The Stress Gauge Fell 0.652 in Four Sessions. The All-In Cost of Borrowing Rose 4.4 Basis Points.
A credit market can be described two ways, and this week they point in opposite directions. The first is the stress gauge: a four-part average of high-yield spreads, investment-grade spreads, the VIX and the MOVE. The second is the all-in cost, which is what a treasurer actually signs - the Treasury yield plus the spread on top of it. Since 6 November the gauge has fallen 0.652, its largest four-session decline in this file, and more than all of that came from the two insurance prices rather than the two default prices. Over the same window the 10yr rose 10.4 basis points, so every borrower's base cost went up and only a fraction was given back in spread. The gauge reads relief; the invoice reads dearer. Ahead: the exact decomposition, why the equity tape never recovered its 5 November level, the 3 basis points that separate investment grade from a one-year high, and the one number that would settle which of the two descriptions is right.
The gauge fell 0.652 in four sessions, the most in this file
Credit & Stress is +0.308. On 6 November it was +0.960. That is 0.652 across four sessions, and across the 48 daily snapshots from 10 September to today no four-session stretch has produced a larger decline. The prior worst was -0.488, ending 30 October, and the third was -0.419, ending 29 October.
The path runs 0.960, then 0.907, 0.457, 0.420 and 0.308: four falls in a row, one of them -0.450 on 10 November, the largest single-session drop in this file, and three small ones of -0.053, -0.037 and -0.112. The 0.112 this session is the second largest of the four falls in the window. The level is the lowest since 28 October, when it printed 0.296.
The volatility legs supplied 105 per cent of it
The sub-index averages four z-scores, so dividing each leg's change by four attributes the move exactly. From 6 November the VIX leg contributes -0.610, the MOVE leg -0.074, high yield +0.022 and investment grade +0.009. The four sum to -0.653 against the reported -0.652; the residual is rounding in the z-scores.
The two insurance legs are 0.684 of the 0.652, or 105% of it, and the two spread legs are minus 5%. That sign is the point: over these four sessions the inputs that measure whether a borrower can pay moved slightly towards stress, and the inputs that measure what hedging costs moved hard away from it. A gauge that weights them equally will report the second as if it were the first.
One caveat belongs in the attribution. The spread legs are not session changes: the OAS leg of the 6 November snapshot carried the 4 November print and today's carries the 9 November one, so the leg labelled four sessions contains three sessions of spread data. That is why its contribution is a print-to-print move rather than a market move. At plus 0.031 against 0.652 it cannot change the conclusion, and the direction - spreads slightly wider, not tighter - is the same on either reading.
The VIX is 17.49, from 28.18 on 6 November - 10.69 points, or -37.9%, in four sessions. The next largest four-session fall in this file is 2.84 points, ending 23 September. The index sits at the 13.9th percentile of its own 3-year range with a z-score of 0.029. The MOVE is unchanged on the session at 92.43; from the 97.92 print of 6 November it is 5.49 lower, at the 43.6th percentile.
The equity tape that produced it has not recovered
The S&P 500 closed at 7,619.38 on 11 November, up 98.45 points or 1.31%, in the 94.1st percentile of its range. The Nasdaq 100 added 2.28% to 28,374.25. Those are the sessions that grind a volatility index down.
They do not restore the level. Against the 5 November close of 7,688.62 the S&P is 167.69 points lower, or -2.18%. The intervening session was 6 November, when the index fell 4.36% - the worst single day in this file - and the VIX printed 28.18, its highest print in the file. The VIX has given back 10.69 points of that since, but it is still 0.88 above the 16.61 it printed on 5 November.
That is the mechanical case for reading the collapse as a spike unwinding rather than as conditions easing. Equity volatility mean-reverts faster than it trends, and a move from the file's high print to its 14th percentile in four sessions is what the unwinding of one risk-off day looks like. It is not evidence that credit conditions improved, and it cannot be, because the VIX has no maturity, no coupon and no issuer.
Spreads tightened and stayed in the top 4 per cent of the year
High-yield OAS is 3.34% (FRED BAMLH0A0HYM2, as of 2026-11-09). The 1yr high is 3.48%, printed on 5 November, so the spread is 14 basis points inside it. The path across the week: 3.40% on 3 November, 3.32% on 4 November, 3.48% on 5 November, 3.42% on 6 November, 3.34% on 9 November. It is 44.4 basis points above its 1yr mean of 2.896% and in the 96.6th percentile of the 263 prints in that window.
Investment-grade OAS is 0.91% (FRED BAMLC0A0CM, as of 2026-11-09), from 0.93% on both 5 and 6 November, so 2 basis points tighter and 3 basis points off the 1yr high of 0.94% set on 16 March. It is 10.6 basis points above its mean of 0.804% and in the 96.6th percentile of 263 prints.
Both spreads sit in the top 3.4% of their own yearly distributions while the stress gauge reports a move to +0.308, the 75th percentile of this file's 48 sessions. That is the same divergence stated in levels rather than changes: the gauge is a z-score of a spread that has spent the year wide, and a z-score of 0.813 on a spread 14 basis points off a high reads as unremarkable. The percentile and the z-score are two descriptions of one print, and only the first is actionable.
Excess spread, high yield less investment grade, is 243 basis points, from 255 on 5 November (3.48% against 0.93%) and 228 on 28 October (3.14% against 0.86%). The ratio of the two spreads is 3.67, from 3.74 on 5 November. Both legs moved together, so the ratio barely moved: this is the whole complex easing off a peak, not one quality decoupling.
The all-in cost rose while the gauge said stress fell
The 10yr closed at 4.998% on 11 November, from 4.894% on 5 November - plus 10.4 basis points. Add the spread and the borrower's rate is 8.338% for high yield against 8.294%, plus 4.4; and 5.908% for investment grade against 5.824%, plus 8.4.
Both rose. The stress gauge fell 0.652 over the same window and now reports the middle of its range. The two statements are not in contradiction: the gauge measures the price of insurance and the compensation for default relative to their own histories, and the all-in cost measures the level of the rate the borrower pays. Insurance got cheaper and money got dearer, over the same four sessions.
The ordering inside it is worth a line. Investment grade's all-in cost rose more than high yield's, 8.4 against 4.4, because its spread gave back only 2 basis points of the base's 10.4 while high yield gave back 6. The better credit is the one whose cost rose most this week, which is what a base-driven move looks like: the spread was not the story, the risk-free rate was.
The base rose on the real leg; inflation compensation is near its year low
Inflation compensation is 2.20% (FRED T10YIE, as of 2026-11-10), from 2.24% on 5 November - minus 4.0 basis points, and in the bottom 2% of the 251 prints in the 1yr window. The window's low is 2.18%, set on 24 June; its high is 2.50%, on 4 May.
The real yield is 2.66% (FRED DFII10, as of 2026-11-09), from 2.60% on 5 November - plus 6.0 basis points, in the 90.8th percentile of the year against a 1yr mean of 2.165% and a high of 2.95% on 5 October. So the part of the rate that has to be earned in goods and services rose while the part that compensates for expected inflation fell to near the bottom of its own distribution.
The three series do not close to the basis point, and the reason matters. T10YIE and DFII10 print rounded to two decimals, they carry different last dates (10 and 9 November respectively against an 11 November nominal), and 10.4 minus 4.0 does not equal 6.0. The snapshot's derived real yield, 2.798%, pairs the 11 November nominal with the 10 November breakeven and is not quoted here for that reason. What survives all three readings is the direction: a nominal rate 10 basis points higher, an inflation premium in the bottom 2% of the year, and a real yield in the top tenth of it.
This is the sharpest tension in the snapshot. Scarcity, the sub-index that owns the real yield, is +1.115 and dominant for a fourth session; Credit & Stress is +0.308. One variable - a real yield 6 basis points higher in four sessions and sitting at the 90.8th percentile - reads as a scarcity signal there and as a credit headwind here, and both readings are arithmetic rather than opinion. What separates them is who pays: scarcity is priced by the holder of an asset, credit by the issuer of a liability.
Funding is quiet, and quiet here means empty
SOFR is 3.88% against an administered rate of 3.90% (FRED IORB, as of 12 November), a spread of minus 2 basis points. Secured money is fixing below the facility that is meant to set the floor, which is the opposite of the plus 7 basis point tilt this column marked on 5 November.
Reverse repo take-up is $0.2bn (FRED RRPONTSYD, as of 10 November), so the facility is holding almost nothing and cannot drain much further. Reserve balances are $2,930.1bn (FRED WRESBAL, as of 4 November - 8 days behind the market marks in this piece), up $36.3bn on the 28 October print of $2,893.8bn.
Initial claims were 189k (FRED ICSA, week ended 31 October), from 195k, in the 3.8th percentile of the 53 weekly prints in the window; continuing claims 1,718k (FRED CCSA, week ended 24 October), from 1,695k, in the 13th. Core CPI is null in this snapshot and is not quoted.
Nothing in the funding or labour data forces a seller. Reserves are rebuilding, the facility is empty, claims are near the bottom of their range, and the secured rate is fixing through the administered one rather than above it. That is the strongest part of the benign case, and it is a statement about plumbing rather than about spreads.
The composite is NEUTRAL for a second session, on Scarcity
The headline is +0.245, from +0.336. The NEUTRAL band runs from -0.35 to +0.35, so the distance to TIGHTENING is 0.105 and to EASING 0.595. The label crossed to NEUTRAL on 11 November after 14 consecutive TIGHTENING sessions, and this is the second session in the new regime. The margin to the upper threshold is 0.105, about 1.8 times the 0.058 mean absolute single-session move in the headline across this file.
The five sub-indices: Rates & Policy +0.865, Scarcity +1.115, Credit & Stress +0.308, Growth & Inflation -0.431, Liquidity -0.630. Coverage is 5/5 sub-indices and Scarcity is dominant for a fourth session.
Read the shape rather than the headline. The two sub-indices that measure the price of money - Rates & Policy and Scarcity - are the two positive ones and both are high. The two that measure activity and liquidity are negative. Credit & Stress sits between them at +0.308, falling for four sessions on an input that is not credit. A composite averaging those five describes a market where money is expensive and activity is soft, which is not the same as a market where default risk is low.
What would confirm it, and what would break it
The read here is that the four-session collapse is the unwinding of the 6 November volatility spike, not an improvement in credit conditions, and that the gauge and the all-in cost will converge on the second. The confirming case is the next OAS prints at or above 3.34% and 0.91% - unchanged or wider - while the VIX stays below 20: insurance cheap, spreads still wide, and the all-in cost for high yield holding above 8.29%, its 5 November level.
The credit read is invalidated if high-yield OAS tightens through 3.14% - the 28 October print - with investment grade below 0.86%. That would put the complex back where it was before the widening began, and the gauge would have been right for a reason it could not see.
The stress read is invalidated from the other side by a VIX above 24 with high yield still above 3.30%, which would make this week a pause inside a volatility regime rather than an exit from one. The VIX printed 28.18 six sessions ago, so the level is not far.
The all-in read is invalidated if the 10yr falls back below 4.90% while the OAS holds, taking high-yield all-in cost under 8.24%: then the base, not the spread, was the whole story and the direction reverses. The funding read is invalidated if SOFR-IORB returns above plus 5 basis points with reverse repo still under $1bn, which would put the quarter-end tilt back on a market that currently has none.
Notes
1. Credit & Stress +0.308 from +0.960 (6 November), -0.652 over 4 sessions; the largest four-session decline in the 48 snapshots from 2026-09-10 to 2026-11-12 (prior: -0.488 ending 30 October, -0.419 ending 29 October). Path 0.960, 0.907, 0.457, 0.420, 0.308; session changes -0.053, -0.450, -0.037, -0.112. Largest single-session declines in the file: -0.450 (10 November), -0.371 (5 November), -0.260 (18 September). Prior level below 0.308: 0.296 on 28 October.
2. Leg attribution divides each leg's z-score change by four. From 6 to 12 November: HY OAS 0.723 to 0.813 (+0.022), IG OAS 0.305 to 0.341 (+0.009), VIX 2.469 to 0.029 (-0.610), MOVE 0.345 to 0.048 (-0.074). Sum -0.653; reported -0.652, the difference being rounding in the z-scores. The OAS legs embed 4 November and 9 November prints respectively, so they span three sessions of spread data inside a four-session window.
3. High-yield OAS 3.34% (FRED BAMLH0A0HYM2, as of 2026-11-09); week path 3.40% (3 November), 3.32% (4 November), 3.48% (5 November, the 1yr high), 3.42% (6 November). 1yr mean 2.896%, low 2.60% (28 August), high 3.48%; n=263; percentile 96.6.
4. Investment-grade OAS 0.91% (FRED BAMLC0A0CM, as of 2026-11-09) against 0.93% on 5 and 6 November. 1yr mean 0.804%, low 0.73% (22 January), high 0.94% (16 March); n=263; percentile 96.6.
5. VIX 17.49 from 28.18 (6 November), -10.69 points or -37.9%; 13.9th percentile, z 0.029, 3-year window. Prior largest four-session fall 2.84 points, ending 23 September. MOVE 92.43 unchanged on the session, from 97.92 (6 November), 43.6th percentile, z 0.048.
6. S&P 500 7,619.38 (+98.45, +1.31%, 94.1st percentile); Nasdaq 100 28,374.25 (+633.46, +2.28%, 81.9th percentile). 6 November: S&P -4.36%, the worst session in the file. Against the 5 November close of 7,688.62 the index is -2.18%.
7. 10yr 4.998% (11 November) from 4.894% (5 November), +10.4bp. All-in high yield 8.338% against 8.294% (+4.4bp); investment grade 5.908% against 5.824% (+8.4bp). All-in is the 10yr plus the OAS print, so the two legs carry different as-of dates by construction.
8. Breakeven 2.20% (FRED T10YIE, as of 2026-11-10) from 2.24% (5 November), -4.0bp; percentile 1.6 of 251 prints, window low 2.18% (24 June), high 2.50% (4 May), mean 2.316%. Real yield 2.66% (FRED DFII10, as of 2026-11-09) from 2.60% (5 November), +6.0bp; 90.8th percentile, mean 2.165%, high 2.95% (5 October). The three series do not reconcile to the basis point: rounded inputs and different last dates. The snapshot's derived real yield of 2.798% pairs an 11 November nominal with a 10 November breakeven and is not used in the text.
9. SOFR 3.88% (NY Fed, snapshot 12 November) against IORB 3.90% (FRED IORB, as of 12 November), minus 2 basis points. Reverse repo $0.2bn (FRED RRPONTSYD, as of 10 November). Reserves $2,930.1bn (FRED WRESBAL, as of 4 November) from $2,893.8bn (28 October), +$36.3bn. TGA is not quoted: the snapshot carries no as-of date for it.
10. Initial claims 189k (FRED ICSA, week ended 31 October) from 195k; continuing claims 1,718k (FRED CCSA, week ended 24 October) from 1,695k. Core CPI is null in this snapshot and is not quoted. Scarcity +1.115, dominant for a fourth session.
11. Composite +0.245 from +0.336; regime NEUTRAL (band -0.35 to +0.35), second session after 14 TIGHTENING sessions; distance to TIGHTENING 0.105, to EASING 0.595. Sub-indices: Rates & Policy +0.865, Liquidity -0.630, Credit & Stress +0.308, Growth & Inflation -0.431, Scarcity +1.115. Coverage 5/5 sub-indices. Mean absolute single-session headline move across the file 0.058. The headline is quoted from data/latest.json, generated 07:00 UTC; the frozen snapshot file for 12 November prints 0.329 for the same date because its Liquidity and Growth legs carry re-marked inputs. The Credit & Stress legs are identical in both, which is why the sub-index quoted throughout this piece is unaffected.
12. Treasury par curve, 10 November close: 2yr 4.73%, 3yr 4.74%, 5yr 4.81%, 7yr 4.89%, 10yr 4.98%, 20yr 5.39%, 30yr 5.35%; 2s10s 25.0 basis points on the rounded prints and 26.8 on the unrounded series used in the Levels line. Source: US Treasury daily par yield curve, 10 November row, from the local ledger.
13. Both charts are generated by script from this piece's own series. The first divides the four Credit & Stress input z-score changes by four; the second adds the 10yr change to each OAS change. Single reported values (the 3.34% and 0.91% OAS prints, the 2.20% breakeven, the 2.66% real yield) are prints for one date, not index levels; percentile ranks are positions inside each series' own history and are not comparable across series.
| TENOR | LEVEL | 1D Δ | Z (3Y) | RANGE %ILE |
|---|---|---|---|---|
| 2Y | 4.800 | -0.030 | +1.73 | 87 |
| 5Y | 4.900 | -0.030 | +2.38 | 90 |
| 10Y | 5.100 | -0.020 | +2.63 | 89 |
| 30Y | 5.460 | 0.000 | +2.19 | 88 |