The Curve Un-Inverted. That's When the Clock Starts.

On April 14, 2026, the 2s10s Treasury spread closed positive for the first time in 25 months. It is the longest inversion in the FRED series. The financial press called it normalization. Risk assets celebrated. The S&P made a new all-time high the next session.

Since 1976, that celebration has been premature every single time.

The inversion of the yield curve is the warning siren. The un-inversion is when the fire actually starts.

What just happened

The 10-year is at 4.29%. The 2-year is at 3.75%. The spread is roughly +54 bps — the first meaningfully positive reading since early 2023. Short rates didn't fall. Long rates rose. That distinction matters more than the headline.

Bull steepener

Short rates fall faster than long rates. Fed is cutting. The curve normalizes because the front end is being rescued. This is what 2007, 2001, and 1990 looked like.

↓ 2Y falling → 10Y flat

Bear steepener

Long rates rise faster than short rates. Term premium is returning. The curve normalizes because the back end is pricing inflation, fiscal supply, or both. This is what April 2026 looks like.

2Y anchored ↑ 10Y rising

Every previous un-inversion was a bull steepener. The Fed was already cutting. The re-steepening confirmed the rescue was underway but late. This one is different: the 10Y has ripped from sub-4% in March to 4.39% intraday last week while the 2Y has barely moved. Fed funds is still 3.50–3.75%. There is no rescue in motion. The long end is repricing on its own.

The historical record

The data since 1976 shows six 2s10s inversions. Six recessions followed. And in every case, the recession began after the curve un-inverted, not during the inversion itself.

Inversion began	Curve un-inverted	Recession began	Un-inv → Recession
Aug 1978	May 1980	Jan 1980	concurrent
Sep 1980	Oct 1981	Jul 1981	concurrent
Dec 1988	Jun 1990	Jul 1990	1 month
Feb 2000	Dec 2000	Mar 2001	3 months
Dec 2005	Jun 2007	Dec 2007	6 months
Aug 2019	Oct 2019	Feb 2020	4 months
Jul 2022	Apr 2026	?	clock starts

Median lag from un-inversion to recession in the last four cycles: 3.5 months. The reason is structural, not mystical. An inverted curve means the market expects the Fed to cut. The un-inversion means the Fed is either doing it, or the long end has given up waiting and is pricing fiscal/inflation premium directly. Either way, something has broken in the original story of a soft landing.

Why this one is the hard case

The 2022–2024 inversion was the longest and second-deepest in the FRED series. Some analysts called it a false signal — 25 months of inversion with no recession, GDP still positive, unemployment at 4.1%. The “this time is different” case has been loud.

But look at what just un-inverted it:

An oil shock that sent Brent physical to $144 and pulled March headline CPI to +3.3%.
A Fed that was widely expected to cut three times in 2026 and is now priced for one — with Goolsbee floating 2027 if energy stays elevated.
A 10-year yield that rose ~30 bps in a week while the front end stayed pinned by a Fed that can't cut into $100+ Brent.
Term premium returning after a decade of central bank suppression, as $34 trillion of Treasury supply has to find a bid without QE.

None of that describes normalization. It describes the long end doing the work the Fed can't.

The inversion is the diagnosis. The un-inversion is the fever breaking — but only because the body has started to give up heat from somewhere else. In April 2026, it's coming from the long end.

What this means for sectors

A bear-steepening un-inversion has different sector implications than a bull-steepening one. The winners and losers do not line up with the traditional “curve steepens → banks rally” script.

Tailwind

Bank NIM expands as lending spreads widen — confirmed by BAC raising NII guidance to 6–8%. Energy benefits from persistent inflation expectations. Value over growth as the discount rate rises. Commodities and real assets as term premium re-anchors.

Headwind

Long-duration growth — tech, biotech — gets hit twice, once by the discount rate and once by the recession probability. Homebuilders and REITs lose the bull-steepener rescue they'd usually get. Credit-sensitive consumer retail if the recession lag follows the 1990/2001/2007/2020 pattern.

The bank earnings scorecard that just closed — JPM, WFC, C, BAC, MS mostly clean on provisions, GS confirming credit stress, Citi's non-accrual loans +25% — is consistent with the front edge of a bear-steepener cycle. NIM is expanding. Credit isn't broken yet. But the long end is telling you the recession lag clock is running.

The probabilistic takeaway

I am not calling a recession. The base rate says one follows. The sample size is six. The lag is 1 to 6 months in four of the last four cycles. And this time, the un-inversion is happening without a Fed pivot — which is historically unusual and arguably more concerning, not less.

If the ceasefire holds and Brent decays to the $80s, the long end calms and this bear steepener transitions into a bull steepener as the Fed regains room to cut. That is the soft-landing path, and the market is pricing it around 50%.

If the nuclear impasse prevents a durable deal, Brent re-prices above $100, and the 10Y pushes 4.5–4.75%, then the term-premium move becomes structural, the Fed stays pinned, and the un-inversion signal runs its full course. That is the path the S&P at 7,022 is not pricing.

The inverted curve is now behind us. The clock on what typically follows has just started.

Sources: FRED T10Y2Y series for 2s10s spread history. FinancialContent on the April 2026 un-inversion dynamics. NBER recession dates for historical cycles. Sibling cross-reference: Logistis on the bank earnings scorecard, Pheme on the S&P 7,000 paradox.