The Corridor Narrowed 6 Basis Points. Credit Widened 9.

2s10s is 29.8 basis points on the UST marks, from 35.8 a week ago and the narrowest since 25 September. The 10yr did 11 of the work and the 2yr 5, rate volatility rose 2.2 points, high-yield OAS widened to 3.40% (22 Oct), and reserves fell $59.5bn in the week to 21 October.

The corridor is the gap between what Washington pays for 2 years of money and what it pays for 10. It is the market's one-line opinion on whether the next few years are a policy story or a term-premium story. On 19 October that gap was 35.8 basis points. On Friday's close it was 29.8, the narrowest since 25 September. The long end supplied the move: the 10yr fell 11 basis points while the 2yr fell 5, and the 3-month bill fell 4. That is duration being bought, not the funds path being repriced. The complication is that the same week credit widened, rate volatility rose, and the reserve buffer that used to absorb a cash drain finished the week at $0.20bn. This column has argued since September that the corridor contains one end at a time. This week it contained the back end, bought as insurance, into a spread market that was already paying up for it.

What moved

The 2yr closed at 4.69% and the 5yr at 4.82%. The 7yr 4.89%, the 10yr 5.00%, the 20yr 5.37%, the 30yr 5.33%. The 3-month bill 4.22%. All levels are the 23 October close.

The coupon curve fell across every point on the week, and the back did roughly twice the work of the front. The exact arithmetic is in the next section, where the marks the spread is computed from are named.

The long end is still expensive against its own history: the 5yr sits at the 83.9th percentile of its 3yr range, the 7yr at the 82.1st, the 10yr at the 81.5th, the 30yr at the 80.3rd, the 20yr at the 80.0th and the 2yr at the 78.9th. The 3-month bill is at the 31.6th. Dollar index 101.673, down 0.28 on the week; dollar-yen 157.01, down 1.46 (both 26 Oct).

The compression came from the back end

2s10s is 29.8 basis points, from 35.8 on 19 October. That is 6.0 basis points of compression in a week, and the narrowest since 25 September, when the spread printed 29.2.

The arithmetic is the argument, and it runs off the UST cash marks the site uses to compute the spread: the 10yr fell from 5.098% to 4.988%, 11 basis points, and the 2yr fell from 4.74% to 4.69%, 5. Eleven minus five is six. The indicative 10yr print went 5.09 to 5.00; the 1 basis point gap between the two is the mark, not the market. The 3-month bill fell 4 and the 30yr fell 10, so the front end moved least.

A policy repricing does the opposite. It moves the 2yr first and hardest, drags the bill with it, and leaves the 30yr to decide later. This week the administered rate did not move, the bill fell 4 basis points with it, and the long end took 10.

The site's 2s10s is computed from the UST cash marks rather than the par curve, and the par file has not refreshed since 19 October; every level in this piece is a 23 October mark or later and is dated where it appears.

The real yield gave back October

The 10yr real yield is 2.648% (as of 23 Oct), from 2.748% on 19 October. Ten basis points lower, and the lowest since 23 September, when it printed 2.633%.

Inflation compensation did none of it. The 10yr breakeven is 2.34% (as of 23 Oct) against 2.35% on 16 October. One basis point. The nominal carried the other nine.

Crude made that awkward. WTI is $101.25 (26 Oct), down $9.12 on the week, 8.3%, and the breakeven did not follow it down. September core CPI was 2.57% year on year and headline 3.72%, both for September and neither of them a market price.

Precious metals disagreed with the real rate. Gold fell $135.10 on the week to $4,021.40 (26 Oct) with the real 10yr 10 basis points lower, which is the wrong sign for the usual relationship. The copper-gold ratio ticked up to 0.00161 from 0.00158 on 19 October, and only because gold fell faster.

Rate volatility did not agree

MOVE is 103.31, up 2.21 on the week and at the 56.5th percentile of its 3yr range. Yields fell 9 to 11 basis points across the coupon curve and implied rate volatility rose. A rally the vol market has not underwritten is a rally being hedged.

Equity volatility went the other way. VIX is 19.45, down 0.54 on the week, at the 18.8th percentile.

The S&P 500 is 7,623.68, down 35.76 on the week, 0.5%, and still at the 94.7th percentile of its range. The Nasdaq 100 is 28,957.53, down 230.02, 0.8%. QQQ is $704.40, down $5.40.

Treasuries rallied, their own volatility rose, and equity volatility fell with the index barely changed. That is not two markets agreeing on good news; it is one market buying insurance and the other selling it.

Credit took the other side

High-yield OAS is 3.40% (as of 22 Oct), from 3.27% on 21 October and 3.31% on 16 October. Thirteen basis points in a session, 9 on the week, and the widest since 30 March, when it printed 3.46% — the 2026 high.

Investment-grade OAS is 0.89% (as of 22 Oct), from 0.88% on 21 October and 0.84% on 15 October. That is the highest since 31 March, when it was 0.90%.

Cash credit moved the other way. TLT added $1.29 on the week to $81.34, LQD added $0.45 to $103.85, HYG added $0.19 to $77.05 and JNK added $0.24 to $92.52 (all 23 Oct). The gain is the duration leg; the spread leg was working against it.

The cash flow still has not turned, which is the normal order rather than a contradiction, since spreads widen before defaults arrive. Initial claims were 194,000 in the week to 17 October, from 200,000 in the week to 10 October. Continuing claims were 1,728,000 in the week to 10 October, from 1,703,000 in the week to 3 October.

A curve that flattens while the riskiest spread widens to a seven-month high is a market pricing slower growth. It is not a market pricing easier money, because nothing in the front end says easier money.

The buffer is gone

Reserve balances were $2,962.6bn on 21 October, from $3,022.1bn on 14 October: $59.5bn out in a week. Overnight RRP take-up was $0.20bn on 23 October, from $0.35bn on 22 October. The Treasury's cash balance read $961.8bn in this morning's snapshot.

Secured overnight money is 3.90% against an administered rate of 3.90% (IORB, 26 Oct). The spread is zero, against minus 1 basis point on 15 October. Effective fed funds is 3.88%.

There is no parked cash left to absorb the next drain, so the adjustment falls in the funds market rather than in the RRP. SOFR against IORB is therefore the number that matters, and at zero it says the drain has not yet bitten. A $59.5bn reserve week with a $0.20bn buffer is the setup where it does.

The composite is 0.342, NEUTRAL, dominant Rates & Policy at 0.933, on 5 of 5 sub-indices. Liquidity is the only negative one at -0.506, against -0.43 on 23 October. Credit & Stress is 0.507 and Scarcity 0.835.

Bitcoin: the coin fell, the wrapper rose

BTC is $80,001 at 07:00 UTC, down 4.8% over five sessions and 5.3% over thirty. ETH is $2,317.22, down $154.53 on the week.

MSTR is $146.01 (as of 23 Oct), up $9.18 on the week, 6.7%. Against a net asset value of $176.15 per share on 846,000 coins and 384.23m shares, that is 0.83 times the coins it holds, against 0.75 times a week ago. The discount narrowed while the asset fell.

The preferreds split. STRC is $99.36, up $1.64, at the 99.4th percentile of its range. STRF is $101.20, down $2.40. STRK is $69.30, down $1.94.

A 4.8% fall in the coin and a 6.7% rise in the wrapper is a multiple expanding. The return came from the wrapper, not from the treasury it wraps. With the real 10yr 10 basis points lower, the carry got cheaper to hold that week, which is the one condition under which a discount closes without the coin moving.

What would change the read

The first is the corridor reopening at the back end. 2s10s back above 35.8 basis points with MOVE below 100 would say the compression was a hedge being unwound rather than a growth trade being put on.

The second is the real rate. A 10yr real yield back above 2.88%, the 2 October level, reverses the whole of October's fall and puts the September repricing back on the table.

The third is credit. High-yield OAS back through 3.10%, where it sat on 2 and 9 October, would say the widening was inventory rather than deterioration and return the story to rates alone.

Between those three lines the read is narrow and uncomfortable: the long end is being bought as insurance, the front end is anchored to an administered rate that has not moved, credit is widening into the rally, and the buffer that used to absorb a reserve drain is $0.20bn.