Hello, and welcome to Regions & Reasons.
Last week went the way I hoped, though not without a scare. The S&P 500 closed at a record on Tuesday, at 7,818.93, and the Nasdaq set records on Monday and Tuesday. On Thursday the AI trade wobbled after a report put OpenAI’s annualized revenue at about $50 billion, some $20 billion below figures that had been circulating, and the Nasdaq fell 1.25%. Friday’s bounce left the S&P 500 up 1.2% on the week.
On rates, the 10-year touched 5.365% on Wednesday, its highest since 2002, and that was the top. Both Treasury auctions cleared well, and the yield ended the week slightly lower, at 5.24%. The Fed’s minutes showed that most officials still expect another hike this year, and Governor Waller said the hikes don’t need to come at consecutive meetings. Markets read that as a hold in October and a hike in December.
On the geopolitical front, Brent jumped about 4% on Thursday after reports that the White House had asked the Pentagon for new strike options against Iran. Trump then said the US won’t attack Iran before the 3 November midterms, while keeping the blockade in place. The weekend undid some of that calm: an attack on Riyadh’s main airport, later claimed by the Houthis, killed 12 people on Saturday; Trump said the US “may” join Saudi strikes on the Houthis and that he would “take a look” at his pledge; and on Sunday Iran called Washington’s reply on reopening Hormuz “insufficient”. Hurricane Isaias also shut about 63% of US Gulf of Mexico oil output before making landfall in the Florida Panhandle on Friday night.
So what changed? The market has learned to live with a 10-year above 5.2%. What it hasn’t faced yet is a bad inflation print with yields this high. That test comes on Wednesday, and the weekend’s news from the Gulf raises the stakes.
The week in five lines
Macro: Wednesday’s US inflation report is the main event; headline CPI is seen rising to 3.6%, mostly on gasoline.
Indices: the S&P 500, my benchmark, sits 0.1% below Tuesday’s record close, and the banks open earnings season on Tuesday.
Rates: the 10-year touched 5.365%, then strong auctions pulled it back to 5.24%. That line separates “earnings in charge” from “bonds in charge”.
Dollar: firm but stuck between 101.75 and 102.54; core CPI decides the next move.
Oil and gold: oil is trading Iran headlines day by day; gold bounced back to the bottom of its old $4,203–4,235 resistance.
What could change everything: core CPI at 0.4% or more, or an escalation in the Gulf; after the weekend, Trump’s no-strike pledge looks less firm.
Last week, honestly
Every edition starts with an honest look back, misses included. Last week’s map got stocks, the dollar and the big call right, oil half right and gold wrong: four hits, one half, one miss.
The lesson: this time the reasons held up as well as the levels. The bond market didn’t need to rally for stocks to rise; it just needed to stop selling off, and it did. My misses were about direction, not levels: oil held its region but I leaned the wrong way, and gold’s break below its region didn’t follow through. My Stock of the Week, Delta, confirmed the fuel story: adjusted EPS of $1.72 missed estimates and fell short of its own $2.00–2.50 guide, and the full-year outlook was cut to $5.10–5.60 from $6.50–7.50. The stock still ended Friday flat, at $82.17.
The big picture: Inflation, banks and the long end
Three things matter this week, and they’re connected. Wednesday’s CPI tells us whether the Fed hikes in December or sooner. The banks tell us whether a 10-year above 5% is starting to hurt dealmaking and credit. And TSMC tells us whether last Thursday’s AI scare was noise.
Start with the bond market, because that’s where the risk was. The 10-year touched 5.365% on Wednesday and the 30-year traded above 5.7%. If there was ever a week for buyers to go on strike, this was it. They didn’t. The 10-year auction cleared 1.7 basis points below where it was trading just before, with indirect bidders, a group that includes foreign official buyers, taking 80.3% against a 74.1% average, and dealers left with just 2.5%. The 30-year sold at 5.618% with 2.54 bids for every bond offered, above the 2.41 average. The Fed’s minutes added that most of the rise in long yields came from real rates, not inflation fears.
Chart: demand at last week’s 10-year and 30-year Treasury auctions against recent averages. Source: US Treasury results via investingLive.
My view is that a hot headline is already in the price, and the market will trade core. The consensus on Investing.com has headline CPI rising 0.6% in September, taking the annual rate to 3.6% from 3.4%, mostly on gasoline, with core at 0.2% and 2.5%. If core comes in at 0.2% or lower, October stays off the table, December stays the base case, and I think stocks can push to new records with the banks and TSMC as the fuel. A 0.3% would change little. At 0.4% or more, an October hike is back in play and I’d expect the 10-year to retest 5.35% quickly.
There’s good reason to take that risk seriously. September’s ISM services prices index jumped to 74.0, its highest in four years, and in Michigan’s survey on Friday, consumers’ one-year inflation expectations rose to 4.7%. Businesses are paying more, and households expect to keep paying more. That’s exactly what a Fed that hiked in September doesn’t want to see.
On geopolitics, Friday’s relief now looks premature. Saturday’s attack on Riyadh’s airport killed 12 people, Trump said he would “take a look” at his pledge not to strike Iran before the midterms, and Iran called the US reply on Hormuz “insufficient”. Oil could open the week with more risk premium, and Brent above $100 keeps feeding the inflation numbers the Fed is watching.
Then there’s earnings season. The banks open it on Tuesday, and the bar looks low: the KBW bank index is down 13% from its August peak, according to Reuters, on fears that higher yields will slow deals and lending. JPMorgan is expected to grow EPS by about 17% from a year ago. If the big banks say the deal pipeline and credit are holding up, the market’s biggest worry about high yields loses some of its bite.
What most people expect, and where I see it differently. The common worry is that stocks at records with a 10-year at 5.2% must be expensive. The numbers say otherwise. FactSet puts the S&P 500’s forward P/E at 19.1, below its five-year average of 19.8 and down from 20.4 at the end of June, because earnings estimates have risen faster than prices. Q3 earnings are now seen up 29.6%, and 84% of the early reporters have beaten. Strategists’ year-end targets still range widely, from Bank of America’s 7,100, set in June, to Ed Yardeni’s 8,400, set in August, while FactSet’s bottom-up target is 9,291. On the Fed, Barclays expects a hold in October and a hike in December, and Goldman’s Jan Hatzius reportedly now sees two more hikes as more likely to be needed than one. Where I differ from the bears: they read the record as a sign of excess. I see a market that has got cheaper on forward earnings, not dearer, because earnings did the heavy lifting.
Chart: S&P 500 forward P/E against its averages, and expected Q3 earnings growth. Source: FactSet Earnings Insight, 9 October 2026.
What would change my mind: core CPI at 0.4% or more, the 10-year closing above last week’s 5.365% high, or Brent breaking above Thursday’s $105.92 high on a new escalation. Any of those would put the bond market, not earnings, back in charge.
Rates & risk
What the bond market is saying: the buyers came back at these levels. The 10-year spent the week between 5.22% and 5.37% and ended 3 basis points lower; the 2-year fell by the same amount, so the curve didn’t move. With October hike odds around 20% and December around 70%, the front end is waiting for CPI. For stocks it’s a green light below 5.30% on the 10-year, a yellow one between 5.30% and 5.37%, and red above that. For the dollar it’s neutral. For gold, real yields near 2.9% are still a headwind, even after easing slightly.
Volatility: VIX at 14.84, its lowest close of the week, with the S&P 500 a whisker from its record. That’s calm going into a CPI week, and it leaves little cushion if core surprises.
Positioning: the latest CFTC figures (released 9 October) show speculators net short about 154,100 S&P 500 contracts, up from about 133,000 in the 25 September release. They kept adding to shorts while the index made records. If CPI and the banks don’t give them a reason, that’s fuel for further gains.
*TIPS real yield as of 8 October, against 2 October (FRED). Yields: Investing.com closes; weekly changes are Macro Atlas calculations. Fed odds: CME FedWatch as reported on 10 October.
The week ahead
Mon Columbus Day: stocks trade, the US bond market is closed. The IMF and World Bank meetings open in Bangkok, with the World Economic Outlook on Tuesday.
Tue JPMorgan, Goldman Sachs, Wells Fargo and Citigroup report. Why I care: the first read on whether a 10-year above 5% is hurting deals and credit. Fed Governor Waller speaks at 09:45 CEST.
Wed CPI at 14:30 CEST: headline seen at 0.6% m/m and 3.6% y/y (0.4% and 3.4% prior), core at 0.2% and 2.5% (0.3% and 2.4%). Why I care: core decides whether October comes back. Bank of America, Morgan Stanley and ASML also report, and the Beige Book is out at 20:00.
Thu retail sales (+0.3% expected, +1.2% prior), PPI (+0.5%, +0.4% prior) and TSMC results. Why I care: a consumer with sentiment at 46.3, and the first big read on AI demand after the OpenAI scare.
Fri Fed Chair Warsh speaks at 05:30 CEST, then industrial production (+0.4% expected, 0.0% prior). Why I care: the Chair’s reaction to CPI, if he gives one.
Chart: economic calendar, 12–16 October 2026, times in CEST. Source: Investing.com, as of 11 October 2026.
Earnings
Chart: earnings this week. Consensus: Investing.com earnings calendar, as of 11 October 2026; company releases for prior-year figures.
Politics
Chart: political events with market relevance, 12–18 October 2026, and key dates beyond. Sources: CBS News, Reuters, Iran International, SCOTUSblog, Euronews, IMF, Newsquawk.
At a glance
Lean is simply which scenario I think is more likely this week. It isn’t advice to buy or sell.
What I’m watching
S&P 500
Closed Friday at 7,811.54, up 1.2% on the week and just 0.1% below Tuesday’s record close of 7,818.93. Last week’s region was never retested: the lowest point of the week was 7,727.59 on Monday.
The region that matters: 7,727–7,765, because that’s where buyers stepped in on Monday and Thursday (lows of 7,727.59 and 7,731.26), and where Thursday’s selloff closed (7,765.36).
If it holds: a close above Tuesday’s intraday record of 7,844.52 would open the way to 7,900, with CPI and the banks as the catalysts.
If it breaks: back towards 7,665–7,700, the top of last week’s region and the 2 October low (7,700.51).
Nasdaq
Closed at 27,366.17, up 0.6% on the week, after record closes on Monday and Tuesday (27,599.89) and a 1.25% drop on Thursday.
The region that matters: 27,070–27,190, because Thursday’s selloff found buyers at 27,066.76 and closed at 27,193.34, right on the 2 October close of 27,190.86.
If it holds: a retest of the 27,600–27,723 record zone, with TSMC on Thursday as the test.
If it breaks: last week’s region, 26,870–27,000, comes back into play, and the AI scare starts to look like a correction.
Dollar
The dollar index ended at 102.21, up 0.3%, and never came close to my 101.0–101.2 region: the low was 101.75 on Tuesday. EUR/USD finished at 1.1202 after dipping to 1.1161 on Monday.
The region that matters: 101.75–101.90, the week’s low and the area the index bounced from on Tuesday and Wednesday.
If the dollar firms: a hot core CPI would bring October hike bets back; above 102.54, Monday’s high, I’d look for 103, with EUR/USD below 1.1160.
If the dollar slips: a soft core and a calm Gulf would do it; below 101.75, back towards 101.0–101.2.
Oil
WTI closed at $91.85, up 0.8% on the week. It dipped to $86.86 on Tuesday, below my region, then spiked to $93.20 on Thursday on reports of new strike plans, before Trump’s pledge took some of the heat out. Brent ended at $104.72.
The region that matters: $88–89, because Wednesday’s close ($88.28) and Thursday’s low ($88.77) mark the floor of the new range.
If it holds: the $93.2–93.7 area, Thursday’s high and the 1 October high, is the ceiling; breaking it would take a new escalation, and the weekend made that more likely.
If it breaks: a de-escalation, the G7 barrels and returning Gulf of Mexico output after Isaias take over, and $86.86, Tuesday’s low, is next.
Gold
Gold rose 1.2% to $4,194 (spot), after breaking my region on Wednesday and touching $4,066. The break didn’t follow through, and Friday’s 1.4% rally took it back to the bottom of the old $4,203–4,235 resistance.
The region that matters: $4,203–4,235, because that’s the region gold broke two weeks ago, and it capped Friday’s rally at $4,207.69.
If it breaks higher: $4,280, where the 28 September selloff began.
If it holds as resistance: back to $4,135–4,140, and $4,066 if real yields rise after CPI.
That’s my map for the week. If it helped, forward it to one person who trades these markets. And if you’d like me to cover something specific, just reply to this email: I read every one.
Have a good week,
Carlos
For information and education only, not investment advice. Trading futures, FX and commodities carries a real risk of loss. Full Disclaimer at themacroatlas.com.
the hot print matters less than the curve reaction, if short end holds the equity sell is just a margin move that gets bought within two days, i write about similar stuff on my substack, let's connect!