Markets · Analysis

Late-Cycle Signals Are Flashing. Canada’s Spending, Deficits, and Household Exposure Say It Isn’t Ready

The indicators that preceded every major downturn in the last 50 years are present. A primary-source assessment of where Canada stands.

March 30, 2026 · 24 primary sources · 9 min read

Toronto financial district skyline at dusk under heavy overcast sky, warm city lights reflected in dark unsettled water

THE RECEIPT

Canada is running non-emergency deficits of 2.2% of GDP during an expansion while the combination of conditions that preceded the last three major downturns is observable across multiple indicators simultaneously.

Federal debt-to-GDP has risen from 29% to 41.2% since the last recession. More than 2 million mortgages are renewing into higher rates — 85% locked when the Bank of Canada rate was at or below 1%. Per-capita GDP is flat. Workforce spending is oriented toward trades while about 60% of Canadian employees face AI-related job transformation. Exports worth 16.8% of GDP flow to a trading partner imposing tariffs. The $16 trillion global repo market that benchmarks Canadian fixed mortgage rates operates with overnight maturity and zero collateral cushion. The fiscal buffer that existed before every previous downturn is diminished. None of this predicts a recession. All of it is on the public record.

That’s the receipt.

The finding above is sourced and verified. Full analysis below.

Full analysis

Continue reading ↓

If you’re one of the million Canadian households renewing a mortgage this year, or you work in a knowledge-economy job that didn’t exist when the last recession hit, the next twelve months are shaped by forces that become visible only when you put several headlines side by side.

Economists use the term “late cycle” to describe the phase where the indicators that historically precede a downturn begin appearing — not all at once, and never in a perfectly predictable order, but in a recognizable cluster. The question this article examines isn’t whether any single indicator is flashing. It’s what the combination looks like, and whether Canada is positioned to absorb what has historically followed.

When these conditions have appeared together in the past, the question that determined outcomes was not whether a downturn arrived — it was how much room the country had to respond. This article examines what the documented record shows about Canada’s room to respond right now.

How to read this article

Green (Documented Facts): Verified claims with source citations.

Blue (Interpretation): Analysis and assessment — labeled as such.

Gold (Context): What both sides tend to leave out.

Grey (Falsifiers): What would change this assessment.

2.2%

Projected federal deficit, % of GDP, FY2025–26 (PBO)

0.6%

Federal surplus entering 2008 recession (Dept. of Finance)

2M+

Mortgages renewing 2025–26 (CMHC)

16.8%

Canadian GDP from U.S. exports (StatCan)

Key Facts — Verified

Federal deficit projected at 2.2% of GDP for FY2025–26 outside of recession or emergency. [1]
Federal surplus was 0.6% of GDP entering the 2008 recession; deficit was 1.7% of GDP entering COVID. [2] [3]
Federal debt-to-GDP: 29.0% in FY2007–08, 31.3% pre-COVID, 41.2% in FY2024–25. [2] [3] [4]
PBO fiscal sustainability room: $46 billion (1.5% of GDP). [5]
Government fiscal anchor: deficits below 1% of GDP starting FY2026–27. Current trajectory: 2.2%. [6] [1]
2+ million mortgages renewing 2025–26. 85% of fixed-rate renewals in 2025 were contracted when the BoC rate was at or below 1%. [7] [8]
Average payment increase: 10% for 2025 renewals, 6% for 2026. The Bank’s 85% coverage figure implies roughly 15% of renewing borrowers may not be able to cover increases from current financial assets alone. [9] [10]
About 60% of Canadian employees could be exposed to AI-related job transformation, with about half in roles where AI may complement rather than replace their work. [11]
Bank of Canada: it is becoming more difficult to find a job in occupations with a larger share of AI-performable tasks. [12]
No identified ESDC program targeting AI-driven white-collar displacement. Federal workforce programs are predominantly trades-oriented. [13] [14]
Exports to the U.S. account for 16.8% of Canadian GDP and support 2.6 million jobs. [15]
Bank of Canada: U.S. tariffs leave Canadian GDP 1.5% lower by end-2026. [16]
U.S. yield curve (2Y/10Y) re-steepened to +0.56 after prolonged inversion. First positive print: August 28, 2024. [17]
Gold up ~103% vs. S&P 500 up ~21% over two years (March 2024–March 2026). Silver up ~183%. [18]
878,000 U.S. households in serious delinquency or foreclosure as of January 2026, up 25% in four months. 80% are FHA loans. [19]

What Late Cycle Looks Like

Late cycle is not a prediction. It is a set of observable conditions that have historically preceded economic downturns. The pattern never repeats exactly, but the cluster is recognizable: credit stress at the margins, capital rotating from financial assets into hard assets, yield curve re-steepening after inversion, a geopolitical supply shock elevating commodity prices, and employment holding steady — the last indicator to turn.

As of March 2026, based on the indicators documented above, each of these conditions is observable. The U.S. 2Y/10Y yield spread has re-steepened to +0.56 after its deepest inversion in decades. [17] Gold has outperformed the S&P 500 by roughly five to one over two years. [18] Silver began outperforming gold in Q2 2024 — a sequence that also appeared in the final phase of the 1979–80 and 2007–08 hard-asset rotations. [18] U.S. mortgage delinquencies have spiked 25% in four months, concentrated in the weakest borrowers. [19] Canadian near-prime delinquency is outpacing subprime year-over-year. U.S. unemployment sits at 4.4% — still relatively low, as it was in late 2007 before the deterioration began. [20]

The Historical Pattern

In 2007–08, the S&P 500 peaked in October 2007. [22] Gold peaked in March 2008, silver the same month. [18] Oil peaked in June–July at $134/bbl. [21] Employment didn’t crack until December 2007, and the NBER didn’t officially call the recession until December 2008 — a full year after it had begun. [22] The consensus said “no recession” while every other signal was already flashing. A broadly similar rotation — equities weakening, then hard assets rallying, then energy spiking — played out in the 1970s, though messier: oil’s shock overlapped the 1973–74 equity decline rather than following it neatly.

The monetary parallel is worth noting. U.S. M2 money supply expanded 40.6% from February 2020 to its March 2022 peak. After the 1971 gold-window closure, M2 expanded 43.4% over four years. [23] Nearly identical magnitude, compressed into half the time. Inflation peaked at 9.1% in June 2022, roughly 28 months after the expansion began. In the 1970s, the first CPI peak hit 12.3% in December 1974 — about 40 months after the gold window closed. [24] The 1970s then produced a second inflation wave, peaking at 14.8% in March 1980. Whether the current cycle produces a second wave is an open question. The conditions — a Middle East energy supply shock arriving on top of residual inflation — are structurally similar.

Canada’s Starting Position

Canada entered the 2008 recession with a federal surplus of 0.6% of GDP and a debt-to-GDP ratio of 29%. [2] It entered COVID at a deficit of 1.7% with debt-to-GDP at 31.3%. [3] Today the deficit is projected at 2.2% of GDP and debt-to-GDP stands at 41.2%. [1] [4] The government’s own fiscal anchor — sub-1% deficits by FY2026–27 — is already more than double its target. [6] The PBO estimates $46 billion in fiscal sustainability room. [5] For context, COVID required $327.7 billion.

This is not the worst fiscal position Canada has occupied. The 1980–81 and 1989–90 deficits were larger as a percentage of GDP. [2] What distinguishes the current moment is the combination: the fiscal position, the household exposure, the workforce misalignment, and the external dependence arriving alongside late-cycle indicators. On every documented dimension we examined, Canada’s starting position is weaker than it was in either 2007 or 2019. No single variable is without precedent. FDI matching 2007 headline levels while the economy underneath is structurally different is a pattern we have already documented. Global credit spreads matching June 2007 with the IMF using the word “complacent” is another.

The revenue side compounds the exposure. Real GDP per capita was unchanged in Q4 2025, and population growth is decelerating under immigration policy shifts. The fiscal projections that underpin the government’s deficit trajectory assume a revenue base that stagnant per-capita productivity and slowing population growth may not deliver.

Where the Money Is Going

Since FES 2024, the PBO has identified $115.1 billion in net new federal spending commitments through 2029–30. [1] Budget 2025 allocates 42% to sovereignty and 36% to bringing down costs. [6] The identifiable workforce programs are overwhelmingly trades-oriented: the Canadian Apprenticeship Strategy, the Union Training and Innovation Program doubled at $75 million over three years, and a Sustainable Jobs stream targeting 30,000 Red Seal workers. [13] [14]

The AI-specific spending the research could identify was modest: CanCode renewed at $39.2 million over two years, plus general digital-skills programs. [14] No centralized budget table breaks out the allocation between clean economy, trades, and AI-adaptation. Meanwhile, Statistics Canada estimates about 60% of Canadian employees could be exposed to AI-related job transformation, with about half in roles where AI may complement rather than replace their work. [11] The Bank of Canada says there is “hard evidence” that jobs in AI-exposed occupations are already harder to find. [12] The OECD has published Canada-specific findings showing demand shifting away from the skills these workers currently hold. The spending and the emerging risk are pointed in different directions.

The Household Exposure

More than 2 million Canadian mortgages are renewing in 2025–26. [7] CMHC reports that 85% of fixed-rate mortgages renewing in 2025 were contracted when the Bank of Canada rate was at or below 1%. [8] The Bank of Canada estimates average payment increases of 10% for 2025 renewals and 6% for 2026. [9] It also says 85% of renewing borrowers could cover expected increases for 12 months or more using financial assets — implying roughly 15% may not be able to cover increases from current financial assets alone. [10] OSFI identifies the GTA and GVA as showing the greatest signs of stress.

The publicly available stress-test frameworks we reviewed model rate shock and employment deterioration separately, not in combination. Near-prime delinquency is already outpacing subprime — the stress is migrating upward through the credit spectrum, not staying contained at the margins.

The transmission mechanism connecting global conditions to Canadian households runs through the repo market. $16 trillion in government bond-backed repos operate globally, half with overnight maturity, roughly 70% of the non-centrally cleared segment with zero collateral cushion. A liquidity shock that forces leveraged players to unwind repo positions spikes government bond yields — which sets the benchmark for Canadian fixed mortgage rates. The Bank of Canada governor said it plainly: the system cannot afford to add financial instability to the mix.

The External Pressure

Exports to the United States account for 16.8% of Canadian GDP and support 2.6 million jobs. [15] The Bank of Canada estimates that U.S. tariffs leave GDP approximately 1.5% lower by end-2026 than baseline. [16] The PBO projects a $12.9 billion annual GDP impact from tariffs and less favourable trading conditions. [1] Canada is dependent on a trading partner simultaneously imposing tariffs and fighting a war that elevates global energy prices. For the deeper treatment, see The Tariff Whiplash.

In plain English: Canada is spending more than it earns during the good times. The indicators that have historically preceded every major economic downturn in the last 50 years are present right now. The money the government is investing is pointed at a different set of problems than the ones that may be arriving. None of this guarantees a downturn. All of it means Canada has less room to manoeuvre if one comes.

Context — What Both Sides Omit

Deficit critics point to the fiscal position but rarely acknowledge that some of the spending — defence, housing, clean energy infrastructure — may be structurally necessary regardless of cycle timing, and that Canada’s debt-to-GDP remains moderate by G7 standards. The government points to employment stability, moderate debt ratios, and the 85% mortgage-resilience figure, but rarely acknowledges that employment is historically the last indicator to turn, that the 85% figure assumes no employment shock, and that its own fiscal anchor is already being exceeded. Both sides are correct about their own data. Neither is engaging with the other’s.

Our Interpretation

Canada’s current combination — fiscal deficits during expansion, household debt exposure concentrated in the mortgage renewal wave, workforce spending misaligned with emerging AI-driven displacement, and trade dependence on a partner imposing tariffs — constitutes a weaker structural starting position than either 2007 or 2019 on every documented dimension we examined. This does not predict a downturn. It documents that if one arrives, the fiscal buffer, household resilience, and workforce alignment that cushioned previous cycles are all diminished relative to their historical baselines. The absence of a centralized federal budget breakdown that makes the allocation between clean economy, trades, and AI-adaptation legible is, in our assessment, itself a gap in the public record.

Counter-interpretation: The deficit spending may be front-loading necessary investment that positions Canada better over a 10-year horizon even if it creates short-term vulnerability. Budget 2025 projects the federal debt-to-GDP ratio stabilizing in the near term and declining from FY2026–27 onward. Employment remains stable at 4.4%. Canada’s debt-to-GDP is still moderate by international standards. The mortgage stress tests introduced after 2008 mean household resilience is higher than raw renewal numbers suggest. The Bank of Canada’s 85% coverage figure indicates most households have buffers, and the Bank’s analysis assumes stable labour markets — meaning the Bank is not itself claiming the system is fragile under current conditions. And late-cycle indicators have produced false signals before — the 2015–16 commodity crash triggered several of these same indicators without a Canadian recession.

What Would Change This

If Canadian employment remains stable through Q3 2026 and mortgage delinquency rates plateau or decline rather than broadening beyond the near-prime segment, the late-cycle thesis weakens significantly.
If the FY2025–26 deficit comes in materially below the PBO’s 2.2% projection, or Budget 2026 demonstrates a credible path to the sub-1% fiscal anchor, the “buffer already spent” framing overstates the risk.
If U.S. tariffs on Canadian steel, aluminum, and automobiles are materially reduced by Q4 2026 and the Iran conflict reaches a ceasefire that reopens Strait of Hormuz shipping, two of the major external pressure vectors diminish simultaneously.
If the government announces a significant federal program targeting AI-driven knowledge-worker displacement, the spending-mismatch finding weakens.
If commodity prices retreat without an accompanying demand shock by late 2026 — as occurred in 2015–16 — the supply-narrative driver behind the current hard-asset rotation may resolve without broader economic deterioration, and the late-cycle signal cluster would register as a false positive.

Sources (24)

No corrections have been issued for this article. Corrections policy