Global Flow Strategy · By Exponential Technology · November 21, 2026
We think the Fed raises rates 25 bps on December 9 to fight a price spike that is already over. Gasoline was the problem, and gasoline is now what is bringing inflation back down. The hike lands after the fix, and households pay on the way up and the way down for a Fed that keeps looking in the rearview mirror.
Our September work said the run in prices was energy and would fade on its own. The metals market reached the same conclusion that month, doubling its bet against gold and flipping copper negative. Both of us were early. The fade began in the first week of November, and the trigger was the election, not the Fed.
What this note covers, section by section:
1. Gasoline Was the Whole Story — why September looked like broad inflation and was not.
2. The October Print Took the Case Apart — the report that landed the same morning as the speech.
3. Prices Fell Before He Spoke — the energy selloff that started with the election.
4. Who Feels It First — how the hike reaches credit cards and home equity lines before it reaches gasoline.
5. The Metals Market Already Voted — what gold, copper and gas positioning said before the Fed met.
6. Bottom Line — what we think happens, what would prove us wrong, and the dates that settle it.

1. Gasoline Was the Whole Story
September looked like broad inflation. It was a gas bill.
We called it. Exponential's Macro Economic Forecasts model had gasoline rising about 10% before the October 14 release, and the government reported 9.1%. Overall prices rose 0.6% on the month.
The number that matters is the one underneath. Core inflation strips out food and energy, because those two move around for reasons that have nothing to do with the economy — a refinery outage, a harvest, a war. Core rose just 0.3%. The gap between 0.6% and 0.3% is the whole story, and we saw it coming.
Rent was never the problem either. Shelter, which is mostly rent and what homeowners would pay to rent their own house, added 0.39% on the month. Rent responds to interest rates and to time — leases turn over slowly, so rate increases work their way through housing over a year or more. Gasoline responds to neither. It responds to crude oil, refining capacity and geopolitics.
That distinction matters because the Fed only has one tool that acts quickly. If the inflation is coming from rent, higher rates help, eventually. If it is coming from gasoline, higher rates do nothing at all — they just make everything else more expensive while the gas price finds its own level.
We will know we were wrong if core inflation climbs above 0.4% while energy starts rising again. Neither is in the data through Friday.
2. The October Print Took the Case Apart
The October inflation report came out the same morning as the speech. The case did not survive the day.
Four numbers told the story, and all four pointed the same way. Energy rose 0.7% in October after climbing 5.2% in September, which is the spike ending. Gasoline went from 9.1% in September to 1.0% in October. Overall prices cooled to 0.4%, while core held at 0.3% for a third straight month. The gap between the two narrowed from 0.3 points to 0.1, which means the thing that had been driving the headline was no longer driving it.
Where we sit on this. Exponential's Macro Economic Forecasts product publishes its inflation numbers eight days ahead of the government and has been running about 50% more accurate than consensus, and they go out eight days before the government publishes its number. Against the three best-known economists in the survey, we are roughly 25% better. That is not a boast so much as a reason to trust the read on the next few paragraphs.
We publish our own scorecard either way. Our October call missed on the headline, where we expected 0.2% and the number came in at 0.4%. Consensus missed too. The surprise came from food and energy rather than from demand — food rose 0.5% after a flat September — and the breakdown between the pieces is the point rather than the round number.
The calendar is the awkward part. The committee votes December 8 and 9, and the November inflation report does not arrive until December 10. That means they will walk into the room without a fresh number in hand, deciding on a July-to-September stretch that has already turned. Central banks are always working from data that describes the recent past. This time the lag is doing real damage, because the thing they are reacting to peaked two months ago.
We would be wrong if November core came in above 0.4% with energy moving back up. That is not what September showed, and it is not what the October print showed either.
3. Prices Fell Before He Spoke
Oil dropped in the week between the election and the speech by the Fed's Waller. Then it kept dropping.
The election did it. Republicans lost Congress on November 3, and in the next session crude fell more than 8% to a two-month low. By Friday, West Texas crude was down about 13% on the week to $90.83 and Brent had fallen to $98.32, one of the sharpest weekly drops of the year.
What traders were unwinding was the war premium — the extra money per barrel that buyers had been paying since the spring in case the conflict with Iran cut off supply. Once an election result suggested a change in policy, that fear got priced out. At the same time the International Energy Agency published a forecast for a 1.7 million barrel a day surplus in the fourth quarter, which is the opposite of a shortage. Fear of scarcity plus a forecast of abundance is a sell signal, and the market took it.
The speech came a week into that move. Waller spoke on November 10 and prices kept falling afterward. Brent settled at $92.44 the following Monday and West Texas at $85.72, with physical crude — actual barrels changing hands rather than paper contracts — falling faster than futures. When the physical market falls faster than the paper market, it usually means real demand is weak, not just positioning.
Then the pump followed. Regular gasoline averaged $3.805 a gallon in the week ended November 16, down 19 cents, with every region of the country lower. That was the second straight week below $4 after sixteen weeks above it. Wholesale gasoline fell 47 cents in the two weeks through November 9, diesel dropped below $6, and natural gas fell about 17% from mid-October. Retail gasoline lags wholesale by a few weeks, so the pump has further to fall even if crude does nothing from here.
What matters is the direction, not the level. Gasoline is still up more than 24% from a year ago, which is what is left of the war premium, but it is coming down and coming down fast. The Fed reacts to where prices are going rather than where they have been. One hopes the Governor still gets the morning paper, because the commodity page would have saved him the trouble.
4. Who Feels It First
A rate hike hits a household budget long before it touches the price of a barrel of oil.
Here is how that works, because the mechanics are where most of this gets lost.
The Fed sets a target for the overnight rate banks charge each other, the funds rate, which has sat between 3.75% and 4.00% since the October 28 meeting. Banks then set their prime rate, which is what they charge their best customers, and it currently sits at 7.00%. Most consumer borrowing is priced as prime plus a margin, or as a benchmark plus a margin, which means it moves automatically when the Fed moves.
Home equity lines of credit are tied directly to prime. They reprice the day after the Fed acts, with no grace period and no negotiation. Credit cards are not formally tied to prime, but they follow within a billing cycle or two. The average card rate is about 21% right now on roughly $1.35 trillion of balances, and a rate increase flows straight through to the interest charge on whatever you already owe.
The math is simple enough. Our expectation is a quarter-point, which costs about $25 a year more for every $10,000 drawn on a home equity line — $250 on a $50,000 line. On credit cards, if balances hold steady, that comes to roughly $3.4 billion a year in extra interest across the country. The illustration worth keeping in mind is the larger move. A half-point — the number the Governor floated — would double every one of those figures: $50 a year per $10,000, $500 on a $50,000 line, and about $6.7 billion a year on cards. That is what the more aggressive version would have cost, on a cushion that could not absorb the smaller one either.
There was no cushion to absorb it. The New York Fed reported on November 10 that household debt reached $19.1 trillion, up $334 billion in the quarter, with card balances up another $22 billion. Five percent of that debt is behind on payments and that share is rising. Wages are growing 3.1% while inflation runs at 3.7%, which means the average household is losing ground before any of this happens.
These are people putting groceries and gas on a card at 21% between paychecks. The funds rate does not drill for oil and it does not generate electricity. What it does is make the borrowing more expensive. A quarter-point lands on the borrower the day after the vote and reaches the gas price last, if it reaches it at all. A half-point would simply land twice as hard.
5. The Metals Market Already Voted
Traders made the argument in contracts before the Fed made it in votes.
This is where Exponential's Futures Flow dataset does the work. It forecasts daily positioning by investor type, separating managed money from producers, merchants and swap dealers, so you can see not just that a market moved but which kind of account was on which side. One of those categories is managed money — hedge funds and commodity trading advisors, the people taking directional bets. Their net position is how many contracts they hold long minus how many they hold short. When that number is negative, more of them are betting on a lower price than a higher one.
Gold shorts doubled, according to Exponential Futures Flow. Money managers were net short at every close in the window, going from 6,287 contracts on November 5 to 12,286 by Thursday and 12,079 on Friday. Gold broke to a one-year low along the way and had its worst week in eight months. That is a market saying higher real rates are coming, and saying it before the vote.
Copper flipped and kept going on the same data. Funds held 3,088 contracts net long on November 5 and have been net short at every close since November 12, reaching 8,338 contracts short by Friday. That came the session after copper hit multi-month lows. Copper is the industrial metal, so a short there is a bet on slower growth — which is what a rate increase into a slowing economy tends to produce.
The one market that went the other way was natural gas, where Exponential Futures Flow showed funds holding 86,835 contracts net long by Friday, a high for the window. That is the AI power demand story we have been tracking since February — data centers need electricity, and a growing share of it comes from gas — and none of this has touched it.
The risk sits in the positions rather than in the argument. Those gold and copper shorts are the largest of the window and they were put on at the lows. If prices turn, those same traders have to buy back what they sold, which is called covering, and covering a large short position can move a market faster than the news that caused it.
6. Bottom Line
We think the December 9 hike lands after the problem it is aimed at has already gone away.
Near term, the metals shorts carry covering risk into year-end. Longer term, the cost to households does not reverse when energy does. A quarter-point move reprices a card balance or an equity line, and it stays repriced until that balance is paid down or refinanced. At a half-point the same arithmetic doubles. Energy coming back down does not undo any of it.
Here is what would change our mind, with dates. The November inflation report on December 10 could show core above 0.4% with energy moving back up, which would reverse this view. A new flare-up in the Gulf could bring the war premium back. A credit event could turn selling into forced selling. Quarter-end rebalancing could add to it.
Near term we are neutral on metals into year-end. Longer term we think the cost of household credit is the story to watch, as a theme rather than a trade.
We will revisit this after the December 10 report.
The hike arrives. The problem has already left.
This material is provided to institutional investors for informational purposes only. It is not investment advice, an offer, or a solicitation. Data are as of the dates shown and are subject to revision. Past performance is not indicative of future results. © Exponential Technology Inc.






Comments