How Long Does "Late Cycle" Actually Last? What History Says — and Why 2026 Is a Genuine Debate

How Long Does "Late Cycle" Actually Last? What History Says — and Why 2026 Is a Genuine Debate

If you've been reading market commentary lately, you've probably noticed strategists throwing around the phrase "late cycle" a lot. It sounds precise, like a stage on a map with clear borders. It isn't. Here's what the term actually means, what history tells us about how long it tends to last, and where the debate over today's market stands — in plain language, for anyone trying to make sense of where we are.

First, an important caveat

Recessions have an official referee. In the U.S., the National Bureau of Economic Research (NBER) Business Cycle Dating Committee determines the exact start ("peak") and end ("trough") of every recession — but only well after the fact, based on a broad basket of economic data rather than a single number like GDP.¹ Historically, that committee has announced recession start dates an average of six to seven months after they actually began.²

"Late cycle," by contrast, has no such referee. No committee declares when it starts. It's a descriptive label that strategists apply in hindsight, based on a cluster of signals — the yield curve flattening or inverting, the Federal Reserve shifting to restrictive policy, valuations stretching, and market gains narrowing to fewer and fewer stocks. That means every duration estimate below is an approximation drawn from how different analysts have retrospectively labeled past cycles — not a hard, agreed-upon date.

What history shows: late cycle typically runs 2 to 3 years

Looking back at the last three full U.S. expansions (which Canada's cycle has historically tracked fairly closely), a pattern emerges: once the classic late-cycle markers show up, that phase has generally lasted somewhere in the neighborhood of two to three years before either sliding into recession or getting cut short by an outside shock.

Cycle Full expansion length Approximate late-cycle window Approx. late-cycle duration

1990s expansion	Mar 1991 – Mar 2001 (~120 months, then record-longest)	~1998 (LTCM crisis, curve flattening) to Mar 2001	~2.5–3 years
Early 2000s expansion	Nov 2001 – Dec 2007 (~73 months)	~2005 (Fed tightening cycle, curve inversion in 2006) to Dec 2007	~2–2.5 years
Post-GFC expansion	Jun 2009 – Feb 2020 (~128 months, longest on record)	Widely dated to 2018 (Fed hiking into restrictive territory, Dec 2018 selloff, 2019 yield curve inversion) to Feb 2020	~2 years (cut short by an exogenous shock — COVID — rather than ending "naturally")
Current expansion	Apr 2020 – ongoing (~6+ years as of mid-2026)	Actively disputed — see below	Undetermined

*Note: expansion length and dating conventions follow the NBER chronology; late-cycle windows are approximate and reflect commonly cited strategist framing rather than an official designation.*³

The 2018–2020 cycle is the clearest illustration of why "2 to 3 years" is a guideline, not a rule. That late-cycle phase was arguably still unfolding — not yet resolved into recession — when COVID triggered the downturn. The cycle didn't die of old age; something external ended it. That's a useful reminder for investors: late-cycle phases don't come with a fixed shelf life. Central banks can sometimes stretch them out by engineering a "soft landing," or an unrelated shock can end one abruptly, well ahead of schedule.

Is 2026 late cycle? Strategists are genuinely split

This is where it gets interesting — and where investors should be skeptical of anyone claiming certainty.

The case for "still early-to-mid cycle": Some strategists point to AI-driven capital spending still accelerating, consensus corporate earnings growth still running strong (above 20% for 2026 in some estimates), and a Federal Reserve that has only recently moved to a cautious hold rather than aggressive tightening. Under this view, the classic overheating signature of late cycle simply hasn't shown up yet.

The case for "already late cycle": Morgan Stanley's late-2025 research lays out this debate explicitly. One camp argues the U.S. economy is already grinding through a genuinely late-cycle environment, marked by persistent labor-market softness that isn't fully responding to Fed rate cuts, with equity gains staying concentrated in a narrow band of AI-related mega-cap stocks rather than broadening across the market.⁴ That narrowing of market breadth is a textbook late-cycle signature.

Technical and sector-rotation evidence: Some technical analysts trace early signs of underlying market weakness back to mid-2025, alongside sector rotation into Energy and Materials and renewed strength in defensive sectors like Consumer Staples — another pattern historically associated with the late innings of a cycle.

The concentration and valuation argument: With AI-related stocks trading near record levels of index concentration, some strategists treat "late cycle" as effectively already underway — the open question isn't whether we're there, but when (or whether) it resolves into something worse.

Putting it together

There's no clean consensus date for when this cycle turned "late." But a reasonable synthesis of current commentary places the onset of late-cycle characteristics somewhere in mid-to-late 2025 — around when labor-market cooling became persistent and AI-driven breadth narrowing became pronounced. That would put the current late-cycle phase at roughly 12 to 18 months old as of mid-2026.

Historically, late-cycle phases have often lasted around 2 to 3 years, but whether the current cycle follows a similar timeline is uncertain. But as the 2018–2020 episode shows, late-cycle phases don't reliably resolve on a predictable schedule — an external shock, or a stumble in AI-related capital spending and monetization, could compress that timeline considerably.

The honest bottom line

This kind of dating exercise is more art than science. Strategists are genuinely divided, and the commentary above reflects the prevailing range of views in the market — not a confident forecast. This article is for informational purposes only, is not financial or investment advice, and shouldn't be read as a signal to change your portfolio. As always, talk to your advisor about how these dynamics apply to your specific situation and time horizon.

Sources

National Bureau of Economic Research, "Business Cycle Dating," and "Business Cycle Dating Procedure: Frequently Asked Questions." nber.org/research/business-cycle-dating
Recession-dating lag estimates (NBER announcements historically trailing actual recession starts by roughly 6–7 months on average, 1979–2021). See NBER FAQ (above) and related academic literature on recession detection.
NBER, "US Business Cycle Expansions and Contractions" (official peak/trough chronology); FRED (Federal Reserve Bank of St. Louis), "NBER based Recession Indicators for the United States" (series USREC).
Morgan Stanley, "Early Surge or Late Grind for U.S. Stocks?" (November 2025), morganstanley.com/insights/articles/2026-economy-business-cycle; Morgan Stanley, "Next Leg of the Bull Market?" (Thoughts on the Market podcast, Mike Wilson); Morgan Stanley, "U.S. Economic Outlook: What's Driving U.S. Growth in 2026?" (November 2025).

This post is for general informational purposes only and does not constitute financial, investment, or tax advice. Speak with a licensed advisor before making investment decisions.

This is not an official publication of Manulife Wealth Inc. The views, opinions and recommendations contained in this publication are those solely of the author and this publication does not express the views, opinions or recommendations of Manulife Wealth. This publication is not an offer to sell, or a solicitation of an offer to buy, any securities. This publication is not meant to provide legal, accounting, account or other advice. As each situation is different, you should seek advice based on your specific circumstances. Please call to arrange for an appointment.