What is driving the global rise in long-term interest rates?

Rising bond yields are likely down to stronger investment demand, more supply of long-duration debt and less absorption by yield-insensitive investors.

Long-term government bond yields have risen sharply in 2026, with United States long-term real rates close to their highest levels in two decades. Such rising yields could have significance well beyond the bond market. One interpretation is that investors are becoming less willing to finance US government debt, whether because of US fiscal policy, concern about Federal Reserve independence or a broader erosion of confidence in US institutions during the second administration of President Donald Trump.

Scrutiny of market functioning has been heightened by the US Treasury’s announcement in August that it would expand long-end liquidity-support buybacks. However, at the opposite extreme, Stephen Miran, the former Chair of the US Council of Economic Advisers, who was appointed to that position by Trump, argues that higher real rates are principally a sign of economic strength, reflecting stronger expected growth rather than deteriorating fiscal or monetary credibility.

Neither interpretation is satisfactory on its own. Four overlapping explanations have dominated the debate. They stress (1) strong demand for capital from fiscal borrowing and AI investment, (2) higher expected real policy rates, potentially reflecting stronger growth (Miran’s argument), (3) Treasuries potentially losing part of their traditional safe-asset premium and (4) growing safe-debt supply running up against increasingly scarce market making and balance-sheet capacity (Caballero, 2026a, 2026b).

If the higher yields principally reflect stronger productive investment and higher potential growth, the increase is relatively benign. If they instead reflect an unanchoring of inflation expectations, deteriorating fiscal credibility or a loss of confidence in the institutions underpinning US government debt, the implications are much more serious. A third possibility is less dramatic but still important: long-term capital has simply become more expensive as investment demand and public and private debt supply rise relative to the willingness of investors to absorb duration. The evidence points mainly at this third explanation.

The first problem with explanations centred on a breakdown in US institutions is that the rise in long-term rates is not uniquely American. US, German, French, Italian, United Kingdom and Japanese 30-year yields rose by 45-79 basis points in the six months to 28 August, with the US in the middle at 58bp. Relative to weekly history since 2006, the US increase has ranked well above average (Figure 1) but has still been less marked than Japan, the UK and France.

This repricing sits within a longer rise that began before the US war on Iran started in late February 2026. Three overlapping channels can lift long sovereign yields: inflation compensation; long-term real yields from shifts in saving, investment or debt supply; and sovereign-specific premia from credit, liquidity or lost convenience value – the liquidity, collateral and scarcity benefits that lower yields.

Since late February, expected policy rates and inflation compensation have risen, while real yields have also moved higher. Since September 2024, the rise has been concentrated more at the long end, consistent with stronger investment demand, greater long-dated debt supply and reduced absorption by relatively yield-insensitive investors. The same channels operate since both of those months, but with different weights.

Inflation and inflation risk

Breakeven inflation – the gap between nominal and inflation-linked bond yields – combines expected inflation with inflation risk and liquidity premia. From 27 February to 28 August, real yields (inflation-linked bond yields) rose by 60bp in the US, 58bp in the UK, 34bp in Germany, 56bp in Japan and 35bp in Australia; breakeven inflation rose by 16bp, 43bp, 28bp, 25bp and 9bp respectively. France’s nominal 10-year yield rose 89bp, but comparable inflation-linked data is unavailable. Real yields accounted for more than half the move in every decomposable market. Inflation compensation contributed more in the energy-importing UK, Germany and Japan, than in the US and Australia

The New York Fed’s July survey put the median modal five-year, five-year-forward inflation forecast at 2.3 percent, with more probability assigned to inflation above 3 percent than below 1.5 percent. Central expectations can remain anchored even when upside inflation risks rise.

Higher real interest rates affecting all debt issuers

Rising yields are not mainly an inflation story. Since September 2024, 30-year yields have risen more than two-year yields in all six markets highlighted in Figure 2, pointing to forces beyond near-term policy expectations. Over the six months to the end of August, two-year yields rose as much as or more than 30-year yields everywhere except Japan; in the US, even the 10-year outpaced the 30-year. The energy shock from the US-Iran war raised near-term inflation risk and the expected policy-rate path. The Japan exception is consistent with monetary-policy normalisation and long-end supply pressures.

Stronger investment demand is one plausible driver behind the longer repricing. According to Miran, higher expected US real rates are partly a sign of stronger potential growth. Caballero (2026a, 2026b) showed how optimism about artificial intelligence can raise consumption and investment before productivity gains arrive. AI is therefore a plausible contributor to high US yields, while defence, energy security and infrastructure add to capital demand elsewhere.

A second driver is the amount of long-dated debt private investors must absorb as yield-insensitive buyers step back. For example, foreign official investors account for declining shares of the US Treasury market; Eurosystem holdings are shrinking as euro-area sovereign issuance rises; and Dutch pension reform is expected to reduce demand for very long bonds and swaps. Large US technology firms are also issuing long-dated euro debt.

The relevant supply variable is not gross issuance alone, but the interest-rate risk from long-dated bonds that price-sensitive investors must absorb. Yield-curve models decompose long yields into expected future short rates and a term premium – the extra return investors require to hold long-dated bonds. But they cannot identify why either component moved: expected short rates are unobservable, estimates vary across models and fiscal or supply shocks can affect expected real rates, inflation compensation and the term premium. A higher estimated term premium therefore does not prove the absorption mechanism; a stable one does not rule out fiscal effects (Adrian et al, 2013; Kim and Wright, 2005; Li et al, 2017; Diercks and Anani, 2024).

Swap rates – the fixed rate exchanged for a floating rate in an interest-rate swap – provide a benchmark outside government bonds. During the sell-off, 30-year swap rates rose almost as much as government yields in the US, euro area, UK and Japan, while the US investment-grade corporate bond-Treasury yield differential changed little. Together, this points to a broader rise in underlying long-term rates, rather than sovereign-specific cheapening or cash-market dysfunction.

Sovereign risk, scarcity and convenience yields

Since September 2024, sovereign-specific factors have contributed to dispersion, but the Bund is a moving benchmark. Against the euro swap rate, Germany’s yield rose 25bp, France’s 33bp and Italy’s fell 30bp (Figure 3). Bund spreads therefore understate France’s deterioration. Most of France’s roughly 115bp yield rise since February reflects the broader euro move. Italy’s relative improvement has coincided with a rating upgrade. France’s deficit remains above 5 percent and debt is heading towards 120 percent; Italy’s deficit is moving toward 2.9 percent, though debt is also rising.

A rise in sovereign yields relative to swaps can reflect either sovereign risk or a loss of ‘convenience yield’ – the liquidity and safety services for which investors accept lower yields. This channel is clearest in the Bund market. Schnabel (2025) noted that 10-year Bunds traded nearly 80bp below swaps at peak QE scarcity, but that spread narrowed from mid-2022. By August 2026, the Bundesbank found no signs of a scarcity premium, while repo measures also show reduced collateral scarcity. The most natural interpretation is that the unusually large quantitative easing-era Bund scarcity premium has unwound as central bank holdings have fallen and the supply of German government debt available to private investors has increased, rather than that German creditworthiness has deteriorated (Schnabel, 2025; Deutsche Bundesbank, 2026; Caballero, 2026b).

A related mechanism appears to have operated in Treasuries, although in a different form. Treasuries were not as scarce as Bunds in the same physical collateral sense, but their convenience yield is also an equilibrium supply-demand price. As the supply of US government debt has risen relative to demand from reserve managers, central banks and other relatively price-insensitive investors, the yield concession investors are willing to accept for holding Treasuries appears to have diminished, particularly at longer maturities. That is consistent with evidence of declining Treasury specialness and the argument that a larger stock of safe debt can reduce scarcity value and raise absorption costs (Caballero, 2026a, 2026b). It does not, by itself, imply a loss of Treasury safe-asset status or a broader loss of confidence in US government debt. The dollar has not persistently weakened when US long yields rise relative to peers.

Conclusion

The rise in long-term yields is unusually large by post-2006 standards, but the evidence does not point to a breakdown in US monetary, fiscal or financial institutions as its principal cause. This does not make the rise benign, and the idea that higher real rates signal stronger potential growth goes beyond what the evidence can establish.

More plausible is a combination of stronger investment demand, greater public and private supply of long-duration debt and reduced absorption by relatively yield-insensitive investors. The weights differ across countries: AI-related investment is a plausible US contributor; Europe combines heavier sovereign borrowing and reduced Eurosystem absorption with changing pension demand; and Japan additionally reflects monetary-policy normalisation.

Fiscal developments and safe-asset premia nevertheless matter. But the evidence points towards a reduction in the scarcity or convenience value of government debt, rather than a generalised increase in sovereign default risk. The Bund’s unusually large QE-era scarcity premium has largely disappeared. Treasury convenience value also appears to have declined as relative supply has risen, particularly at longer maturities, but that is not the same as a loss of Treasury safe-asset status or declining confidence in US government debt.

The implication is therefore that higher real rates should not be dismissed as evidence of stronger growth, but neither do they signal that the US Treasury market or US macroeconomic institutions are approaching a structural crisis. They are principally the price required to clear a global market in which demand for capital and the supply of duration have increased, while some investors that previously absorbed that duration irrespective of price have stepped back.

Source : Bruegel