PROTOTYPE — throwaway, hand-authored. This is the existing weekly macro report (2026-W21) re-authored in a deeper, IB/World-Bank-style format to align on direction before it gets specced. Numbers are verbatim from data/macro/regime/2026-W21.json; the narrative spine, cross-currents, and concept walkthroughs are the new layer under evaluation.
Macro Intelligence · Weekly · Rung 1 (4-bloc regime)

The world isn't tightening — except where it is

2026-W21 · week ending 23 May 2026 · US · China · Euro area · Japan

Read at: thesis narrative mechanism data
What's different from the current weekly: the current report is organized bloc-by-bloc (a scanner's shape) and names mechanisms like "M13" assuming you know them. This version leads with a thesis, builds the story around a central tension and cross-bloc currents a per-bloc scan can't see, and makes every concept expandable on demand — so the same artifact reads as a 30-second brief or a full teaching walkthrough.

The thesisL0

Almost no central bank is tightening, and three of the four major blocs are disinflating or growing softly — a backdrop that quietly favours duration. The dissonant one is the United States, where the standing regime still reads late-cycle and overheating (growth above trend, inflation re-accelerating to 3.8%) even as the manufacturing cycle is bottoming. The week's real question: is that US inflation pop a supply-driven head-fake the on-hold Fed can look through, or the first crack in the soft-landing consensus the rest of the world is leaning on?

The central tension: a late-cycle US with a troughing factory sectorL1

The US is sending two signals that don't usually sit together. Its standing classification is overheating / late-cycle overheating / late-cycle — above-trend growth with above-target inflation. Cyclicals and inflation beneficiaries can still run, but the central bank is biased to tighten, so duration and high-multiple growth are the exposed trades. — GDP was last reported at 3.3% (up from 3.0%) and headline CPI re-accelerated to 3.8% (from 3.3%), with core PCE core PCE — the Fed's preferred inflation gauge: personal-consumption prices excluding food and energy. It matters more than headline CPI because the Fed's reaction function keys off it. ticking up to 2.9%. Yet manufacturing tells the opposite story: the PMI PMI — Purchasing Managers' Index. Above 50 = the sector is expanding vs last month; below 50 = contracting. A timely read on activity. is 48.7 — still below 50, but rising from 48.0.

That "below 50 but rising" combination is its own mechanism, and it is the opposite of a deepening slowdown. It fires M13 — contraction easing, the trough channel, not M6 (deepening contraction), which would need a sub-50 PMI that is falling.

M13 Contraction easing — why a sub-50-but-rising PMI is a trough signal, not a slowdown
  1. Manufacturing PMI is below 50 but rising.
    Activity is still contracting in absolute terms (below 50), but the rate of contraction is easing — the second derivative has turned up, which historically clusters near cyclical lows.
  2. Forward-looking investors start to price the trough.
    Markets discount the turn before it completes; an improving-but-sub-50 PMI often coincides with the low in cyclical risk assets, well before the PMI itself crosses back above 50.
  3. Early-cycle cyclicals and small-caps tend to lead off the low; defensives' relative edge fades.
    The names most beaten down in the contraction have the most operating leverage to even a marginal improvement, so they re-rate first when the deterioration stops.
Helps
early-cycle cyclicalssmall-capsEM equity
Hurts
defensives (relative)long govt bonds, if recovery firms
Caveat (misfire condition): a one-month uptick that doesn't persist is a head-fake; the turn needs follow-through, and a sub-50 level still means activity is contracting in absolute terms — this is an inflection signal, not an all-clear.
House-view hook — empty. This is the textbook default; Stephen's divergence (if any) gets baked in here during the judgment-layer pass.

So the tension resolves into one watch-item: inflation. With the policy rate held at 4.5% and the 2s10s 2s10s — the gap between the 10-year and 2-year government bond yields. Positive = normal upward-sloping curve; negative ("inverted") = a classic late-cycle/recession warning. Currently +0.43, but it narrowed this week from +0.49. curve still positive at +0.43, there is no tightening impulse in the data yet — only the conditions for one. With the 10-year at 4.56% against 3.8% headline inflation, real yields real yield — the yield after subtracting expected inflation — what a lender actually earns in purchasing power (≈ nominal yield − expected inflation). Only modestly positive here, which keeps financial conditions from being truly restrictive despite the 4.5% policy rate. are only modestly positive. The soft read (interpretation, from the news): a Bloomberg note ties sagging global factory activity to a third month of a "war-induced energy crunch" — if the CPI pop is energy/supply-driven rather than demand, it's exactly the kind the Fed is more willing to look through. That distinction is the whole ballgame for the week.

→ Switch to the full walkthrough to expand the M13 transmission chain and the cross-bloc currents.

Cross-currents: the linkages a per-bloc scan missesL1

The standing-regime layer reads each bloc in isolation. The value is in how they push on each other this week — and they point the same way: a strong-dollar, disinflating-rest-of-world setup.

Japan is the cleanest disinflation signal. Core CPI cooled sharply to 1.1% (from 1.4%), firing M2 — downside inflation surprise. That lets the market price a slower BOJ normalization path; but the yen sits at 159 to the dollar, so a slower BOJ plus a higher-for-longer Fed keeps the carry trade carry trade — borrowing in a low-rate currency (the yen) to buy higher-yielding assets elsewhere. It unwinds violently when the funding rate rises or the funding currency strengthens — a key fragility when USD/JPY is this stretched. intact but fragile.

M2 Downside inflation surprise — the disinflation tailwind, step by step
  1. Core inflation falls more than the prior reading.
    A faster-than-expected cooling in the persistent part of inflation tells the market the central bank has more room to ease.
  2. The market prices an earlier or deeper easing path.
    Lower inflation removes the reason to keep rates high, so traders pull forward expected cuts into the curve.
  3. Real yields fall.
    The expected policy rate drops while inflation expectations stay anchored, so the inflation-adjusted return demanded by lenders declines.
  4. A lower discount rate re-rates long-duration assets upward; long bonds rally.
    Future cash flows are now discounted less harshly, lifting growth equities most; bond prices rise mechanically as their yields fall.
Helps
long-duration growth equitieslong-dated govt bondsgoldutilities, REITs
Hurts
JPY vs easing-laggard peersbank net-interest margins
Caveat: misfires if disinflation is driven by demand collapse rather than supply normalization — then the growth scare dominates and risk assets fall despite lower yields.

China is exporting disinflation and holding its fire. CPI is pinned at 0.0% and the loan prime rate is held at 3.0% — near-zero domestic prices with unused policy room. The credit impulse credit impulse — the change in the flow of new credit into an economy. A rising impulse front-runs stronger demand for commodities and cyclical goods, with a lag. China's is the swing factor for the whole EM/commodity complex. channel (M11) is dormant precisely because the LPR didn't move — but a near-zero-CPI economy holding rates flat is the setup where an eventual cut would matter most, globally.

The euro area is the one place a tightening impulse is being voiced. CPI is 2.1% — essentially on the ECB's target — and the deposit rate is held at 2.15%, so no mechanism fires from the hard data. But the soft read cuts against the global grain: ECB Governing Council member Stournaras was quoted this week saying a hike "may be inevitable to keep credibility." That's interpretation, not data — but it makes the euro area the singular hawkish outlier in a world that is otherwise on hold or easing.

The through-line: a US that can't fully relax on inflation, a rest-of-world that is disinflating, and a dollar index at 99.3 — that combination pressures the yen carry, gives China every reason (and room) to ease eventually, and leaves the euro area's lone hawk pushing the other way. The duration tailwind from M2/M13 is real but conditional on the US 3.8% print not forcing the Fed's hand.

The standing pictureL1

Each bloc's always-on growth × inflation × policy classification — the regime headline, independent of whether a change-mechanism fired this week.

United States

overheating / late-cycle

Growth above trend (GDP 3.3%), inflation above target and re-accelerating (CPI 3.8%, core PCE 2.9%), policy on hold at 4.5%. Cyclicals and inflation beneficiaries can still run; duration and high-multiple growth are the exposed trades.

● M13 contraction-easing active (mfg PMI 48.7, rising)

China

deflation watch / soft

CPI 0.0%, GDP decelerating to 5.2% (from 5.4%), LPR held at 3.0%, mfg PMI just above the line at 50.3. A low-pressure backdrop with policy room unused — the textbook setup for an eventual easing move to matter.

○ M11 credit-impulse dormant (LPR unchanged)

Euro area

on-target / stalling

CPI 2.1% (essentially at target), deposit rate 2.15%, GDP barely positive at 0.2%, PMIs fractionally above 50. Quiet on the hard data — but the lone voice floating a hike (soft read).

○ no mechanism active · 10Y feed gap (M9/M10 can't evaluate)

Japan

disinflating / early-normalization

Core CPI cooled to 1.1% (from 1.4%), headline 1.4%, policy held at 0.5%, unemployment low at 2.7%. The just-begun normalization is not being pressed; yen weak at 159.

● M2 disinflation-surprise active (core CPI −0.3pp)

What to watch — the prints that would flip the readL1

The dataL3

Every number above traces here, verbatim from the regime object. ⚠ = "as of" older than ~60 days — akshare's calendar functions return the latest actual print, so several US hard series are mid/late-2025 reads sitting next to current (May-2026) inflation and yields. This is a known freshness issue, surfaced here on purpose rather than hidden.

United States

IndicatorCurrentPriorAs of
CPI YoY3.8%3.3%2026-05-12
Core CPI (m/m)0.30.22025-08-12
Core PCE2.9%2.8%2025-08-29
Policy rate4.5%4.5%2025-07-31
GDP3.3%3.0%2025-08-28
Mfg PMI48.748.02025-09-02
Svc PMI52.050.12025-09-04
Unemployment4.2%4.1%2025-08-01
10Y yield4.564.572026-05-22
2Y yield4.134.082026-05-22
2s10s slope0.430.492026-05-22
Dollar index99.3299.192026-05-22

China

IndicatorCurrentPriorAs of
CPI YoY0.0%0.1%2025-08-09
Policy rate (LPR)3.0%3.0%2026-05-20
GDP5.2%5.4%2025-07-15
Mfg PMI50.350.42026-04-01
Unemployment5.2%5.4%2026-04-01
10Y yield1.751.742026-05-22
USD/CNY6.796.802026-05-23

Euro area & Japan

IndicatorCurrentPriorAs of
EU · CPI YoY2.1%2.0%2025-09-02
EU · Deposit rate2.15%2.15%2025-07-24
EU · 10Y yieldfeed gap (akshare covers US+China only)
EU · EUR/USD1.161.162026-05-22
JP · CPI YoY1.4%1.5%2026-05-22
JP · Core CPI1.1%1.4%2026-05-22
JP · Policy rate0.5%0.5%2025-07-31
JP · GDP / Mfg PMI / 10Yfeed gap (no akshare JP function)
JP · USD/JPY159.15158.892026-05-22

Concept libraryL3

The evergreen layer: timeless definitions, written once and linked from anywhere in the period read. The inline dotted termsLike this — click any dotted term in the text to expand its definition in place. above pull from here. In production this is one shared file across all weeks, and where the house-view sign-off lives.

nominal yield
The headline interest rate on a bond, before subtracting inflation.
real yield
The yield after subtracting expected inflation — what a lender actually earns in purchasing power. ≈ nominal yield − expected inflation.
discount rate
The rate used to convert a future cash flow into today's value. A higher discount rate makes far-off cash flows worth less now.
duration
How sensitive an asset's price is to interest rates. "Long-duration" assets (long bonds, fast-growing stocks whose profits are years away) move most when rates change.
policy rate
The interest rate a central bank sets directly (Fed funds, ECB deposit rate, China's LPR). It anchors all other rates.
rate path
The market's expectation of where the policy rate goes over the next year or two — the whole expected trajectory, not just today's level.
2s10s
The gap between the 10-year and 2-year government bond yields. Positive = normal; negative ("inverted") = a classic late-cycle/recession warning.
term premium
The extra yield investors demand to hold a long bond instead of rolling short ones — compensation for tying money up and for issuance/inflation risk.
PMI
Purchasing Managers' Index. Above 50 = expanding vs last month; below 50 = contracting. A timely read on activity.
credit impulse
The change in the flow of new credit into an economy. A rising impulse front-runs stronger demand for commodities and cyclical goods, with a lag.
carry trade
Borrowing in a low-rate currency (e.g. the yen) to buy higher-yielding assets elsewhere. Unwinds violently when the funding rate rises or the funding currency strengthens.
net interest margin
The spread a bank earns between what it charges on loans and pays on deposits. Higher short rates / steeper curves tend to widen it.
cyclicals
Sectors whose earnings rise and fall with the economic cycle — industrials, materials, energy, consumer discretionary.
defensives
Sectors whose demand is steady through the cycle — staples, utilities, healthcare. They outperform when growth slows.
disinflation
Inflation that is still positive but falling (prices rising more slowly) — distinct from deflation, where prices actually fall.