Washington’s growing debt may be doing more than increasing the amount taxpayers owe. It may also be increasing the price the government must pay to borrow it.
A new working paper from the Congressional Budget Office estimates that the rising supply of longer-term Treasury debt associated with projected federal borrowing could add roughly 50 basis points, or half a percentage point, to the 10-year Treasury term premium over the next decade, compared with a scenario in which federal debt remained unchanged relative to the economy.
It is an obscure-sounding consequence of a considerably less obscure problem: The government has to find somebody willing to buy all that debt.
CBO’s latest official budget projections put debt held by the public at 101 percent of gross domestic product in 2026. By 2036, it is projected to reach 120 percent, exceeding the previous record set just after World War II. Annual deficits, meanwhile, are projected to grow from $1.9 trillion this year to $3.1 trillion in 2036. CBO’s budget outlook lays out that trajectory.
The new paper suggests that increasing the supply of debt needed to finance those deficits can itself put upward pressure on longer-term interest rates.
And, in a particularly inconvenient bit of financial physics, markets may begin charging the government for tomorrow’s borrowing before Washington actually does it.
The Price of Finding a Buyer
The paper, “Treasury Duration Supply and Term Premiums,” was written by CBO analyst Michael McGrane and published as Working Paper 2026-11.
McGrane examined what happens to the “term premium,” essentially the additional compensation investors demand for holding a longer-term security rather than repeatedly investing in short-term debt.
Long-term bonds expose investors to more interest-rate risk. If the government issues more long-term securities, private investors must absorb more of that risk. Some investors will do so only at a better price, meaning lower bond prices and higher yields.
McGrane modeled that relationship using Treasury data stretching from February 1975 through March 2025.
His baseline result is striking.
A five-percentage-point increase in what the paper calls “Treasury duration supply” produces an estimated 2.8-basis-point increase in the 10-year term premium. At Treasury’s historical average maturity, that increase in duration supply corresponds roughly to a one-percentage-point increase in the debt-to-GDP ratio.
One percentage point of additional debt, therefore, does not simply add more debt. According to the model, it also makes investors demand slightly greater compensation for holding longer-term Treasury securities.
One percentage point isn’t particularly frightening.
Nineteen percentage points begins to attract attention.
CBO’s official February budget outlook projects public debt rising from 101 percent of GDP this year to 120 percent in 2036. The projection is available here.
Applying the working paper’s model to CBO’s projected debt trajectory, McGrane estimates that the resulting increase in Treasury duration supply could raise the 10-year term premium by approximately 50 basis points over the next decade compared with a scenario in which debt-to-GDP remained at its current level.
That does not mean CBO is forecasting the 10-year Treasury yield will simply rise half a percentage point. Treasury yields are influenced by inflation expectations, monetary policy, economic growth and numerous other forces.
It means the model identifies federal debt supply itself as a source of upward pressure.
Borrowing to Pay for Borrowing
That matters because interest is already one of the fastest-growing pieces of the federal budget.
CBO projects federal net interest expenditures will more than double, from approximately $1 trillion in 2026 to $2.1 trillion in 2036.
As a share of the economy, net interest rises from 3.3 percent of GDP to 4.6 percent. By 2036, CBO projects interest expenditures will nearly equal the entirety of federal discretionary spending. Those projections appear in CBO’s 2026 budget outlook.
That creates the makings of an unpleasant feedback loop.
Large deficits require additional borrowing. Additional borrowing increases the quantity of Treasury securities investors must absorb. Greater duration supply can increase the premium investors demand for holding longer-term government debt. Higher borrowing costs then increase federal interest expenditures, contributing to future deficits that must themselves be financed.
The government, in other words, can eventually encounter the fiscal equivalent of paying interest on the consequences of paying interest.
McGrane’s research also suggests Washington doesn’t necessarily have to wait for the additional debt to arrive before markets react.
The model examines what would happen if investors expected Treasury duration supply to increase five years in the future by an amount corresponding roughly to another percentage point of debt-to-GDP.
The estimated 10-year term premium rises immediately by about one basis point, then gradually increases until reaching the full 2.8-basis-point effect when the additional duration actually enters the market.
Shorter-term securities initially show essentially no response because they will mature before the expected increase occurs. As the future issuance approaches, however, the effect begins appearing in shorter maturities as well.
Put more simply: investors can start charging Washington today for debt they expect Washington to issue tomorrow.
It Isn’t Just How Much Treasury Borrows
There is another wrinkle hidden inside the paper that matters for Treasury debt management.
McGrane does not simply measure the number of dollars Treasury has borrowed. His measure incorporates both the amount of privately held debt and its weighted-average maturity.
That means Treasury duration supply can increase in two ways: Washington can borrow more money, or it can issue debt with longer maturities.
That turns the government’s choice between bills, notes and bonds into something more consequential than administrative housekeeping.
Short-term borrowing leaves Treasury more exposed to refinancing risk and changes in short-term interest rates. Longer-term borrowing can lock in financing for years or decades but forces private investors to absorb more duration.
Treasury is currently navigating that problem on a very large scale.
In August, the department estimated it would borrow $739 billion in privately held net marketable debt during the July-through-September quarter and another $628 billion during October through December. Treasury’s borrowing estimate is available here.
Its August quarterly refunding alone included $58 billion of three-year notes, $42 billion of 10-year notes and $25 billion of 30-year bonds. Treasury said the $125 billion package would refinance about $96.3 billion of maturing privately held securities while raising approximately $28.7 billion in new cash from private investors. Treasury detailed the refunding here.
Those maturity decisions are made through Treasury’s quarterly refunding process, where debt managers regularly reassess how the government’s enormous financing requirement will be distributed across the maturity spectrum.
McGrane’s findings suggest those choices can have consequences for the government’s eventual borrowing costs.
Then There Is the Federal Reserve
Complicating matters further is the world’s largest monetary-policy institution sitting on the other side of the Treasury market.
The Federal Reserve used large-scale purchases of Treasury securities after the 2007–2009 financial crisis and again during the pandemic precisely in part to remove long-duration securities from private investors and push longer-term rates downward.
McGrane finds that Federal Reserve purchases helped slow the growth in Treasury duration held by private investors, particularly during the early 2010s and around the pandemic recession.
The Fed has since ended its post-pandemic balance-sheet runoff. Beginning in early 2026, it resumed purchases of shorter-term Treasury securities as part of its effort to maintain ample bank reserves.
By July 1, the Fed held approximately $4.49 trillion in Treasury securities, an increase of $303 billion since balance-sheet runoff ended in December 2025. The Fed reported purchasing nearly $250 billion in Treasury bills during the first half of 2026, including reserve-management purchases and reinvestments of proceeds from maturing mortgage-backed securities. The Federal Reserve’s July 2026 Monetary Policy Report provides the balance-sheet details.
The distinction between bills and longer-term securities is important.
McGrane’s research is fundamentally about duration, not merely dollars. A dollar of three-month Treasury bills does not expose investors to the same interest-rate risk as a dollar locked into a 10- or 30-year security.
That means fiscal policy, Treasury debt management and Federal Reserve balance-sheet policy can all push on different parts of the same machinery.
There Are Important Caveats
The 50-basis-point estimate should not be mistaken for an official CBO interest-rate forecast.
McGrane’s paper is part of CBO’s Working Paper Series, which the agency uses to publish technical research and encourage outside review. The paper explicitly states that it has not undergone CBO’s normal review and editing process and that its conclusions should not be interpreted as CBO’s views.
There is also meaningful uncertainty surrounding the model.
McGrane’s baseline analysis excludes periods in which short-term interest rates were constrained near zero because the type of term-structure model used in the paper has known forecasting problems under those conditions.
When those zero-lower-bound years are included, the estimated effect becomes smaller.
Under the alternative specification, a one-percentage-point increase in Treasury duration supply raises the 10-year term premium by 0.42 basis points rather than 0.56. The modeled increase corresponding approximately to another percentage point of debt-to-GDP eventually produces about a 2.1-basis-point increase rather than 2.8 basis points.
McGrane argues the baseline may nevertheless be more useful for evaluating fiscal policy because the zero-rate period coincided with massive Federal Reserve Treasury purchases, making it difficult to isolate the effects of Treasury borrowing from monetary policy.
Nor can the 50-basis-point estimate simply be multiplied by the entire federal debt to produce a taxpayer price tag. Treasury securities mature and refinance at different times and carry different maturities, and McGrane is estimating the effect on the 10-year term premium, not a uniform half-percentage-point increase applied instantaneously to every dollar Washington owes.
The number is consequential enough without giving it mathematical assistance it didn’t ask for.
The Debt Has to Go Somewhere
The larger finding is more straightforward.
Washington does not borrow in a vacuum.
CBO expects the federal government to run trillions of dollars in annual deficits throughout the coming decade. Debt held by the public is projected to reach 120 percent of GDP by 2036, while annual net interest expenditures climb past $2 trillion. Those projections are detailed in CBO’s budget outlook.
Treasury must finance those deficits by issuing securities, and someone ultimately has to hold them.
McGrane’s research suggests that as Washington asks private investors to absorb ever greater quantities of interest-rate risk, those investors may demand increasingly greater compensation for doing so.
The debt, in other words, may not merely be getting larger.
The growing pile may be helping determine the interest rate charged on the next shovel-full.
