Federal spending per person in Canada, 1926 to 2025

This issue covers federal government spending per person each year from 1926 to 2025, after inflation, in 2025 dollars. It is the second of four issues on federal finances: last week covered the surplus or deficit, and revenue and the federal debt follow. It comes in four images: the chart, the yearly change, the Party Scoreboard (Liberal vs Conservative / PC) and the PM Scoreboard (the average under each Prime Minister).

On the charts, each year is shaded by the federal party in office for most of that year: red for Liberal, blue for Conservative and Progressive Conservative (PC). Diagonal stripes mark minority governments.

Spending per person in 2025 dollars was about 24 times as high in 2025 as in 1926, so the scoreboards do not compare the parties' average level, which would mostly show when each party governed. They compare how fast spending per person grew each year.

In brief. In 2025 the federal government spent $544.4 billion, or $13,083 per person, on Statistics Canada's national accounts basis (spending in the calendar year on programs, transfers and interest, not counting its own new buildings and equipment). In 2025 dollars, that compares with $534 per person in 1926, a peak of $7,840 in 1944, during the Second World War, and $17,660 in 2020, during the pandemic, the highest of the century. Over the century, spending per person grew at a steady average rate of 3.3% a year. It grew 3.9% a year in Liberal years and 1.9% in Conservative and PC years, a gap of 2.1 percentage points ("points" from here on; figures are rounded, and gaps are worked out before rounding) that is well within what chance could produce.

Some tests shrink the gap and others widen it. Leaving out the war and the drops after it (1939 to 1948) narrows it to 1.3 points, and leaving out only the war years (1939 to 1945), which keeps the drops of 1946 to 1948, puts Conservative and PC years ahead, 1.9% against 1.2%. Leaving out the Depression and six later recession years (9 Conservative years and 3 Liberal ones) widens it to 3.7 points, and crediting each year to the government in office two years earlier widens it to 3.0 points. Over the whole span the median year (the middle one when the years are ranked by growth) grew at almost the same pace under both, 1.9% and 1.7%; since 1948 it was 2.0% against 0.4%. Shading and ranks show who was in office, not who caused the result.

The numbered markers give context, not credit or blame:

We marked 14 of the many events that budget documents, the Library of Parliament, Health Canada or Statistics Canada tie to federal spending: five world events and nine federal measures, five enacted under Liberal governments and four under Conservative and PC governments. Four of the measures aimed to restrain spending (the 1984, 1990 and 2012 measures under Conservative and PC governments and the 1995 budget under a Liberal one) and five expanded it (four enacted under Liberal governments and the 2009 Economic Action Plan under a Conservative one). Two of the Liberal measures began under the next government: federal hospital insurance was enacted in April 1957 under Louis St-Laurent and began in 1958 under John Diefenbaker's PC government, and federal medicare was enacted in 1966 under Lester B. Pearson and began in 1968 under Pierre Trudeau. The November 1984 economic statement, by Finance Minister Michael Wilson of the new Progressive Conservative government, falls in a year counted as Liberal, because Liberals held office for most of it. We considered other events and did not mark them, because we could not confirm their details on an official page or because they sit too close to another marker to read.

The chart shows federal spending in Statistics Canada's national accounts, the international system used to measure every government on the same basis, by calendar year. It counts what the federal government buys (including the yearly wear on its buildings and equipment), the money it pays to people (such as Old Age Security, the Canada Child Benefit and Employment Insurance), its transfers to provinces, territories and Indigenous governments (such as the Canada Health Transfer), its subsidies and aid, its grants toward others' buildings and equipment and one-time items such as settlements, and the interest on its debt. It leaves out the government's own new buildings and equipment. From 1961, federal revenue minus this spending is exactly the surplus or deficit charted last week. Before 1961 the older figures count a narrower set of grants toward others' capital, but in 1961, the first year both cover, the two agree: $22 million in each.

It is not the spending total the government announces. That figure, total expenses, comes from the Public Accounts for fiscal years that run from April to March, under Canadian public sector accounting standards: $547.3 billion for April 2024 to March 2025. Adding up the national accounts quarters for the same twelve months gives $532.5 billion. The two follow different accounting rules: for example, the Public Accounts total includes $4.0 billion of net actuarial losses. The calendar year matters too: the national accounts put 2024 spending at $562.1 billion, because January to March 2024, a quarter that belongs to the previous fiscal year, includes one-time settlements with First Nations (see below). We use the national accounts because they are the only series that runs every year from 1926 to 2025 on one calendar basis.

Each year's spending is converted to 2025 dollars with the Consumer Price Index and divided by the population. Using the GDP price index instead gives a similar party result, a gap of 1.7 points instead of 2.1 (see below). From 1926 to 1960 the figures come from Statistics Canada's older national accounts, and from 1961 from the current ones. In 1961 they differ by 0.2%; over 1961 to 1986 the current figures run between 0.2% below and 4.3% above the older ones. To keep each year's growth on one basis, the growth for 1961 (and its change in the share of GDP) compares 1961 with 1960 in the older figures, and later years use the current figures.

The notes below describe the main moves, with the reasons official sources give where they give one. Figures are per person in 2025 dollars.

The second chart shows each year's growth in spending per person after inflation. In 78 of the 99 years, the change falls between a drop of 5% and a rise of 10%. The large spikes are the war years, with rises of 112% in 1940 and 132% in 1942, and 2020, a rise of 57%. The largest drops followed the war: 21% in 1945, 33% in 1946, 36% in 1947 and 23% in 1948; after the pandemic, spending fell 14% in 2021 and 18% in 2022. All of them fall in Liberal years. The war rises and the drops after them account for part of the gap between the parties' averages: leaving out 1939 to 1948 shrinks it from 2.1 to 1.3 points. The pandemic rise and the drops after it roughly cancel out, so leaving out 2020 to 2022 barely changes the gap.

The Party Scoreboard counts the growth for each year from 1927 to 2025: 99 years, 68 Liberal and 31 Conservative and PC (1926 has no year before it in the data). Spending per person grew in 51 of the 68 Liberal years (75%) and 18 of the 31 Conservative and PC years (58%).

The averages on the board are compound: the steady yearly rate that gives the same total change over the years counted. Spending per person grew 3.3% a year across all years, 3.9% in Liberal years and 1.9% in Conservative and PC years. The median, the middle year when the years are lined up from lowest to highest, was 1.9%, 1.9% and 1.7%. A few extreme years pull an average a long way; the median is less affected.

The board adds four test rows. Without the war and the drops after it (1939 to 1948), all in Liberal years, Liberal years average 3.2% and Conservative and PC years 1.9%. Without the pandemic years 2020 to 2022, 4.0% and 1.9%. Program spending, which leaves out interest on the debt, grew 4.6% a year in Liberal years and 2.2% in Conservative and PC years. Spending as a share of GDP rose by 0.13 percentage points a year on average in Liberal years and 0.10 in Conservative and PC years (for example, from 15.00% to 15.13% of GDP).

The gap between the parties' average growth is 2.1 points (3.94% against 1.87% before rounding). We tested it several ways.

Chance. To see how large a gap chance alone produces, we shuffled the party labels across the years. Of a million shuffles, half gave a gap of about 2.5 points or more, and one in ten about 5.8 or more. In a separate run of 100,000 shuffles, gaps at least as large as the board's came up 57% of the time. Neighbouring years resemble each other, so we also kept each party's unbroken stretches in office together and tried all 462 ways of assigning the parties to those 11 stretches: a gap as large came up in 205 of them (44%). Both say the gap is within what chance could easily produce. In the brackets below, each figure is how often chance alone gave a gap at least as large; the higher it is, the more ordinary the gap.

The simple average. A simple average of each year's percent change is pulled up by large swings: a rise of 100% followed by a fall of 50% averages 25% a year, although spending ends where it started. That is why the board uses the steady yearly rate. On the simple average, Liberal years grew 6.0% a year and Conservative and PC years 2.0%, a gap of 4.0 points (40% of the shuffles, 47% of the stretch arrangements).

The war. Leaving out 1939 to 1948, the war and the drops after it, gives 3.2% against 1.9%, a gap of 1.3 points (46% of the shuffles, 56% of the stretch arrangements). Leaving out only 1939 to 1945, which removes the war years (including the drops of 1939 and 1945) but keeps the drops of 1946 to 1948, puts Conservative and PC years ahead: 1.9% a year against 1.2% (77% of the shuffles, 83% of the stretch arrangements). Leaving out the war, the drops after it and the pandemic years 2020 to 2022 gives 3.2% against 1.9%, a gap of 1.3 points (30% of the shuffles, 58% of the stretch arrangements).

Recessions. Leaving out instead the Depression years 1930 to 1935 (all but 1930 under R.B. Bennett's Conservative government) and the recession years 1981, 1982, 1991, 1992, 2009 and 2010 widens the gap to 3.7 points (3.9% against 0.3%; 39% of the shuffles, 21% of the stretch arrangements). The years left out are 9 Conservative years and 3 Liberal ones. In the early 1930s prices fell, so spending that rose modestly in dollars rose faster after inflation: in 1931 federal spending rose 12% in dollars and 23% per person after inflation.

Decades. Compared with each decade's own average, Liberal years grew 0.15 points faster than their decade and Conservative and PC years 0.33 points slower, a gap of 0.48 points (89% of the shuffles, 58% of the stretch arrangements).

Eras. The gap is not the same in every era:

Policy takes time. A budget passed in one year can shape spending in the next. Crediting each year to the party in office the year before widens the gap to 2.7 points (4.1% against 1.4%; 46% of the shuffles, 29% of the stretch arrangements); with a two-year delay, to 3.0 points (4.2% against 1.2%; 40% and 25%).

Other measures. Measured as a share of GDP, the gap is small and well within chance: spending rose 0.13 points of GDP a year in Liberal years and 0.10 in Conservative and PC years (96% of the shuffles, 93% of the stretch arrangements). Program spending per person, without interest, grew 4.6% against 2.2%, a gap of 2.4 points (57% of the shuffles, 45% of the stretch arrangements). With the GDP price index in place of the Consumer Price Index, 3.5% against 1.9%, a gap of 1.7 points (65% and 48%).

The older national accounts. On the older basis for 1962 to 1986, the gap in those years is 5.5 points instead of 4.9, and the full-period gap 2.1 points either way. In those 25 years alone, only 4 of them Conservative, gaps as large came up in 1% of the shuffles on the older basis and 3% on the current one, but in 2 of the 10 stretch arrangements on both: unusual when years are shuffled one by one, not unusual once each party's stretches in office are kept together.

Volatility. Spending per person changed by 11.5% a year on average, up or down, in Liberal years and 3.7% in Conservative and PC years, mostly because the war and pandemic swings fall in Liberal years.

What is left: spending per person grew faster on average in Liberal years than in Conservative and PC years. But the gap over the full span is within chance, and so is every full-span test above. It shrinks to 1.3 points without the war and the drops after it, reverses when only the war years are left out, and is small as a share of GDP; it widens to between 2.7 and 3.7 points when the Depression and later recession years are left out or each year is credited to the government in office a year or two earlier. The middle years are close over the whole span, though less so since 1948. Spending also depends on wars, recessions, the pandemic, prices and population, and the comparison cannot separate those from who held office. That describes who was in office when spending moved. It does not show that either party caused it.

The PM Scoreboard sorts each term by its steady yearly growth in spending per person after inflation over the whole term, from the slowest to the fastest. Rank 1 is the slowest growth, not the best or worst result: each term's spending reflects wars, recessions, the pandemic and policy.

Most rows are a term: one unbroken stretch as Prime Minister, however many elections it spans. Each year counts in proportion to the days in office that year, so 1957 counts partly for Louis St-Laurent and partly for John Diefenbaker. The "Years counted" column adds up those shares. The board starts in 1927, the first year of growth in the data, so Arthur Meighen, whose last term ended in 1926, has no row. The 15 ranked rows are 13 terms and 2 combined rows. A combined row repeats terms already listed; it is not another government.

The ranked terms, from the slowest growth to the fastest, with each term's years and the chart markers that fall within it (the two combined rows are described below):

On the simple average of yearly changes, the slow end of the ranking is the same, but large swings move the fast end: Mackenzie King's third term ranks last at 15.9% a year, R.B. Bennett 13th at 9.1%, and Justin Trudeau's term 9th instead of 7th, at 4.8%.

Two Prime Ministers have a combined row, shaded grey, which pools the years of the terms it names: Mackenzie King (his second and third terms, 5.7% a year; his first term ended in 1926, before the first year of growth) and Pierre Trudeau (both terms, 5.4%).

Four rows are listed below the ranking without a rank: Joe Clark (a fall of 2.0% a year), John Turner (a rise of 6.0%), Kim Campbell (a fall of 1.3%) and Mark Carney (a fall of 6.1%, the 2025 drop after the one-time First Nations settlements, most of them in early 2024). Each rests on less than one year of data, and the yearly figures they draw on also cover months when someone else was in office. Mark Carney took office in March 2025 and was in office for 0.8 of 2025, the latest full year of data.

The right-hand column of the PM Scoreboard names the law that created each federal benefit to people that has its own line in Statistics Canada's spending table, counted in the term in which the law received royal assent. Ten laws pass that test (the table's line for family and youth allowances names two benefits, so both laws count):

Under this test, seven of the ten laws were enacted under Liberal governments and three under Conservative and PC governments, close to the Liberal share of the years on the board (68 of 99, or 69%); counting one law per line instead gives six to three. We set the test before counting. Some of these benefits later replaced others: the Child Tax Benefit replaced family allowances in 1993, and the Canada Child Benefit replaced the Canada Child Tax Benefit and the Universal Child Care Benefit in 2016. Creating a benefit is not the same as spending growth: a new benefit can replace an older one, and most federal spending goes to transfers to provinces, purchases and interest, which the column does not cover.

Left out for every term:

Comments give context. They do not assign credit or blame. Sources for every comment are in the source list at the end of the web version of this issue.

Federal spending for 1926 to 1960 is total current expenditure of the federal government in Statistics Canada's archived national accounts (table 36-10-0177-01); from 1961 it is federal general government expenditure (table 36-10-0477-01), the four quarters of each year added together. Program spending is that spending minus interest on the debt. Prices are the Consumer Price Index, all items, yearly average (table 18-10-0005-01), with 2025 as the base year. Population is the same series as the GDP per person issue (tables 17-10-0063-01 to 1970 and 17-10-0005-01 from 1971, joined in 1971). The share of GDP uses the same GDP figures as the surplus or deficit and total tax burden issues. The GDP price index test uses the GDP per person issue's figures. Each year's growth is the percent change in spending per person after inflation from the year before. The averages are compound: each year's growth is converted to a logarithmic scale, on which a rise and the fall that undoes it cancel out, averaged, and converted back to a yearly percent.

More on the method: How the charts work

Who we are: About Gov Scoreboard

Spot an error? Write to [email protected].

Adapted from Statistics Canada, Tables 36-10-0177-01, 36-10-0477-01, 18-10-0005-01, 17-10-0063-01, 17-10-0005-01, 36-10-0202-01, 36-10-0104-01, 36-10-0699-01, 36-10-0277-01 and 36-10-0137-01 (retrieved 2026-10-27). This does not constitute an endorsement by Statistics Canada of this product.

Dates in office: House of Commons, Procedure and Practice (4th ed., 2025), Appendix 6.

Gov Scoreboard