The 2026 Trustees Reports landed this month, and the headline was the runway, as it always is: the Old-Age and Survivors trust fund now pays full scheduled benefits only until the fourth quarter of 2032, a quarter earlier than last year, after which the money coming in covers 78 percent of what is owed. The combined Social Security funds reach that line in 2034, at about 83 percent, and Medicare's hospital fund in 2033, at about 89 percent (2026 Trustees summary, via CRFB). Those are the numbers everyone quotes.
Here is the number almost no one quotes. The Old-Age and Survivors and Disability funds together held about $2.56 trillion at the end of 2025, down from $2.72 trillion a year earlier, and every single dollar of it was invested in exactly one kind of thing. Not stocks. Not corporate bonds. Not real estate or private equity or a share of anything that grows. Just special-issue United States Treasury bonds, which earned an effective 2.6 percent in 2025 (AARP on the 2026 report; SSA effective rates).
Now set that beside two other enormous pools of public money. Norway's Government Pension Fund Global holds about $2.2 trillion and keeps roughly 71 percent of it in the stock market, owning a piece of nearly every listed company on the planet. Saudi Arabia's Public Investment Fund holds more than a trillion and uses it to buy Uber, a Premier League club, and to build cities in the desert. All three are national funds, filled with a nation's money, meant to serve a nation's people. Two of them behave like the most aggressive endowments on earth. The third is forbidden by law from buying a single share. This is about that difference, and about who it serves.
What Social Security is allowed to own
The restriction is not a preference or a habit. It is a statute. Section 201(d) of the Social Security Act, at 42 U.S.C. 401(d), says the trust funds may be invested "only in interest-bearing obligations of the United States or in obligations guaranteed as to both principal and interest by the United States" (Cornell Law, 42 U.S.C. 401). Medicare's hospital fund carries the identical language at 42 U.S.C. 1395i (Cornell Law, 42 U.S.C. 1395i). The Congressional Research Service states the practical effect plainly: government securities "are the trust funds' only investment option under current law" (CRS R45709).
What the funds actually hold are "special issues," bonds the Treasury sells to no one but the trust funds. They are non-marketable, which means they never trade, and they are redeemable on demand at par, so the fund can hand one back for its face value plus accrued interest at any time without a penny of loss (SSA on special issues). Their interest rate is set by a formula: the average market yield on all Treasury securities that are more than four years from maturity, rounded to the nearest eighth of a percent (Cornell Law, 42 U.S.C. 401). That formula is why the fund earned 2.6 percent in 2025 while newly issued Treasuries were paying north of 4 percent: the fund's rate is a lagging average of a long book, not a fresh-market yield (SSA new-issue rates).
So the largest retirement pool in the country holds the safest asset in the world and earns the yield that safety pays. Hold that thought against what the other two funds do with the same kind of money.
Norway owns a slice of the whole world
Norway's fund was worth about 21.3 trillion kroner at the end of 2025, roughly $2.2 trillion, which makes it the largest sovereign wealth fund on earth and nearly four times the size of Norway's entire economy (Norges Bank Investment Management). At year-end it held 71.3 percent of its money in listed equities, 26.5 percent in fixed income, and the small remainder in unlisted real estate and renewable-energy infrastructure (NBIM 2025 results).
The scale of that equity position is hard to overstate. The fund "holds 1.5 percent of all the world's listed companies," with stakes in roughly 7,200 firms across almost every country and industry, and its largest single holdings at year-end were NVIDIA, Apple, and Microsoft (NBIM, the fund). It is, functionally, an index of global capitalism owned by five and a half million Norwegians. And it has been paid like one: an annualized return of 6.64 percent since 1998, 4.34 percent after inflation and costs, and a 15.1 percent gain in 2025 alone, about $247 billion in a single year (NBIM returns; NBIM 2025 results).
The money came out of the ground. Norway pours its state petroleum income into the fund, has done so since the first transfer in 1996, and then binds its own government with a fiscal rule: the state may spend only the fund's expected real return, estimated at about 3 percent a year, never the capital itself (Norwegian Petroleum, management of revenues). The point of the rule is written into the fund's own mission, to make sure "both current and future generations of Norway get to benefit from our oil wealth" (NBIM, about the fund). Remember that phrase, future generations, because it is the whole reason Norway can do what Social Security cannot.
Saudi Arabia builds a country with it
Saudi Arabia's Public Investment Fund is the same idea pointed in a different direction. It held somewhere between about $913 billion and $941 billion at the end of 2024, crossed a trillion dollars in its mid-2025 annual report, and is reckoned by outside trackers as the fourth largest sovereign fund in the world (PIF 2024 results; Semafor). In April 2025 it raised its own 2030 target to $2.67 trillion (The National).
Where Norway buys a quiet slice of everything, the PIF takes loud, concentrated stakes and builds things outright. It wholly owns the giga-project companies remaking the Saudi map, NEOM, Red Sea Global, Qiddiya, Diriyah, and ROSHN (PIF giga-projects). It put $3.5 billion into Uber in 2016 and took a board seat (PIF on Uber). It owns about 58 percent of the electric-car maker Lucid after roughly $8 billion of investment (Lucid investor relations), 80 percent of Newcastle United (PIF on Newcastle), the whole of the LIV Golf venture, a $4.9 billion mobile-games company, and up to $45 billion committed to SoftBank's Vision Fund (Savvy Games; PIF on SoftBank). Analysts describe it not as a savings fund at all but as "a stabilisation and development fund," an instrument for diversifying the economy away from oil (SWP Berlin). It is a national checkbook run like a venture firm that happens to be building a country.
Same idea, opposite mandate
Line them up and the contrast is almost absurd. Norway's national fund owns one and a half percent of every public company in existence. Saudi Arabia's national fund owns a soccer team and a linear city. America's national retirement fund is not permitted to own one share of one company anywhere. Three pools of public money, and the American one is the only one legally fenced into the single asset that grows the slowest.
The gap that fence creates is not small, and it compounds. Norway has earned about 6.6 percent a year for a quarter century. Social Security earned 2.6 percent last year. Four points a year, compounded across a working life, is the difference between a dollar and roughly four dollars. So the obvious question is the one this whole comparison is built to ask: why does the richest country on earth run its largest fund on bond yields while smaller nations run theirs on the market?
Why the American fund is different
The first answer is the statute, but the statute is the mechanism, not the reason. The reasons underneath it are three, and they are worth stating fairly because each one is real.
The deepest reason is structural, and it is the phrase from Norway's mission. Norway and Saudi Arabia are investing surplus wealth, money the state pulled out of the ground and does not need to spend now, deliberately set aside to compound for future generations. A fund like that can hold 71 percent stocks and ride out a crash, because it has decades before it needs the money. Social Security and Medicare are the opposite kind of thing. They are pay-as-you-go insurance, where this year's payroll taxes flow straight out as this year's benefit checks (CRS R45990). There was never a great invested pool; the trust fund is a buffer, not a nest egg, and in 2025 that buffer paid out more than it took in, about $1.61 trillion against $1.45 trillion (AARP on the 2026 report). You do not put next month's rent in the stock market, and a fund that owes benefits in the near term and is already shrinking is closer to next month's rent than to Norway's grandchildren.
The second reason is political, and the country has actually had this argument. In the 1990s the Advisory Council on Social Security, chaired by the economist Edward Gramlich, studied investing part of the trust fund in equities, one plan proposing up to 40 percent of reserves in a broad market index (CRS 97-81). President Clinton took it further in his January 1999 State of the Union, proposing to invest a portion of the fund in the stock market "just as any state or local government, or private pension does," with equities capped at a small share of the total (Clinton 1999 address). It did not happen, and the objections are why. The Government Accountability Office warned that a government holding that much stock "would have a stronger incentive to actively exercise the voting rights of its sizable stock portfolio," and that the fund would face "tremendous political pressures to steer" its investments toward social or political goals rather than pure return (GAO AIMD/HEHS-98-74). Federal Reserve Chairman Alan Greenspan testified against it directly, saying he doubted "it is possible to secure and sustain institutional arrangements that would insulate, over the long run, the trust funds from political pressures" (Greenspan testimony, March 3, 1999). The public agreed, rejecting the idea 52 to 34 in a Pew survey that January (Pew Research). The fear was simple: a fund that big, voting that many shares, becomes the government owning corporate America, and neither party wanted to hand Washington that lever.
The third reason is that the guarantee itself argues for safety. Social Security is a promise the law makes to pay a defined benefit. Norway's fund can lose 15 percent in a bad year and shrug, because it owes no one a fixed check. A fund backing a legal guarantee cannot, and the special-issue Treasury, redeemable at par with no possibility of capital loss, is the asset that makes the guarantee keepable in every market (SSA on special issues).
The captive lender, and where the money actually went
There is a sharper way to tell this story, and it deserves telling because it is true, not because it is a conspiracy. When the trust fund runs a surplus, that surplus does not sit in a vault. It is lent to the rest of the government. The fund buys special-issue bonds, and by the Social Security Administration's own description, "the cash exchanged for the securities goes into the general fund of the Treasury and is indistinguishable from other cash in the general fund" (SSA trust fund FAQ). The government then spends that cash on whatever the government spends on, and leaves behind a bond paying 2.6 percent. The GAO says it directly: Treasury "borrows the cash from Social Security's surplus to pay for other government expenses," which "reduces the amount Treasury would otherwise need to borrow from the public" (GAO AIMD-96-30R). For decades, counting that surplus inside the unified budget made the reported federal deficit look smaller than it was (Bipartisan Policy Center).
Put that way, the critique writes itself, and it is essentially an arbitrage. The government borrowed from a captive lender, the one lender legally required to buy its bonds, at a conservative rate, spent the proceeds on general operations, and paid back only the interest. Meanwhile the fund forwent the equity premium a Norway-style portfolio would have earned. The retiree's money financed the government at 2.6 percent while the market ran at three times that, and the difference stayed with the borrower.
That is the fair version of the accusation. Here is the fair version of the answer, because both belong in an honest ledger. The bonds are not fictional. They are, in SSA's words, "backed by the full faith and credit of the U.S. Government" and "just as safe as U.S. Savings Bonds" (SSA trust fund FAQ). The roughly $2.8 trillion is legally owed to the fund and will be repaid with interest; the Committee for a Responsible Federal Budget, no cheerleader for the program's finances, calls the "borrowed and never paid back" story a myth, noting the full sum "is still owed to the Social Security trust fund under current law" (CRFB). When a president stood in West Virginia in 2005 and called the trust fund "just IOUs," the fact-checkers pushed back that the framing was misleading, because the IOUs must be repaid with interest like any Treasury bond (FactCheck.org; PolitiFact). And the redemption is real in a way that matters: when the fund cashes those bonds to pay benefits, the rest of the government has to find the money, by taxing more, spending less, or borrowing from the public (CRS RL33028). The IOU is a claim the Treasury genuinely has to honor.
So both things are true at once. The surplus really was used to fund general government, and the obligation really is binding. What the fund gave up was the opportunity cost, the gap between 2.6 percent and the market, and it gave that up partly on purpose, because a guaranteed benefit cannot ride the market's swings, and partly because the law and the politics never allowed anything else. Whether you call that prudence or a raid depends on where you stand, but the plumbing is not in dispute, and it is published.
The mirror at the center of this series
Here is why this fund belongs in a series that has spent months reading the tax returns of foundations, universities, and hospitals. Every one of those private institutions invests exactly the way Norway and Saudi Arabia do. The Howard Hughes Medical Institute runs a $27 billion endowment through public and private markets and pays its investment chief like a hedge-fund manager. The university endowments hold their oldest equity positions for decades and compound them untouched. Kaiser's profit or loss is decided each year by what its enormous portfolio did. America's private institutions behave like sovereign wealth funds, heavy in equities and alternatives, earning the market's return and keeping it.
America's actual public fund, the one that belongs to every worker in the country, is the single large pool forbidden from doing the same. The foundation earns the market. The kingdom earns the market. The endowment earns the market. The retiree's collective reserve earns the bond yield, lends its surplus to the government at that yield, and shrinks a little more each year. It is the same well this series keeps finding underneath every institution that holds money, the same public and private markets everyone else is invested in. The difference is only who is allowed to lower a bucket into it, and how deep. Norway and the Broad Institute reach the water. Social Security, by law and by design, is handed a receipt.
Related reading
- The Runway and the Ruler: the 2026 Trustees numbers in full, and the worked example of what a max earner's contributions would have become if they had been invested.
- The Working Ledgers: the single market underneath every foundation, pension, and endowment in this series.
- How Nonprofit Hospitals Actually Make Money: private institutions whose bottom line, like a sovereign fund's, is set in the markets.
- University Endowments: the private funds that invest exactly the way the public one is forbidden to.
Fact-check notes and sources
- The 2026 Trustees figures (OASI full benefits through Q4 2032 then 78 percent payable; combined OASDI through 2034 at about 83 percent; Medicare Hospital Insurance through 2033 at about 89 percent; combined reserves of about $2.56 trillion at the end of 2025, down from about $2.72 trillion, on income of about $1.45 trillion against cost of about $1.61 trillion): from the 2026 OASDI and Medicare Trustees Reports as summarized by the Committee for a Responsible Federal Budget and AARP, read alongside the reports themselves (the OASDI report at ssa.gov/OACT/TR/2026).
- The investment mandate (trust funds may hold only US Treasury obligations, per Social Security Act section 201(d) at 42 U.S.C. 401(d), and section 1817 at 42 U.S.C. 1395i for Medicare's hospital fund; that government securities are the only option under current law): Cornell Law 42 U.S.C. 401, 42 U.S.C. 1395i, and CRS R45709.
- Special issues and the interest rate (non-marketable, redeemable at par plus accrued interest; the rate formula tied to the average yield on marketable Treasuries more than four years to maturity; the 2.6 percent effective rate in 2025 and 2.512 percent combined in 2024 against a 2024 new-issue rate of 4.271 percent): SSA on special issues, SSA effective rates, and SSA new-issue rates.
- Norway's Government Pension Fund Global (about $2.2 trillion, 21.3 trillion kroner, at year-end 2025; 71.3 percent equities, 26.5 percent fixed income, the small remainder in real estate and renewable infrastructure; ownership of about 1.5 percent of the world's listed companies across roughly 7,200 firms with NVIDIA, Apple, and Microsoft the largest holdings; 6.64 percent annualized and 4.34 percent real since 1998, and 15.1 percent in 2025; petroleum funded since 1996; the roughly 3 percent fiscal spending rule): NBIM, the fund, NBIM 2025 results, NBIM returns, NBIM about the fund, and Norwegian Petroleum.
- Saudi Arabia's Public Investment Fund (about $913 to $941 billion at the end of 2024, crossing $1 trillion in its mid-2025 report and reckoned the fourth largest sovereign fund; the 2030 target raised to $2.67 trillion in April 2025; wholly owned giga-project companies NEOM, Red Sea Global, Qiddiya, Diriyah, and ROSHN; the $3.5 billion Uber stake, about 58 percent of Lucid after roughly $8 billion, 80 percent of Newcastle United, the LIV Golf venture, a $4.9 billion games acquisition, and up to $45 billion to the SoftBank Vision Fund; its character as a development fund): PIF 2024 results, Semafor, The National, PIF giga-projects, PIF on Uber, Lucid IR, PIF on Newcastle, Savvy Games, PIF on SoftBank, and SWP Berlin. The 2023 LIV and PGA Tour framework agreement remained unconsummated as of 2026.
- The 1990s equity-investment debate (the Gramlich Advisory Council studying up to 40 percent of reserves in a market index; Clinton's January 1999 proposal; the GAO objections on corporate-share voting and political pressure; Greenspan's March 1999 testimony against; the 52-to-34 public rejection): CRS 97-81, Clinton 1999 address, GAO AIMD/HEHS-98-74, Greenspan testimony, and Pew Research.
- The captive-lender mechanism and the rebuttal (that surplus cash goes into the general fund and is spent, per SSA and GAO; that the unified-budget treatment made deficits look smaller; that the bonds are backed by full faith and credit and the roughly $2.8 trillion is legally owed; the 2005 "just IOUs" framing and the fact-checkers' response; that redemption forces the general fund to raise cash; and that the system is pay-as-you-go so there was never a pre-funded pool): SSA trust fund FAQ, GAO AIMD-96-30R, Bipartisan Policy Center, CRFB, FactCheck.org, PolitiFact, CRS RL33028, and CRS R45990.
This post is informational and historical, not political advocacy or financial advice. All figures are reproduced from the cited government reports, statutes, official fund disclosures, and public record. Public officials and institutions are described from the documented public record as nominative fair use, with no affiliation implied.
