Regime Check — August 7, 2026: The Freeze Converts, B Takes the Lead

Builds-on: regime-check-july-11-2026 Related: how-inflation-dies-the-empty-reservoir (weights re-cut; tracker updated), failure-cascade-index, two-economy-gauge, the-hidden-bank-and-channel-1-yen-carry-unwind, the-shadow-balance-sheet-nikkei-1-65t-and-the-spv-layer, china-oil-buffer-thesis-audit, demand-destruction-or-strategy-the-china-import-cut-adjudicated, apartment-glut-and-the-multifamily-lender-tell, the-involution-import-open-weight-deflation-and-frontier-pricing-power, cyclical-20-and-the-ai-capex-mask

The corridor (July 14 – August 7) is complete. Its last and largest print: July payrolls -23,000 — the first negative print of the cycle — against +83-95K expected, with May revised 129K→63K and June 57K→20K (-103K combined). The three-month reality: +63K, +20K, -23K. The labor freeze the vault has tracked since May didn't thin further; it converted. Per the pre-committed reaction function published July 30 and restated August 5 ("sub-25K or negative print → B takes the lead"), the weights move: A 38→33, B 36→42, C 16, D 9. For the first time since the scenario frame was built, the break is the most likely ending.

1. Payrolls, scored against the pre-commitment

• The crack condition fired on the payrolls leg: -23K headline (BLS, CNBC), negative revisions re-confirming the QCEW-style downward pattern — demand-side-audit-may-2026's revision thesis is now fully rehabilitated after being "corrected" June 10 and half-restored July 11. The June-10 lesson (never trust a strong headline against the revision trend) is the cycle's most expensive tuition, now paid in full.
• Composition tells the two-economy story to the end: government -53K (the fiscal pump now negative), retail and leisure/hospitality soft (confirming ADP's -11K L&H — the last pillar from June's "73% of gains" duo is now shedding), healthcare slowing. Wages +3.2% YoY — the lowest since May 2021. All four reservoir pumps (wages, savings, credit, transfers) now read failing simultaneously.
• The formal stage-5 tripwires did NOT fire — and the reason is itself diagnostic. U-3 fell to 4.1%, but only because participation fell again; Bloomberg's headline said it plainly: "Employers unexpectedly shed jobs, unemployment rate falls". Claims are benign (199K initial, continuing 1.801M vs the 1.95M tripwire). Sahm stays mechanically suppressed. Read: this is hiring cessation plus attrition, not synchronized firing — stage 5's opening act, not its main event. The unemployment-side dials can't see a contraction that proceeds by people exiting the labor force.
• The market's reaction is the 1998 branch announcing itself: stocks rose on the negative print (Nasdaq +0.9%) as September hike odds collapsed 55%→40%; 2y -8bp to 4.16%, 10y to ~4.61%. Bad news is good news again — the rescue-expectation regime that historically precedes terminal melt-ups. Note the double edge: the print that raises B's weight also defused the September policy-error trigger. B's probability rose while one of its trigger paths softened — the mass moved toward the demand-rolling and funding-event channels.

2. The corridor scorecard (July 14 – August 7, complete)

3. Developments formalized since July 11 (the between-prints research)

• Channel 1 is live — all three signals fired. Yen touched 163.99 (July 23), Japan funded its defense by selling Treasuries, and the first coordinated US-Japan yen intervention since 1998 followed (July 31/Aug 1, ~$36.6B, NY Fed selling euros). Yen ~155 post-intervention; the decay clock on this round is running (April-May's round bought six weeks). The 1998 precedent — coordinated intervention preceding the carry snap by three months — defines the 1998 branch: a rescued first break producing emergency easing and one terminal melt-up before the real top. Assigned ~⅓ of B's mass. Key structural difference from 1998: the BoJ is hiking, so a rescue-driven melt-up cannot be Japan-funded — it would require the Fed's own balance sheet, i.e., the D machinery. See the-hidden-bank-and-channel-1-yen-carry-unwind.
• The Iran binary dissolved into a toll-duopoly oscillation. The Versailles MoU collapsed in late July (Iran banked concessions, declared the strait closed; the US reimposed the blockade and its own 20% "Guardian of the Strait" toll). August 16 is retired as a watch item. The war is now a permanent whipsaw band — A-food, C-poison — with continuous miscalculation tail instead of dated binaries. Oil: Brent ~$85 on deal-hopes round three.
• The shadow balance sheet got its census and its smoking gun. Nikkei: $1.65T off-book across five hyperscalers (Meta ~$420B, Oracle ~$273B), 8x growth in four years — and Meta guaranteed Hyperion investors against losses, collapsing the "is it really their risk" debate. Japanese institutions are structurally significant holders of the funding stack precisely as repatriation reverses the flow — the scissors: the buyer base leaving as issuance peaks. See the-shadow-balance-sheet-nikkei-1-65t-and-the-spv-layer.
• The China import cut adjudicated as distress-enabled strategy; the restock bid (XLE floor) downgraded to contingent, with Test A (throughput vs imports) resolving Aug-Sep and the buffer's first non-zero draw (~940 kbd June) on the clock.
• Sentiment cracked while positioning held (AAII bears 42% with the index near highs), and the doom narrative graduated from monetized fringe to institutional adoption (48% of BofA managers naming AI DC debt the top systemic risk) — the recognition phase, on schedule.

4. Weights — re-cut per the pre-commitment

Reading appended to the tracker in how-inflation-dies-the-empty-reservoir.

4b. B decomposed — the five modes of the break (added 2026-08-07, same day)

B's 42 points are not one path. Decomposition, with each mode's share, mechanism, and discriminating dial:

The modal single answer: the caught break (B-1998). The evidence is already performing it — stocks rallying on a negative payroll print is the rescue reflex operating in real time, the backstop ladder has been warming all summer, and the suppression apparatus retains deep capacity. Most-likely composite sequence: a break attempt in the September-October cluster (triggered by marks, carry, or funding — the trigger modes chain into the catch), caught within days, melt-up into early 2027, and the completed break arriving when the rescue meets the 2027 issuance wall. Practical translation: expect maximum whipsaw, not a clean crash — the environment that punishes conviction bears and rewards written rules. The uncaught modes (credit + marks + carry ≈ 18 of B's 42) are the true-crash mass; attrition (10) is the no-drama recession; the catch (14) is the crash postponed at compound interest.

5. Portfolio — rules status

• The TIPS rotation rule is half-armed, not executed. The crack branch fired on the payrolls side, but the rule's own dials (U-3 >4.6-4.8%, Sahm >0.30, claims >1.95M) remain unfired — the participation mechanics suppress them, which the rule's authors (us) did not fully anticipate. Amendment, logged now: the rotation executes on EITHER (a) a second consecutive negative payroll print, OR (b) claims >1.95M / U-3 ≥4.4%, PROVIDED breakevens ≥2.2% at execution. Breakevens were 2.24% Aug 3; today's rally needs a fresh read before any action. Partial rotation only, per the standing analysis (2026-08-05): the sleeve's B-damage is opportunity cost, and the A/D half of the board (42 combined) still owns the position.
• STIP-over-SCHP tilt for new money stands — though today's print starts the clock toward "last hike is in," after which extending duration locks 2.4%+ real yields.
• XLE: hold; Test A (China restock) is the dated exit condition, resolving Aug-Sep. The Iran oscillation band and the war's toll-duopoly floor keep the insurance live.
• Household base: unchanged and validated — 66% index target-date, 6.25% mortgage lock, energy costs structurally deleted, no leverage. The book was built for exactly this tape.

6. What to watch — the autumn cluster, now with B in the lead

1. Breakevens vs 2.2% — daily now. The rotation's arming condition; a fresh read is the first action item.
2. Claims, weekly — the series that converts "attrition" into "firing"; 1.95M continuing is the line.
3. September FOMC (16-17) — hike odds 40% and falling; the meeting that either completes the policy error or begins the 1998-branch rescue.
4. BoJ September meeting + yen 160 retest — the intervention decay clock; Takata's 1.25% is live.
5. The October block: Anthropic public S-1 (~Sept) then debut at a reported $1T; SpaceX lockup + first-ever earnings call; Q3 prints (Tesla carries the SpaceX markdown; hyperscaler FCF). The forced-truth-telling cluster.
6. Hyperscaler debt absorption — the digestion-vs-spiral line (a pulled deal, an undersubscribed book) now carries more of B's probability than the Fed does.
7. Epiq August filings + FDIC Q2 small-bank data — the failure cascade's breadth test and the triple-loaded tier.
8. China Test A (imports vs throughput, monthly customs) — the XLE adjudicator.
9. Midterm-season fiscal noise — D's pre-emptive variant; watch for stimulus trial balloons as the labor data headlines worsen.

7. Scorecard

Right: The pre-commitment discipline itself — the weights moved on a written rule, not a reaction; "frozen, and thinning faster" (July 11) → conversion called by trajectory; the revision-pattern rehabilitation; the breakevens divergence dial (the bond market saw this print coming at 2.22% three weeks ago); ADP's L&H crack as the payroll preview; the corridor calendar (every load-bearing event landed inside the mapped window).

Wrong / instructive: The stage-5 tripwires were mis-specified — all unemployment-side, blind to a participation-suppressed contraction; amended above. June-10's confidence in the labor-strength correction was the cycle's worst read (now twice reversed). The Aug 16 binary was retired by events within two weeks of being set — deadline-shaped analysis keeps losing to oscillation-shaped reality.

The one-line regime statement, updated: The grind ended in July; the economy is now contracting by attrition behind a participation mask, the Fed's hike case just died, the rescue posture is warming up, and the only question the autumn answers is whether the first break gets caught — and funded by whom.

Sources

• BLS Employment Situation, July 2026 · CNBC jobs report · Bloomberg: employers shed jobs, unemployment falls · Fox Business: wages 3.2%
• CNBC: September hike odds tumble 55→40 · Quartz: S&P rallies on the miss · TheStreet: market close Aug 7
• TradingEconomics: claims 199K / continuing 1.801M · WikiFX: yen ~155 post-intervention, oil after the crash · Yahoo: oil sinks on Iran hopes, yen gains after joint intervention
• Corridor prints: see regime-check-july-11-2026 sources plus per-doc citations in the Related links above