Tuesday, 25 August 2026

Market Regime

High-yield spreads tightened 4bps on the week to 2.69%, and the credit widening the 2026-08-24 entry flagged as the one input pulling against the classification did not extend into a third print — the sequence that would have made credit directional is retired rather than confirmed, and the record's only cross-asset dissent is gone with it. The curve moved in parallel: 5bps at both the 2Y and the 10Y, leaving the 10Y–2Y spread unmoved at 0.46%. The composition of the long-end move flipped from the prior level's split — real yield carried +3bps of it against +2bps of inflation compensation, where a week earlier breakevens did nearly all the work. A nominal 4.70% financed by a rising real rate is a tighter setting for long-duration equity than the same nominal print financed by expected inflation, which is the one input here that got less friendly, not more. VIX at 15.13 sits at the low end of the calm band rather than the upper reach it occupied on the prior read. Initial claims at 206K carry a 2026-08-15 stamp and are the only labor observation on the sheet under eight weeks old. FRED market prints are as of 2026-08-24.

Key Macro Reads

Regime Assessment

Measured: every input the model classifies on now agrees, and the disagreement that had been accumulating in credit has cleared. Inferred: a classification with no dissenting input is easier to state and harder to learn from. When the equity inputs, the volatility input and the credit input all point the same way, the read carries no information about which of them the model is actually leaning on, and the answer only becomes observable when one of them breaks.

Where the risk now concentrates is participation. The model's own qualifier on breadth is marginal, and 579 of 977 names above their 200-EMA leaves a narrow buffer to the edge of the healthy band — a few percentage points of deterioration, not a regime-scale event, is enough to make breadth the first input to fail on its own. That is the input worth watching for anyone tracking whether this classification is durable.

The rate move argues the other way at the margin. A flat 10Y–2Y spread means this was a shift in level with no change in slope, so it carries no fresh recession signal; but the real-yield share of the move is the channel that actually transmits to equity discount rates, and it moved against the tape rather than with it. One week of that is a data point.

The labor and policy block still carries a 2026-07-01 stamp — Fed Funds 3.63%, unemployment 4.1%, payrolls 158.9M, housing starts 1,239K. Those readings sit roughly eight weeks behind the market prints beside them and cannot arbitrate anything current, and the sole recent labor observation is a single weekly claims figure.

Fourteen prints, one environment. No volatility expansion, no credit event, no breadth breakdown has occurred inside the sample. That is coverage, not calibration, and the distinction cannot be resolved until the tape produces a condition the classifier could get wrong.

What Would Invalidate

• Breadth at 59.3% (579/977) retains the model's marginal flag. A fall out of the healthy band makes participation the first equity input to fail unaided and is the cleanest refutation available from this block.
• VIX at 15.13 sits in the calm band. A sustained move out of that band removes the one condition every print in this record has shared.
• SPY at 763.46 is +7.2% over its 200-EMA of 712.01. Sustained compression of that gap alongside weakening breadth pressures the classification on equity inputs alone, with no help needed from rates or credit.
• HY at 2.69% tightened 4bps and broke the prior widening run. A resumption — particularly a move of similar or larger size in the opposite direction — restores credit as the dissenting input and re-opens the question the tightening closed.
• The real 10Y at 2.38% rose 3bps while the breakeven at 2.32% rose 2bps. Continued real-yield-led increases, rather than a return to breakeven-led ones, would work against duration-sensitive equity even with the nominal 10Y stable at 4.70%.

Forward Catalysts

• This block carries no dated calendar events; the catalysts below are data-refresh conditions, not scheduled releases.
• The next update to the 2026-07-01-stamped block — Fed Funds, unemployment, payrolls, housing starts — is the first thing capable of arbitrating the labor and policy picture with current data. Until it lands, the classification rests on market prints alone.
• The next weekly claims print tests whether 206K and its −6K WoW move were signal or a single draw from a noisy series.
• The next HY reading determines whether the −4bps is a turn or an oscillation inside the prior widening pattern.
• The next real-10Y and breakeven prints show whether the composition flip in the long-end move persists past one week.

Status

RISK-ON — model's authoritative read for 2026-08-25.

--