Canada's total tax burden, 1926 to 2025

This issue covers Canada's total tax burden as a share of the economy (gross domestic product, or GDP), from 1926 to 2025. The total counts every tax, levy and mandatory contribution collected by federal, provincial and local governments and by the Canada and Quebec Pension Plans (CPP and QPP). It comes in four images: the chart, the yearly change, the Party Scoreboard (Liberal vs Conservative / PC) and the PM Scoreboard (the average yearly change 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.

Taxes are harder to pin on one government than they look. The total includes provincial and local taxes, which Ottawa does not set, and a recession can move the share without any tax changing. So this issue also tests the party comparison seven ways.

In brief. Taxes took 13.1% of GDP in 1926 and 34.2% in 2025: about 13 cents of every dollar the economy produced then, and 34 cents now. The peak was 35.8% in 1998. Measured against household income instead of the whole economy, the same taxes came to 42.1% in 2025. Almost all of the rise since 1961 came from provincial and local taxes and the CPP and QPP: federal taxes were 14.4% of GDP in 1961 and 14.7% in 2025. On average, the share rose 0.15 percentage points a year in Liberal years and 0.35 in Conservative and PC years. That order holds against each decade's average and without the Depression and war years. Count federal taxes alone, measure against household income, or look only at the last 20 years, and it reverses. None of these gaps is larger than chance could easily produce. Shading and ranks show who was in office, not who caused the result.

The numbered markers give context, not credit or blame:

The chart adds up every tax and mandatory contribution that Statistics Canada records, from every level of government, and divides it by GDP, the value of everything produced in Canada that year. It counts:

It leaves out fines and donations, which are not taxes; prices for services people choose to buy, such as transit fares and tuition; and the royalties and interest that governments earn.

In 2025 these taxes came to $1.1 trillion, against a GDP of $3.2 trillion. Both are in current dollars, so prices cancel out. The GDP changes in this issue are in current dollars too, so they differ from the after-inflation growth in our GDP per person issue. A change in the share is measured in percentage points ("points" from here on): a move from 33.0% to 33.5% is a rise of 0.5 points.

There is more than one fair way to measure a tax burden, and they give different numbers.

The share of GDP is the measure the Organisation for Economic Co-operation and Development (OECD), a group of 38 mostly wealthy countries, uses for every member, so Canada can be compared with others. The OECD puts Canada at 34.9% in 2024 (a provisional figure), 21st highest of its 38 members and a little above the OECD average of 34.1%. Our figures, built from Statistics Canada's tables, are within about half a point of the OECD's for every year from 1999 to 2022 (0.52 points at most, in 2000). Our 2024 figure, 34.4%, is lower mainly because Statistics Canada revised GDP up in November 2025, after the OECD's figures were compiled.

The same taxes can also be measured against household income: personal income (wages, business and investment income, and government benefits), plus the provincial payroll taxes on employers, such as Ontario's Employer Health Tax, which personal income leaves out. Measured that way, they came to 42.1% in 2025. That reads higher because it sets taxes that businesses pay, such as corporate income tax, against household income alone. It is not the rate any one household pays.

Others measure it differently again. The Fraser Institute, which also counts business taxes as paid by families, estimates 41.9% of the average family's cash income for 2025 (43.2% for families of two or more, the basis of its Tax Freedom Day). The Canadian Centre for Policy Alternatives measures against a broader income that also counts corporate profits, employer benefits, capital gains and inheritances, and estimates 36.7% for 2022. Similar taxes, different yardsticks.

We use the share of GDP for the charts and the scoreboards for three reasons. It is the standard used for every OECD country. It can be checked against the OECD's published figures. And its yearly changes follow taxes more closely: the household measure also moves with government benefit payments, which swell household income in a downturn. In 2020, the household measure fell 2.3 points while tax revenue barely changed, because pandemic benefits raised household income. The Party Scoreboard shows the household measure as a test.

In 1926, taxes took 13.1% of GDP. The low point was 12.1% in 1930.

The Great Depression pushed the share up. From 1929 to 1933, GDP in current dollars fell 41.8% while tax revenue fell 20.7%, so taxes took 17.5% of a much smaller economy by 1933. Property taxes, mostly local, fell only about 11% in dollars. Ottawa also raised taxes: the federal sales tax, cut to 1% in 1930, went back to 4% in 1931 and up to 6% in 1932; corporate income tax was raised in 1931, 1932 and 1933, and personal income tax in 1932 and 1933.

The Second World War raised the share from 16.4% in 1939 to 23.3% in 1941. Ottawa brought in an excess profits tax with its September 1939 war budget and raised it to 75% of profits above a prewar standard in 1940, added a national defence tax in 1940, and raised income taxes sharply from 1940 to 1942. In 1943 it began collecting income tax in the year the income was earned, forgiving half of the 1942 tax on earned income, and the share rose 1.7 points that year. The share reached 25.5% in 1946, a level not seen again until 1967. War production ended and the economy stalled, while imports and consumer goods came back, and with them customs and sales taxes.

It then fell to 20.6% by 1949. The economy grew quickly, the excess profits tax ended for profits earned after January 1, 1948, and income and excise taxes were cut in 1947 and 1949. The Korean War brought new defence taxes in 1950 and 1951, including a sales tax increase from 8% to 10%, and the share rose to 23.7% in 1951.

From 23.3% in 1961, the share rose to its peak of 35.8% in 1998. Three things drove most of the rise:

The share fell to 30.9% in 2011. Budget 2000's Five-Year Tax Reduction Plan restored full indexation and cut personal and corporate income tax rates; corporate profits fell in 2001; the GST was cut from 7% to 5% in 2006 and 2008; and the general corporate rate was cut again, to 15% by 2012. Personal income tax fell from 13.9% of GDP in 1998 to 11.8% in 2002. Payroll contributions went the other way: CPP contribution rates rose from 5.6% in 1996 to 9.9% in 2003.

Since 2011 the share has risen again, to 34.2% in 2025. It jumped 1.6 points in 2015, when GDP fell 0.2% and high earners moved income into 2015 ahead of a new 33% federal top rate in 2016, and 1.4 points in 2020, the first year of the COVID-19 pandemic, when GDP fell 4.0% while tax revenue held almost steady (down 0.1%).

The mix of taxes has changed even more than the total. In 1926, 87% of all taxes were sales, excise, customs, property and other taxes on goods and business, mostly built into prices. Income tax on people was 3%, corporate income tax 5%, and estate duties, payroll contributions and licences the remaining 5%.

By 2025, personal income tax was 37% of all taxes, payroll contributions (the CPP and QPP, Employment Insurance and workers' compensation) 15%, corporate income tax 13%, sales, excise, property and other taxes 33%, and taxes withheld on income paid to people outside Canada 1%. So while the total tax burden grew to 2.6 times its 1926 level, income tax on people plus payroll contributions went from 0.6% of GDP in 1926 to 17.9% in 2025: about 30 times.

Federal taxes alone were 6.4% of GDP in 1926. They climbed during the Second World War to 19.4% in 1943, and were 14.4% in 1961. In 2025 they were 14.7%.

So nearly all of the rise since 1961 (10.6 of 10.9 points) came from provincial and local taxes and the CPP and QPP. Federal taxes were 62% of all these taxes in 1961 and 43% in 2025. Federal governments do not set provincial income or sales tax rates or municipal property taxes.

The second image shows how much the share rose or fell each year, in points. It rose in 54 years and fell in 45.

The five largest rises:

The five largest falls:

The share often rises in recessions, because the economy shrinks faster than tax revenue: in 2009, GDP fell 5.2% and tax revenue 2.2%, and the share rose 1.0 point. It can rise in a war, when new taxes pay for it. And it falls when the economy grows faster than tax revenue, as in 1948. A rise is not always a tax increase, and a fall is not always a tax cut.

The Party Scoreboard counts the years 1927 to 2025: 99 yearly changes, 68 in Liberal years and 31 in Conservative and PC years. 2026 is not counted: it has no full-year figure yet.

Across all 99 years, the share rose by 0.21 points a year on average; the median was +0.06 points. The median is the middle value when the years are lined up from lowest to highest. A few extreme years can pull an average a long way; the median is less affected.

In Liberal years, the average was +0.15 points and the median −0.04. In Conservative and PC years, the average was +0.35 and the median +0.25.

The board adds two rows as tests. "Federal taxes only" counts just the federal government's own taxes, as a share of GDP: +0.10 points a year in Liberal years and +0.05 in Conservative and PC years. "All taxes (% of household income)" divides the same total by household income instead of GDP: +0.27 in Liberal years and +0.19 in Conservative and PC years. Both reverse the order of the averages in the first row. The board shows averages for these two rows. On medians, household income reverses the order too (+0.24 and 0.00), but federal taxes alone do not (−0.06 and +0.05).

The gap in the first row is 0.20 points a year (0.15 against 0.35). We tested it seven ways.

Timing. Taxes rose faster in some decades than others. Measured against each decade's own average change, Liberal years rose 0.06 points a year slower than their decade's average and Conservative and PC years 0.12 faster: a gap of 0.18 points.

Two big episodes. The Depression years under R.B. Bennett, 1931 to 1935, were Conservative years, and the share rose 0.96 points a year on average as GDP collapsed and taxes were raised. The Second World War years under Mackenzie King, 1939 to 1945, were Liberal years, and it rose 0.91 points a year as wartime taxes were added. Leave out the Depression years alone and the averages are +0.15 for Liberal years and +0.23 for Conservative and PC years; leave out the war years alone, +0.06 and +0.35; leave out both, +0.06 and +0.23.

Recent years. Since 1981, Liberal years averaged +0.03 points and Conservative and PC years +0.26. In the 20 years from 2006 to 2025, 10 for each party, the order reverses: +0.12 points in Liberal years and +0.03 in Conservative and PC years.

Policy takes time. A budget can take a year or more to change what people pay, so we also credited each year's change to the party in office the year before. The averages barely move: +0.15 points for Liberal years and +0.35 for Conservative and PC years. With a two-year delay, +0.16 and +0.33.

Federal taxes only. Counting only federal taxes, the order reverses: +0.10 points a year in Liberal years and +0.05 in Conservative and PC years.

Household income. Measured against household income instead of GDP, the order also reverses: +0.27 points in Liberal years and +0.19 in Conservative and PC years.

Chance. The yearly change is noisy: about half of all years move by half a point or more, and 28 of the 99 by a full point or more. So averages over 68 and 31 years can differ by a few tenths of a point by chance alone. To check, we shuffled the party labels across the years 10,000 times. Gaps at least as large as the first row's came up in about 38% of the shuffles; for the federal-only row, 84%; for the household-income row, 81%.

What is left: on the main measure, the share of GDP rose a little faster in Conservative and PC years. That holds against each decade's average, without the Depression and war years (the gap narrows without the Depression and widens without the war), since 1981 and with a one- or two-year delay. On federal taxes alone, against household income, and over the last 20 years, it rose a little faster in Liberal years. None of these gaps is larger than chance could easily produce. Over the whole century, the 21-point rise splits almost evenly: 10.2 points came in Liberal years and 10.8 in Conservative and PC years, though Liberals governed more than twice as many years. That describes who was in office when the share moved. It does not show that either party caused it.

We did not test against the United States, as our GDP per person issue did: each country sets its own taxes, so a foreign benchmark does not separate a government's choices from world conditions the way it does for growth.

The PM Scoreboard sorts each term by its average yearly change, from the lowest to the highest. Rank 1 is the lowest figure (a fall, or the smallest rise), not the best result: whether a lower or higher tax burden is better is a judgment for readers.

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 yearly changes start in 1927, so terms that ended before then have no data: Arthur Meighen's two terms and Mackenzie King's first. King's combined row therefore pools his second and third terms. That is why the title starts in 1927.

The 15 ranked rows are 13 terms plus 2 combined rows. A combined row repeats terms already listed; it is not another government.

Rank 1 goes to Jean Chrétien: the share fell 0.23 points a year, from 34.9% in 1993 to 32.7% in 2003, with the Five-Year Tax Reduction Plan of 2000 and a fall in corporate profits in 2001, while CPP contribution rates rose. Next come Mackenzie King's second term (−0.19), Louis St-Laurent (−0.06) and Paul Martin (−0.01).

The highest is R.B. Bennett's, at +0.87 points a year, in the Great Depression: GDP fell much faster than tax revenue, and the sales tax and income taxes were raised. Next come Lester B. Pearson (+0.66), in the years the CPP and QPP began and provincial taxes rose, and Pierre Trudeau's second term (+0.58), with the National Energy Program's charges and the recessions of 1980 and 1981-82. Brian Mulroney follows (+0.55), with partial deindexing (from 1986 until full indexation returned in 2000), unemployment insurance moving to premiums alone and the 1990-92 recession.

Some of this holds on the other measures and some does not. R.B. Bennett and Lester B. Pearson also have the two largest rises against household income. On federal taxes alone, the four highest rows above rise by much less: Pearson +0.01, Pierre Trudeau's second term +0.29, Mulroney +0.24 and Bennett +0.38. At the other end, the four lowest rows also fall on federal taxes alone, but three of them rise against household income: Jean Chrétien +0.21, Louis St-Laurent +0.23 and Paul Martin +0.42. Among the ranked rows, the lowest against household income is Stephen Harper's (−0.35); on federal taxes alone, the lowest is Mackenzie King's second term (−0.34) and the highest is his third term (+0.66), which spans the Second World War.

Mackenzie King and Pierre Trudeau each have a combined row, shaded grey, which weights each term by its years counted. Pierre Trudeau's terms show +0.13 and +0.58, and +0.26 together; his first term is 11.1 of his 15.4 years. King's second and third terms show −0.19 and +0.41, and +0.28 together.

Four rows are listed below the ranking without a rank: Joe Clark (−0.28), John Turner (−0.18), Kim Campbell (−0.62) and Mark Carney (−0.28). 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 2025, the latest full year of data, and was in office for 0.8 of that year. His row joins the ranking once the 2026 figure is out.

The right-hand column names the same kinds of events as our GDP per person issue: the recessions Statistics Canada dates and three of its four shocks (the Great Depression, the Second World War and the COVID-19 pandemic; the 1920-21 depression comes before these data), on every term they overlap by three months or more. The same test applies to every Prime Minister, so some rows are blank.

The column does not list tax changes. Official cost estimates exist for some measures but not others, and they are not made on one basis across 100 years, so a list would test some Prime Ministers' measures more fully than others'. The largest tax changes are described above, by year. 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.

The taxes come from Statistics Canada's government accounts: the current tables for 1961 to 2025 and an archived set that runs from 1926 to 1986. The two agree on the 1961 tax total to within 0.3%, but today's GDP figures are 2.8% larger than the old ones, because later revisions widened what GDP counts. So the 1926 to 1960 shares are multiplied by 0.976 (the 1961 share in the current tables divided by the 1961 share in the archived ones) to meet the current figures. Each year-to-year change inside the older piece keeps its shape. The federal-only and household-income rows are joined the same way, each with its own ratio.

Recent years get revised, and 2025 is preliminary.

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 36-10-0477-01, 36-10-0484-01, 36-10-0104-01, 36-10-0112-01, 36-10-0161-01, 36-10-0178-01, 36-10-0179-01, 36-10-0180-01, 36-10-0181-01, 36-10-0152-01, 36-10-0202-01 and 18-10-0005-01 (retrieved 2026-10-02). 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