Think of interest as just another government spending program

This piece covers government interest payments, and US government interest payments in particular, with the hope of offering the reader a fundamentally new way of understanding them. I will argue the payment by government of interest (on liabilities denominated in its currency) is best thought of as a government spending program, conceptually similar to Social Security. Like Social Security, the "interest payment program" as I will call it here, has an open-ended budget. Currently, the government spends through this program at a rate of a little over $1T per year.

First, I will cover how government spending programs work in general. Then I will outline how interest payment works operationally, and why the most straightforward approach is to understand the government's payment of interest as being like any other government spending program. Then I will outline how economists, politicians, and the media talk about about government interest payments. Finally, I will share some brief thoughts on why this program persists and conclude.

How do government spending programs work in general?

Every government has a lot of spending programs. In the US, in no particular order, we have Social Security, Medicare, Medicaid, military research programs, the armed services, programs to fund and organize scientific research, the Supplemental Nutrition Assistance Program (SNAP), the private health insurance subsidies associated with the Affordable Care Act (Obamacare), foreign aid programs, foreign diplomacy programs, war manufacturing, education-related programs, and thousands of other programs with spending components. The way these programs are authorized in the US is that Congress passes a bill. That bill (1) describes the program, (2) delegates its implementation to some government authority, and (3) sets the parameters of its budget. Some programs require the budget to be renegotiated annually, some allot a budget for a longer period, and others provide a permanent budget that extends into the future. Some budgets are specified in absolute dollar terms, while others are specified in terms of a formula. Every category of spending has specific instructions attached to it outlining what it can, or in many cases must, be spent on. The executive branch carries out the program but is bound by the constitution to "take Care that the Laws be faithfully executed", which includes strictly abiding by the budget parameters of the program: the executive may spend neither more nor less than the law prescribes on each program.

So for example, we have the National Science Foundation (NSF), which (1) fosters scientific research through grants and other activities, (2) is its own independent agency, and (3) has an annually negotiated budget that prescribes dollar amounts for each category of programming. This is termed "discretionary" spending, because the dollar amount is annually negotiated at the discretion of Congress, typically after the President proposes an annual budget.

On the other hand we also have programs like Social Security, which (1) provides pensions (recurring cash transfers) to seniors, families after the death of an income-earner, and the disabled, (2) is administered by the Social Security Administration, also an independent agency, and (3) has a budget that is open-ended in terms of total annual dollar amount, but specifies instead the terms and conditions under which spending is authorized: the more people who qualify, the higher the dollar amount will end up being spent through Social Security. Programs like this are called "mandatory", because the bills that created the program have permanent budgets that continue into the future with no additional annual action required from Congress.

How does spending on the interest payment program work?

Similar to how the administration of science funding and Social Security are delegated to independent agencies of the US Federal Government, the "interest payment program" is also delegated to an independent agency, the Federal Reserve System, commonly called just "the Fed". There are some caveats here: the Federal Reserve System has a complicated structure, and technically speaking only the Federal Reserve Board, which directs the twelve Federal Reserve Banks and the seven board members of which comprise a majority on the Federal Open Market Committee (FOMC), is an independent agency of the US Federal Government. These other subunits are all government instrumentalities and are part of the overall System, which in nearly every aspect acts as a single organization, and for the purpose of this discussion not much is lost by speaking of the entire System—"the Fed"—as an independent agency.

The job of determining government policy concerning the interest rates paid on most government liabilities is delegated to the Fed. This is the key set of policy decisions that determine the overall level of spending on the interest rate program.

Interestingly, there are some government liabilities on which interest rate policy is prescribed directly by law or delegated to a different authority, and thus falls outside of the policymaking purview of the Fed. The most salient example is coins and paper currency. Congress has permanently pegged interest on these at 0%—indeed, it's kind of baked into the instrument. Less known examples are Series I and Series EE Savings Bonds, which are non-marketable securities meant as safe savings instruments for the everyday public. The interest rates for these are straightforwardly set by the Treasury, a government department. Combined, however, these categories of instruments that the Fed does not manage interest rate policy on is a fairly small portion of the government's total issued liabilities.

Most of the government's liabilities fall into one of two categories: Treasury securities ("Treasurys"), which are issued by the US Treasury, and exchange settlement balances ("reserves") issued by the Fed. A mix of law and convention have the Treasury issue new securities in lock step with its net spending, as well as to replace maturing securities. The Fed then choses how many of those securities to pull out of circulation by purchasing them from the open market, effectively replacing them with the Fed's own zero-duration interest-bearing government liability, exchange settlement balances called "reserves" or "reserve balances"; indeed, this is the main way that "reserves" come into existence. Similarly, the Fed can do the reverse of this operation and replace reserves with Treasurys again by selling previously purchased Treasurys back into the open market. (Don't get confused by the word "reserve"—it's just the name of a particular US government liability issued by the Fed and there's no further underlying meaning to it.)

In the case of reserves, the Fed's zero-duration government liability, the Fed both issues them and directly sets the interest paid on them in a straightforward manner, via public announcements from time to time. For example, yesterday December 18, 2024, the Fed announced it would pay 4.4% annual interest on reserve balances starting the next day. The Fed pays this interest by directly issuing new reserves, though from an overall government accounting and macroeconomic standpoint, it doesn't really matter which government subunit technically makes the payment.

Most of the interest paid through the interest payment program, though, is paid by the Treasury on securities also issued by the Treasury. Interest rate policy on these Treasurys is still the purview of the Fed. In its current framework, the Fed has chosen to directly set only the very short term interest rate via announcements from time to time, and for the most part takes a hands-off approach to rates paid on these longer term government liabilities by market participants. This effectively sets the Treasury market up as a prediction market in which market participants bet on the Fed's future announcements of the very short term interest rate. This framework is not prescribed by law, and indeed the Fed has adopted various frameworks over its history.

Thus, the Fed can operationally and legally set the interest rate paid on government liabilities of all durations. (The only major constraint is that it is not legally permitted to set a negative interest rate on reserve balances.) More broadly, it decides the policy framework by which government liabilities of all durations are set. For example, in the current framework, the Fed directly announces the very short term rate and then has relatively wide ranges in which it likes to see the level of interest paid on Treasurys of each duration, and when market rates fall outside of these ranges, such as in a moment of financial instability, it uses its ability to retire and "unretire" Treasurys of various durations to manage rates back to ranges it finds appropriate. This is a common framework which many other central banks also employ some versions of. Other central banks have ended up using quite different frameworks. Japan, for example, in 2016 shifted to a framework called yield curve control in which it really does just announce rates (or narrow ranges) along the entire maturity spectrum and then relatively effortlessly keeps them there.

Going back to first principles, of course, Congress could revoke this policymaking authority from the Fed entirely and simply legislate various rates, or assign the authority to a different body, or stop issuing Treasury securities and let them all roll over into zero-duration government liabilities instead (like reserves), or any number of other permutations. In this sense, it is really Congress that ultimately decides how interest rates are set, which instruments are interest-eligible, and who decides how much of each instrument to issue. Historically we've seen many of these permutations. There are plenty of historical examples in which governments, if they paid interest at all, simply paid a pre-specified level of interest that was hard-coded into law or other long-term policy. At the time that the Bank of England was created, for example, the law that created it effectively specified an interest rate of 6%. During World War II, under essentially the same legal structure we have today, the Fed set interest rates for terms up to a year at 0.375% ("three eighths of one percent") and directly set the entire yield curve. In recent decades, a number of central banks have set (very slightly) negative interest rates for sustained periods of time, including those in Europe, Japan, and Australia. In the extreme, it's certainly within Congress's (or even the Fed's) power to eliminate the payment of interest on new securities entirely. All of these are policy choices open to the government.

This is why I refer to this complex system of payment of interest on government liabilities as simply the government's "interest payment program". It's not conventional naming, but it describes what's actually happening much more clearly and neutrally. The interest rates paid on government liabilities are to the interest payment program what, say, monthly entitlement amounts are to Social Security: within an overall uncapped annual budget, they are the key policy-determined variable that sets the total level of spending. In fact, Social Security and the interest payment program are actually substantially similar. They are both programs for making regular payments to a specific segment of society: loosely speaking, old people and rich people, respectively.

What's the traditional way to talk about government interest and why is it lacking?

The traditional way to talk about government interest payments frames them as something the government is forced to do, something it operationally can't not do. In this framing, a government taxes its people and collects that money in a big pile. Then it goes about spending that money. And then as the big pile of money is dwindling, it panics, and goes around to banks, rich people, and other countries' governments to borrow their money so that it can continue to spend. It can shop around for the best rate, but at the end of the day it's not in a position to negotiate: it has to accept whatever the best rate is that the market has to offer.

The main error in this story is the error of projection: we are imagining that the government operates this way because we operate this way. In this telling (1) money is an outside entity created by forces outside our comprehension or control, (2) to spend money we have to first have money, (3) if we can't get money but still want to spend money, we need to borrow money, and (4) if we can find anyone who will lend us money at all, they—not we—determine the rate of interest we will pay. You could argue this logic applies to a government when it borrows currency issued by other governments, but this logic simply does not apply to a government operating within the very monetary system that it oversees. The dollar itself is a government liability, the same type of thing that the US government's "debt" (Treasury securities) is. Every dollar that the government "borrows", it first issues: there is no where else that dollar could have possibly originally come from. It "borrows" by swapping one of its liabilities (reserves) for another (Treasury securities). The operational reality bears no relation to the story.

Regardless, in the conventional framing, the level of interest that the government must pay is "set by the market". Some also speak of the government occasionally "intervening" in the market to "artificially control" interest rates. This is in a sense a small step in the right direction, because while it needlessly problematizes it, it at least acknowledges the government's policy discretion over interest rates. The fairly deep problem with this framing is that it is an irreconcilably incorrect description of the actual operations that determine how much interest a government pays, described in the previous section. In particular, while the government can select the short term interest rate and leave the long term interest rate up to the market as it currently does, it essentially has no choice but to set the short term interest rate, because there is no other meaningful place for that rate to come from. To illustrate with an example, in the 1970s the Fed adopted a different framework heavily influenced by a contemporary fad in the economics profession and attempted not to set even the overnight rate, but they quickly realized they weren't really able to not set it: when banks had a regulatory need for more reserves than the Fed was giving them, the Fed basically had to lend them reserves at some rate—the alternative was to intentionally cause pointless mass failures of perfectly solvent banks—and there's no one who could possibly decide what rate the Fed would lend reserves to banks at other than the Fed. As alluded to in this Bank of England explainer, academic theories that posit that the government determines some money supply and the market determines the overnight interest rate based on supply and demand for money are disconnected from the operational realities of actually-existing money systems.

Economists and policymakers who have taken the time to understand central bank-Treasury operations and have moved past the childish understandings of money that still pervade these professions do exist, but they often simply retreat to another line of argument to defend essentially the same framing. Yes, they might say, the government can peg its whole yield curve to 0% or 5% or any other value it choses, but letting the central bank determine interest rate policy is an essential part of how governments maintain low inflation, therefore the government really must behave as if it had no control over interest rates. In this view, there is always some unobservable level of interest that will result in stable inflation; any level below this will cause rising inflation and any level above this will cause falling inflation. This is very much economic and political orthodoxy.

There are two problems with even this story though.

One is that economists tend to way overstate to the public how effective changing the short term interest rate from time to time is as an inflation management policy. Because this tool's actual track record at hitting inflation targets, both in the US and globally, is actually quite poor, economists tend to fall back on counterfactuals about how much worse inflation would be without it. It's hard to find evidence for this either. All of the world's worst inflations and hyperinflations have happened while central banks were free to set their interest rates, and typically while they were setting them quite high. Meanwhile, attempts to raise inflation and promote spending by keeping interest rates near zero all along the maturity spectrum in the US for about a decade (and about 30 years in Japan) coincided with inflation deemed by economists to be too low. It is not even clear that interest rates can't have the opposite of their theorized effect. Indeed, most people experience higher interest rates as inflation, since it directly results in higher monthly payments on a car or house, for example. The relationship between interest rates and measured inflation is an interesting theoretical and empirical question, which I may explore in future pieces, but it is a complex topic with significant ambiguity.

The other problem is that even if interest rates are a potential tool for managing inflation, the argument for treating interest as something government is essentially forced to pay requires not just that that they are a potential tool, but that they are the only effective tool. But that is silly on its face. There are many tools the government has at its disposal, from its budgeting process, to maintaining large distributed commodity buffer stocks, to adopting a price rule rather than quantity rule on a portion of its spending, to macroprudential credit regulation, to housing and healthcare policy, to tax policy, to coordinated procurement, to any number of other tools it already has.

To frame the interest payment program as being forced on the government by the market, then, requires not just a belief in current economic orthodoxy, but in a particularly extreme version of that orthodoxy that says that not only are interest rates a potential tool but (1) they are a really good and powerful tool and (2) they are the only workable tool. Any deviations from this double hardline immediately make the framing substantially dishonest. Many people would argue that ending Social Security or Medicare or science funding would be quite bad—even catastrophic—but no one speaks as if "the market" is "forcing" the government to maintain these programs or that "the market" is determining Social Security pay rates. Everyone regardless of their opinion of these programs agrees that they are government programs that exist at the pleasure of Congress. They argue in (relatively) honest terms about the costs and benefits of these programs to society, how best to structure them, and what appropriate pay rates are. This is how we should argue about the interest payment program, in honest terms, about its benefits and drawbacks, not in terms that attempt to recast it as something entirely outside of a government's conceivable policy space.

Why does the interest payment program exist and persist?

There are two main reasons that the interest payment program might persist.

I'll start with the really simple one. Just like most old people like Social Security, most rich people like the interest payment program. Banks like it, because (by regulatory necessity) they have more assets on which they earn interest than they have liabilities on which they pay interest, and moreover, when rates are higher overall they tend to manage to keep a higher spread between the rate they pay and the rate they are paid, which is nice for them. Companies with large cash reserves (typically Treasurys) also like it. And rich people in general like it even if they own hardly any Treasurys because the returns on all investments have to compete with Treasurys, so they can demand higher returns from the companies they own shares in as well. Larger companies that have significant debt are generally able to pass on higher borrowing costs to their customers in the form of higher prices—especially because their competitors are also facing the same higher borrowing costs—so their management and shareholders tend to find it still overall a good bargain.

The other potential reason that I have to keep an open mind to is that maybe the interest payment program is useful and effective. Maybe it is a good tool for managing inflation, and our society has made it sacrosanct and rhetorically cordoned it off as a wise convention. This is certainly the story told by economists. And it's certainly true that economic theory suggests this policy, that it is a sacrosanct social value (among a certain class of people at least), and that government operational practice implements it. Economists explain the causality among those three facts this way: economic theory shows what a great idea the interest payment program is objectively, which causes it to be sacrosanct, which causes it to be implemented by government. In that story, economists are impartial scientists who have discovered a universal truth and help the government technocratically implement what is best for society. My own best stab at the causality is somewhat different: the fact that it is sacrosanct among a certain class of people has profoundly influenced both economic theory and government practice. My own (admittedly limited) understanding of the history of economics, the history of the development of monetary policy, and the history of inter-elite struggle within the US align better with this interpretation.

Conclusion

This piece argues that the government interest payment program is a government spending program like any other. The only question is, just as with any other government program, are the societal costs worth the societal benefits? Plenty of people will continue to argue "yes". My own opinion is that that is far from clear, and that the interest payment program probably just sticks around for the same reason the US's large corn subsidies stick around decade after decade despite widely acknowledged overproduction of inedible corn: some mix of habit and the fact that the people getting the checks sure do like getting the checks. I know I do!