“Vote early. Vote often’ ~William Hale Thompson. mayor of Chicago from 1915-1923 and 1931-1935
I will admit it. I hate when we change the clocks. Despise it. I still can’t wrap my head around night beginning at 5 ( soon to be 4) pm, and knowing winter is imminent drags me down. I even have a S.A.D. light; no cue of it works, but Brookstone and Instagram swear it does, so I bought one. This year, there is an additional piece: cold weather means more indoor time, which translates into fewer distractions (Fore!!) from reading the news and worrying about the fate of the world and my children's future. I mean, there are only so many Netflix series to watch, and I watching the Jets is even more painful.
Yes, first-world problems, I know. Ah well, I’ll soldier through.
**THE AI BUILD IS A FUNDING STORY, AND CREDIT IS PRICING IT.**I own no builders; the VIX upside is my hedge. I’M WRONG IF high yield prints under 320bp with the VIX near 18.**A DEMOCRATIC CONGRESS BUYS OVERSIGHT, NOT A SMALLER DEFICIT.**The long end keeps its term premium; I stay short it. I’M WRONG IF Wednesday’s refunding cuts coupon sizes and the 30-year rallies 15bp.**IN FINANCIALS I OWN THE RESERVE CUSHION, NOT THE WALLET.**Long Citigroup and JPMorgan, short PayPal; mostly one bet on payrolls. I’M WRONG IF payrolls beat and December odds top 85%.**THE OIL PREMIUM IS ON NOTICE TODAY.**I’m still long Brent upside into a post-vote Iran decision. I’M WRONG IF Brent settles at or under $110 today, resetting the clock.
Kalshi closed Sunday with Democrats at 93.2% for the House and 71% for the Senate; Polymarket runs a point or two richer on both.
The combined contracts put a full sweep at 70%, a Democratic House with a Republican Senate at 27% and a Republican sweep at 6% (
Kalshi).I treat the House as done and put the Senate at 75%, four points over the market.
There are signs of both a red riptide and a blue tsunami, and they’re different things.
The riptide is underneath, pulling Republican voters out of the water. Rural voters have swung about 20 points toward Democrats since the 2024 House vote and men about 13, and registered Democratic Hispanics are out-voting registered Republican Hispanics early in all 17 states checked, after trailing in 10 of 17 at this point in 2024.
The tsunami is the part you can see from the beach. Early ballots are running 70% over 2022 in Wisconsin, 43% in Maine and 31% in North Carolina.
A riptide alone wins the House. It takes the wave to carry Texas, Ohio, or Kansas. Why am I a little more confident than the poll?
Engagement looks like a wave year. Gallup has 56% of adults giving ‘quite a lot’ of thought to the election, the highest it has measured at this point in a midterm, against 54% in 2018, 55% in 2010 and 52% in 1994, and all three of those produced big seat changes (
Gallup). G. Elliott Morris finds the early-vote ‘engagement gap’ in Texas, Georgia and North Carolina running ahead of the Democrats’ 2018 edge (Strength In Numbers).The last round of NYT/Siena polls, taken 23 October to 1 November, has the Democrats ahead in all seven Senate races it polled, from 11 points in Michigan to 1 in Iowa.
The swing is broad. The generic ballot averages D+8.9 against a House popular vote of R+2.6 in 2024, and independents have moved about 16 points. About 51.4 million early ballots were in across 47 states by Sunday afternoon.
The seat math is narrower than the national number:
Republicans hold the Senate 53 to 47, so Democrats need a net four. North Carolina flips in the base case and Maine gets them to 49.
That leaves two of five toss-ups in states Trump carried by 11 to 16 points: Texas, Alaska, Ohio, Iowa and Kansas.
Polling error is correlated, so those five tend to go one way together, and a uniform 4- or 5-point miss toward the GOP (within the 2016 and 2020 range) hands them all back. It’s the 25% I’m not giving the Senate.
Two late wrinkles:
Rep. Haley Stevens, who lost the Michigan primary, endorsed Republican Mike Rogers over Abdul El-Sayed on Sunday (
NBC News). With El-Sayed up 11 in the Siena poll, I doubt it moves the race. As an aside, her behavior has probably ended her political career.In Alaska, a judge threw out the Justice Department’s bid to delay certifying the Senate result, and the department has appealed (
CNN). The suit asks the court to block a Peltola certification while a second Daniel J. Sullivan stays on the ballot (Alaska Public Media). It only matters if Alaska is the 51st seat, which is exactly the night it was filed for.
Tuesday night, the early tell comes from Florida, Georgia and North Carolina, which count fast. A Democrat running 3 or 4 points behind the polls in Georgia or North Carolina early is the first sign the red-state toss-ups break Republican. In the House, TX-15 and TX-35 are the first hard test of whether the polled Hispanic swing is real.
WHAT A DEMOCRATIC HOUSE DOES (I’m treating it as certain).
High probability:
**Subpoenas, early and specific.**Rep. Robert Garcia, the ranking Democrat in line to chair Oversight, has named Stephen Miller, Kristi Noem, Corey Lewandowski, Kash Patel and acting Attorney General Todd Blanche as early witnesses, with Epstein’s estate and financial ties, Jared Kushner’s investments and companies ‘working with the White House’ on pricing and data as the topics (Bloomberg Government). The last item is the one I’d watch from a portfolio: an investigation of a company is a headline risk to that stock, whatever comes of it.**A Hegseth impeachment vote.**Republican Thomas Massie has already promised to force one after the midterms, and Jason Crow and Chris Deluzio want him out over the conduct and cost of the Iran war (CNN). The House can impeach on a simple majority; removal needs 67 senators, which no outcome on Tuesday delivers.**The war-powers vote passes again, and sticks.**The House passed Gregory Meeks’s Iran resolution 215 to 208 in June with four Republicans. A Democratic House passes it easily. A veto still stands.**No second tax bill.**Nothing major gets through either way, so the fiscal path into 2028 is the one already written. The election changes who holds the gavel and leaves Wednesday’s borrowing arithmetic alone, and my long-end short is built on that arithmetic.
Likely:
**The purse becomes the weapon.**The stopgap runs to 11 December and the lame-duck Congress is still Republican. From January the House writes the money. Iran operations, ICE and the DHS budget are where riders and holds go, and a shutdown fight in early 2027 is more likely than not.**Gulf money gets a hearing.**The new book by Matthew Cole alleges $2 billion of Gulf state money routed toward Jared Kushner, which he calls ‘completely insane’ (SpyWeek). It’s a ready-made Oversight file, and it lands in the middle of the Qatar-run Iran talks.
Possible:
**Impeachment of the President.**Pramila Jayapal has been noticeably cooler on it than on Hegseth. I’d put it at the low end of possible in 2027. And not something I would recommend, if for some reason I was asked.
WHAT CHANGES IF THEY TAKE THE SENATE TOO (my 75%).
High probability, given a sweep:
**Confirmations go through Democrats.**Every cabinet replacement (a Hegseth successor, a confirmed attorney general), every judge, and every Fed board seat needs a Democratic majority from January. On the Fed, I think that’s mildly hawkish at the margin: the White House loses the easy path to a more pliant board.**Joint resolutions reach the President’s desk.**War powers on Iran, disapproval of Gulf arms sales and termination of tariff emergencies can all pass both chambers. None survives a veto, so their value is the record and bargaining power in spending talks.
Likely:
**Russia sanctions and Ukraine aid move.**It’s the one area with enough Republican support to make a veto expensive. The cyberattacks on both countries’ air-raid apps after the Kherson truce lapsed this weekend won’t hurt that case (Euronews).**Senate committees get subpoena power too.**Armed Services on the Iran war’s cost and Banking on crypto and family ventures are the two I’d expect first.
Possible, and the one I trade:
**Iran moves before January.**A Congress that can’t stop a strike can still make one harder after 3 January. Netanyahu pressed Trump on Saturday to strike right after the vote, and Trump told him he wants the talks to run (Axios). Gulf governments are asking for an off-ramp (’We are clearly seeking a political off-ramp’, the UAE’s Anwar Gargash told the Manama Dialogue) and expect a new round anyway (Bloomberg). I think a Democratic sweep shortens the window and raises the odds of action in the lame duck. It’s my case for keeping Brent upside through the vote, and today’s settle decides whether I can.
The market has traded a divided Congress as benign for most of my career, and the seasonals agree.
I wouldn’t lean on a table of 19 observations in a market with a war and a hiking Fed in it. Steve Bannon is already promising a Republican ‘civil war’ after Tuesday (FT). That fight belongs to the party and doesn’t change the bond math.
The main story is still who pays for the AI build, and since Thursday the question has reached credit. The second is what the bond market did with the scare, which matters more to my duration short than anything else this week. Oil is the third, and today’s settle is its own event.
High yield printed 326bp for 29 October, 12bp wider in a session and 1bp through the line I use for stress, on a print that predates Friday’s index drop. One basis point on one print is thin, and I hold the read loosely until the 30 October number lands. The direction fits everything around it:
Alphabet’s negative free cash flow, Amazon’s trailing minus $42 billion and Intel’s memory-cost guide all landed in one week.
A five-year swap on Oracle reached a record near 119bp.
The borrowers are paying for the build, and the balance sheets that never have to ask got paid on Friday: Berkshire, with a Treasury bill portfolio and no construction program, closed near a five-month high.
Now the bonds. The 10-year round-tripped the September hike, and the breakdown of that move is the most useful chart I saw this weekend.
Almost the whole rally since the 5 October peak came out of the real yield; breakevens gave back a single basis point. The market cut its growth expectations and left its inflation expectations alone. I’m still short the long end for that reason: a growth scare can pull the long end lower for a while, but nothing in that split says the inflation premium is leaving.
The inflation side is still building underneath.
MUFG’s work has information processing equipment adding close to 0.18 points to core PCE inflation, the most in the series, as the AI build pulls in hardware.
The New York Fed’s multivariate core trend turned back up in September.
Apollo’s Torsten Slok runs a rule-based model that needs the funds rate at 6.75% by the end of 2027, about 275bp of hikes, to get CPI back to 2%.
I don’t believe the Fed does anything close to that. What matters for my book is the shape of the risk: a growth scare in real yields sitting on top of an inflation trend that hasn’t turned. If the growth scare fades, the long end sells off with nothing to hold it.
THE CASE AGAINST.
The other explanation is positioning.
Morgan Stanley’s prime book shows hedge funds have cut more than 8% of their combined long and short positions since June, the biggest de-grossing since at least 2017.
A flush that size, in names that crowded together, is also consistent with Friday: a 7.20% semiconductor day with HYG down only 0.19%.
On that reading the credit print is noise at the edge of a forced sale, and it fades as the selling ends.
The two aren’t mutually exclusive: a positioning flush can expose a funding problem, and a funding problem forces the flush. What separates them is the next week of credit and the tech shorts.
If the 30 October high yield print comes in over 325bp and stays there through Wednesday’s refunding while the VIX sits around 20, credit is pricing funding and I’m right.
If it comes back under 320bp while QQQ short interest starts to fall, I was watching a de-gross, and the credit view loses its reason before Friday’s payrolls can give it one.
I’d put it at 60/40 my way, and the 1bp margin is what stops it going higher.
**10:00, ISM manufacturing.**Consensus is 54.7 against 54.5. A hot print with firm prices paid puts the December hike back in play, now roughly seven in ten after the mid-80s on Wednesday, and that takes the front end away from my bank longs’ worst case. A soft one adds to the growth scare in the real yield, and for my duration short that’s the wrong direction.**After the close, Apple and Palantir.**Apple is expected to earn about $1.99 a share on $113.4 billion of revenue. It’s the least capital-intensive of the mega-caps, so a clean print separates a spending problem from a demand problem and gives the whole group a bid. It’s also the named test for the funding read, and the event that could prove me early.**Friday, October payrolls.**Consensus is +55,000 against +29,000, with unemployment at 4.2%. Lighthouse Macro’s work on the revisions is why I’d trust the first print less than usual, unless it comes with upward revisions to the prior two months (Lighthouse Macro).
2025’s first prints added 1.21 million jobs, and the same twelve months now show 116,000. If the labor market is that much weaker than its headlines, Citigroup’s reserve against a 5.3% unemployment rate is the right cushion to own, and the reason I own it.
Wednesday is the quarterly refunding, against a $628 billion borrowing estimate for the quarter and up to $6 billion of buybacks in 10-year to 20-year paper. It’s the test for the duration short, and the 20-year still yields more than the 30-year going into it.
Two charts on the cost side. Building materials are running far ahead of consumer prices, copper above all, and the Fed’s own conditions index tightened through September.
Tighter conditions and input costs that keep climbing are a stagflation mix. It doesn’t change any line today; it’s why I’d rather keep the long end short than the front end.
THE REST OF THE TAPE.
The PBoC opened a fresh one-year swap facility worth about $15 billion so 31 brokers and asset managers can swap illiquid assets for liquid ones to buy stocks, after China’s worst equity selloff in almost a decade (
SCMP).Nvidia, AMD and Broadcom put $5 billion into Meta’s $25 billion share sale last week, buying stock in a customer that raised the money partly to buy their chips (Paul Kedrosky).
OpenAI raised less than half of its $30 billion target, at about a $1…
