As a sovereign bond selloff sweeps across global markets, the world's largest sovereign wealth fund, Norway's Government Pension Fund Global, has dropped a bombshell: reducing its allocation to global government bonds.
On September 1, Norges Bank Investment Management (NBIM) formally wrote to the Norwegian Ministry of Finance, proposing to cut the government bond weight in the fund's benchmark bond index from 70% to 50%. If the proposal ultimately gains approval from both the Norwegian Ministry of Finance and parliament, the consequences would be immediate: U.S. Treasuries would gradually decline from their current 34.1% share of the fund's fixed-income portfolio to 21.9%, equivalent to trimming roughly $75-80 billion in U.S. debt — the most significant change resulting from this bond investment strategy overhaul.
Why is NBIM making such a dramatic move at this particular moment? And why does it appear so specifically targeted at reducing U.S. Treasury holdings?
First, it's important to clarify that NBIM did not suddenly decide to overhaul its bond investment strategy. Rather, the move comes in response to two questions posed by the Norwegian Ministry of Finance on February 25: First, whether the three roles of bond investments remain useful — reducing overall portfolio volatility, providing liquidity, and capturing risk premiums — and how various factors should be weighted. Second, how to evaluate the composition of the bond benchmark index, including which markets and sectors should be included, along with related weighting principles.
After six months of careful deliberation, NBIM has proposed three major changes to its debt market investment strategy: First, reducing the government bond portion of the bond index from 70% to 50%; Second, switching from GDP-based to market value-based weighting for the government bond segment; Third, incorporating mortgage-backed securities (MBS) and non-government U.S. fixed-income assets into the bond index. Each recommendation implicitly reflects the dramatic shifts occurring in today's world.
NBIM provided justifications for each change in its letter. On the first point, bonds serve as the fund's primary and most important tool for reducing overall return volatility, and NBIM believes a 50% government bond allocation is sufficient to meet liquidity needs, even during periods of financial market turbulence.
The second point deserves particular attention. NBIM has proposed a sharp pivot regarding the fiscal realities of debt-issuing nations — abandoning GDP weighting in favor of market value weighting, on the grounds that high government debt is now a common feature of developed economies rather than a phenomenon unique to a few countries. Previously, the fund allocated bonds based on GDP, meaning larger economies attracted more bond purchases. That approach was essentially a "lending based on strength" logic, designed to prevent the fund from automatically buying excessive amounts of debt from heavily indebted nations. But wouldn't switching to market value weighting simply mean buying more from whoever issues the most debt?
That's precisely why NBIM introduced its third recommendation — expanding the entire fixed-income pool to include institutional bonds (MBS) and corporate bonds, which would naturally dilute the government bond allocation. Notably, U.S. corporate debt issuance is now approaching parity with Treasury issuance. In recent years, net U.S. Treasury supply has held steady at roughly $2.1-2.2 trillion annually, while U.S. investment-grade and corporate credit bond issuance has been accelerating rapidly — over $800 billion in net issuance last year, and $1.6 trillion this year. Based on capital expenditure plans, corporate bond issuance could reach $2 trillion next year. To put that in perspective, next year's U.S. corporate net issuance would be roughly equivalent to the net issuance of U.S. Treasuries.
Economist Mohamed El-Erian has noted that traditional buyers and holders of U.S. Treasuries are facing mounting pressure. While the amount Norway's fund plans to reduce is not large enough to shake the U.S. Treasury market, its symbolic significance matters more — it signals that traditional, stable buyers of U.S. debt are becoming less reliable.
For nearly a century, the classic 60% stock / 40% bond portfolio has been regarded as the "holy grail" of asset allocation. Vanguard's data shows that from 1926 to 2021, a 60% U.S. stock / 40% U.S. bond portfolio delivered an average annualized nominal return of approximately 8.8%. But today, long-dated U.S. Treasuries are experiencing their darkest decade on record: U.S. Treasuries with maturities of 15 years or longer have posted an average annual return of -2% over the past decade — only the second time since records began in 1936 that a ten-year period has produced negative returns. Meanwhile, U.S. stocks have averaged 15% annual gains over the same period, and commodities have risen 11%.
The turning point came in 2020, during the largest monetary expansion in human history. Before 2020, U.S. Treasuries delivered an average annual return of 9% over ten-year periods. Since the start of 2021, the most popular long-dated Treasury ETF — the iShares 20+ Year Treasury Bond ETF — has fallen a cumulative 37%. The bond market is no longer the safe haven it once was.
This has triggered a historic shift in global capital flows — "selling U.S. bonds, buying U.S. stocks." Strategists at Deutsche Bank noted in a recent report that foreign investors have shifted capital toward U.S. equities rather than government bonds for the first time since the 2008 financial crisis. Deutsche Bank data shows that over the past year through March 2026, U.S. stocks attracted a record $600 billion in net inflows, surpassing investments in government and agency bonds "by the largest margin in history."
Even global asset management giant BlackRock has adopted a strategy of "overweighting U.S. stocks and underweighting U.S. Treasuries," citing optimism about earnings growth driven by AI infrastructure investments, while arguing that the reliability of long-dated Treasuries as a diversification tool has diminished. In an August 31 report, BlackRock wrote: "In the new environment, the reliability of long-dated bonds as a portfolio diversification tool has declined."
This "buy stocks, sell bonds" trend is by no means unique to Norway. In January, Denmark's Akademiker Pension pension fund publicly liquidated its U.S. Treasury holdings, with its chief investment officer stating bluntly: "We need to find more resilient alternatives." Meanwhile, fellow Danish pension giant Pension Danmark continued adding to U.S. equities in the second quarter, with holdings reaching $12 billion — a 17% quarter-over-quarter increase — and its top ten holdings are exclusively U.S. tech giants. (This article presents objective data and information only and does not constitute investment advice.)
In fact, national pension funds around the world are actively allocating to U.S. tech stocks. A 2026 study by the Bank for International Settlements (BIS) quantifies this trend: U.S. pension funds' fixed-income allocation has fallen from roughly 40% in the early 1980s to about 10% in 2023, while developed European economies have seen allocations drop from approximately 35% in the early 2000s to under 20%. The BIS paper also highlights another shift in long-term funds' asset allocation — allocations to alternative assets such as private equity, real estate, and private credit have been steadily rising. U.S. state and local pension funds' alternative allocations have grown from under 10% in the early 2000s to over 30% by 2024.
Interestingly, Norway's Government Pension Fund Global is also planning a shift toward private equity. On September 1, NBIM sent two letters simultaneously — one addressing bond strategy, the other addressing geopolitical risks in equity investments. In the latter letter, NBIM discussed the increasing frequency of geopolitical risks and the risks of high stock concentration (a result of the fund's index-based strategy, where market-cap weighting automatically channels substantial capital toward a handful of U.S. tech giants). The proposed solution: a strong recommendation to increase allocations to 'unlisted assets' (i.e., private equity and private assets).
The reasoning is straightforward: the dividends of globalization are fading, and the era of effortlessly earning outsized returns by simply buying U.S. Treasuries and U.S. tech stocks is over. However, NBIM believes that by extending investment horizons and gradually channeling capital into unlisted corporate assets, the fund can gain exposure to distinctly different risk and return sources, thereby reducing the excessive concentration risk of tech giants in public equity indices.
And change is already quietly underway. In late 2025, three Danish pension funds — Pension Danmark, AP Pension, and Akademiker Pension — jointly committed €220 million to establish the private equity fund ETNA, dedicated to investing in European defense, cybersecurity, and critical infrastructure.
As the world's steadiest long-term capital begins to redraw the boundaries between bonds, equities, and private assets, we may well be standing at the crossroads of a new era in asset allocation.
Comments