdings can also influence their holdings, but any such regulatory changes are not planned in the near future (OECD, 2023[45]).

Among investment funds, MMFs, open-ended funds (OEFs), and hedge funds contribute to market liquidity, but can also be the cause of instability. MMFs, who are often major investors in short-term government securities, may quickly sell assets to fulfil investor redemptions in periods of market stress. OEFs face liquidity mismatches due to their daily redemption promises and holding of bonds, which can result in fire sales at the onset of crises to avoid selling later and sustaining large losses. Hedge funds, which may be required to rapidly unwind their leveraged positions in times of heightened volatility, can also exacerbate selling pressures.29

Sovereign issuers regard the non-bank financial sector’s involvement in bond markets as a contributor to liquidity but note its potential to amplify market shock sensitivity.30 These entities invest in government bonds under a different regulatory framework than banks. Moreover, as they are not primary dealers, they are not obligated to maintain liquidity, and they offer two-way pricing at all times. Thus, in times of heightened volatility, they can sometimes add to market instability. This became apparent in 2020, when central banks stepped in to purchase a large number of government securities, and during the UK 'mini-budget' crisis in Autumn 2022, which was exacerbated by the leveraged positions of pension funds.31

If the non-bank financial sector becomes an even more significant holder of government bonds, supporting continued liquidity in times of heightened market stress is crucial. These liquidity challenges can also create financial stability concerns, with central banks and regulators considering steps to prevent the need for direct interventions to stabilise markets, as was required in 2020.32 This is vital for sovereign issuers, as fire sales of government securities could result in higher cost of borrowing or more drastically prevent them from accessing liquidity during crises.

Liquidity in the secondary market for government bonds is essential to support primary market access and minimise sovereign borrowing costs over the long run. Therefore, liquidity conditions are among the most important factors that are actively monitored by sovereign DMOs (OECD, 2018[46]). The share of DMOs that reported a decline in liquidity fell from its record high level of 63% in 2022 to 19% in 2023 (Figure ‎1.23). Sovereign issuers assessed geopolitical risks and the slowdown in global trade as the main factors negatively impacting liquidity in 2023. Unlike 2022, geopolitical risks outweighed macroeconomic uncertainties in terms of their impact on liquidity (Figure ‎1.23, Panel B).

Liquidity in the repo and derivative bond markets in 2023 improved in a larger number of OECD countries compared to those experiencing a decline, a notable shift from the pattern for 2022 (Figure ‎1.23, Panel C). These markets experienced increased liquidity mainly due to QT, increasing the free float (i.e. the share of debt held by the market) and therefore the availability of government securities for use in repo and derivative hedging. However, some issuers with negative net borrowings did not report improvements in liquidity in their markets, implying that their reduced borrowings may have offset the increase in collateral released by central banks into the market for repo and derivative transactions.33

DMOs assess market liquidity using multiple indicators such as bid-ask spreads, market turnover, volatility, and auction-related metrics. In 2023 these data points have presented somewhat conflicting signals (Figure ‎1.24).

Bid-ask spreads have nearly returned to pre-pandemic levels. In the US, bid-ask spreads have been stable, but are still slightly wider than pre-pandemic levels; in the euro area, the bid-ask spreads of the 30-year benchmark tightened significantly; in Japan, they widened in 2023, particularly at the long end of the curve with the BoJ’s more flexible approach to yield curve control; whilst in the UK they tightened at the long end.

Volatility in the OECD area remains above pre-pandemic levels, notably at the long end of the yield curve. Many OECD countries noted that structural changes in markets have contributed more to higher volatility, as well as ongoing macroeconomic and geopolitical uncertainties, and a more price sensitive investor base. Despite this heightened volatility and greater borrowing needs, bond auctions in the OECD were predominantly oversubscribed, indicating robust investor interest.34

Auction metrics are of great importance as sovereign issuers mainly borrow through auctions where, in most cases, only primary dealers can directly participate.35 This structure contrasts with conventional exchanges where participants directly trade, as seen in equity and derivative markets. Primary dealers buy bonds in auctions, operate in the interdealer market, and engage with clients mainly through over-the-counter operations.36

Thus, the role of primary dealers in buying and distributing bonds is vital for liquidity. As they are mostly banks (Figure ‎1.25, Panel A),37 they face stricter post-GFC regulations under Basel III, leading to increased balance sheet constraints that could impact their warehousing capability (Figure ‎1.25, Panel B). With the obligation to maintain capital against settlement risk, a rise in trading volume elevates capital requirements. In scenarios where dealers lack adequate capital or cannot promptly secure it, their ability to trade is constrained. This is evidenced in the United States, where liquidity dropped when the balance sheets of primary dealers in treasuries were heavily utilised (Duffie et al., 2023[47]).38 Under the previous QE regime, these constraints were less of a factor, but they are exacerbated when volatility is higher, and in particular, when having to absorb the extra supply created by QT.

Despite heightened volatility and borrowing requirements, primary dealers have remained able to fulfil their roles. Most OECD countries reported stability in primary dealers' market-making capacity (Figure ‎1.25, Panel C), with downturns reported only in periods of heightened market stress. The approach of selecting dealers to promote competition and enhance market activities has proven effective.39 To address the current demanding context, DMOs in 8 OECD countries are adjusting dealers’ performance evaluation standards, increasing their privileges in 4 countries, and reducing their obligations in 2.

Common measures adopted to help support market liquidity include regular taps of existing securities, securities lending facilities, enhanced market communication and buyback/switch operations (these were deemed the most effective measures – Figure ‎1.26).40 Recent initiatives include the US’s new buyback program which starts in 2024 (US DMO, 2023[48]),41 and the Swiss DMO's plan to build a government bond portfolio for collaterised money markets.

Tapping lines when rates are higher brings immediate cost but accrues longer-term benefits. In periods of higher rates, tapping bonds trading below par means that the government accrues more debt than the funds it raises, increasing net debt levels. However, tapping leads to lower interest expenses due to depressed liquidity premiums, fulfilling DMOs’ long-term mandates of aiming for low borrowing costs under controlled risks.

Security lending enables DMOs to stabilise yield curves, standardise repo rates, supply collateral and cash for market making, reduce bid-ask spreads, and access the repo market for their own cash management purposes. The main objective is to aid the market without becoming a dominant player, with large issuers often being more active. Large interventions are generally limited to exceptional cases. The benefits of these facilities are widely considered to outweigh their operational costs (e.g. IT, human resources and accounting).

Anticipating possible market changes due to QT, some sovereign issuers are discussing a wider adoption of central clearing in government bond markets. This involves a Central Clearing Counterparty (CCP) acting as the intermediary to both buyers and sellers. Currently, government bond markets, unlike their repo and derivative counterparts, are mostly not centrally cleared. In France, Germany, the United Kingdom and the United States, only a negligible volume of transactions are centrally cleared; in Italy and Japan, transactions are centrally cleared for inter-dealer trades, but not for dealer-to-client trades (FSB, 2022[49]).

Central clearing offers the benefits of decreasing the pressure on primary dealers’ balance sheets and ensuring access to central bank aid in the event of a crisis. By netting opposing transactions, a CCP can reduce settlement obligations, evidenced by the potential reductions of up to 53% in the United Kingdom (Baranova et al., 2023[50]), and up to 70% in both the United States (Fleming and Keane, 2021[51]) and Canada (Chen et al., 2022[52]) during the pandemic. CCPs' ability to utilise central bank reserves and security lending facilities further enables effective trading during such periods, acting as a backstop for defaults and transaction clearing. Additionally, higher pre- and post-trade transparency in combination with central clearing can enable all-to-all trading in government bond markets. This framework is currently only present in Israel (Kutai, Nathan and Wittwer, 2023[53]) but it could further increase liquidity in other markets (Chaboud et al., 2022[54]).

However, structural changes in government bond markets, such as greater central clearing and all-to-all trading, require careful consideration due to the potential risks they pose to the current system. Firstly, some market participants often prefer direct dealer relationships for advantageous pricing and discreet handling of large-scale trades, with primary dealers incentivised to engage even in times of acute stress due to customer loyalty. Secondly, these reforms could lead to a concentration of risks in CCPs. Thirdly, while reforms could enhance dealers' intermediation capacity, this additional balance sheet room might be diverted to other business areas, potentially affecting government bond market liquidity. Thus, structural reforms should be approached cautiously, incrementally, with long-term aims and with comprehensive market consultation, as is the current practice of sovereign issuers. The recent announcement in the US by the SEC that all trading in US Treasuries should be centrally cleared by 2026 is a huge development in the largest and most important bond market in the world.

Against the backdrop of changing market conditions and investor dynamics, particular attention needs to be given to monitoring and supporting the liquidity in government securities markets. Sovereign debt management offices should monitor market conditions closely and remain vigilant in using a variety of tools to support liquidity, including through enhanced market communication, tapping existing securities and conducting buyback and switch operations, as well as providing security lending facilities,


[20] Acalin, J. and L. Ball (2023), “Did the U.S. really grow out of its World War II debt?”, National Bureau of Economic Research, Vol. Working Paper No. 31577, http://www.nber.org/papers/w31577.

[66] Bank of England (2022), “Asset Purchase Facility: Gilt Sales”, Market Notice 1 September 2022, https://www.bankofengland.co.uk/markets/market-notices/2022/september/apf-gilt-sales-market-notice-1-september-2022 (accessed on 16 January 2024).

[50] Baranova, Y. et al. (2023), “The potential impact of broader central clearing on dealer balance sheet capacity: a case study of UK gilt and gilt repo markets”, BoE Staff Working Paper No. 1,026, https://www.bankofengland.co.uk/-/media/boe/files/working-paper/2023/the-potential-impact-of-broader-central-clearing-on-dealer-balance-sheet-capacity.pdf.

[13] Belgian Debt Agency (2023), Bons d’Etat, https://www.debtagency.be/fr/productsbeinfo.

[30] Belhocine, N., A. Bhatia and J. Frie (2023), “Raising Rates with a Large Balance Sheet: The Eurosystem’s Net Income and its Fiscal Implications”, IMF Working Paper 23/145, https://www.imf.org/en/Publications/WP/Issues/2023/07/07/Raising-Rates-with-a-Large-Balance-Sheet-The-Eurosystems-Net-Income-and-its-Fiscal-535549.

[28] Bell, S. et al. (2023), “Why are central banks reporting losses? Does it matter?”, BIS Bulletin 68, https://www.bis.org/publ/bisbull68.htm.

[35] BIS (2023), “Central bank asset purchases in response to the Covid-19 crisis”, Committee on the Global Financial System Papers No. 68, https://www.bis.org/publ/cgfs68.pdf.

[21] BIS (2023), Central bank policy rates, https://www.bis.org/statistics/cbpol.htm (accessed on 20 December 2023).

[3] Bjellerup, M. and J. Rådahl (2023), “Debt Office Commentary: Long-term conditions for debt management”, Swedish National Debt Office, Vol. 2023 Number 3, https://www.riksgalden.se/contentassets/ea257dc12453430186b47e2bcfe20376/debt-office-commentary-long-term-conditions-for-debt-management.pdf.

[68] Blanchard, O., Á. Leandro and J. Zettelmeyer (2021), “Redesigning EU Fiscal Rules: From Rules to Standards”, Economic Policy, Vol. 36/106, https://doi.org/10.1093/epolic/eiab003.

[32] BoC (2022), “Understanding quantitative easing”, Bank of Canada, https://www.bankofcanada.ca/2022/06/understanding-quantitative-easing/#:~:text=QE%20helps%20stabilize%20the%20economy,interest%20they%20pay%20to%20bondholders.

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

[9] Boitreau, S. and L. Secunho (2020), “Retail investors in government debt: Can fintech bring about cheaper, more inclusive programs?”, World Bank Blog, https://blogs.worldbank.org/psd/retail-investors-government-debt-can-fintech-bring-about-cheaper-more-inclusive-programs.

[23] BoJ (2023), “Statement on Monetary Policy October 31”, https://www.boj.or.jp/en/mopo/mpmdeci/mpr_2023/k231031a.pdf.

[31] Castro, M. and J. Jordan-Wood (2023), “The Fed’s Remittances to the Treasury: Explaining the ’Deferred Asset”, Federal Reserve Bank of ST. Louis, https://www.stlouisfed.org/on-the-economy/2023/nov/fed-remittances-treasury-explaining-deferred-asset.

[54] Chaboud, A. et al. (2022), “All-to-All Trading in the U.S. Treasury Market”, New York Fed Staff Report No. 1036, https://www.newyorkfed.org/research/staff_reports/sr1036.

[52] Chen, J. et al. (2022), “Potential netting benefits from expanded central clearing in Canada’s fixed-income market”, Bank of Canada Staff Analytical Note 2022-8, https://www.bankofcanada.ca/2022/06/staff-analytical-note-2022-8/.

[22] Cohen, B., H. P. and D. Xia (2018), “Term premia: models and some stylised facts”, BIS Quarterly Review, September 2018, https://www.bis.org/publ/qtrpdf/r_qt1809h.htm.

[42] CRS (2023), “Bank Capital Requirements: Basel III Endgame”, Congressional Research Service, https://crsreports.congress.gov/product/pdf/R/R47855.

[47] Duffie, D. et al. (2023), “Dealer Capacity and U.S. Treasury Market Functionality”, New York Fed Staff Report No. 1070, https://doi.org/10.59576/sr.1070.

[62] EC (2023), “Commission adopts report on the functioning of the Money Market Funds Regulation (MMF)”, News Article, https://finance.ec.europa.eu/news/commission-adopts-report-functioning-money-market-funds-regulation-mmf-2023-07-20_en.

[25] ECB (2023), “Overall amount of unrealised losses in euro area banks’ bond portfolios contained”, Press Release, https://www.bankingsupervision.europa.eu/press/pr/date/2023/html/ssm.pr230728_1~4d466b8b80.en.html.

[33] ECB (2023), “Quantitative tightening: rationale and market impact”, European Central Bank Speech, https://www.ecb.europa.eu/press/key/date/2023/html/ecb.sp230302~41273ad467.en.html.

[43] Eren, E., A. Schrimpf and F. Xia (2023), “The demand for government debt”, BIS Working Paper No. 1105, https://www.bis.org/publ/work1105.htm.

[55] Escolano, J. (2010), “A Practical Guide to Public Debt Dynamics, Fiscal Sustainability, and Cyclical Adjustment of Budgetary Aggregates”, IMF Technical Notes and Manuals, https://www.imf.org/external/pubs/ft/tnm/2010/tnm1002.pdf.

[60] European Parliament (2023), “Debt sustainability analysis as an anchor in EU fiscal rules”, Economic Governance and EMU Scrutiny Unit, https://www.europarl.europa.eu/RegData/etudes/IDAN/2023/741504/IPOL_IDA(2023)741504_EN.pdf.

[44] Fang, X., B. Hard and K. Lewis (2023), “Who holds sovereign debt and why it matters”, BIS Working Papers No. 1099, https://www.bis.org/publ/work1099.pdf.

[63] FCA (2022), “Resilience of Money Market Funds”, UK Financial Conduct Authority, https://www.fca.org.uk/publications/discussion-papers/dp22-1-resilience-money-market-funds.

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

[51] Fleming, M. and F. Keane (2021), “The Netting Efficiencies of Marketwide Central Clearing”, New York Fed Staff Report No. 964, https://www.newyorkfed.org/research/staff_reports/sr964.

[41] FSB (2023), “Promoting Global Financial Stability”, 2023 FSB Annual Report, https://www.fsb.org/wp-content/uploads/P111023.pdf.

[64] FSB (2023), “Thematic Peer Review on Money Market Fund Reforms: Summary Terms of Reference and request for public feedback”, Annoucement, https://www.fsb.org/2023/08/thematic-peer-review-on-money-market-fund-reforms-summary-terms-of-reference-and-request-for-public-feedback/.

[65] FSB (2022), “Enhancing the Resilience of Non-Bank Financial Intermediation”, Progress report, https://www.fsb.org/2022/11/enhancing-the-resilience-of-non-bank-financial-intermediation-progress-report-2/.

[49] FSB (2022), “Liquidity in Core Government Bond Markets”, Report, https://www.fsb.org/2022/10/liquidity-in-core-government-bond-markets/.

[67] Garcia-Macia, D. (2023), “The Effects of Inflation on Public Finances”, IMF Working Paper WP/23/93, https://www.imf.org/en/Publications/WP/Issues/2023/05/05/The-Effects-of-Inflation-on-Public-Finances-533099.

[8] 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.

[39] IMF (2023), “Foreign Exchange Reserves”, IMF Data, https://data.imf.org/?sk=e6a5f467-c14b-4aa8-9f6d-5a09ec4e62a4.

[40] IMF (2023), “IMF Sovereign Debt Investor Base for Advanced Economies”, IMF Data, https://www.imf.org/-/media/Websites/IMF/imported-datasets/external/pubs/ft/wp/2012/Data/_wp12284.ashx.

[2] IMF (2023), World Economic Outlook (October 2023), https://www.imf.org/external/datamapper/NGDP_RPCH@WEO/OEMDC/ADVEC/WEOWORLD (accessed on 1 December 2023).

[10] Italian Ministry of Economy and Finance (2023), “Press releases”, MeF Website, https://www.mef.gov.it/en/ufficio-stampa/comunicati/2023/index.html.

[53] Kutai, A., D. Nathan and M. Wittwer (2023), “Exchanges for Government Bonds? Evidence during COVID-19”, Bank of Israel, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3882548.

[59] Lybek, T. (2023), “Hungarian Monetary Policy Operations Before, During, and After the Pandemic: Hungary”, IMF, https://www.elibrary.imf.org/view/journals/018/2023/005/article-A001-en.xml.

[11] Ministry of Finance of Japan (2023), “Debt management report”, MoF website, https://www.mof.go.jp/english/policy/jgbs/publication/debt_management_report/index.htm.

[12] National Savings & Investments (2023), , Annual report, https://nsandi-corporate.com/performance/annual-reports.

[16] OECD (2024), OECD Economic Outlook, Interim Report February 2024: Strengthening the Foundations for Growth, OECD Publishing, Paris, https://doi.org/10.1787/0fd73462-en.

[18] OECD (2023), Government at a Glance 2023, OECD Publishing, Paris, https://doi.org/10.1787/3d5c5d31-en.

[26] OECD (2023), OECD Economic Outlook, Volume 2023 Issue 1, OECD Publishing, Paris, https://doi.org/10.1787/ce188438-en.

[1] OECD (2023), OECD Economic Outlook, Volume 2023 Issue 2, OECD Publishing, Paris, https://doi.org/10.1787/7a5f73ce-en.

[34] OECD (2023), OECD Economic Surveys: Israel 2023, OECD Publishing, Paris, https://doi.org/10.1787/901365a6-en.

[24] OECD (2023), OECD Sovereign Borrowing Outlook 2023, OECD Publishing, Paris, https://doi.org/10.1787/09b4cfba-en.

[45] OECD (2023), Pensions at a Glance 2023: OECD and G20 Indicators, OECD Publishing, Paris, https://doi.org/10.1787/678055dd-en.

[4] OECD (2022), OECD Economic Outlook, Volume 2022 Issue 2, OECD Publishing, https://www.oecd-ilibrary.org/economics/oecd-economic-outlook/volume-2022/issue-2_f6da2159-en.

[38] OECD (2019), OECD Sovereign Borrowing Outlook 2019, OECD Publishing, Paris, https://doi.org/10.1787/aa7aad38-en.

[46] OECD (2018), OECD Sovereign Borrowing Outlook 2018, OECD Publishing, Paris, https://doi.org/10.1787/sov_b_outlk-2018-en.

[15] Portuguese Treasury and Debt Management Agency (2023), “Government Debt Statistics”, IGCP, https://www.igcp.pt/en/1-4-399/statistics/government-debt/.

[19] Rawdanowicz, Ł. et al. (2021), “Constraints and demands on public finances: Considerations of resilient fiscal policy”, OECD Economics Department Working Papers, No. 1694, OECD Publishing, Paris, https://doi.org/10.1787/602500be-en.

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

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

[27] Reis, R. (2013), “The Mystique Surrounding the Central Bank’s Balance Sheet”, American Economic Review 103(3), pp. 135–140, https://doi.org/10.1257/aer.103.3.135.

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

[5] TBAC (2023), “Explaining the recent market moves across the Treasury yield curve”, TBAC Presentation, https://home.treasury.gov/system/files/221/TBACCharge1Q42023.pdf.

[6] TBAC (2023), “Outlook for demand for US Treasuries”, TBAC Presentation, https://home.treasury.gov/system/files/221/TBACCharge2Q42023.pdf.

[7] UK DMO (2024), “Minutes of annual consultation meetings with GEMMs and gilt investors held at HM Treasury on 29 January 2023”, The UK Debt Management Office’s financing remit 2024-25, https://www.dmo.gov.uk/media/rihkgdsg/sa300124.pdf.

[29] UK Office for Budget Responsibility (2023), “Economic and fiscal outlook”, https://obr.uk/efo/economic-and-fiscal-outlook-november-2023/.

[48] US DMO (2023), “Treasury’s Current Views on the Operational Design of a Regular Buyback Program”, The Department of the Treasury, https://home.treasury.gov/system/files/221/TreasurySupplementalQRQ32023.pdf.

[61] US SEC (2023), “SEC Adopts Money Market Fund Reforms and Amendments to Form PF Reporting Requirements for Large Liquidity Fund Advisers”, Press Release, https://www.sec.gov/news/press-release/2023-129.

[17] US Treasury (2023), Upcoming Auctions, https://www.treasurydirect.gov/auctions/upcoming/ (accessed on 1 November 2023).

[14] US Treasury Department (2023), “Securities issued in TreasuryDirect”, Fiscal Data Treasury Website, https://fiscaldata.treasury.gov/datasets/securities-issued-in-treasurydirect/.

The Borrowing Outlook survey collects gross borrowing requirements, redemption and outstanding debt amounts with a breakdown of these items by maturity, currency, interest rate types and Environmental, Social and Governance (ESG)-labelling (i.e. sustainable bonds). It also collects data on DMOs’ holdings, NextGenerationEU loans and country-specific methodological aspects. It uses the core definition of sovereign debt, 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.

The coverage of institutions by debt statistics varies from public sector to central government. Public sector represents the broadest institutional coverage, as it includes local governments, state funds financial and non-financial public corporations as well as central government debt. The 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 an instrument of the central authority of a country, except for 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 purpose of estimations.

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 to trade in the secondary markets, and which are not. Another reason is to enable 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 difficulties.

  • The 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 in 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 the calendar year amounts to how much the DMO needs to issue in nominal terms 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 required to finance the current budget deficit. While the refinancing of redemptions is a matter of rolling over the same exposure as before, NBR refers to new exposure in the market, or new borrowing.

  • 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 redeemed.

  • Redemptions refers 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.

  • Total OECD area denotes the following 38 countries: Australia, Austria, Belgium, Canada, Chile, Colombia, Costa Rica, Czechia, 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. The euro (EUR) is the official currency of 20 out of 27 EU Member countries. These countries are collectively known as the euro area. The euro area countries are Austria, Belgium, Croatia, 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.

  • 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.

  • 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 2003 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, with the data sourced from LSEG.

  • All figures use 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 2023 and 2024, 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 2023, 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 to estimate some missing data.

  • Gross and net borrowing requirements presented in the report follow the standardised method and are net of cash management-related issuances.

  • Both the 2023 OECD Survey on Primary Market Developments and the 2023 OECD Survey on Liquidity in Secondary Government Bond Markets were carried out in October 2023.

  • Inflation-linked securities are instruments with coupon and/or principal payments which are linked to an inflation index. The data 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 SONIA or Euribor. For example, medium and long-term floating rate notes (FRNs, colloquially known 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 (ATM) figures follow the same coverage described at the beginning of this Annex, with country specific methods detailed below:

    • 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: Liabilities under the Support to mitigate Unemployment Risks in an Emergency and Next Generation EU programmes (which would raise the figure at 7.3) are excluded.

    • Japan: The Ministry of Finance announces ATM, based on Fiscal Year, not Calendar Year. Figures from 2006 to 2022 exclude saving bonds. The figure for 2023 is estimated and includes saving bonds.

    • Netherlands: Numbers are based on outstanding T-Bills and bonds and do not include outstanding commercial paper.

    • New Zealand: Figures include only marketable securities (excluding non-marketable securities held by the Reserve Bank of New Zealand and the Earthquake Commission). However, it does include securities held by the RBNZ that were purchased under the large asset purchase program and government bond repurchase.

    • Sweden: End of year figures. Time to maturity (uplifted amount at current exchange rate) in years. Includes Government bonds; Inflation-linked bonds; Public bonds, foreign currencies; Green bonds; and T-bills.

    • 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.

This chapter adopts a variation of the methodology outlined by Escolano (2010[55]) and uses the equation below to capture the change in debt-to-GDP ratio between time t and t-1:

d-dt-1= ip1+y* dt-1- π1+y*dt-1- g1+g*dt-1- pb+sf


  • d is the central government debt stock from the Borrowing Survey expressed as a ratio to GDP from OECD countries (2023[1])

  • ip is the effective interest rate, expressed as general government gross interest expenses from OECD (2023[1]) in time t as a ratio of d in time t-1

  • y is the nominal GDP growth rate from OECD (2023[1]) between time t and t-1

  • π is inflation, defined as the change in GDP deflator from OECD (2023[1]) between time t and t-1

  • g is the real GDP growth rate from OECD (2023[1]) between time t and t-1

  • pb is the general government's primary balance as a ratio of GDP, both from OECD (2023[1])

  • sf denotes the stock-flow adjustments

For the OECD area decomposition, this chapter uses the equation above with the following variables:

  • d is the sum of each country's central government debt stock divided by the sum of each country’s GDP

  • ip is the sum of each country's general government gross interest payments divided by the sum of each country’s d at time t-1

  • y is the growth rate of the sum of each country’s nominal GDP

  • π is the sum of each country’s π, weighted by GDP

  • g is the sum of each country’s g, weighted by GDP

  • pb is the sum of each country’s general government primary balance, divided by the sum of each country’s GDP

  • sf denotes the stock-flow adjustments

Country figures were converted into US Dollars using the year-end exchange rates, sourced from LSEG.

The debt-to-GDP growth rate encompassed all OECD countries, while factor calculations were limited to countries with available data. For example, interest expense computations excluded CHL, COL, CRI, ISR, MEX, and TUR, representing around 3% of OECD central government marketable debt. Stock-flow adjustments absorbed discrepancies arising from the concurrent use of central and general government data, as well as from data gaps.

Effective Interest Rates (EIR) represent the ratio of interest payments to debt stock and are estimated under accrual accounting using a combination of data sources. Annual inflation and policy rate data up to 2022 and projections for 2023 are sourced from the OECD Economic Outlook (OECD, 2023[1]); bonds’ annual yield to maturity at issuance (YTM) and outstanding amounts as well as foreign exchange (FX) rates are sourced from LSEG; and outstanding amount of debt by country and by type of instrument are sourced from the Borrowing Outlook Survey. The annual interest rates, inflation rates, and YTM are converted to daily values to account for the fact that bonds are issued and redeemed throughout the year. This approach ensures that interest payments are accurately apportioned to reflect the actual period during which the bonds remain outstanding.

Fixed Rate and Short-Term Instruments


Variable Rate Instruments

EIRiy= d=0DySid×YTMiDy+Fid d=0DySid

Inflation-linked Instruments



EIRty: Effective Interest Rate for debt instrument i for year y.

YTMi : Annual Yield to Maturity at issuance of debt instrument i.

Sid: Outstanding stock of debt instrument i in day d.

Dy: Number of days in the year y.

IY: Inflation rate for year y.

Fid: Daily floating interest rate for variable rate debt instrument.

The debt stock experiences intra-year variations owing to occurrences such as bond maturities and buybacks. In this context, each bond issuance, inclusive of re-openings, is considered a distinct unit until there is a discernible decrease in the outstanding amount of the bond line. When a reduction in any specific bond line is observed, its effect is systematically apportioned across both the original issuance and any subsequent re-openings in the day of the event. This allocation is executed in proportion to their respective outstanding amounts at that time.



Sb,id: Outstanding stock of debt instrument i in the day d of the event for instrument i from bond line b.

Rbd: Decline amount in the outstanding amount for the bond line b in the day of the event d.

The instrument-level EIR is estimated using a stock-weighted average method. This approach involves multiplying the annual effective yield of each debt instrument by its daily average outstanding stock in the year. Subsequently, for each country and type of instrument, the yields are aggregated and then divided by the debt stock of each instrument sourced from the Borrowing Outlook Survey.

EIRt,cy=i,dYTMi,t,cy×Si,t,cd,y i,dSi,t,cd,y ; EIRcy=tEIRt,cy× ϕt,cytϕt,cy ; EIRty=t,cEIRt,cy× ϕt,cy× Xcyt,cϕt,cy×Xcy ; Wty=EIRty× cϕt,cy× Xcyt,cEIRt,cy×ϕt,cy× Xcy


EIRt,cy: Weighted average Effective Interest Rate for the instrument type t in year y and country c.

EIRcy: Weighted average Effective Interest Rate for the country c in year y.

EIRty: Weighted average Effective Interest Rate for the instrument type t in year y for the OECD area.

ϕt,cy: Debt stock for the instrument t in country c in year y.

Xcy: Foreign exchange rate between country c's domestic currency and USD on the last day of year y.

Wty: Portion of the interest rates attributed to the instrument type t

Primary sovereign bond market data are based on original OECD calculations using data obtained from LSEG that provides international security-level data on new issues of sovereign bonds. The data set covers bonds issued by emerging market sovereigns in the period from 1 January 2007 to 31 December 2023 and includes both short-term and long-term debt. Short-term debt (“bills”) is defined as any security with a maturity less than or equal to 365 days but no less than 30 days, as bill issuances with a maturity less than 30 days are considered to be done for cash management purposes and excluded from calculations. The data provides a detailed set of information for each bond issue, including the proceeds, maturity date, interest rate and currency structure.

The definition of emerging markets used in this report is consistent with the IMF’s classification of Emerging Markets and Developing Economies used in its World Economic Outlook. The regional definitions are also those used by the IMF, while the income categories used (high income, low income, lower middle income, upper middle income) are defined by the World Bank according to GNI per capita levels.

A number of bonds have been subject to reopening. For these bonds, the initial data only provide the total amount (original issuance plus reopening). To retrieve the issuance amount for such reopened bonds, specific data on the outstanding amount on each reopening date for the concerned bonds have been downloaded separately from LSEG. As the reopening data only provide amounts outstanding, the outstanding amount on the previous date is subtracted from the outstanding amount on that given date, in order to obtain the issuance amount on each relevant date. These calculated issuance amounts are converted on the transaction date using USD foreign exchange data from LSEG. To ensure consistency and comparability, the same method is used for all bonds, including those which have not been subject to reopening.

Exchange offers and certain bonds in the dataset have been manually excluded when they did not have any identifier (ISIN, RIC or CUSIP) and when they have not been able to be manually confirmed by comparing with official government data.


← 1. Fiscal deficits denote government fiscal funding needs, and NBR represents market borrowing beyond refinancing. Discrepancies between NBR and government fiscal deficits may arise when governments finance their fiscal needs through means other than issuing debt. Such alternative methods encompass the utilisation of existing cash reserves, asset sales (such as real estate and state-owned enterprises), or the acquisition of different liabilities, including accruing arrears or obtaining loans. Overborrowing of USD 1.6 trillion in 2020, driven by forecasting uncertainties from the pandemic, resulted in abnormal cash accumulation. However, in 2021 and occasionally in 2022, excess cash was used to smooth debt issuances in times of rising yields.

← 2. As debt values are converted to USD, a stronger dollar magnifies the United States' debt size with the currency effect accounting for 40% of the growth in the country’s share.

← 3. Quantitative tightening is the process where central banks reduce the size of their balance sheets by selling off bonds or by not fully reinvesting the proceeds of maturing bonds, effectively withdrawing liquidity from the financial system and increasing the market holdings of bonds.

← 4. According to the OECD 2023 Survey on Liquidity in Government Bond Secondary Markets the demand for linkers declined in New Zealand and the United Kingdom and rose in Chile, Hungary, Iceland, Mexico, Portugal, Spain and Türkiye.

← 5. According to the OECD 2023 Survey on Primary Market Developments, these are Australia, Costa Rica, Hungary, Latvia, Lithuania, Mexico, Netherlands, New Zealand, Portugal, Slovenia, Sweden and Türkiye.

← 6. This represents the average across the 38 OECD member countries. The OECD aggregate is predominantly shaped by major issuers, particularly G7 countries, potentially obscuring unique circumstances in the 38 member countries. The simple average treats all countries uniformly.

← 7. This approximation likely underestimates the short-term impact of inflation on debt levels. Inflation also increases tax revenues in the short-term, thereby reducing net borrowing requirements. This partially explains improvements in net borrowing requirements which coincided with a rise in inflation in the period of 2021-23 (refer to Figure ‎1.1A).

← 8. Forecasts for the 2022-24 period indicate entirely positive real yield curves exclusively in Canada, Denmark, Norway, and the United States.

← 9. The pandemic's impact on EU public finances has intensified discussions on reforming EU fiscal rules. The European Commission, after a public consultation, published orientations for reforming these rules, emphasising the role of debt sustainability analysis (DSA) in assessing fiscal risks (Blanchard, Leandro and Zettelmeyer, 2021[68]; European Parliament, 2023[60]). The Commission's proposed DSA-based fiscal adjustment paths, covering at least four years, aim to ensure a downward public debt trajectory and compliance with the 3% of GDP fiscal deficit limit.

← 10. Breakeven inflation is defined as the difference between the yield on a zero-coupon default-free nominal bond and on a zero-coupon default-free inflation-linked bond of the same maturity.

← 11. The likelihood of this scenario decreases with declining inflation, as central banks may opt to reduce policy rates sooner.

← 12. The OECD currently forecasts a soft landing for its member countries (OECD, 2023[1]; OECD, 2024[16]).A soft landing in a monetary tightening cycle refers to the successful slowing of economic growth by central banks to curb inflation without triggering a recession.

← 13. Not all OECD countries are covered in this figure due to data availability in the OECD Economic Outlook (OECD, 2023[1]).

← 14. Accrual accounting, employed in the OECD National Accounts and by the majority of OECD countries, records interest expenses as they accumulate over time, irrespective of the actual timing of coupon and principal repayments. Consequently, the effective interest rate for a particular year is calculated as a weighted average of the yield to maturity, adjusted for any associated indices or rates for index-linked or floating rate debts, of all outstanding debt.

← 15. Inflation typically increases nominal tax bases, thereby boosting tax revenues (Garcia-Macia, 2023[67]). This increase often counterbalances the inflationary effects on the interest costs of inflation-linked bonds.

← 16. According to the OECD 2023 Survey on Primary Market Developments, 65% of OECD sovereign issuers with an inflation-linked programme reported the programme was responsible for an increase in their debt portfolio's maturity.

← 17. In fact, some linkers’ programmes originated in the 1980s, serving as an important method for market funding during a period when central banks didn’t have operational independence to conduct monetary policy.

← 18. This figure more directly reflects potential increases in interest payments, differing from the proportion of debt maturing or being refixed. Most of the debt due has already been refixed at higher rates, indicating the probability of lower future borrowing costs (assuming rates fall). Additionally, short-term instruments have been refinanced at these higher rates. Lastly, countries with significant debt refixing or refinancing in the next three years but low debt-to-GDP ratios face limited exposure to increased borrowing costs, like Poland with a 30% debt-to-GDP ratio and nearly half its debt maturing or due to be refixed by the end of 2026.

← 19. This calculation presupposes that the YTM of fixed-rate issues in 2024-26 will reflect their 2023 YTM.

← 20. Accounting practices for securities held for monetary policy purposes vary among central banks (OECD, 2023[26]). The US Federal Reserve and the BoJ account for these securities at historic cost accounting, delaying the impact of market valuation changes until the securities are sold. These banks only report unrealised valuation changes for transparency purposes. The ECB has a flexible accounting principle, allowing the country’s central banks to choose between amortised cost and current market price valuation – most national central banks in the euro area opt for amortised cost. Other central banks, including those in Australia, Canada, New Zealand, Sweden, Switzerland, and the United Kingdom, apply mark-to-market accounting to their monetary policy securities. Central banks not using mark-to-market accounting for their holdings may still face losses when actively selling securities purchased at low rates during a period of comparably higher rates.

← 21. For instance, when incurring losses, the Fed accumulates them in the form of a "deferred asset". This asset’s value is deducted when the Fed has positive net income. At one point, this asset value is cleared, and the Fed recommences remittances to the Treasury (Castro and Jordan-Wood, 2023[31]).

← 22. This selection excludes countries where the central bank's holdings in 2022 were less than 10% of the national debt and which did not respond to the relevant question in the OECD 2022 and 2023 Survey on Liquidity in Government Bond Secondary Markets. The specific question asked was: "Please indicate the share of the central bank's holdings of national government debt in total, distinguishing between your central bank's holdings and those of foreign central banks as of 30 June 2023 (2022)".

← 23. The central banks of Australia, Hungary and Israel are also adopting a passive QT approach (RBA, 2022[58]; Lybek, 2023[59]; OECD, 2023[34]; BIS, 2023[35]). While central banks in New Zealand and Sweden are actively selling bonds as part of their QT programmes. (RBNZ, 2022[56]; SRB, 2022[57]). The Bank of Korea did not engage in QT due to the small size of its now discontinued QE programme (BIS, 2023[35]).

← 24. Where central banks are actively selling bonds as part of their QT programme, this directly increases the level of debt and duration for the market to absorb. Where QT is passive, central banks are no longer taking debt and duration out of the market via reinvestments, meaning the net level of debt and duration for the market to absorb will increase assuming the same issuance profile by sovereigns.

← 25. The ATM of the government securities sold by the BoE under QT was nearly 20 years.

← 26. Foreign exchange reserves are mainly allocated to government securities and overnight or term deposits at the relevant central bank.

← 27. The term "price insensitive seller" reflects that the central bank sells government bonds not based on yield considerations, but to uphold its policy mandate of price and financial stability. However, the sale of these bonds is not entirely insensitive to the price offered. For example, the BoE conducts bond auctions with predetermined minimum bid prices and may choose to decrease the supply in these auctions if there is notably weak demand, with such reductions being deferred to future auctions, ensuring the BoE achieves its intended reduction in the bond stock (Bank of England, 2022[66]).

← 28. According to the OECD 2023 Survey on Liquidity in Government Bond Secondary Markets, this includes Austria, Canada, Finland, Iceland, Ireland, Israel, Italy, Japan, Korea, Latvia, Lithuania, Portugal, Slovakia, Switzerland and the United Kingdom.

← 29. Hedge funds' interaction with government bond markets varies significantly across countries (FSB, 2022[49]). For instance, in 2020, US hedge funds were heavily involved in the cash-futures basis trade, leveraging positions to profit from small price differences. However, market volatility led to mark-to-market losses and fire sales. Conversely, in the United Kingdom and euro area, hedge funds focused more on relative value trades like on-the-run/off-the-run arbitrage.

← 30. According to findings from the OECD 2023 Survey on Liquidity in Government Bond Secondary Markets.

← 31. The latest SBO edition examines the UK’s mini-budget crisis (OECD, 2023[24]), and the Financial Stability Board (FSB) provides an in-depth analysis of the 2020 government bond fire sale (FSB, 2022[49]).

← 32. The increased presence of the non-banking financial sector in the government bond market has prompted action from international organisations, financial regulators, and central banks. In 2023, the US updated its MMF regulations to improve liquidity and reporting management (US SEC, 2023[61]). The European Commission, likely to follow suit, has identified the need for further adjustments in its evaluation of the MMF Regulation (EC, 2023[62]). The UK is advancing its reform agenda, as evidenced by a recent public consultation on MMFs (FCA, 2022[63]). Concurrently, the FSB is conducting an extensive review of reforms in the global non-banking financial intermediation sector to ensure enhanced financial stability (FSB, 2023[64]; FSB, 2022[65]).

← 33. According to the OECD 2023 Survey on Liquidity in Government Bond Secondary Markets, this group comprises Canada, Denmark, Ireland, Netherlands, Norway, and Portugal.

← 34. The bid-to-cover ratio is calculated by dividing the total amount of bids received in a bond auction by the amount of bonds being offered.

← 35. Sovereign issuers engage in borrowing not only through auctions but also via syndications to end investors and tender offers. The volumes transacted through these methods vary considerably between countries, most notably between smaller and larger issuers.

← 36. In the United States, Principal Trading Firms (PTFs) are also active in inter-dealer markets. They trade with their own capital, primarily engaging in proprietary, electronic, and algorithmic trading. They act as intermediaries in markets dealing with government bonds, derivatives, and other securities, playing a crucial role in liquidity provision and price discovery.

← 37. According to the OECD 2023 Survey on Liquidity in Government Bond Secondary Markets, the number of primary dealers ranges from 4 in Norway to 32 in Germany, averaging 12-13. Only 13 countries currently include non-bank entities as dealers.

← 38. A US Fed’s study showed that when dealer capacity utilisation is approximately 20%, there is negligible estimated effect on market illiquidity from further capacity increases, but as capacity utilization escalates from 40% towards 80%, the market's estimated illiquidity surges by about three standard deviations above the level anticipated based on volatility alone (Duffie et al., 2023[47]).

← 39. According to the OECD 2023 Survey on Liquidity in Government Bond Secondary Markets, pprimary dealers (PD) are selected based on the following criteria: maintaining a competitive environment (in 22 countries), secondary market activity (22), distribution capacity of PDs (21), PD system stability (17), geographical distribution (13), size of the borrowing requirement (13), and warehousing capacity (8).

← 40. A tap issuance or reopening is a method by which a sovereign issuer sells additional amounts of a previously issued bond, increasing its supply without creating a new maturity or coupon rate. By adding volume to an existing issue, especially off-the-run bonds, tap issuances enhance market liquidity, facilitating easier trading and price discovery for those securities.

← 41. The US Treasury's planned buyback program, set to start in 2024, focuses on liquidity support and cash management, involving regular operations across nine security buckets with a quarterly maximum buyback amount of USD 30 billion for liquidity support and up to USD 120 billion for cash management in the first year, subject to market conditions and operational adjustments.


This work is published under the responsibility of the Secretary-General of the OECD. The opinions expressed and arguments employed herein do not necessarily reflect the official views of the Member countries of the OECD.

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.

The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

Note by the Republic of Türkiye
The information in this document with reference to “Cyprus” relates to the southern part of the Island. There is no single authority representing both Turkish and Greek Cypriot people on the Island. Türkiye recognises the Turkish Republic of Northern Cyprus (TRNC). Until a lasting and equitable solution is found within the context of the United Nations, Türkiye shall preserve its position concerning the “Cyprus issue”.

Note by all the European Union Member States of the OECD and the European Union
The Republic of Cyprus is recognised by all members of the United Nations with the exception of Türkiye. The information in this document relates to the area under the effective control of the Government of the Republic of Cyprus.

Photo credits: Cover © gyn9038/Getty Images.

Corrigenda to OECD publications may be found on line at: www.oecd.org/about/publishing/corrigenda.htm.

© OECD 2024

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.