A U.S. 10-year yield near 5% may reveal more than a debt-supply problem. The institutions that once absorbed long government bonds are changing too.
The First Number I Check Every Day
I monitor dozens of indicators every day to understand financial markets and equity markets. But there is one number I usually check first: the U.S. 10-year Treasury yield.
It is not because the 10-year yield explains everything. But few market prices compress so many forces into one number: growth expectations, inflation, monetary policy, fiscal credibility, the cost of capital and the discount rate applied to financial assets.
That is why the latest move caught my attention.
On September 11, the official U.S. Treasury 10-year par yield reached 4.96%. Intraday, it traded as high as 4.979%.
It had approached the level I have watched most closely since the global interest-rate repricing began in 2022: 5%.
This is not an academically defined threshold. It is my own market reference point.
For some time, I had assumed that a sustained move materially above 5% would be difficult. One reason was uncertainty in the labor market. My expectation was that, before long-term yields could remain much higher, weaker employment or growth expectations would eventually begin to offset inflation and fiscal pressure.
But that assumption now deserves another look.
While researching European sovereign-bond markets, I came across a development that changed the way I thought about the long end of the yield curve: Dutch pension reform.
That reform led me to a deceptively simple question.
What happens when institutions that once had structural reasons to own very long-dated government bonds no longer need to own as much of them?
From there, the question became larger.
What if long-term yields are rising not only because governments are borrowing more, but because the institutions that historically bought that debt are changing?
<Source: Yahoo Finance. September 11, 2026>
This article advances a non-consensus hypothesis rather than claiming an established causal relationship. Where the evidence is strong, I will treat it as evidence. Where the data remain incomplete, I will keep the argument explicitly as a hypothesis.
The Missing Half of the Bond Debate
The supply side of the developed-market bond problem is familiar.
Governments need to finance fiscal deficits, ageing societies, defense, industrial policy, energy systems and infrastructure. The AI investment cycle adds another enormous claim on savings through semiconductors, data centers, electricity generation and grids.
But every bond sold by a government must be held by someone.
Central banks are no longer expanding their balance sheets as they did after the global financial crisis. Banks face their own capital and balance-sheet constraints. Foreign reserve managers have strategic as well as financial considerations. Pension systems are changing.
Part of a long-term government-bond yield reflects the expected path of short-term rates. Another part is the term premium: the extra return investors demand to lock up capital and accept inflation and interest-rate risk over many years.
If the natural buyer base becomes less willing to own 20-, 30- or 40-year debt, the market may need a higher yield to attract a different buyer.
That does not prove pension reform caused the recent rise in yields.
It establishes a mechanism worth taking seriously.
Why Pension Design Can Move the Long End
A defined-benefit pension, or DB plan, promises a retirement payment in advance. Those payments may extend decades into the future. The fund therefore has a natural reason to own assets whose value moves with those long-dated obligations.
This is liability matching.
Liability-driven investment, usually shortened to LDI, organizes the portfolio around that objective. Duration is the sensitivity of a bond or portfolio to changes in interest rates. Long pension liabilities often create demand for long-duration bonds and interest-rate swaps because those instruments can help the assets and liabilities respond more similarly when long rates move.
A defined-contribution pension, or DC plan, works differently. Contributions are defined, but the final retirement income depends on investment returns. The portfolio can have more freedom to hold equities, private assets and other growth investments.
That does not mean DC systems never buy long bonds. Older participants may move toward fixed income, and retirement products can create new demand for duration. But the automatic link between a collective long-dated promise and an ultra-long asset can become weaker.
The OECD’s Global Debt Report 2026 finds a clear shift in the pension funds it surveyed away from debt securities with maturities beyond 20 years and toward the five-to-20-year range. It identifies the transition from DB to DC plans as one contributing factor, alongside ageing and liquidity needs. https://www.oecd.org/en/publications/global-debt-report-2026_e9d80efd-en/full-report.html
The conclusion should remain narrow.
Pension-system design does not dictate every portfolio decision. But it can change the marginal buyer of long-duration sovereign debt.
The Dutch Experiment Looked More Dramatic at First
As tackled some of the points in the last analysis, the Netherlands provides an unusually useful case because the pension system is both enormous and deeply involved in long-duration hedging. https://www.geomarketsignal.com/2026/09/europes-hidden-bond-risk-what-happens.html
Under the new Dutch pension framework, investment policy becomes more differentiated by age. Younger members can carry more investment risk. Older participants and pensioners retain a stronger need for fixed income, but their preferred maturity can be shorter than the duration required under the old collective system.
In October 2025, De Nederlandsche Bank reported that market analysts expected the transition to reduce pension-fund holdings of government bonds with maturities beyond 25 years and interest-rate swaps by roughly €100–150 billion. https://www.dnb.nl/en/general-news/background-2025/new-pension-contract-implications-for-international-interest-rate-markets/
That number initially looked like powerful evidence for the thesis.
But it required two immediate corrections.
First, the estimate combines government bonds and swaps. It is not a forecast of €100–150 billion in bond sales alone.
Second, it was a market expectation, not an observed transaction.
The causal evidence is stronger than the forecast. A DNB working paper studied an earlier Dutch regulatory change that made pension and insurance liabilities more sensitive around 20 years and less sensitive at longer maturities. The institutions reduced their longest-duration holdings and increased exposure around 20 years. The resulting demand shift steepened the long end of the curve, while banks helped absorb the change. https://www.dnb.nl/en/publications/research-publications/working-paper-2023/769-long-term-investors-demand-shifts-and-yields
That is evidence that regulation can affect government borrowing costs through institutional demand.
Then the first 2026 transition data complicated the story again.
DNB’s spring financial-stability review found that sales of long-dated swaps after the first pension transitions were just over €10 billion—smaller and more gradual than anticipated—and had not caused significant disruption in the swap market. By June 30, 34 pension funds with €589 billion of assets had converted to the new system, out of total Dutch pension assets of €1.72 trillion. https://www.dnb.nl/en/general-news/news-2026/pension-transition-and-pension-fund-funding-ratios-2026q2/
My conclusion therefore had to change.
Dutch reform does not prove that a sudden pension shock is driving developed-market yields higher today. It does show that the structure of the long-end buyer base can change—and that markets may adjust more gradually than the most dramatic estimates imply.
Higher Yields Can Recreate Buyers—Without Financial Repression
There is a tempting policy response to a shrinking natural buyer base: require pension funds to buy more government debt.
That would be the wrong starting point.
European pension institutions operate under the prudent-person principle. Their investments must serve the long-term interests of members and beneficiaries. A pension fund’s first obligation is to retirees, not to lowering the borrowing cost of the government that regulates it.
If a fiscally weak state uses regulation to create captive domestic buyers, it risks building a sovereign–pension nexus: a feedback loop in which retirement security becomes more exposed to the credit of the same state seeking cheaper financing. It is analogous in concept, though not identical, to the sovereign–bank nexus exposed during the euro crisis.
There is a better question.
Have long-term yields risen far enough that sovereign bonds can once again satisfy both sides of the pension mandate—return and certainty?
For a pension fund, the goal is not simply the highest nominal return. It is the risk-adjusted ability to pay future benefits.
Long-duration demand can therefore return voluntarily through several channels: DB liability matching, less-leveraged LDI structures, lifecycle allocations that move DC savers toward fixed income near retirement, annuitization, and inflation-linked sovereign bonds that match real spending needs.
The key is alignment.
Long bonds should be attractive because they help secure retirement outcomes—not because governments need a buyer.
Then There Is Europe
The United States has serious fiscal challenges of its own. But its financial architecture is fundamentally different from Europe’s.
The United States issues debt in the world’s dominant reserve currency through one enormous federal Treasury market. At June 30, 2026, marketable Treasury debt stood at approximately $31.1 trillion. The U.S. Treasury describes its market as the deepest and most liquid in the world, supported by a vast repo, futures and derivatives ecosystem.
The euro area has one currency but many sovereign borrowers, fiscal positions and yield curves.
That produces a different portfolio question.
Who should the European pension fund buy?
German Bunds? French OATs? Italian BTPs? EU-Bonds issued by the European Commission? Or U.S. Treasuries with the currency exposure hedged back into euros?
The distinction can be stated simply.
The United States has a duration-demand problem. Europe has a duration-demand problem and a sovereign-credit problem.
That does not mean U.S. debt is free of fiscal risk. It means the U.S. offers one federal curve and one federal credit, while euro-area investors must choose among national balance sheets that share a currency but not a common fiscal position.
Pension reform can create demand for duration but it cannot manufacture credit quality.
Europe Has Pension Capital, but No Treasury
Europe does not lack safe assets.
German Bunds remain a core benchmark. Other highly rated sovereign and supranational bonds provide substantial safe collateral. The European Commission has also built a larger common issuance program through EU-Bonds and EU-Bills.
The problem is comparability with the Treasury system: scale, uniformity, liquidity, benchmark depth, common fiscal backing and market infrastructure.
As of July 31, 2026, the Commission reported approximately €792 billion of EU-Bonds outstanding, plus about €45 billion of EU-Bills. The bonds span benchmark maturities from three to 30 years. EU-Bonds qualify as Level 1 high-quality liquid assets, carry a 0% risk weight under the applicable bank framework and are eligible as ECB collateral. https://commission.europa.eu/strategy-and-policy/eu-budget/eu-borrower-investor-relations/investor-presentation_en
Their credit is strong. The European Commission describes EU borrowing as a direct and unconditional obligation of the European Union, supported by the EU budget. The market infrastructure is also improving: dealers quote EU-Bonds on electronic platforms, and an EU repo facility allows primary dealers to source eligible bonds and support secondary-market liquidity.
But this remains much smaller than the Treasury market, and its long-term political and institutional architecture is still evolving.
Europe has pension capital but it does not yet have a Treasury equivalent.
The Highest Yield Does Not Automatically Win
It would be easy to look at a higher Treasury yield and conclude that European pensions should buy U.S. debt.
That comparison is incomplete.
A Dutch pension fund may buy a Treasury in dollars, but it pays its beneficiaries mainly in euros. If the dollar falls, the currency loss can overwhelm the yield advantage. The fund can use an FX hedge—a contract that reduces currency risk—but the hedge has a cost determined by interest-rate differentials, forward pricing, market balance sheets and conditions in cross-currency funding markets.
That cost changes through time.
A U.S. Treasury yield therefore cannot be compared directly with a euro-denominated bond yield when the investor’s liabilities are in euros.
This is why I cannot conclude that Treasuries are already a superior solution for European pension funds. The U.S. offers exceptional depth, liquidity and collateral value. EU-Bonds offer euro liability matching and strong regulatory treatment without a dollar hedge.
The contest is not a simple yield ranking.
It is a comparison between the unmatched depth of the U.S. market and the currency-matched potential of a deeper, more permanent European common safe asset.
The Actual Flows Reject the Most Dramatic Story
The strongest version of the thesis would say European pension funds are already fleeing into Treasuries.
The available evidence says the opposite.
In 2025, Dutch pension funds sold a net €30 billion of U.S. securities, including €18 billion of U.S. government and corporate bonds. They bought €27 billion of European debt, including €20 billion of European sovereign bonds, with purchases concentrated in Germany and Spain while French sovereign exposure declined.
Some of that activity reflected ordinary rebalancing. It does not prove a permanent preference for Europe.
But it is decisive counterevidence against the claim that the Treasury market has already won Europe’s pension savings.
What may be emerging instead is differentiation inside Europe.
Funds can remain in euro-denominated debt while becoming more selective about sovereign credit. That is exactly why Europe’s problem differs from America’s. A common currency does not eliminate the choice among national fiscal risks.
Miran’s Architecture and the Competition for Savings
This question brought me back to remarks Stephen Miran made at the Hudson Institute in April 2025.
In the official White House transcript, Miran argued that the United States provides the dollar and Treasury securities that serve as reserve assets for the global trading and financial system. He also connected trade, foreign savings housed in dollar securities and U.S. financing conditions, saying that demand for dollars had kept borrowing rates low. https://www.whitehouse.gov/briefings-statements/2025/04/cea-chairman-steve-miran-hudson-institute-event-remarks/
He did not identify European pension funds as a policy target.
The defensible question is broader.
Could one consequence of the architecture he described be a stronger U.S. ability to attract foreign long-term savings when other developed markets face fragmented sovereign risk?
The answer is not yet visible in Dutch pension flows. But the strategic stakes are growing.
The competition over artificial intelligence is also a competition over semiconductors, electricity, data centers, grids and defense capacity. All require capital. If governments and companies must finance that buildout simultaneously, the ability to attract long-term savings becomes a source of economic power.
Europe may possess part of the savings required for the next industrial cycle while financing part of that cycle somewhere else.
That is an interpretation, not a demonstrated capital-flow outcome.
It is also why this research connects directly to GMS’s earlier work on U.S. Treasury pressure and the expanding supply of Big Tech’s long-duration AI debt. Governments are not competing for capital in isolation. Corporations building the next industrial system are bidding for it too.
Stablecoins show how policy can create a different type of structural Treasury buyer. Under the U.S. reserve framework, payment stablecoins must be backed one-for-one with permitted liquid assets such as dollars and short-term Treasuries. Growth in that system could support bills and other short-duration instruments. It does not solve the buyer question for 10- or 30-year debt. https://www.whitehouse.gov/fact-sheets/2025/07/fact-sheet-president-donald-j-trump-signs-genius-act-into-law/
What I Will Watch Next
The key signal is not whether European pension funds own U.S. Treasuries. Many already hold dollar fixed income.
The stronger signal would be major European funds systematically increasing allocations to long-duration Treasuries while maintaining or increasing the FX hedges on that dollar exposure.
That would show a willingness to pay the hedging cost in exchange for Treasury liquidity, credit quality and depth.
The evidence must be more specific than a headline increase in foreign holdings. I would want to know:
whether the purchases are concentrated in 10-, 20- or 30-year maturities;
whether they exceed the growth of total pension assets;
whether currency hedges rise with the dollar allocation;
whether domestic sovereign or EU-Bond exposure declines;
and whether Treasuries and EU-Bonds are being bought together rather than treated as substitutes.
Public data do not yet provide that full picture. In particular, the latest Dutch flow release does not isolate a clean stock of long-duration Treasury holdings and associated hedge ratios.
This is important because the hypothesis should be tested by actual portfolios—not by assuming the destination in advance.
The Capital Flows Will Decide
The evidence supports a narrower conclusion than the most dramatic version of this argument.
Pension regulation can change the demand for duration. The transition from DB toward DC structures can weaken some of the mechanisms that once created natural buyers of very long-dated sovereign debt. The Dutch experience provides unusually useful evidence that these institutional changes can matter for the long end of the yield curve.
But Europe’s problem cannot be solved simply by asking pension funds to buy more government bonds.
Pension reform can create demand for duration.
It cannot manufacture credit quality.
That distinction stands out in a monetary union in which German, French, Italian and other sovereign bonds share a currency but not the same fiscal balance sheet.
It also explains why the comparison between U.S. Treasuries and common European bonds deserves close attention.
The Treasury market offers a combination of depth, liquidity, collateral value and institutional permanence that Europe has not yet replicated.
But a European pension fund does not invest in dollars and retire its beneficiaries in dollars.
Its liabilities are primarily in euros.
Once currency risk and the cost of hedging dollar exposure are taken into account, there is not enough evidence at present to conclude that U.S. Treasuries provide European pension funds with a superior alternative to EU-Bonds.
The more revealing evidence may therefore come from what the pension funds themselves do next.
If major European pension funds begin systematically increasing allocations to long-duration U.S. Treasuries while hedging that dollar exposure, the signal would be much stronger.
But that is not what the evidence clearly shows today.
For European pension funds, the emerging competition may ultimately be between the unmatched depth of the U.S. Treasury market and the currency-matched potential of a deeper, permanent European common safe asset.
In an era in which governments must finance ageing societies, defense, energy infrastructure and a new AI-driven industrial cycle, the ability to attract long-term savings may itself become a source of national economic power.
I began this research with a question about yields.
I ended up with a question about who controls the world’s savings.
For now, the evidence does not tell us which market will win.
The capital flows will.