Subscribe

Labour · 3 min read

Youth unemployment's biggest climb in 50 years — without a recession

Canada's youth unemployment rate rose from 10.0% in 2022 to 13.8% in 2025. We checked it against the full StatCan record back to 1976: every bigger 3-year rise in that history rode a recession. This one hasn't.

Published 3 August 2026 · Open Data Canada Source Statistics Canada, table 14-10-0017-01
Chart: +38%, youth unemployment rate rose from 10.0% in 2022 to 13.8% in 2025, shown as two bars with an arrow marking the increase.
The CNE said it received nearly 50,000 applications ahead of its 2026 job fair, as Ontario youth unemployment topped 15%. Chart: Open Data Canada, Statistics Canada Table 14-10-0017-01.

Canada's youth (15-24) unemployment rate climbed from 10.0% in 2022 to 13.8% in 2025, a 3.8-point rise. On its own, that number invites an obvious question: how unusual is this, historically?

We pulled Statistics Canada's full youth unemployment series back to 1976 — 50 years of annual data — and checked every 3-year window in it, not just the recent one. The answer: this climb is real, and it is structurally unusual. Every 3-year rise in youth unemployment bigger than 2022-25's, going back half a century, happened during a defined recession — the early 1980s (+6.5 points), the early 1990s (+6.2 points), the 2008-09 financial crisis (+3.6 points), and COVID (+8.9 points, the sharpest on record). 2022-25 is the largest 3-year climb in the entire series that isn't tied to one.

No recession has been officially declared for 2022-2025. The C.D. Howe Institute's Business Cycle Council — Canada's conventional arbiter of recession dates — said as recently as September 2025 that a technical two-quarter GDP dip in mid-2025 still didn't meet its bar.

Line chart of Canada's youth unemployment rate, 1976-2025, with recession years shaded and the 2022-25 rise highlighted as the biggest climb outside any shaded band.

It's not just a national story

Comparing 2019 to 2025 province by province, youth unemployment worsened almost everywhere. Ontario and British Columbia rose the most, both by 4.3 points — Ontario from 11.6% to 15.9%, BC from 9.1% to 13.4%. Alberta rose 3.4 points, from 12.1% to 15.5%. Only two provinces — Nova Scotia and Newfoundland and Labrador — saw youth unemployment ease, and only slightly (−0.7 and −1.1 points). Quebec has the lowest rate in the country at 9.6%, but even it rose, from 8.3%.

Dumbbell chart comparing youth unemployment rate by province, 2019 vs 2025, sorted by 2025 level. Ontario and British Columbia show the largest increases; Nova Scotia and Newfoundland ease slightly; Quebec stays lowest.

What's driving it is contested

A Bank of Canada deputy governor has pointed to immigration volume — specifically the surge in non-permanent residents and international students — as a possible structural factor colliding with a slower hiring cycle. That theory dominates public discussion of the numbers too: in the highest-upvoted Reddit threads on the topic, the most common complaint isn't the headline number, it's that entry-level jobs that used to go to teenagers increasingly don't. None of this is proof of causation — the unemployment rate itself doesn't say why it rose — but it's the dominant public read on it.

2020 remains the worst single year for youth unemployment on record, at 20.1%. But 2022-2025 is the most structurally unusual pattern in the data: a rise this size with no downturn behind it, landing hardest in Ontario and British Columbia, and easing almost nowhere.

Sources and data

Sources

Methodology. The "biggest 3-year rise" claim was checked exhaustively: every valid 3-year window (year t minus year t-3) in the 1976-2025 series was computed and classified as recession-linked if it overlapped 1981-82, 1990-91, 2008-09, or 2017-2020 (the window containing the COVID spike). 2022-25 (+3.8 points) is the largest window that overlaps none of those. The Fraser Institute's 57% figure is a headcount statistic (290,000 to 437,000 unemployed youth, as published in the source PDF) and is not directly comparable to the 3.8-point / roughly 38% relative rate increase used throughout this piece — both are legitimate ways to describe the same underlying rise, but they answer different questions and should not be quoted interchangeably. Note the study's own rounded figures (290,000 → 437,000) compute to roughly 51%, not 57%; the 57.0% is the study's own stated figure, reproduced here as published, likely reflecting unrounded underlying counts.