Government debt: the price of getting out

Who holds government debt? · Episode 6 of 7

A fund’s sale meets a buyer, moves a price and can change the terms of public borrowing. Between private urgency and future coupons, timing determines the bill.

On 9 April 2025, the US Treasury found its buyers. In the middle of a bond-market shock, bids for a reopening of its ten-year note amounted to 2.67 times the quantity awarded, up from 2.59 a month earlier. Yet the yield at which the government borrowed had risen. The same security, due in February 2035, now offered a new buyer 4.435%, compared with 4.310% at the 12 March auction. Its coupon remained 4.625%. ^[1][2]

Those figures describe different parts of the transaction. Bid-to-cover measures the quantity of bids against awards. Yield describes the return available at the accepted price. The coupon determines contractual payments. An auction can attract more bids while becoming more expensive for the issuer. The two Treasury results sheets show exactly that combination. ^[1][2]

Earlier episodes followed dealers’ balance sheets, leveraged buyers and margin calls when markets move. A sale prompted by a margin call now leaves the seller’s balance sheet. It encounters another constraint: someone must agree to take the security, in a particular quantity and at a particular price.

Two delays separate that private decision from the government’s interest bill. Transactions must first affect the terms on which the state can raise money. Those terms must then enter its borrowing contracts. A disruption lasting a few hours and a rate increase lasting several years can leave very different fiscal footprints. ^[4][15]

How much can the next buyer take?

The outstanding stock of debt describes the size of a market. It says much less about how much a hurried seller can offload in the next few seconds. Daily turnover answers another question: how much trading actually took place, possibly including repeated transfers of the same security. Market depth describes the quantities offered for purchase or sale at particular prices. ^[3]

Consider two simplified order books, with quantities expressed in millions of euros of face value. An investor needs to sell 40 million. In the first book, buyers will take 20 million at a price of 100 and another 20 million at 99.80. The seller receives 39.96 million. In the second, the best bid still reads 100, but it is available for only 5 million. The remaining bids sit at 99.80, 99.50 and 99.20. The same sale now raises 39.78 million.

The difference is €180,000. Neither the portfolio nor the need to sell has changed. The available buying capacity has. The average execution price falls from 99.90 to 99.45 per 100 of face value. The final portion trades lower still, at 99.20. This is a hypothetical snapshot of one trading venue. Real markets are fragmented, quotes can be withdrawn, and new buyers can arrive. The model isolates a mechanism; it does not estimate an observed market impact.

The next trade takes place in a changed environment. Intermediaries have added risk to their inventories. Other participants have learned that a sizeable position is looking for an exit. Anticipating further sales, they may lower their bids or reduce the quantities quoted. Transaction prices can also feed into valuations, reduce the assessed value of other investors’ collateral and prompt additional selling decisions. How far that feedback travels depends on contracts and valuation practices, as well as the size of the initial order. ^[7][16]

Liquidity therefore has a time dimension. A portfolio may appeal to many investors over a month and to very few in the next twenty minutes. A dealer bridging that gap must fund the waiting period, carry price risk and have access to instruments that can hedge it. When all three become harder, a lower price may be required to attract the next available balance sheet. ^[7]

The coupon keeps its schedule

A fall in the price of an outstanding bond first creates a market loss for its holder. To connect that price with yield, take a hypothetical ten-year bond: €100 of principal, a €4 annual coupon and repayment of principal with the final coupon. When investors require a 4% annual yield, the price is 100. At a required yield of 5%, the present value of those payments falls to €92.28. ^[4]

Each of the ten coupons remains €4. Principal repayment remains €100. A buyer who pays less receives a higher yield on the same promised payments. The exact price change in this example is −7.72%. The calculation assumes annual coupons and a valuation date immediately after a coupon payment. A semiannual bond would require the appropriate different conventions.

The seller crystallises the loss by selling. An investor who holds on needs the resources to wait for the scheduled payments, subject to the issuer meeting its obligations. For the government, a resale between investors leaves the outstanding bond’s coupon unchanged. It does, however, create a new market reference when the government next seeks funding. ^[4][15]

A prospective buyer compares a new bond with securities already available. If similar bonds offer 5%, selling a new 4% bond at 100 requires some other attraction: a different maturity, greater liquidity, particular contractual features or another benefit. For a comparable instrument, the issuer will generally have to adjust the coupon or the price.

That comparison links the secondary market, where existing securities trade, to the primary market, where governments receive fresh money. It does not require every holder to sell. Trades involving a fraction of the stock can reveal a marginal price against which the next issue must compete. ^[4][15]

Finding the imprint of UK forced sales

The difficult part is establishing that constrained selling actually moved the price. Higher yields can also reflect stronger expected inflation, a different monetary-policy path or a reassessment of fiscal risk. Even the term premium, compensation for the risk of holding a longer bond rather than rolling short investments, has to be estimated. A displayed yield does not reveal the contribution of each cause. ^[9]

Britain’s autumn 2022 crisis offers a revealing case. In research published by the Bank of England in July 2024, Gabor Pinter, Emil Siriwardane and Danny Walker compare transactions in gilts of similar maturity at the same time, differentiated by their prior exposure to LDI funds. These funds invest to match pension obligations. The authors report relative price discounts of 6.87% on 27 September and 7.77% on 11 October for heavily exposed securities. These are estimates of selling pressure, conditional on the study’s identification method. ^[5]

The choice of comparison matters. Comparing a very long bond with a short one would mix selling pressure with their different sensitivity to interest rates. Looking at closer neighbours helps separate the mechanisms. It still requires the assumption that, without the sales, the groups would have followed sufficiently comparable paths. The researchers discuss and test the robustness of that assumption. ^[5]

The 6.87% figure cannot be translated into a 6.87-percentage-point increase in the UK government’s borrowing rate. It measures a relative price effect in a particular sample and specification. Converting it to yield would require each bond’s cash flows and maturity. Converting it to a fiscal cost would then require the issuance conducted during the disruption and a credible estimate of the price that issuance would have achieved without it.

That final distinction is essential to an investigation of public costs. A fund’s loss and an additional government expense have different obligors and different payment schedules. Adding them together would treat every private capital loss on a government bond as a public bill. A market loss can start a transmission process. Its size does not establish the eventual fiscal result.

A separate study by Pinter, published in March 2023, documents sharply higher transaction costs and greater dispersion in prices during the crisis, particularly for some smaller trades and intermediaries. It adds another dimension: exit terms depend on access to the market. An average price can conceal substantially worse executions for less well-placed sellers. ^[6]

Buyers have reasons to wait

A more attractive price can draw in an insurer, a cash-rich fund or an investor able to finance a position. That response takes time. A buyer must assess the risk, secure any required approvals and arrange settlement. It may also believe the next wave of selling will offer a better entry price. The prospect of a bargain tomorrow can delay the purchase that would stabilise the market today. This is a possible economic mechanism, not a motive attributed here to a named institution.

The Bank of England’s exploratory exercise, published on 29 November 2024, examines the other side of that wait. In its scenario, sales use 70% of banks’ appetite for temporarily warehousing gilt risk before requiring additional compensation through wider bid-ask spreads. The percentage refers to risk appetite in the exercise, not to the consumption of 70% of regulatory capital. ^[7]

The central scenario still has buyers to meet the selling. But its outcome depends on hedging conditions and on which securities are being offered. A bank that can neutralise some interest-rate risk is more willing to take a bond it plans to resell. If hedging markets also deteriorate, a cheaper bond may no longer be enough to persuade it to buy. ^[7]

The implication is to look at available buyers rather than simply potential ones. A large balance sheet, already committed to other risks or constrained by internal limits, may offer little extra capacity at the critical moment. A smaller investor with room to act can set the next trade’s price. The list of a sovereign’s largest creditors and the list of its buyers in a disruption answer different questions.

The US auction that still cleared

April 2025 provides a useful counterexample. The New York Fed’s analysis published on 2 April 2026 describes a sharp deterioration in Treasury liquidity after the tariff announcements, followed by a quick recovery. It examines the most recently issued two-, five- and ten-year notes on BrokerTec. Its sample ends on 27 February 2026, so it does not measure conditions in October. ^[3]

In his 9 May 2025 speech, Roberto Perli, manager of the Fed’s System Open Market Account (SOMA), described relatively stable repo funding and dealers that remained able to intermediate. According to the New York Fed Desk’s estimates, cash-futures basis trades did not experience the dislocation seen in March 2020. The Federal Reserve’s April 2025 Financial Stability Report likewise found that the market continued functioning without some of the severe strains of earlier episodes. ^[8][18]

Return to the February 2035 note. Awards excluding the Federal Reserve portfolio’s add-on were close to $39 billion at each auction. Bids rose from $100.948 billion in March to $103.944 billion in April. Bid-to-cover improved, while the Treasury accepted a yield 12.5 basis points higher, equivalent to 0.125 percentage point. ^[1][2]

The price excluding accrued interest fell from $102.513731 to $101.490492 per $100 of face value. On an identical $39 billion nominal amount, the price difference would reduce clean proceeds by about $399.1 million. That is a normalised comparison calculated for this article, not an estimate of damage caused by funds. Remaining maturity shortened by a month, and much other information changed between the auction dates. The actual awards differed slightly: $39,000,114,800 in March and $39,000,018,800 in April, excluding SOMA. ^[1][2]

Why did the coupon stay at 4.625% while the yield increased? Both auctions reopened the existing issue identified by CUSIP 91282CMM0. The additional securities inherited its payment schedule and maturity. The adjustment occurred through price: the premium above 100 shrank, leaving the Treasury with less cash per unit of principal it would eventually repay. Borrowing can become more expensive without a higher coupon appearing on the bond. ^[1][2]

The example challenges the common inference that rising yields demonstrate an absence of buyers. Buyers may be present precisely because the new price offers better compensation. Assessing funding conditions requires quantities, prices and the functioning of the markets that allow investors to hold the securities.

From market prices to the French interest bill

For France, Agence France Trésor reported a TEC10 of 4.79% on 9 October 2026. The index represents the yield on a hypothetical government bond with exactly ten years remaining. It differs from the weighted average yield on the year’s medium- and long-term issuance, reported at 3.55% as of 30 September, and differs again from the contractual cost of the outstanding debt portfolio. ^[10][11]

Market prices move first: an OAT, a French government bond, changes hands and its valuation moves. Financing follows: an old bond matures, a deficit needs funding or cash management requires fresh resources. The fiscal accounts record the charges: coupons, interest and other charges are recorded according to their payment dates and accounting treatment. ^[14][15]

France’s refinancing calendar illustrates the timing. French marketable central-government debt had an average remaining life of eight years and 158 days on 30 September 2026. That average describes how repayments are distributed. It does not provide eight years of protection before a rate shock first matters. Some securities mature much earlier, new financing needs arise immediately, and short-term borrowing has to be renewed more frequently. ^[10][15]

At the opposite extreme, instantly applying the latest ten-year yield to the entire debt stock would produce a fictional interest bill. Outstanding fixed-rate contracts retain their coupons. Sensitivity depends on redemptions, new funding requirements, the share of short-term instruments and securities whose payments can vary, including with inflation. ^[4][14][15]

In its 29 September release, AFT outlined €340 billion of medium- and long-term issuance, net of buybacks, for 2027. It noted that repayments at those maturities were expected to increase by €19.4 billion relative to 2026. Tomorrow’s borrowing therefore depends partly on yesterday’s contracts. This remains a financing plan. The €340 billion is neither the deficit alone nor simply the increase in the stock of debt. ^[12]

The same release reported bids amounting to 2.5 times supply at medium- and long-term auctions since the beginning of 2026. As in the US case, that ratio establishes substantial bidding at the terms offered. It does not set the price investors will accept at the next sale. ^[12]

A short disruption can leave a ten-year payment trail

The duration of a market disturbance and the duration of its consequences can differ. Consider another entirely hypothetical model. A government keeps its debt stock at €100 billion, refinancing exactly €10 billion at the beginning of each year with new ten-year bonds issued at par. The baseline rate is 4%. To isolate the effect, additional interest is financed without further borrowing and therefore does not increase the model’s debt stock.

Suppose the rate rises to 5% only when the first year’s refinancing occurs. That new tranche adds €100 million to annual coupon payments. Market rates then return to 4%. Subsequent issues regain the lower rate, but the first tranche keeps its coupon for ten years. The market disruption has ended; the contract preserves its imprint.

Now suppose the 5% rate persists through all ten refinancings. Each year adds another €100 million of annual cost. The extra annual bill reaches €500 million in year five and €1 billion in year ten. These figures describe the additional charge in each particular year. The cumulative ten-year totals are different: €1 billion for the temporary shock and €5.5 billion for the persistent one.

Timing is decisive. A disruption that ends before an issue is sold may leave little direct trace in that borrowing. The same disruption encountered when money has to be raised may be embedded in the price or coupon for years. Turning this mechanism into a historical cost estimate would require the dates, amounts and exact terms of actual issuance.

The model also explains why a subsequent market recovery does not automatically erase the bill. An issuer can consider buybacks or exchanges, but holders also value a coupon above prevailing rates. Buying that promise back has a price. Refinancing does not allow the state to retrieve earlier market conditions at no cost. ^[4][15]

The AFT’s 2025 sensitivity scenario

An AFT simulation reproduced in the Senate’s 2026 budget report, filed on 24 November 2025, gives a French scale to this gradual transmission. It assumes a permanent one-percentage-point increase across the entire yield curve relative to its reference scenario. The reported additional interest charge is €3.2 billion at one year, €23.5 billion at five years and €33.5 billion at nine years. ^[13]

The breakdown shows short-term debt responding quickly, while longer maturities become more important over time: €1.3 billion from short-term rates and €1.9 billion from medium- and long-term rates at the first horizon, compared with €2 billion and €31.5 billion at the last. The figure reproduces only the three published horizons. It does not manufacture the intervening years. ^[13]

This is a 2025 simulation. It describes a sustained whole-curve shock, not an intraday spike in the ten-year yield. Its figures are not additional expenses already committed as of 11 October 2026. The public passage does not provide all the conventions needed to reconcile them with a new budget forecast. Adding them to AFT’s projected €72.9 billion budgetary debt charge for 2027 would be unjustified: the documents represent different exercises. ^[12][13]

Issuance premiums and discounts add another timing distinction. A reopening can raise less cash today for the same face value repayable tomorrow. Accrual accounts spread premiums and discounts over the life of the borrowing, whereas budgetary reporting follows a different treatment. Any cost comparison must state its accounting basis, even when the story begins with a simple movement in yields. ^[14]

Buying time before borrowing again

A debt manager has options: distribute issuance over time, use cash holdings, adjust maturities and diversify buyers. The IMF–World Bank debt-management guidelines emphasise the trade-off between medium- to long-term cost, refinancing risk and liquid resources available during disruption. ^[15]

Longer borrowing can reduce how frequently a particular tranche returns to the market. Depending on the yield curve and investor demand, that protection may require more compensation at the outset. Shorter borrowing may be easier to place at a given moment but exposes the issuer sooner to future rates. A cash buffer provides flexibility and can carry a cost. No choice simultaneously removes every constraint. ^[15]

Market functioning and fiscal fundamentals interact while remaining analytically distinct. A worrying budget outlook can legitimately change the price of risk. Synchronous sales can add temporary pressure. An ill-suited issuance strategy can lock in unfavourable terms. Assigning an additional cost requires following each step rather than reading a single explanation into the ten-year yield.

On 6 October 2026, the IMF reiterated that hedge funds contribute to liquidity and price discovery in normal conditions while potentially amplifying stress through leverage and crowded positions. The industry association MFA responded that uniform restrictions could reduce participation and raise governments’ funding costs. That second claim is the sector’s policy position, not an independent measurement of the cost of a reform. ^[16][17]

The underlying trade-off concerns the resilience of demand. A leveraged buyer may improve a market today and have to sell tomorrow. Requiring more stable resources can reduce some fragility while changing the price at which that buyer participates. Comparing the effects requires an assessment of the specific rule, its implementation and possible substitutes. An intermediary’s usefulness does not, by itself, settle the question of how it behaves under stress.

Three propositions can ultimately be established separately: a sale encountered less available capacity; the marginal price moved; and new terms entered a government issue. The UK research supports the transmission from selling pressure into prices. The US auctions show funding continuing at a different yield. The French repayment calendar explains how that yield gradually enters the public accounts. Estimating the precise causal fiscal cost of an episode requires linking those propositions in the same market at the same dates.

When private buyers take longer to arrive than the state and its creditors can wait, another balance sheet may enter the trade: the central bank’s. The final episode will examine what it buys, which risks it accepts and how intervention redistributes the time available to everyone else.

Method and limitations

Sources checked on 12 October 2026. l0g models are hypothetical; UK findings are historical estimates or an exploratory scenario. No confidential transaction records or interviews were obtained. The references below allow readers to trace the observations and conventions used.

Sources and documents

1. U.S. Treasury · Treasury auction results · 9 April 2025Published / updated : 9 April 2025. Period / scope : Auction: 9 April 2025; settlement: 15 April.https://www.treasurydirect.gov/instit/annceresult/press/preanre/2025/R_20250409_2.pdf
2. U.S. Treasury · Treasury auction results · 12 March 2025Published / updated : 12 March 2025. Period / scope : Auction: 12 March 2025; settlement: 17 March.https://www.treasurydirect.gov/instit/annceresult/press/preanre/2025/R_20250312_2.pdf
3. Federal Reserve Bank of New York · Treasury Market Liquidity Since April 2025Published / updated : 2 April 2026. Period / scope : 2 January 2020–27 February 2026; focus on April 2025.https://libertystreeteconomics.newyorkfed.org/2026/04/treasury-market-liquidity-since-april-2025/
4. SEC · Investor.gov · Fixed Income Investments: When Interest Rates Go Up, Prices of Fixed-Rate Bonds FallPublished / updated : 26 June 2013. Period / scope : General fixed-rate bond mechanics.https://www.investor.gov/introduction-investing/general-resources/news-alerts/alerts-bulletins/investor-bulletins-86
5. Bank of England · Staff Working Paper 1089 · Fire sales of safe assets · Gabor Pinter, Emil Siriwardane, Danny WalkerPublished / updated : 26 July 2024. Period / scope : UK transactions, September–October 2022.https://www.bankofengland.co.uk/-/media/boe/files/working-paper/2024/fire-sales-of-safe-assets.pdf
6. Bank of England · Staff Working Paper 1019 · An anatomy of the 2022 gilt market crisis · Gabor PinterPublished / updated : 31 March 2023. Period / scope : UK gilt crisis, September–October 2022.https://www.bankofengland.co.uk/working-paper/2023/an-anatomy-of-the-2022-gilt-market-crisis
7. Bank of England · System-wide exploratory scenario exercise: final reportPublished / updated : 29 November 2024. Period / scope : Hypothetical exploratory stress scenario over two weeks.https://www.bankofengland.co.uk/financial-stability/boe-system-wide-exploratory-scenario-exercise/boe-swes-exercise-final-report
8. Federal Reserve Bank of New York · Recent Developments in Treasury Market Liquidity and Funding Conditions · Roberto PerliPublished / updated : 9 May 2025. Period / scope : April 2025 episode.https://www.newyorkfed.org/newsevents/speeches/2025/per250509
9. Federal Reserve Bank of New York · Treasury Term Premia: 1961–PresentPublished / updated : 12 May 2014. Period / scope : Analytical framework; not a current term-premium reading.https://libertystreeteconomics.newyorkfed.org/2014/05/treasury-term-premia-1961-present/
10. Agence France Trésor · Chiffres clefs · dette négociable et maturitéPublished / updated : no single date established on the page. Period / scope : Debt stock and maturity: 30 September 2026; TEC10: 9 October 2026.https://www.aft.gouv.fr/fr
11. Agence France Trésor · L’OAT TEC 10 · définition de l’indicePublished / updated : no single date established on the page. Period / scope : Definition of the constant-maturity ten-year index.https://www.aft.gouv.fr/fr/oat-tec-10
12. Agence France Trésor · Besoins et ressources de financement de l’État en 2027 et point sur 2026Published / updated : 29 September 2026. Period / scope : 2027 financing plans and late-September 2026 update.https://www.aft.gouv.fr/fr/publications/communiques-presse/29092026-besoins-et-ressources-financement-letat-en-2027-et-point
13. Sénat · commission des finances, d’après AFT · PLF 2026 · Engagements financiers de l’État, rapport n°139, annexe 12Published / updated : 24 November 2025. Period / scope : AFT conventional simulation published in 2025; 1-, 5- and 9-year horizons.https://www.senat.fr/rap/l25-139-312/l25-139-312_mono.html
14. Agence France Trésor · Principaux chiffres · charge de la dettePublished / updated : 21 November 2025. Period / scope : 2019–2024 outturns; last displayed update.https://www.aft.gouv.fr/fr/principaux-chiffres-budget-etat
15. IMF / World Bank · Revised Guidelines for Public Debt ManagementPublished / updated : 1 April 2014. Period / scope : Debt-management principles, 2014 revision. Cover of the official document: 1 April 2014.https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/np/pp/eng/2014/_040114.pdf
16. FMI · Hedge Funds Improve Market Functioning, but Can Also Amplify StressPublished / updated : 6 October 2026. Period / scope : October 2026 analysis; historical datasets vary.https://www.imf.org/en/blogs/articles/2026/10/06/hedge-funds-improve-market-functioning-but-can-also-amplify-stress
17. Managed Funds Association · Statement on IMF report on hedge funds and financial stabilityPublished / updated : 6 October 2026. Period / scope : Industry statement, 6 October 2026.https://www.mfaalts.org/statement/mfa-statement-on-imf-report-on-hedge-funds-and-financial-stability/
18. Federal Reserve Board · Financial Stability Report, April 2025 · Asset ValuationsPublished / updated : no single date established on the page. Period / scope : April 2025 report; early-April market episode.https://www.federalreserve.gov/publications/April-2025-financial-stability-report-Asset-Valuations.htm