The interest bill

The loan did not change. That was the part the Werribee couple found hardest to explain to their families. In 2021 a hospital theatre nurse and a warehouse supervisor fixed the bulk of their mortgage for two years at a rate starting with a one. They did not borrow another dollar after that. They did not miss a payment, refinance for a boat, or extend for a renovation. And yet when the fixed term rolled off in 2023, the monthly repayment on exactly the same debt was hundreds of dollars higher, because the fixed period had ended and the same principal now had to be carried at the price money cost in 2023 rather than the price it had cost in 2021.

Nothing about their behaviour explained the increase. The increase was the arrival, all at once, of a cost that had always been coming — the cost of holding debt that was borrowed cheaply and had to be re-priced later. The Commonwealth of Australia is now living through the national-scale version of that experience, and for the same reason. The Australian Government borrowed enormous sums during the pandemic at some of the lowest interest rates in the country's history. Those borrowings are now being refinanced at today's rates. The bill is arriving on debt that was already incurred.

The analogy is useful and it is also limited, and the limits matter enough to state up front. A sovereign government that borrows in its own currency is not a household. It has no bank that can foreclose, no fixed retirement date by which the loan must be extinguished, and no requirement to ever repay the principal in full rather than roll it forward. It can, in extremis, be financed by its own central bank in a way no family can. But the interest still has to be paid, in cash, out of the same budget that funds Medicare and the Australian Defence Force. That part of the analogy holds exactly.

What happened: a trillion dollars of debt, and a bill that arrives on all of it

In August 2026, the face value of Commonwealth Government Securities on issue passed one trillion dollars for the first time (Australian Office of Financial Management data; widely reported, Aug-2026). That is the gross debt figure — the total value of bonds and notes the Commonwealth has sold to investors and must eventually repay or refinance. It is not the same as net debt, which subtracts the government's own financial assets and is a smaller number, and it is not a measure of whether the budget is in surplus or deficit this year. It is simply the size of the stock.

Where that stock came from is a question this site takes seriously, because the honest answer cuts across party lines. The single largest addition to Commonwealth debt in modern history was made in 2020 and 2021, under the Morrison government, in response to the COVID-19 pandemic — JobKeeper, the coronavirus supplement, cash-flow boost payments for small business, and the tax and health measures that accompanied them. The Commonwealth deficit peaked at around $134 billion (approx.) in 2020-21, the largest in the country's history (Final Budget Outcome, 2020-21). That borrowing was supported by the Labor opposition of the day, criticised at the time mainly for being too small or too narrow rather than too large, and it is the direct ancestor of most of the interest bill discussed in this article.

So the distinction has to be drawn precisely. The current government did not cause the bulk of the debt stock; it inherited it. It did not cause the global rise in interest rates that has re-priced that stock; that was driven by post-pandemic inflation and by the tightening cycles of the world's major central banks, above all the United States Federal Reserve. What the current government does own is the management of the debt from May 2022 onward — how much it added, how much of a revenue windfall it banked rather than spent, and what it has done, or not done, about the structural gap that determines whether the stock keeps growing. Those are separable questions, and they get separate answers below.

The line this article is about has a formal name in the budget papers: public debt interest, or PDI, reported in Budget Paper No. 1, Statement 6 (Expenses and Net Capital Investment). It is the accrued cost of servicing Commonwealth Government Securities — the coupons paid to bondholders, plus the amortised effect of issuing bonds above or below their face value. On the published figures, PDI is around $27.1 billion in 2025-26, equal to about 3.5 per cent of the Commonwealth's operating expenses (Budget papers, 2026-27 edition). It is projected to pass $38 billion by 2028-29, and to reach approximately $42.2 billion by 2029-30 — about 4.7 per cent of operating expenses (Budget papers, forward estimates).

Across the medium term, Treasury's projections put the average annual growth of public debt interest at around 8.8 per cent a year over the decade. On the government's own numbers, that makes it the fastest-growing major spending item in the federal budget (Budget 2026-27; Parliamentary Budget Office medium-term projections). Not the largest. The fastest-growing.

What "fastest-growing" does to a number over time

Growth rates are easy to nod at and hard to feel, so it is worth doing the arithmetic in the open. A line growing at 8.8 per cent a year doubles in a little over eight years. Compounding is why the interest bill behaves differently from most of the budget: it does not need a new decision, a new program or a new election commitment to get bigger. It gets bigger by arithmetic, in the background, while the political system argues about everything else.

The near-term climb is steeper still than the decade average, and the published numbers say so. Going from roughly $27.1 billion in 2025-26 to about $42.2 billion in 2029-30 is an increase of around 56 per cent in four years — a compound rate closer to 11 to 12 per cent a year (author's calculation on the budget figures above). The decade-average figure of 8.8 per cent smooths that out because the later years of the projection assume yields eventually stabilise. The front half of the decade is where the re-pricing does most of its work, and the front half of the decade is now.

Translated into household terms — the way most people actually judge a number — an interest bill of $27.1 billion is roughly $2,500 a year for every Australian household, and $42.2 billion is roughly $3,900 a year per household (author's calculation using ABS household counts of around 10.9 million, approx.). No household writes that cheque directly. It comes out of the same tax revenue that would otherwise fund something, or out of further borrowing. Set against another line entirely: total defence funding for 2026-27 is around $63.4 billion (Budget 2026-27). By 2029-30, on the published projections, the Commonwealth will be spending about two-thirds as much servicing its debt as it spends on its entire defence effort today.

The evidence: why the bill climbs even in years when borrowing slows

This is the part that is almost never explained properly, and it is the analytical heart of the topic. Most people assume the interest bill goes up because the government is borrowing more. That is only half true, and in the years immediately ahead it is the less important half. The interest bill is rising mainly because debt already borrowed is being re-priced. Understanding the mechanism takes about four paragraphs, and it changes how the whole debate reads.

How the Commonwealth actually borrows

The Australian Office of Financial Management (AOFM) is the agency inside the Treasury portfolio that raises and manages Commonwealth debt. It does not take out a loan. It sells securities — mostly Treasury Bonds, which pay a fixed rate of interest (the "coupon") twice a year for a set term and then repay their face value on a set maturity date. A bond issued in 2020 with a 1 per cent coupon and a 2031 maturity pays that same 1 per cent every year until 2031, no matter what happens to interest rates in between, and then repays its face value in full.

That fixed-coupon structure is the key to everything that follows. It means the Commonwealth's interest cost in any given year is not the current market interest rate applied to the whole trillion-dollar stock. It is the weighted average of the coupons on every individual line of debt outstanding — a blend of very cheap old paper and much more expensive new paper. The AOFM publishes exactly this in its investor materials as the weighted average cost of debt on the portfolio, and that average has been rising steadily as the mix shifts (Australian Office of Financial Management, investor presentations and annual reports, various).

The AOFM also deliberately spreads maturities across many years — the weighted average term to maturity of the Commonwealth's bond portfolio has been maintained in the order of seven years (approx.; AOFM disclosures). That is a genuine risk-management achievement, and it is why Australia is not exposed to a sudden re-pricing of the entire stock in a single bad year. But it cuts the other way too: a long maturity profile means the re-pricing is slow, staggered and unavoidable. It arrives in instalments, year after year, and it cannot be cancelled.

The refinancing mechanism, in plain English

During 2020 and 2021, the Reserve Bank of Australia cut the cash rate to a record low of 0.10 per cent, set an explicit target for the three-year government bond yield (initially 0.25 per cent in March 2020, lowered to 0.10 per cent in November 2020, and discontinued in November 2021), and bought government bonds in the secondary market (Reserve Bank of Australia, monetary policy decisions, 2020-2021). Yields across the curve fell to the lowest levels ever recorded in Australia. The AOFM issued at enormous volume into that environment — Treasury Bond issuance in 2020-21 ran into the order of $200 billion or more (approx.; AOFM annual report, 2020-21). Much of that paper carries coupons in the order of 0.25 to 1.5 per cent (approx.).

Those bonds are now beginning to mature. When a bond matures, the Commonwealth must repay its face value to the holder. It does not usually do this out of tax revenue; it does it by issuing a new bond and using the proceeds — the process called refinancing, or rolling the debt. But the new bond has to be sold at the yield investors demand today. Australian government bond yields have spent recent years in the order of 4 to 4.5 per cent for longer maturities (approx.; Reserve Bank of Australia, capital market yields), rather than the sub-1 per cent levels of 2020-21.

So a billion dollars of debt that cost the Commonwealth roughly $10 million a year to carry at a 1 per cent coupon costs roughly $40 million a year once it has been refinanced at 4 per cent. The debt has not grown by a cent. The government has made no new decision. The nation has bought nothing new. The annual cost of that same billion dollars has simply quadrupled, because the fixed period ended.

The interest bill can rise sharply in a year when the government borrows nothing new, because the cost is arriving on debt that was already incurred.

This is why the projections show interest expense climbing even in scenarios where deficits narrow. Two separate forces push the line upward at once. The first is any new borrowing to fund a continuing deficit, which adds to the stock. The second — the one that operates regardless — is the steady replacement of cheap pandemic-era paper with expensive current paper as it matures. The second force is already locked in by contracts signed years ago. No budget decision taken in 2026 can undo it. It can only be paid.

Where the household analogy breaks down, honestly

The Werribee couple's experience is a true illustration of the mechanism, and it would be dishonest to leave it there without stating the differences, because a reader who later discovers them will rightly wonder what else was left out.

A sovereign government issuing debt in its own currency differs from a household in at least five ways that matter. It never has to repay the principal on a fixed schedule the way a mortgage amortises; it can roll debt forward indefinitely so long as investors will buy. It cannot be forced into default on its own-currency obligations by a creditor, because it controls the currency of settlement — which is the substantive reason ratings agencies treat Australian dollar debt as they do. Its "income" is a tax base attached to an economy that grows in nominal terms every year, so a debt stock that is merely stable shrinks steadily as a share of GDP without a dollar being repaid. Inflation erodes the real value of fixed-coupon debt, transferring value from bondholders to the borrower — a fact that is uncomfortable but arithmetically true. And unlike a household, the Commonwealth's borrowing can itself affect the size of the income supporting it, if the money is spent on things that raise long-run productive capacity.

None of those differences make the interest bill imaginary. The cash still has to be found each year from the same pool that funds everything else. The market still sets the price at which each new bond is sold, and no government can legislate that price. And the mechanism the couple experienced — a fixed rate ending, the same debt re-pricing upward — operates identically at both scales. The right conclusion is not that a sovereign faces no constraint. It is that the constraint is about opportunity cost and room to manoeuvre, not about solvency. That distinction is the difference between analysis and alarmism, and this article stays on the analysis side of it.

The trajectory, drawn

What the line buys

Every other large item in the federal budget purchases something a citizen can point to. The Pharmaceutical Benefits Scheme buys medicines. The Age Pension buys income for retirees. Defence buys ships, aircraft and personnel. Even lines that attract fierce argument — industry subsidies, consultancies, grants programs — at least deliver a good or a service that can be inspected and judged.

Public debt interest delivers nothing. It is a transfer from Australian taxpayers to the holders of Commonwealth Government Securities: superannuation funds, domestic and offshore banks, insurers, central banks and other institutional investors. That transfer is entirely legitimate — it is the contracted price of money the country genuinely needed and genuinely used — but it is not a service, and no Australian's life is improved by the payment itself. This is why the line deserves separate treatment rather than being folded into general "spending" debates. Arguing about whether a program is worth its cost is a meaningful argument. There is no equivalent argument available here.

It is worth saying clearly what this article is not claiming, because that boundary is where most public debate about government debt goes wrong. It is not claiming Australia cannot pay. It is not claiming a debt crisis is coming. It is not claiming bond markets are about to lose confidence in Commonwealth paper; the evidence, including a AAA rating from all three major agencies and consistently well-covered AOFM auctions, points the other way. The claim is narrower and, precisely because it is narrower, harder to dismiss: a rapidly growing, non-discretionary, service-free expense is displacing discretionary, service-delivering expenses, without any decision being taken by anyone.

The critique: the one line no minister can reprioritise

Consider what a treasurer can actually do in a budget round. Almost every large spending program has a lever attached to it. Eligibility rules can be tightened. Indexation arrangements can be adjusted. Prices paid to providers can be capped. Programs can be paused, means-tested, folded into other programs, or ended. These decisions are politically painful, sometimes brutally so, but they exist. They can be made, defended, legislated and measured.

Interest on Commonwealth Government Securities has no such lever. The obligation arises from securities already issued under long-standing Commonwealth borrowing law, at coupons already contracted, held by investors who bought them in good faith. There is no eligibility test to tighten and no provider price to negotiate. The Commonwealth pays what it agreed to pay, on the dates it agreed to pay it. In the language of the budget process, it is a first-claim expense: it comes off the top, before the contest for everything else begins.

The comparison that lands

The sharpest way to see the problem is to set the interest line beside the program this government has worked hardest to slow. In April 2023, National Cabinet agreed a target of moderating National Disability Insurance Scheme cost growth to 8 per cent a year by 1 July 2026, down from growth that had been running considerably higher — at times well into double digits and in some years past 20 per cent (approx.) — a target examined in detail in a separate audit on this site. Bending that curve required an independent review, a National Cabinet negotiation with every state and territory, contested legislation, and a sustained and genuinely difficult political effort against real human stakes.

Public debt interest is projected to grow at around 8.8 per cent a year over the decade. On those two numbers, the interest bill is projected to grow faster than the ceiling the government negotiated for the scheme it has fought hardest to restrain — and unlike the NDIS target, there is no review, no National Cabinet, no legislation and no lever available. The comparison should be made carefully: the two figures come from different exercises, the NDIS number is a policy target rather than a forecast outcome, and the NDIS remains far larger in dollar terms. But the structural point survives every one of those qualifications. Enormous political capital was spent to bend one curve to 8 per cent. Another curve is bending itself upward at a similar or faster rate, and no amount of political capital can touch it.

Crowding out, with nobody choosing it

The word economists use for this is crowding out — a fixed claim on a finite budget squeezing out other uses. What makes the interest line distinctive is that the crowding out is silent. There is no announcement, no press release, no parliamentary vote and no minister who owns the decision, because there is no decision. Between 2025-26 and 2029-30, public debt interest rises from about 3.5 per cent of Commonwealth operating expenses to about 4.7 per cent (Budget papers). Put another way, the share of every dollar of federal spending consumed by interest moves from roughly one dollar in twenty-nine to roughly one dollar in twenty-one.

That shift of a bit over one percentage point of total expenses sounds modest, which is exactly the danger in it. Expressed in money, it is the difference between $27.1 billion and $42.2 billion — around $15 billion a year of additional first-claim expense by the end of the decade, on the published projections. Fifteen billion dollars a year is a serious sum in any portfolio. It is the kind of money that funds a substantial health measure, a meaningful lift in defence capability, or a visible tax cut. It will instead fund coupon payments on bonds, and it will do so without appearing as a single decision in a single budget that anyone can be held to account for.

Every other budget line has a lever attached. This one has a contract.

The deeper point: the price of no buffer

The most serious version of the critique is not about this year's $27 billion or even next decade's $42 billion. It is about what those numbers do to the country's capacity to respond to the next emergency.

The reason Australia was able to respond to COVID-19 at the scale it did in 2020 is that it entered the pandemic with comparatively low public debt and a low interest burden. That fiscal space was not an accident and it was not free — it was the accumulated product of decades of budget positions, across governments of both persuasions, that left room on the balance sheet. It was spent, appropriately, on a genuine emergency. It has not been rebuilt.

The next shock is unknowable in its nature and its timing — a pandemic, a regional security crisis, a severe global recession, a sequence of natural disasters, or something not currently on anyone's list. What is knowable is the arithmetic of the starting position. A country entering a crisis with a trillion dollars of debt and a rising, non-discretionary interest bill has less room than a country entering it with half that. It does not have no room; Australia's debt-to-GDP ratio remains low by peer standards and its AAA rating is intact. But room is a matter of degree, and the degree has narrowed.

Attribution matters here too, and it should be stated with precision. This government did not create that narrowing; it inherited most of it. What it presided over was five years in which the structural gap between ongoing spending and ongoing revenue was not closed — a gap examined in full in a separate audit on this site — which means the debt stock has not been placed on a declining path in nominal terms. What it can fairly be held to is the counterfactual: on the record, its decisions reduced the trajectory relative to what it would otherwise have been. That is a real defence, and the next section gives it its full weight.

What they got right, and the honest caveats

The single most important thing this government did about the interest bill was not a debt policy at all. It was a decision about what to do with money that arrived unexpectedly.

Across its first budgets, the government received extraordinary upgrades to expected tax receipts, driven by high iron ore and coal prices and a labour market running at near-fifty-year-low unemployment. It banked the large majority of those upgrades to the bottom line rather than matching them with new permanent spending. In the October 2022 Budget the government stated it had banked around 99 per cent (approx.) of the receipts upgrade since the pre-election fiscal outlook (Budget Speech, 25 October 2022; Budget Paper No. 1, Statement 3, October 2022-23). The share banked fell in later budgets as new commitments absorbed more of the improvement, but the overall pattern across the period was one of materially more restraint than the alternative of spending the windfall in full.

That decision has a direct and measurable effect on the subject of this article. Every dollar of windfall revenue banked rather than spent is a dollar the AOFM did not have to raise by issuing a bond, and therefore a dollar on which no coupon is ever paid, in perpetuity, until that bond would have matured. The two consecutive underlying cash surpluses that followed — $22.1 billion (approx.) in 2022-23 and $15.8 billion (approx.) in 2023-24, per the Final Budget Outcomes of September 2023 and September 2024 — reduced planned issuance below what the pre-election projections had assumed. Gross debt is a trillion dollars today. Absent that restraint, it would be higher, and so would the interest bill in every year of the projection. This must be said plainly, because it is true and because the critique above is only credible if the credit is given first.

Three further points belong on the government's side of the ledger.

First, debt management. The AOFM's practice of maintaining a long weighted average term to maturity — in the order of seven years (approx.) — and of issuing across a wide range of maturities means Australia is not exposed to a cliff-edge refinancing event. Other sovereigns with shorter portfolios re-price far more of their stock in any given year. This is an operational strength, built up across administrations of both parties, and it is doing real work in dampening the very mechanism this article describes.

Second, yields. The single largest driver of the rising cost of new issuance is the global level of interest rates, which followed post-pandemic inflation and the tightening cycles of major central banks, and which is anchored by the United States Treasury curve. This is not a domestic policy variable. It is a condition Australian governments of any stripe operate inside. Blaming a treasurer for the yield on a ten-year bond is a category error.

Third, the counterfactual on the original borrowing. The pandemic borrowing was not a discretionary indulgence. It was an emergency response, supported across the parliament, and by the standard of what it prevented it was cheap. A government that had chosen a smaller response in order to protect its interest line in 2029 would have been making a straightforwardly worse decision, and the people who paid for it would have been workers and small businesses in 2020, not bondholders in 2029.

The caveats on the credit are equally honest. The banking of the revenue upgrades slowed after the first budget. The surpluses were driven substantially by commodity prices and employment rather than by structural repair, and they ended. Gross debt has continued to rise through the period, passing the trillion-dollar mark in August 2026. And restraint that reduces a trajectory is not the same as restraint that reverses it. On the published projections, the interest line rises in every year of the forward estimates. The government's decisions made it rise more slowly than it otherwise would have. They did not make it fall.