The Big Idea
The past and future path of federal interest expense
Stephen Stanley | October 25, 2024
This document is intended for institutional investors and is not subject to all of the independence and disclosure standards applicable to debt research reports prepared for retail investors. This material does not constitute research.
The recently released federal budget for FY2024 shows that the government paid almost $900 billion in interest during the year, an increase of over $200 billion, or 35%, from FY2023. The interest burden added up to 3.1% of GDP, the highest proportion of the economy since 1996. Concerns about the fiscal outlook have become a pressing issue in financial markets, especially during an election campaign during which both presidential candidates have offered agendas that would further expand deficits and debt. Fed easing may ease some pressure on interest expense, but pressure could ratchet back up within a few years.
Historical perspective
The federal debt has surged in recent years. From 1960 until the Global Financial Crisis in 2008, federal debt never exceeded 50% of GDP. Federal borrowing shot up in the years after 2008, reaching over 75% of GDP by 2016, after which it leveled off until Covid hit in 2020. The unprecedented federal government spending initiatives designed to counteract the effects of the pandemic pushed the debt-to-GDP ratio by almost 20 percentage points in one year. Since then, federal debt as a percentage of GDP has teetered just below 100%. The figure for the just-completed FY2024 was 98.1% (Exhibit 1). My budget projections would likely take the figure above 100% in FY2026 for the first time since the immediate aftermath of World War II in the 1940s.
Exhibit 1: Federal debt nearly equals GDP
Source: CBO, Treasury Department, BEA.
The era of zero interest rates helped to cushion the impact of this ascending debt. Borrowing rates remained quite low throughout the period from the Global Financial Crisis to the pandemic. From early 2009 until the pandemic, the average interest rate on marketable Treasury debt remained between 2% and 3% (Exhibit 2). Then, when Covid hit and the Fed returned its policy rate to the zero bound, the government’s average borrowing rate fell even further, breaking below 2% in 2020 and remaining there until late 2022. Even after the Fed pushed its fed funds rate target above 5% in 2023, the average interest rate for Treasury has remained below 3.5%, reflecting the trillions in long-term debt borrowed at ultra-low rates during the stretches of the Fed’s zero rate policy.
Exhibit 2: A rising average rate on outstanding marketable federal debt
Source: Treasury Department.
As a result, the absolute costs of ballooning federal debt were masked for years. Lately absolute costs have risen sharply (Exhibit 3). Note that debt was surging from 2007 through 2021 relative to GDP and yet the interest expense rose only modestly, thanks to the rock-bottom interest rates. Perhaps this helps to explain the temporary popularity of “Modern Monetary Theory,” which essentially argued that the federal government could borrow massive amounts of money to invest in the economy without concern about the swelling debt burden. This view sounded appealing, offering what amounted to a free lunch, and appeared to be borne out by the interest costs data right up until inflation broke out and forced the Fed to hike rates dramatically in 2022. However, the parabolic upward move in recent years in interest costs represents the vulnerability of a highly leveraged government to swings in borrowing costs.
Exhibit 3: Rising federal government interest expense
Source: Treasury Department.
Looking ahead
With federal government interest costs surging from almost $400 billion in FY2021 to nearly $500 billion in FY2022 to $666 billion in FY2023 and $896 billion, an extrapolation of the current trend suggests a serious fiscal problem.
Thankfully, with inflation coming down, the Fed has been able to begin cutting rates. Economists expect substantial rate cuts over the next year or two. Given the muted nature of average borrowing costs, a key question is how much of a difference will lower rates make for the federal government’s interest burden.
A first pass at answering this question is surprisingly simple. Assume for the sake of simplicity that the average borrowing rate for Treasury coupon securities remains steady (I will come back to this question later). This is not exactly correct, but it is a reasonable approximation.
That leaves the rolling over of Treasury bills as the main swing factor for borrowing costs. As of the end of September, bills totaled $6.004 trillion and floating rate notes constituted another $616 billion. Rounding to $6.6 trillion, we can easily calculate the sensitivity of interest expense to changes in rates. Borrowing $6.6 trillion at 5% would cost $330 for a year. At 4%, the interest expense would fall to $264 billion, and at 3%, the government would pay $198 billion in interest.
So, each percentage point drop in average short-term interest rates would save the government about $66 billion. Thinking about how the likely evolution of the FOMC’s fed funds rate target over the next year, assuming an average rate for Treasury bills of 4% for the 12-month period from October through September 2025 seems like a reasonable assumption, so rate cut should trim around $60 to $70 billion from the annual interest expense.
The other key variable relevant to the calculation is the projected borrowing need for the current fiscal year. Unfortunately, it looks like the deficit will remain close to $2 trillion. Just to keep the arithmetic simple, assume that the government will need to borrow $2 trillion to cover next year’s deficit. Assuming that the average borrowing rate over the next year is about 4% (we have just established that this is a reasonable assumption for bills and, given the current level of rates along the Treasury coupon curve, I would argue that 4% is as good an assumption as any), borrowing another $2 trillion would be an additional $80 billion in interest ($2 trillion times 4%).
Thus, a very rough set of assumptions suggests that the federal government’s interest expense over the next year will rise somewhat, reflecting the $80 billion carrying costs of net new debt minus the $66 billion in savings from lower short rates.
Applying that arithmetic forward another year, it seems reasonable to assume that short rates could decline again in FY2026. Most economists peg the endpoint of the current Fed easing cycle at somewhere in the neighborhood of 3%. If the average borrowing rate for bills and FRNs drops by another percentage point in FY2026, then the same result would fall out: interest expense would be slightly higher.
Just to return to the simplifying assumption for Treasury coupons, our back-of-the-envelope calculation is probably a bit too optimistic. The average interest rate on outstanding Treasury notes at the end of September was 2.73%, while the average for bonds was 3.20%. Most likely, the borrowing rate for new coupon debt over the next year will be closer to 4%. So, total interest expense from the rolling over of about $2.1 trillion in maturing coupon debt could add another $10 to $20 billion to the FY2025 interest expense (that range reflects new borrowing rates for coupon debt of 50 to 100 basis points, respectively, above the averages for all outstanding debt cited above).
Thus, in all, it feels like the interest expense may inch up modestly in FY2025, perhaps by around $25 billion or so. And we may see a similar modest rise in FY2026. This would be a temporary respite, achieved only by seeing a substantial drop in short-term interest rates. Once we get to FY2027, presumably, there will no longer be a tailwind from falling interest rates, and the interest burden will begin to rise more rapidly again, perhaps by close to $100 billion per year if interest rates are generally steady and annual budget deficits continue to run in the neighborhood of $2 trillion and by more if rates back up significantly.