Governments rarely spend less. What they do instead is quietly stop paying for some things so they can start paying for others. In November 2025, Canada did this on an unusually large scale — and published enough detail that you can see exactly what was traded for what.
The federal government collects about $511 billion a year and spends about $547 billion. The gap — roughly $36 billion in the most recent completed year — is the deficitThe amount a government spends beyond what it collects in a year, covered by borrowing. Different from the debt, which is all past deficits added together., covered by borrowing.
Those are the numbers a budget moves around. When you hear that a programme is being cut by $200 million, that is roughly one twenty-seven-hundredth of annual spending. Almost nothing in a budget is large relative to the whole. What matters is direction and pattern, not size.
It also helps to know that spending divides into two kinds, because the entire story here turns on the difference.
Operating is the cost of running things day to day: salaries, office space, consultants, software licences, grants handed out each year. If you stop paying, it stops immediately.
Capital is money spent on things that last: ports, transit lines, military equipment, housing. A bridge costs a fortune to build and then serves people for forty years.
Economists generally treat borrowing for capital as more defensible than borrowing for operating, on the reasoning that you are buying an asset rather than covering a shortfall. Whether a given item is genuinely capital is often where the argument lives.
One more piece of vocabulary, because you will trip over it otherwise. Governments count money two ways. CashWhat actually leaves the government's bank account in a given year. is what leaves the bank this year. AccrualSpreads the cost of a long-lived asset across the years it is used, rather than booking it all in the year it is paid for. spreads the cost of something long-lived across the years you use it — so a billion-dollar bridge lasting forty years appears as roughly $25 million a year, not a billion up front. Both are honest. They answer different questions. The budget uses both, and switching between them without noticing is the most common way to misread it.
Take roughly $60 billion out of running the government over five years, and put roughly $280 billion into building things.
The government's own phrase for this is "spend less to invest more." The savings exercise was called the Comprehensive Expenditure Review — every department was asked to find cuts, and the results were published inside the budget.
So this is not austerity in the usual sense. Total spending goes up. What changes is the mix: less on operating, far more on capital. The budget states the target plainly — growth in day-to-day programme spending falls from about 8% a year over the past decade to around 0.5% a year. Meanwhile capital investment nearly doubles, from $32.2 billion in 2024–25 to $59.6 billion in 2029–30.
The budget's stated reason is trade. The United States has been raising tariffs and reorganising its supply chains, and Canada sends most of its exports there. The budget describes this in unusually blunt language: "This is not a transition. It is a rupture."
The response has two halves. Build the physical capacity to sell to other countries — ports, rail, corridors — and cut the cost of running the government to help pay for it. Whether that diagnosis is correct is a political judgement. What follows is what it produced.
That distinction matters more than almost anything else here. Roughly $6.95 billion of the $12.84 billion in 2028–29 savings comes from the government spending less on itself: fewer managers, less office space, fewer outside consultants, more automation. Programme reductions are about $5.4 billion. Public service headcount falls too — reporting on the budget put it at roughly 40,000 fewer positions than the 2024 peak, though the budget itself describes it as bringing the public service back in line with population growth rather than giving a number.
In July 2025, the Minister of Finance and the President of the Treasury Board sent a letter to every minister. Find savings, they said, and gave escalating targets: up to 7.5% of your budget in 2026–27, up to 10% the year after, up to 15% by 2028–29. Proposals were due on 28 August. Ten weeks later the results appeared in the budget.
But not every department got the same target. Seven were protected at 2%, and the budget gives a written reason for each. This is the most revealing part of the whole exercise, because it is a government ranking its own priorities in public, in its own words.
| Target | Who | Reason given |
|---|---|---|
| 2% | National Defence, Canada Border Services Agency, RCMP | "Essential mandate to protect our sovereignty and secure our borders" |
| 2% | Indigenous Services, Crown-Indigenous Relations | "To protect services to Indigenous peoples" |
| 2% | Women and Gender Equality Canada | "To protect the essential mandate" |
| 2% | The three federal research funding agencies | "To protect the essential role of the federal research granting councils" |
| 15% | Everyone else — Global Affairs, Environment and Climate Change, Health, Natural Resources, Immigration, Transport, Veterans Affairs, Statistics Canada, Parks Canada, Fisheries, Heritage, and the rest | No individual reason given. 15% was the default. |
The budget says "up to 15 per cent," not "15 per cent." Departments proposed their own savings against that ceiling, so few will land exactly there. When this page says a department was "in the 15% group," it means the ceiling it was handed — not the cut it took.
Before the chart, one thing has to be corrected, because it would otherwise mislead you badly.
Two of the largest-looking lines in the budget are not cuts at all. The RCMP appears to be losing $1,521.6 million a year by 2028–29 and Veterans Affairs $1,064.0 million. In fact these are actuarial adjustmentsA revised estimate of what an organisation will owe people decades into the future — here, pensions and medical benefits. Changing the estimate changes the accounting today without changing what anyone is paid this year.: revised estimates of future pension and medical-benefit obligations, triggered by changes such as reimbursing medical cannabis at $6.00 a gram instead of $8.50. The budget's own footnotes give the real cash effect as $50.8 million and $110.0 million.
Left uncorrected, the RCMP would look like the biggest cut in government. It isn't. The chart below uses the cash figures for those two.
Housing and infrastructure, $1,799 million. Older infrastructure funds and mortgage-agency programmes are being wound down and folded into a new agency called Build Canada Homes. Much of this is consolidation rather than money disappearing.
Innovation and science, $1,257 million. Several business-support funds are not being renewed, including the Net Zero Accelerator, which the budget says had declining demand. Support to the Canada Foundation for Innovation is reduced, and the small-business lending programme moves to a Crown bank.
Global Affairs, $1,148 million. Mostly foreign aid rather than administration — $861 million of it is programme money. Named targets are global health funding, contributions to international development banks, and country-by-country aid programmes. The budget states the goal directly: returning the foreign aid budget to where it stood before the pandemic.
Employment and social development, $782 million. Programmes merged, more automation, less office space, fewer consultants.
Immigration, $619 million. Temporary housing support for asylum seekers reduced in line with falling claim numbers, settlement services narrowed for economic immigrants, and a co-payment added to the health programme for newcomers.
The budget describes every reduction as removing waste, duplication, or programmes that have finished their job. Sometimes that is obviously true — a programme moved to another department has not vanished.
Sometimes it is contestable. "Declining demand" for a clean-industry fund is a claim, not an audited finding, and the department making the claim is the one being asked to find savings.
This page does not try to score each one. There is no neutral test for whether a particular cut is efficiency or abandonment, and pretending there is would turn a set of numbers into an argument in disguise.
The $280 billion goes into four buckets over five years. Notice, before the chart, an asymmetry that shapes everything you can know here: the cuts are published department by department, year by year, in a detailed annex. The investments are described as four headline totals in a narrative chapter. That makes the two sides genuinely hard to compare, and the imbalance is not accidental.
Infrastructure, $115 billion. Water systems, transit, Indigenous and municipal infrastructure, and trade corridors. A $5 billion fund over seven years goes into ports, airports and rail, attached to a stated goal of doubling overseas exports within a decade — about $300 billion more trade.
Productivity and competitiveness, $110 billion. Most of this is tax measures rather than cheques. A new "productivity super-deduction" lowers the effective tax rate on new business investment from 15.6% to 13.2%. There is research and development tax support, money for artificial intelligence and electric vehicles, a $5 billion fund over six years for industries hit by tariffs, and a $2 billion fund for critical minerals — the metals used in batteries, electronics and defence equipment, where Canada is among the world's top five producers of ten of them.
Defence and security, $30 billion. Canada reaches NATO's 2% targetNATO members agree to spend at least 2% of their economy on defence. A newer pledge raises the goal to 5%, counting broader security and infrastructure spending, by 2035. this year, five years ahead of schedule, and states a path to the 5% pledge by 2035. Alongside it: 1,000 new border officers and 1,000 new RCMP personnel.
Housing, $25 billion. A new federal agency, Build Canada Homes, plus the elimination of federal sales tax for first-time buyers on homes under $1 million. For scale: the federal mortgage agency estimates Canada needs to build 430,000 to 480,000 homes a year to restore 2019 affordability, against roughly 280,000 today.
On 1 April 2025 the government cancelled the consumer carbon price — roughly 18 cents a litre at the pump. In its final year that charge raised $13.55 billion, and returned $15.60 billion to households through the Canada Carbon Rebate. Separately, the lowest income tax bracket is being reduced from 15% toward 14%, which the budget says benefits 22 million people and saves a two-income family up to $840 a year.
Neither shows up as a spending cut or an investment, because money not collected is neither. But between them they move more than the entire savings exercise does.
No department was told to cut climate programmes. But read through the list of individually named programme terminations and one theme appears more often than any other. This is a pattern that emerges from the document rather than being stated in it, which is part of why it drew relatively little attention.
Programmes that pay people and companies directly to cut emissions are shrinking. Tax credits and carbon pricing aimed at industry are being kept. The government's position is that this is a shift from subsidy to market signal — letting prices do the work instead of grants — not a retreat from the goal.
That is a real argument with genuine support among economists. It is also unfalsifiable for now, because the thing that would settle it is emissions data that does not yet exist.
A quieter effect is worth naming. Statistics Canada's savings plan includes collecting some data less often, substituting statistical modelling where possible. The capacity to measure whether any of this worked is itself inside the cuts.
Against all of that, wildfire money is going up. After the record-breaking 2023 fire season and the second-worst season on record in 2025, federal spending has expanded across prevention, detection and firefighting.
| Programme | Amount | What it does |
|---|---|---|
| Wildfire Resilient Futures Initiative Natural Resources Canada | up to $150.7M | Prevention and mitigation, expanding the FireSmart programme. Includes $48M over four years for research, demonstration sites, and bringing Indigenous fire knowledge into practice. |
| Resilient Communities through FireSmart | $104M announced | Cost-shared with ten provinces and territories; more than $78.5M committed between 2023–24 and 2027–28. |
| Parks Canada wildfire preparedness | $47.8M | Controlled burns, clearing flammable vegetation, equipment and trained crews. |
| Wildland fire knowledge programme | $45.7M | 30 research projects across the country. |
| Aerial firefighting fleet | not separately published | Federally funded aircraft leased and sent to provinces during heavy fire periods, targeted at the 2026 season. |
Look at what kind of spending this is. Every line is adaptation — reducing the damage once fires start. The mitigation spending, which would reduce how often extreme fire weather happens in the first place, is in the previous section, being cut.
That is less a contradiction than a revealed preference, and the logic is not hard to follow. Adaptation produces visible results inside a single term of government: a community that didn't burn, a fleet of water bombers you can photograph. Mitigation produces diffuse results over decades, and only if most other countries do it too. A government under fiscal pressure will reliably pick the first.
The same logic explains something readers sometimes ask about glaciers. Nothing in this budget touches glacier loss, and nothing could. Glaciers respond to global temperature, which responds to global emissions, of which Canada is a small share. There is no line item for it — only the mitigation column, which just contracted.
Everything above is drawn from documents published in November 2025. Several important things are outside what those documents can support, and it is worth being explicit about which.
The reductions begin in the 2026–27 fiscal year. Every figure here is a plan. Budgets get amended, cancelled programmes get restored, and departments routinely miss savings targets. Read all of it as stated intent, not as record.
Reductions come from a detailed annex listing every department and agency. Investments come from four headline totals in a narrative chapter. Anything said here about investments with more precision than a bucket total was found in passing mentions inside the cuts annex — which means that list leans toward whatever the cuts document happened to cite as an offset. It is not a full survey of new spending.
The $280 billion investment figure is accrual; the budget gives $450 billion as the cash equivalent. Departmental cuts are accrual except the two corrected in section four. Every comparison drawn here holds one method constant, but be careful comparing across the two.
The wildfire numbers come from press releases, not budget tables. Announcement amounts often bundle previously committed money and get re-announced, so treat those totals as approximate. The public service headcount figure comes from news reporting rather than the budget itself.
Whether a 15% reduction at Global Affairs means less aid delivered, or the same aid delivered more cheaply, cannot be known from a budget. The same applies to every line in this piece. That evidence arrives years later, in departmental results reports and audits — and sometimes not at all.
Primary government documents first. Everything in the charts and tables comes from the first four.