So you thought that watching defence spending was complicated. Well, Australia’s new National Defence Strategy displays three different ways of counting it.
There’s the Defence portfolio’s appropriation from government; the appropriation from government plus Defence’s own-source revenue; and, newly introduced, Defence’s interpretation of the NATO measure of defence spending. These numbers differ from one another considerably. While each serves a purpose, use of all three hasn’t made the National Defence Strategy (NDS) a markedly more transparent document. This article examines each as it appears in the NDS, with a view to informing future analysis.
The government’s measure of headline appropriation
Table 1 on page 85 of the NDS (reproduced above) presents the new ‘10-year funding model’ for Defence. The table tells us how much the government plans to appropriate for the core agencies in the Defence portfolio in each of the future financial years indicated. The core agencies are Defence itself (the department and services), the Australian Signals Directorate (ASD), the Australian Submarine Agency (ASA) and the Australian Naval Nuclear Power Safety Regulator (ANNPSR). When the Defence Delivery Agency is spun off from Defence in 2027, it’ll almost certainly have its appropriation fall under this total as well.
This is the source of the headline figure of A$887 billion of total spending over the decade. It’s also the measure used for long-term spending plans since the 2016 Defence White Paper. You can think about this as defence spending as seen from Cabinet: it’s the part of that year’s tax revenue and borrowing that the government is willing to give Defence legal authority to spend.
Defence’s measure of planned expenditure
In practice, Defence can always spend a little more than what the government gives it. Defence, like many government agencies, collects money by providing goods and services to third parties – selling off assets, charging rent, that sort of thing. This adds roughly A$1 billion a year on top of the appropriation from the government. You can think about this combined figure as defence spending as seen from Russell: the total amount of money that becomes available in a given financial year for Defence and its other portfolio agencies to use to do their jobs.
That means paying the workforce, sustaining existing capabilities, acquiring new capabilities, and day-to-day activities. Table 3 in the NDS Budget Factsheet (reproduced below) presents the new 10-year funding model in terms of the government appropriation plus Defence’s own-source revenue (OSR). This is why it asserts otherwise mysterious 2035–36 expenditure of $113.2 billion while the headline funding for the period set out in the NDS provides only $112.1 billion for that year. The gap is covered by money that Defence is expected to make.
NATO’s measure of defence spending
Finally, love it or hate it, there’s the NATO measure. The NDS just once refers to the bespoke method NATO uses to calculate and compare defence spending among its member states. Using the NATO measure, it says, ‘Defence funding is now projected to rise to approximately 3.0 per cent of GDP by 2033–34′. Unlike the headline appropriation or headline-plus-own-source revenue, which are well-defined legally, the precise way in which Defence has interpreted NATO’s measure is unclear. This is partly because, since the government began to mention spending according to NATO’s measure of defence spending in September, Australia’s NATO-definition defence spending has never been quoted as a dollar figure. (NATO, by the way, provides a comprehensive annual dataset of its member states’ defence spending in both local currency and US dollars.) As a GDP ratio, it is thus at the mercy of a fickle denominator.
We don’t know what exactly the government now thinks Australia’s GDP will be in 2033–34. Let’s be conservative and assume that nominal GDP growth is 5.25 percent in 2025–26 (as the latest Treasury predictions say) and then ranges between 3 and 5 percent a year. That would put nominal GDP for 2033–34 in a range between A$3.71 trillion and A$4.31 trillion. Then 3.0 percent gives us a hypothetical NATO-definition defence spending of A$111.0 billion to A$129.5 billion. But NDS 2026 plans A$106.7 billion for the headline appropriation in 2033–34, leaving a shortfall of A$4.3 billion to A$22.8 billion.
Explaining that gap is in fact not too difficult, but we may still regret the lack of a clear public account from the government of how it’s produced its numbers. The best we have is a high-level explanation provided by Defence’s chief finance officer, Steven Groves, who was asked to clarify the matter at Senate Estimates on 9 October 2025. He replied as follows:
At a high level, [the NATO measure] includes the additions over and above the funding to [Defence, ASD, and ASA]—the direct funding. It includes the full cost of military retirement superannuation. That’s currently reflected as an administered expense within the Defence financial statements and [portfolio budget statements (PBS)]. It includes the full cost of Defence housing; it includes naval infrastructure investments, done, for instance, by the [Australian Naval Infrastructure] organisation; it includes pensions and compensation payments to veterans that are funded out of the Department of Veterans’ Affairs, DVA, but doesn’t include medical costs; it includes some elements out of the Australian Border Force to include what is considered a US Coast Guard equivalent; and, finally, it includes some elements of the Australian intelligence community.
His explanation might not be current. Groves mentioned Australian Naval Infrastructure (ANI) – the government-owned company which partly manages Osborne Naval Shipyard – as specifically included in the NATO calculation. But ‘investment in Australian Naval Infrastructure for the Osborne Naval Shipyard’ is specifically excluded from the additional investment contained in the 2026 NDS. Does that mean all or part of ANI is excluded from the funding added to the NDS headline appropriation, but included in the NATO calculation? Or has the NATO calculation been updated since last October, when Groves indicated it might be? We don’t know.
Still, the money to fill the gap is probably there. A brief look at the last PBS indicates the ‘full cost’ to government of Defence’s payouts for military superannuation is roughly A$4.5 billion, and financial statements of the Department of Veterans Affairs’ show another A$9.1 billion being paid out in personal (non-medical) benefits over 2025–26. This is all what’s called ‘administered’ funding, meaning the government agency that spends it doesn’t get to control what it’s spent on – which is why Defence’s administered spending isn’t counted as part of the headline appropriation. It would be useful to find in the next PBS a table setting out exactly how this calculation is done.
While the least clear in composition, the NATO measure is currently the figure with the greatest political importance. The White House continues to exhibit a strong interest in seeing US allies increase their defence spending. The target set by NATO – 3.5 percent of GDP on defence spending by 2035 – uses NATO’s bespoke metric. It’s only fair that, if America’s NATO allies are being held to an arbitrary standard, its non-NATO allies should be held to the same arbitrary standard. Whether Australia can meet it is another question. Achieving 3.0 percent by 2033 is still quite a march from achieving 3.5 percent by 2035.