Oh Canada!!
morning musings 7.16.26- views from the island
“Who controls the past controls the future: who controls the present controls the past.”George Orwell, 1984
The sun came up brass this morning. Not gold. Brass. Over Toronto it went the colour of a nicotine ceiling. Over the Boston Seaport it hung behind the Ferris wheel like a bulb wrapped in gauze. From Jersey City the Empire State smeared into an orange-brown haze that photographers three years ago called once-in-a-lifetime and now file under Tuesday. 836 fires burning across Canada as of the CIFFC count, 194 of them flatly ‘out of control’ CIFFC. Ontario and Quebec exhaling the north woods into the Lower 48 on a wind shift, and half the Northeast woke up unable to see the end of the block. Pennsylvania went Code Red. Massachusetts alerted the whole state. The sky over Brattleboro turned the brown of an old photograph.
Which is fitting, because at 9pm tonight the President goes primetime to tell us what we did and did not see in an election that finished six years ago.
He has been coy about the contents. ‘Really big news,’ he told the Oval on Tuesday, ‘because without free and fair elections, you don’t have a country’ AP. The topic is ‘election security,’ the venue is the whole country, and the subtext could not be less subtle: he lost Georgia in 2020 by 11,779 votes, he asked the Secretary of State on tape to ‘find’ him 11,780, and he has never once let it go. Earlier this year the FBI raided the Fulton County elections office and carted off 2020 materials, with the sitting Director of National Intelligence flying to Atlanta to watch the warrant get served CNN. Gabbard has since departed. Her acting replacement, per the President, ‘may find out some things about the rigged elections.’ Ossoff, whose seat is on the ballot, calls it ‘reheating debunked conspiracy theories and launching bizarre new lies because he fears losing the midterms’ CNN. Every audit, several of them run by Republicans and one by the President’s own former Attorney General, found no fraud that mattered.
Did I miss the part where we agreed the past is now negotiable? Orwell put it in a slogan and we turned it into a policy. Control the present, seize the records, and the past becomes whatever tonight’s teleprompter says it was.
The rhyme runs south, too. Nicolás Maduro sits in a Brooklyn cell, 160-plus days now, having been lifted out of Caracas in January by Delta Force and flown to a Manhattan courthouse to answer a narco-terrorism indictment Rtrs. He calls himself a ‘prisoner of war.’ The trial hearing just slid to July 22, partly because MetLife has a World Cup final to host on the 19th and you cannot very well move a deposed head of state through Iran-war traffic during a tournament. On Tuesday a Miami judge handed three tortured Americans a $314mm default judgment against him and the ‘Cartel of the Suns’ MercoPress. And the throughline, the one the administration has never bothered to hide, is oil: we ‘run’ Venezuela now, sanctions lifted, barrels flowing, a 50mm-barrel deal and billions booked Rtrs. An election in Caracas that observers called fraudulent, a Nobel laureate sidelined, a Chavista installed as interim president, all of it laundered into a supply story. Set up for a clean conscience, perhaps.
So we have smoke. Real smoke over the northeast, and the other kind everywhere else. Which brings me, at last, to a tape that is doing the exact same thing: printing clean at the index level while the thing underneath it burns.
State of Play: the S&P closed 7,572.40 Wednesday, a whisker (48bp) under the 7,620.90 cycle high FMP. Apple all-time high, the megacaps up 3-4% carrying the whole thing on their backs. And underneath that placid surface, the highest-beta leadership group of the entire cycle is in freefall. Do not let the number at the top fool you about the fire at the bottom.
TL;DR ·This is a rotation, not a de-risk, and the honest version has two tensions the sell-side is smoothing over. One, the Fed is talking hawkish (Warsh ‘no tolerance,’ Cook ‘prepared to act,’ ~50% July-hike odds) into confirmed disinflation, so the 29-30 July FOMC is genuinely two-way and the bond market is leaning dovish against it. Two, credit at HY OAS 266bp is either ‘benign’ or ‘late,’ and it is the single print that decides whether the memory crash stays quarantined or becomes the whole tape. Non-consensus: stop trading ‘chips’ as one basket, do not chase financials beta at 52-week highs, own a cheap credit hedge while a 16 VIX pays almost nothing to insure. The one number that changes everything is HY OAS through 300.
My Read
I think the crowd is making the same mistake it made 48 hours ago, in the opposite direction, and calling it analysis both times. On Tuesday ‘chips’ ripped as one. Wednesday and into Thursday ‘chips’ crashed as one, KOSPI down 6.4% in a ‘Black Thursday’ that tripped a sell sidecar, SK Hynix and Samsung off double digits, Micron down 8% BBG. And the sell-side will tell you today either that the TSM blowout means ‘buy every chip on the dip’ or that the Korean bloodbath means ‘AI is cracking.’ Both are wrong for the same reason: they treat a bifurcation as a monolith.
Here is what actually happened. TSM printed a beat-and-raise, Q3 revenue guided $44.6-45.8bn against $43.1bn consensus, full-year growth lifted to ‘slightly above 40%,’ and capex raised to $60-64bn with another $100bn pledged for Arizona PDB. That is a company telling you leading-edge compute demand is accelerating and it is spending more to meet it. In the same 24 hours, commodity memory got a genuine, nameable supply shock: China’s CXMT opened an $8.55bn DRAM capacity IPO threatening the Samsung-Hynix-Micron oligopoly, the US is weighing fresh HBM export curbs, and the group had gone parabolic into all of it. Compute is a demand-and-pricing-power story that is strengthening. Memory is a supply-and-competition story that is turning down. They are not the same trade, and NVDA held (+0.33%, -1.1% premarket) while Micron drowned. Own the split, do not buy the basket.
My variant on the macro is narrower and, I think, more useful than the ‘melt-up continues’ consensus. The index is at the highs on the narrowest leadership of the cycle, mega-cap records papering over a leadership-group collapse. That configuration historically precedes chop, not clean breakouts. And the two things a 16 VIX is not paying you to worry about are both binary: a memory-supply wall (the SK Hynix ~$26bn Nasdaq listing drops into this rout tomorrow) and an Iran blockade that is now kinetic, with the US disabling a tanker near Kharg and expanding strikes into northern Iran while Brent sits calm in the mid-80s BBG Rtrs. I am not short the index. I am saying the paid-for risk is in the fractures, not the level.
Confidence the rotation read is right: moderate. What kills it: HY OAS through 300, or NVDA breaking its 50-day with the S&P also down. Either one and ‘rotation’ becomes ‘de-risk’ and the whole long book is wrong at once. That is not a hedge-your-language caveat. It is the trade.
What Matters Today
Retail sales (June), jobless claims (~217k survey vs 215k prior), and Philly Fed (~12.7 vs 10.3) all cross at 8:30 ET PDB. This is the soft-landing test: cooling inflation plus resilient-not-hot activity is the ideal mix, and a hot retail print into an oil bid is the thing that reintroduces the hawkish-Fed tail the two soft CPI/PPI prints just marked down. Netflix after the close (consensus EPS ~$0.80, revenue ~$12.6bn) is the consumer-and-advertising read. And the President’s 9pm address is a macro event whether the desk wants it to be or not, because ‘new findings’ about election security land straight on the midterm-legislation push and, tangentially, on how much political oxygen is left for anything fiscal.
And here is the piece nobody has circled: tomorrow is a $2.83tn SPX+SPY options expiration T1A, and it lands on the exact same day as the SK Hynix listing and the FITB print. SK Hynix lists on the Nasdaq, ~$26bn, into a memory tape in freefall. A clean, well-bid print says the crash was a positioning flush. A sloppy one confirms the downcycle and the equity-supply-wall risk in a single trade. Stack a record memory-supply listing, a regional-bank NIM tell, and a near-$3tn opex on one Friday and you have the single densest catalyst cluster of the month. That is the near-term arbiter, not anything a Fed governor says between now and the 29th.
Worth noting on the macro: the Beige Book (7/15) diffusion just printed its highest since mid-2022, activity ‘slight to moderate’ across 11 of 12 districts, with contacts flagging elevated fuel-cost and inflation uncertainty Daily Shot. That is the tell that lets Warsh and Cook talk hawkish into two soft inflation prints without contradicting themselves. Activity is firm, the disinflation is real, and the inflation-uncertainty is elevated all at once. Which is why the meeting is two-way and not the ‘hold locked’ the sell-side keeps writing.
Movers That Matter
TSM ADRs -4.8% premarket on a blowout PDB. So what: this is the whole thesis in one ticker. The market used the best AI-compute print of the cycle as exit liquidity, which is a positioning tell, not a fundamentals tell. Compute is fine. Memory is the problem. If you sold TSM here you sold the demand and kept the supply glut.
UNH +7.5% premarket, FY adjusted EPS guided to $19.50-20.00 from above $18.25, Mizuho calling it ‘higher than the bull case’ BBG. So what: it dragged the whole managed-care complex up (Humana +4.5%, Centene +4.4%, Molina +3.5%, CVS +2.9%, Elevance +3.0%) and is the reason XLV is not red on a risk-off-ish tape. Medical-cost-ratio recovery is the surprise, and a genuine one.
PGR -9.4%, combined ratio 90.0 vs 88.8 estimate, cat losses, plus a JPM downgrade to Neutral FMP. So what: this is the financials tell hiding under the bank-earnings euphoria. P&C underwriting is where the margin scare lives, and the question for the sleeve is whether it is PGR-idiosyncratic or the leading edge of a claims-inflation problem that spreads to TRV and ALL. KIE -2.1% and Chubb -2.5% say the group is already discounting the second.
Financials
Everyone is long the banks and the banks are at the highs, which is precisely why ‘long XLF’ is not an idea this morning. It is a 52-week-high beta bet dressed as conviction. XLF closed 56.56, at the very top of its range, RSI cluster overbought across the money-centers (BAC 73.4, BNY 76.4, STT 74.8) and the whole group extended double digits above the 50-day FMP. The edge is not the sector. The edge is the dispersion inside it, and the dispersion is unusually wide.
The winners: capital markets and asset gathering. Goldman printed EPS $20.98 against $14.40 estimate, net revenue +39%, the CEO out calling the AI boom ‘early innings’ BBG. Morgan Stanley solid but muted (+0.12%), which tells you the trading strength is now largely priced. BlackRock +6.6% on a Q2 beat and record $15.3tn AUM, the fee engine re-rating FMP. USB record revenue on the BTIG deal, NII $4.36bn beat, guide raised. IAI (broker-dealers) at a fresh 52-week high. This is where the earnings actually showed up.
The losers: P&C insurance (the PGR scare, KIE -2.1%, CB -2.5%) and the private-credit-adjacent book. ARCC holds at 18.92, roughly 19% below its high, non-accruals at stressed BDCs around 4.2% and grinding toward the ~6% line where a redemption story becomes a realized-loss story Financials Daily. BCRED took $3.8bn of March redemptions and met them via a $400mm sponsor injection, not organically. That is the tail nobody has in the bank multiples. Payments lagged too, MA below its 200-day, V and MA both red.
The level that matters: 2s10s at ~+41bp (2Y ~4.13, 10Y 4.545 FMP) is a bull-steepening, the clean NIM tailwind, but the KRE relative-value reactivation the framework flags sits at +60bp. We are 19bp of curve away, not at it. And the AOCI line to respect is 10Y through ~4.75%, where AFS marks start to bite the regional cohort (roughly $1bn of industry unrealized loss per 5bp above 4.50%). At 4.545 that is a watch item, not a live problem, but a bear-steepening led by the long end selling off would hit regionals through the AFS book even as the curve looks ‘friendlier.’ Steeper for the wrong reason is not a tailwind.
So what for positioning: own the dispersion, not the beta. Long the capital-markets and broker complex against the P&C underwriting scare. Stand down on regionals until the curve confirms at +60bp or KRE clears 76.84 on volume, because everything below that is buying momentum at resistance into a group that discounts a curve re-flattening faster than any other. FITB reports tomorrow BMO and is the next regional NIM confirm-or-deny. Full expressions behind the wall.
Sectors SPDR, 15 Jul close, FMP
XLC +1.73%: communication services led the tape, mega-cap breadth doing real work here.
XLY +0.95%: discretionary firm, consumer not cracking at the index level.
XLF +0.69%: financials at a fresh 52-week high, leadership but extended.
XLRE +0.17%: real estate a shade green, quietly ignoring the CRE maturity wall.
XLP +0.06%: staples flat, no defensive bid worth the name.
XLV 0.00%: healthcare pinned flat, UNH’s premarket rip lands in today’s tape not yesterday’s.
XLI -0.22%: industrials soft, GE beat then sold on profit-taking.
XLB -0.28%: materials heavy, China Q2 GDP at 4.3% below target is not helping.
XLE -0.79%: energy lower even with Brent firm, the equity refusing to chase the barrel.
XLU -1.03%: utilities the second-worst, the AI-power-demand darling taking a breather.
XLK -1.11%: tech the worst, and that is the memory rout bleeding through the sector line.
The read: leadership rotated to comms, discretionary and financials; tech and utilities lagged. Green Dow, red Nasdaq, benign credit. That is a rotation signature, not a de-risk. If XLK’s weakness spreads to XLC and XLY, the story changes.
Flows & Positioning
The single most important thing on the tape is a disagreement between three markets that cannot all be right. Equities are cautious (Nasdaq futures -0.7%, high-beta on pace for its worst month since the GFC). Credit is not: US HY extended its rally for a second session on the soft prints, HY OAS 266bp and positive on the month BBG. And Middle East sovereign spreads blew out ~20bp in a week to 402bp over Treasuries, the widest since October 2022 BBG. Credit-is-calm, equity-is-nervous, sovereign-is-screaming. One of them is mispricing the macro, and history says the one that moves last is HY. Watch it, do not assume it.
Now the part the 16 VIX is hiding, and it is the whole variant thesis in four charts. First, the vol is partly manufactured. SPX is leaning positive gamma, dealers hedging against the trend and pinning the index in the 7,500-7,600 range, and the notional that enforces that pin, $2.83tn of SPX+SPY, expires tomorrow T1A. Mechanical suppression, not organic calm, and a chunk of it rolls off in 24 hours.
Second, vol-control implied equity exposure sits near the top of its multi-year range T1A. That is systematic length, and systematic length is the fuel for a downside cascade, because vol-control funds are mechanical sellers into a vol spike. The VIX > 22 kill switch is not an abstraction; it is the level where that chart starts selling for you whether you want it to or not.
Third, positioning is stretched long and undefended. Combined equity portfolio allocations are near record highs even though surveyed sentiment is only modestly bullish Topdown. Read that carefully: people are not euphoric, they are just all-in. There is no dry powder and no de-gross, which is bullish until the day it is not, because a fully-allocated book has nothing left to buy the dip with.
Fourth, the skepticism is building where you would expect it, in single names: median short interest across the Russell 3000 has surged to roughly 5%, the highest in the series DB. Two-edged, that. Squeeze fuel on any rip, but also a market quietly betting against the breadth even as the index prints highs.
Layer the cross-asset piece on top. Japanese demand for Treasuries is ebbing per Wednesday’s TIC data, a structural term-premium headwind for the long end that supports the sticky-30Y thesis BBG. Korea’s FSC is suspending single-stock leveraged-ETF listings and tripling the deposit requirement after the chip carnage, a momentum unwind off a KOSPI down ~25% from its peak but still +61% YTD, a flush not a collapse. And for all the equity caution, SPY pulled $2.72bn and QQQ $2.37bn of daily inflows in the prior session PDB. Nobody is de-grossing yet.
So what: the melt-up sits on mechanical vol suppression that partially rolls off tomorrow, over elevated systematic length that turns into forced selling on a vol reset, on top of record-high allocations with no cushion, with short interest rising underneath. That is not a sturdy tape. It is a pinned one, and pins release. The pain trade if it extends is being underweight mega-cap and foundry. The pain trade if it does not is being long anything at a 52-week high into that positioning. Both are live. The hedge is cheap for a reason, and the reason is on these four charts.
Paid Bridge
Behind the wall today: the full book marked to the live tape, six expressions with entries, stops, targets and the print that falsifies each. The two financials trades are the dispersion, long the broker-dealer complex against the P&C underwriting scare, plus the regional-bank stand-down with the exact curve trigger to arm it. The cross-asset book, the duration short on trim watch with the level that trips the trim today, and the Iran energy expression with the Brent print that arms it. The credit hedge the whole long book depends on, sized as ballast, not as a view. A decision tree branched on the 29-30 FOMC and the one credit print that changes the regime. Conviction ranking, and the watchlist into tomorrow’s SK Hynix listing, FITB, and the $2.83tn opex.
Paid subscribers · morningmusing.com
Keep reading with a 7-day free trial
Subscribe to Morning Musings- Thoughts from Dauber Island to keep reading this post and get 7 days of free access to the full post archives.