Long bonds stop listening to the central bank
The US 30-year hit 5.56% and the French OAT outyielded the US Treasury: when the central bank buys and the long end ignores it, the problem is no longer about…
English · Original discussion in Spanish · Published
The day the central bank buys and the bond ignores it
The US thirty-year closed on 28 September at 5.56% on the Treasury par curve, its highest since the turn of the century, twelve days after the Federal Reserve raised rates for the first time since 2023. That same Monday, a few hours apart, the French ten-year stood at 4.72% on the official index of the Banco de Francia (Bank of France), territory last seen in 2008, with the spread against the Bund hovering around 110 basis points. Two countries, two central banks and a question nobody had been asking out loud until now: at what point does the buyer of last resort buy while the long end shuts it out?
Since 2008 it has been taken for granted that intervention always works. 2008 confirmed it, 2012 —where the promise alone was enough—, 2020 and the 2022 scare over British gilts. The script was identical: the market seizes up, the central bank buys, the price recovers. The suspicion is that this script carries a validity condition that is almost never spelled out.
Illiquidity or insolvency: two accidents that look the same
In a financial accident, someone needs liquidity. A leveraged fund loaded with the very bond, a market maker without balance sheet, an institution with duration losses watching deposits walk out: all sell at once and there is no buyer on the other side. There the central bank is the textbook solution, because what is missing is balance sheet, and balance sheet is the only thing it can manufacture without limit.
In a fiscal accident, what is in doubt is the issuer. The investor sells because they no longer believe the deficit trajectory, or because they suspect the way to square it will be to repay them in a currency worth less. And here buying can make things worse: if the central bank prints to prop up the bond price, it confirms the very suspicion that triggered the selling. The real rate can fall, but at the cost of higher implied inflation or term premium.
The underlying distinction is the old one between illiquidity and insolvency, applied to the issuer whose bond serves as collateral for everything else. Walter Bagehot wrote down in 1873 the rule every central bank claims to trinc: lend without limit, at a penalty rate and against good collateral. It works because the problem lies with the borrower, not the guarantee. In a fiscal accident the guarantee is the government bond itself; the central bank would have to lend against collateral whose quality is precisely what the market is questioning. Bagehot takes for granted the one thing missing here.
The convenience yield that evaporates
The yield on a long bond breaks down into what the market expects from official rates and the term premium. There is a third component the literature has measured for years and which rarely appears in the usual debate: the convenience premium, what the investor accepts to forgo in exchange for holding the safest and most liquid asset in the system. The issuer can charge it as long as there is no substitute. When investors find alternatives and stop overpaying, the bond loses part of its safe-haven treatment. April 2025 was a partial rehearsal: the Bund and gold acted as havens while the US bond fell.
France: a spread that opens up in two sessions
Using the series of the Banco de Francia's TEC10 index minus the Bund, the spread closed at 111 points on 29 September, 117 on the 30th and 130 on 1 October, the day the budget was presented; screens showed 137 intraday. The breakdown matters: on the 30th the OAT barely moved (4.74 to 4.75) and it was the Bund that fell, from 3.63 to 3.58 — money seeking Germany rather than people fleeing France. On the 1st it changed: the French ten-year rose 15 points to 4.90 with the Bund almost flat at 3.60. That is outright selling of French paper.
On 7 October the ten-year closed fourteen points lower, at 4.72, after Marine Le Pen raised her target to 140,000 million in net savings by 2032 and added that, once France regains control of its accounts, it will be necessary to talk to the ECB so that it helps lower what the country pays to borrow. The governor of the Banco de Francia, Emmanuel Moulin, responded on France Inter. Then came Christine Lagarde, at a CNN event in Bangkok: "there are particular circumstances that make some countries more affected, and France is an example. For that case, the first and main solution lies at home, in France".
The gradient of contagion, with Teleborsa closes and the ten-year spread against the Bund: France went from 136.2 on Friday to 147.5 on Monday and 151.4 on Tuesday; Italy from 106.9 to 111.9 and 112.6; Spain from 60.0 to 62.2 and 63.3. In two sessions France widens 15 points, Italy less than 6 and Spain just over 3. It is a French shock being transmitted only halfway.
The TPI and its fine print
On Thursday the 29th the ECB talked and did not buy. The scenario of a TPI (the ECB's Transmission Protection Instrument) for everyone but France is no longer an oddity: it is shared by JP Morgan, UBS, Capital Economics, ING and Robin Brooks. JP Morgan also describes the cheap version, the one that costs not a single euro: the ECB could "signal the possibility of activating the TPI for countries affected by contagion, even if current repercussions remain too small to justify actual intervention". A wall of words at the press conference.
The serious obstacle for France is not being under the excessive deficit procedure —the Commission concluded on 3 June that effective action had been taken, and the procedure remains open but suspended— but the other condition: halting purchases if the tension responds to the country's fundamentals. And a rise explained by the public accounts is exactly that.
Borrowing in a currency you do not issue
On Thursday, a single session showed what the argument had already spelled out on the 1st. The French premium closed at 175 points, after touching 177.7 intraday, its highest since November 2011, and the OAT came to pay more than the US ten-year Treasury, something that had not peine since 2013. All of this with the United States showing worse numbers: a deficit the CBO projects at 5.8% of GDP and a president who has said out loud that inflation will pay off the debt.
The underlying explanation was given by Paul De Grauwe in 2011, at the height of the euro crisis. A eurozone country borrows in a currency it does not issue; when investors get scared, nobody guarantees there will be a buyer of last resort, and a liquidity crisis can turn into something else.
On 28 September, at the White House, Trump told Time: "Certain levels of inflation will also pay off that debt very rapidly". Two weeks later, the Nobel Prize in Economics went to Robert Hall and Michael Woodford, and the scientific report contains a paragraph describing the same mechanism in model language: the price level rises until it devalues in real terms the government's nominal debt.
How much does the long end have to rise while the central bank is buying before we stop calling it a scare and start calling it something else?
The US thirty-year closed on 28 September at 5.56% on the Treasury par curve, its highest since the turn of the century, twelve days after the Federal Reserve raised rates for the first time since 2023. That same Monday, a few hours apart, the French ten-year stood at 4.72% on the official index of the Banco de Francia (Bank of France), territory last seen in 2008, with the spread against the Bund hovering around 110 basis points. Two countries, two central banks and a question nobody had been asking out loud until now: at what point does the buyer of last resort buy while the long end shuts it out?
Since 2008 it has been taken for granted that intervention always works. 2008 confirmed it, 2012 —where the promise alone was enough—, 2020 and the 2022 scare over British gilts. The script was identical: the market seizes up, the central bank buys, the price recovers. The suspicion is that this script carries a validity condition that is almost never spelled out.
Illiquidity or insolvency: two accidents that look the same
In a financial accident, someone needs liquidity. A leveraged fund loaded with the very bond, a market maker without balance sheet, an institution with duration losses watching deposits walk out: all sell at once and there is no buyer on the other side. There the central bank is the textbook solution, because what is missing is balance sheet, and balance sheet is the only thing it can manufacture without limit.
In a fiscal accident, what is in doubt is the issuer. The investor sells because they no longer believe the deficit trajectory, or because they suspect the way to square it will be to repay them in a currency worth less. And here buying can make things worse: if the central bank prints to prop up the bond price, it confirms the very suspicion that triggered the selling. The real rate can fall, but at the cost of higher implied inflation or term premium.
The underlying distinction is the old one between illiquidity and insolvency, applied to the issuer whose bond serves as collateral for everything else. Walter Bagehot wrote down in 1873 the rule every central bank claims to trinc: lend without limit, at a penalty rate and against good collateral. It works because the problem lies with the borrower, not the guarantee. In a fiscal accident the guarantee is the government bond itself; the central bank would have to lend against collateral whose quality is precisely what the market is questioning. Bagehot takes for granted the one thing missing here.
The convenience yield that evaporates
The yield on a long bond breaks down into what the market expects from official rates and the term premium. There is a third component the literature has measured for years and which rarely appears in the usual debate: the convenience premium, what the investor accepts to forgo in exchange for holding the safest and most liquid asset in the system. The issuer can charge it as long as there is no substitute. When investors find alternatives and stop overpaying, the bond loses part of its safe-haven treatment. April 2025 was a partial rehearsal: the Bund and gold acted as havens while the US bond fell.
France: a spread that opens up in two sessions
Using the series of the Banco de Francia's TEC10 index minus the Bund, the spread closed at 111 points on 29 September, 117 on the 30th and 130 on 1 October, the day the budget was presented; screens showed 137 intraday. The breakdown matters: on the 30th the OAT barely moved (4.74 to 4.75) and it was the Bund that fell, from 3.63 to 3.58 — money seeking Germany rather than people fleeing France. On the 1st it changed: the French ten-year rose 15 points to 4.90 with the Bund almost flat at 3.60. That is outright selling of French paper.
On 7 October the ten-year closed fourteen points lower, at 4.72, after Marine Le Pen raised her target to 140,000 million in net savings by 2032 and added that, once France regains control of its accounts, it will be necessary to talk to the ECB so that it helps lower what the country pays to borrow. The governor of the Banco de Francia, Emmanuel Moulin, responded on France Inter. Then came Christine Lagarde, at a CNN event in Bangkok: "there are particular circumstances that make some countries more affected, and France is an example. For that case, the first and main solution lies at home, in France".
The gradient of contagion, with Teleborsa closes and the ten-year spread against the Bund: France went from 136.2 on Friday to 147.5 on Monday and 151.4 on Tuesday; Italy from 106.9 to 111.9 and 112.6; Spain from 60.0 to 62.2 and 63.3. In two sessions France widens 15 points, Italy less than 6 and Spain just over 3. It is a French shock being transmitted only halfway.
The TPI and its fine print
On Thursday the 29th the ECB talked and did not buy. The scenario of a TPI (the ECB's Transmission Protection Instrument) for everyone but France is no longer an oddity: it is shared by JP Morgan, UBS, Capital Economics, ING and Robin Brooks. JP Morgan also describes the cheap version, the one that costs not a single euro: the ECB could "signal the possibility of activating the TPI for countries affected by contagion, even if current repercussions remain too small to justify actual intervention". A wall of words at the press conference.
The serious obstacle for France is not being under the excessive deficit procedure —the Commission concluded on 3 June that effective action had been taken, and the procedure remains open but suspended— but the other condition: halting purchases if the tension responds to the country's fundamentals. And a rise explained by the public accounts is exactly that.
Borrowing in a currency you do not issue
On Thursday, a single session showed what the argument had already spelled out on the 1st. The French premium closed at 175 points, after touching 177.7 intraday, its highest since November 2011, and the OAT came to pay more than the US ten-year Treasury, something that had not peine since 2013. All of this with the United States showing worse numbers: a deficit the CBO projects at 5.8% of GDP and a president who has said out loud that inflation will pay off the debt.
The underlying explanation was given by Paul De Grauwe in 2011, at the height of the euro crisis. A eurozone country borrows in a currency it does not issue; when investors get scared, nobody guarantees there will be a buyer of last resort, and a liquidity crisis can turn into something else.
On 28 September, at the White House, Trump told Time: "Certain levels of inflation will also pay off that debt very rapidly". Two weeks later, the Nobel Prize in Economics went to Robert Hall and Michael Woodford, and the scientific report contains a paragraph describing the same mechanism in model language: the price level rises until it devalues in real terms the government's nominal debt.
How much does the long end have to rise while the central bank is buying before we stop calling it a scare and start calling it something else?
Also available in: Deutsch
Summary of a discussion on Burbuja.info - Foro de economía, actualidad y política., translated from Spanish and reviewed before publication. Read the full discussion (17 replies).
More summaries
-
French bond hits 5% and Japanese savings dry up: the fuse of the system
The French bond breaks the 5% barrier and Japanese savings stop financing the Western deficit: auctions replace monetary policy and the system loses its…
-
72-hour ultimatum to Sánchez remains unfulfilled after 487 days
The 72-hour ultimatum to Sánchez was ignored and the deadline renewed itself. The timeline of failed predictions of his downfall goes back to 2014.
-
Sánchez: alternation becomes 'regression' if Vox enters
Spain's PM Sánchez says that with Vox 'we are no longer talking about alternation, we are talking about regression'.
-
Queues form for the iPhone priced over €2,000
A post shows the queue at the Apple Store on Passeig de Gràcia at 6 a.m. to buy the iPhone Duo without a reservation. The price exceeds €2,000.
-
Maricarmen's landlords: 'She paid 199 euros, we pay a 1,300 mortgage'
An interview reveals the case of a tenant with an old rent of 199 euros against a 1,300 euro mortgage. The gap between legacy rents and market rates is under debate.
-
Remax agent says Spain's property market has stalled
A Remax agent reports that offices aren't being let, rental supply has fallen and sales are slowing due to high prices and uncertainty. Here's what's behind it.