I've never trusted a forecast that came with a guarantee attached. The honest ones always hedge — a chance of rain, a probability of a storm, never a promise. And yet the people who are actually good at reading weather rarely need the full forecast to act. A few clouds. A smell in the air. A change in the wind nobody else notices yet. They move before the instruments catch up, because waiting for certainty was never the point. Noticing was.

This issue is built on that same instinct, aimed at an economy that keeps handing us the same kind of incomplete picture. The 10-year Treasury just touched its highest level in nineteen years, and the 30-year hasn't traded this high since 2002. A government agency built to backstop housing is now buying mortgage bonds by the tens of billions to hold a rate down that the bond market has already decided should be higher — a lever pulled by one hand while another hand, at the Fed, pulls the opposite one for the opposite reason, in the same month, ahead of the same election. Underneath a stock index sitting near its highs, most of the stocks in it are quietly weak — a divergence some analysts haven't seen since right before the dot-com bubble finished bursting. None of that is a single, clean signal. All of it is a barometer pinned somewhere it shouldn't be.

What convinced me this issue needed to exist wasn't the data, though. It was noticing how much of the real cost of all this doesn't show up in any of it. A half-empty plane at a four-figure fare isn't in the inflation numbers. A farmer losing money on every acre while the land itself gets more expensive isn't in a headline about GDP. A grocery bill quietly split into four payments doesn't register as strain anywhere except in a lender's late-payment data, months after the fact. The instruments we're all trained to check are telling us less than we think they are, at exactly the moment we need them to tell us more.

That's the whole argument this issue makes, cover to close: imperfect data is still actionable data. You don't need a perfect forecast to know when to turn the boat around — you need to notice the pattern early enough, and trust it enough, to act while everyone else is still checking the instruments and finding nothing wrong. This issue tries to be the noticing. What you do with it is yours to decide, same as it's always been.

I'll note an irony I didn't plan and can't take credit for: as I sit here finishing this draft, an actual nor'easter is moving up the coast — the same storm system behind the flight cancellation you'll read about a few pages from now, in a section written days before this note. I didn't build that coincidence into the metaphor. The weather built it into the metaphor for me. Sometimes the barometer doesn't need help making the point.

A real storm doesn't announce itself as an inconvenience and leave it there. It knocks down power lines, and the outage that follows starts fires in dry brush a hundred miles from where the wind actually blew. It pushes water over a levee that held fine the year before. It takes the roof, the foundation, the things a family spent a decade paying for, in the time it takes the sky to go from gray to black. Nobody who has lived through one needs the metaphor explained twice. And nobody who takes that risk seriously waits for the wind to pick up before calling the insurance agent — the policy gets bought on a clear afternoon, precisely because the cost of being unprepared for a storm that arrives is so much higher than the cost of preparing for one that doesn't. Consider this issue the same kind of policy, bought early, on a day that still looks clear to almost everyone else.

The clouds are visible. Let's get into it.

The Fed has one lever for long-term interest rates: buy or sell, and let the price find its level. This year, someone else picked up a second lever and started pulling it the other way.

On January 8, the President posted that he was directing Fannie Mae and Freddie Mac to buy $200 billion of mortgage bonds, and FHFA Director Bill Pulte — who also chairs both companies — confirmed they'd be the ones buying. Four days later, FHFA quietly raised each firm's mortgage-bond portfolio cap from $40 billion to $225 billion — more room than the $200 billion order actually required. That's not incidental. It's a government regulator handing government-controlled companies the standing capacity to lean on a market price whenever it's asked to again.

Call it what it is: this is the government manipulating the mortgage market. Not in some narrow legal-brief sense — in the plain sense everyone already understands: a thumb on the scale, applied by the party that controls the scale.

You've seen this exact shape before, just wearing a different sign. Walk onto almost any car lot and you'll see "0% APR" or "$1,000 below invoice" in the window. The bank didn't wake up and decide lending you money is free. The manufacturer paid that interest cost on the back end, out of its own pocket, so the number you see looks better than the number the actual credit market would have handed you on your own merit. Your risk profile didn't improve. The true cost of borrowing that money in the open market didn't move at all. What changed is that a party with deep pockets and a reason to move the product decided which number you'd be allowed to see. Somebody still pays the difference — it just shows up somewhere you're not looking, baked into the sticker price, spread across every other buyer, or carried on a balance sheet nobody hands you a statement for.

That's the GSE playbook, run on a house instead of a car — except here, there's a wrinkle the dealership analogy doesn't cover. A manufacturer eating an interest cost is spending its own money. The federal government doesn't have its own money. It has taxpayer money, and when that isn't enough, it has borrowed money — and this program is being run by an institution that is, itself, $41 trillion in debt and climbing. The GSEs' capacity to "eat the difference" ultimately traces back to a government balance sheet that is already the borrower of last resort for its own deficit, now also volunteering to absorb mortgage-market losses on top of it.

That matters because this specific lever hasn't clearly worked. Mortgage rates were 5.98% in late February, before the Iran conflict pushed oil and yields higher. They're 7.26% now. The GSE buying compressed the mortgage-Treasury spread for a while — briefly under 190 basis points — but that spread is back near 219 basis points today, roughly where it started. $200 billion committed against a $13.5 trillion mortgage market is about 1.5% of it: enough to move a spread temporarily, not enough to overrule a bond market that just sent the 10-year Treasury to its highest level in nineteen years. A lever that gets pulled, costs real money, and mostly reverses within months is the fiscal equivalent of a car dealer's buydown that expires the moment the promotion ends — except taxpayers, not shareholders, are the ones who financed the promotion.

Here is where the two threads of this issue meet. The national debt passed $40 trillion this August. Interest expense alone ran $963 billion for just the first ten months of this fiscal year — up 14% year-over-year, now the government's second-largest budget line, ahead of Medicare, and equal to roughly 70% of the entire Social Security outlay. Every dollar the government commits to a program like this one is a dollar borrowed at today's rates, on top of a debt pile already this large, in a year when this newsletter's own opening story is that borrowing costs just hit a two-decade high. Debt taken on now doesn't disappear if the program doesn't work. It compounds, at whatever rate the bond market is currently demanding — which right now is the highest rate in a generation.

There are only two ways that bill eventually gets paid, and only one of them is honest. The first is future taxpayers: higher taxes, or fewer services than they'd otherwise have gotten — a straightforward, at least visible, transfer from people who aren't old enough to vote on any of this yet to people who already spent the money. It's regressive and often unfair, but it runs through Congress, on the record, with a name attached to the vote.

The second path is currency debasement, and calling it "paying the bill" is generous to it. Nobody votes on it. It isn't authorized, itemized, or debated on the floor of Congress — it's executed unilaterally, on the Fed's balance sheet, by an institution nobody on that balance sheet ever elected, then discovered by everyone else after the fact, in the price of groceries, rent, and everything else a currency buys. That is taxation without representation in the most literal sense still available to a modern economy: the people paying it never got a vote, and most of them will never be told a tax was levied on them at all. It is, historically, the more common ending — and it is the less honest one, not the more honest one, precisely because no one has to stand up and vote for it. President Trump has said as much directly: in a Time interview reported October 1, he said "certain levels of inflation" could pay down the national debt "very rapidly." He isn't wrong about the mechanism — only about the word for it. Paying a debt back in dollars worth less than the dollars borrowed isn't repayment. Call it what it actually is: theft of purchasing power, dressed up as a de facto default where bondholders, savers, and wage-earners all pay the difference even though no bond technically goes unpaid. CBO's own long-term outlook warns that a growing debt burden risks undermining confidence in the dollar itself, raising the cost of borrowing for the government and everyone else who borrows in dollars alongside it.

There's a second, quieter version of the same bet playing out in how the government borrows in the first place. Treasury under Secretary Scott Bessent has kept debt issuance heavily weighted toward short-term bills — T-bills under one year now make up roughly 22% of marketable debt as of this past summer, above the 15–20% range the Treasury's own borrowing advisory committee has recommended since 2020 — because short-term rates run meaningfully below long-term ones, a gap of roughly 120 basis points as of early August. It's the same math as a homeowner choosing an adjustable rate over a fixed one to save on the payment today. It works, right up until the loan comes due and rates haven't cooperated. A growing share of this debt doesn't lock in for thirty years; it has to be rolled over and repriced within months. Layer in a president who has openly floated inflation as a tool for shrinking the debt's real value, and the two risks compound each other: debt that resets fast, married to a stated appetite for the exact kind of inflation that would force each reset to land at a worse rate than the last one. That's not a cloud on a distant horizon. It's the setup where, once the first bad roll happens, the next one is already baked in behind it.

This isn't confined to one mortgage-bond program, either. This administration has increasingly routed policy around Congress on multiple fronts at once — tariffs set unilaterally under emergency-authority statutes, spending redirected without an appropriations vote, and now a housing-market lever pulled by directive rather than legislation. This newsletter doesn't know whether that pattern is designed to minimize the say of the people's elected representatives or is simply a byproduct of how this administration prefers to govern. It doesn't need to know. The practical effect is the same either way: decisions that used to require a floor vote increasingly don't, and the people those decisions land on have correspondingly less voice in them, intentional or not.

Debt monetization doesn't require a directive to happen, either. It happens just as effectively through artificially suppressed interest rates, and through the Federal Reserve acting as buyer of last resort — expanding its own balance sheet through quantitative easing to absorb debt the open market won't take at the price the government wants to pay. It's worth naming what the Fed actually is while making that point: not a normal federal agency funded through the ordinary appropriations process, and not a private company either — a hybrid, quasi-governmental structure whose governors are nominated and confirmed but whose actual balance-sheet decisions, month to month, face no vote at all. Functionally, it's a lever with the same distance from direct democratic consent as the GSEs this section opened with.

And "temporary" has a track record worth remembering. The Fed's balance sheet sat around $0.9 trillion before the 2008 financial crisis. Quantitative easing was sold as an emergency measure, one Fed chairman at the time describing the purchases as fully reversible once the economy recovered. By the mid-2010s the balance sheet had grown to roughly $4.5 trillion and never came close to returning to its pre-crisis size. Then came 2020: another round of buying took it to nearly $9 trillion at its peak. Even after the most aggressive unwind in the Fed's history, it sits at $6.7 trillion today — more than seven times its pre-2008 level, seventeen years after the emergency that supposedly required it. Temporary, it turns out, is temporary until it isn't.

This matters for a debate this newsletter sees play out every time prices rise: whether the blame belongs to "greedy corporations" or to a currency that buys less than it used to. The two aren't mutually exclusive, and the honest answer, per the data, is that it depends on the year. Kansas City Fed research on the 2021–2022 recovery found that corporate profits accounted for roughly 126% of that period's inflation in year one — genuinely outsized, and a real instance of pricing power doing damage — but by year two, cost increases rather than profit-taking drove most of the rise, with profits' overall two-year contribution actually landing below the historical average for a post-recession recovery. That data point cuts against a purely monetary explanation for 2021. It cuts just as hard against a purely corporate-greed explanation for the run of inflation since. Roughly 33 million small businesses, and millions more larger firms, did not coordinate a simultaneous, multi-year, national price hike. The more parsimonious explanation most years is that a currency debased through borrowing, suppressed rates, and Fed balance-sheet expansion buys less than it used to — and a business facing its own rising costs has to charge more just to stand still, the same as anyone drawing a paycheck. Some of what gets called corporate greed is real, and the data above proves it happened at least once, recently, at scale. Most of what gets called corporate greed, most years, is a currency problem wearing a corporate face.

That's not an abstraction that lands evenly. Currency debasement is regressive by construction. Someone in the upper-middle class or above holds assets — a home, equities, a business — that were already, this issue's whole point, appreciating faster than the inflation eating away at cash. Debasement is an inconvenience to them, absorbed by portfolios built for exactly this. Someone at the bottom of the K holds a paycheck, not a portfolio. Wages don't reprice as fast as the dollar loses value, and a family with no cushion doesn't get to wait out a currency cycle — they eat it in real time, at the register, every week, in groceries that cost more with a paycheck that doesn't. A policy sold as help for "affordability" that runs through a $41 trillion balance sheet, on a program that has mostly reversed, ends up financed by the people it was least able to help in the first place.

Who benefits in the meantime is not a mystery, either. Rocket Companies rose more than 9% the day of the announcement, UWM Holdings more than 13%, PennyMac 6%. Lenders and the households who already qualify for a loan get a real, if partial and temporary, subsidy. Everyone else gets the bill later, with interest, unevenly distributed by design.

There's a second irony worth sitting with. The Federal Reserve Act's mandate isn't just employment and stable prices — it names "moderate long-term interest rates" as a third, explicit goal, the same lever the GSEs are now being aimed at directly, by an office that isn't the Fed. In September, the Fed hiked to fight inflation. The same month, the GSEs were told to buy harder to push a specific long-term rate down. Two arms of the same government, working the same variable in opposite directions, in the same season, ahead of the same midterm election, financed by a debt load that doesn't care which direction either lever is pointed.

The mechanism itself isn't partisan — any administration holding these levers could pull them the same way, and the newsletter would name it the same way. What's not up for debate is the arithmetic: a government with no money of its own spent borrowed money on a fix that mostly didn't hold, and someone is going to pay for that borrowing eventually — through taxes, or through a currency that buys a little less every year it goes unaddressed. History says the second one is more likely. Either way, it isn't the government that pays first. Watch the spread. Watch who's still standing when the bill comes due.

One government lever wasn't the only place this pressure system showed up this month. Read the barometric pressure over the airline industry right now and it's falling just as fast — forty thousand feet up, a different kind of thumb on the scale was written into every fare on the board.

In numbers: FHFA's mortgage-bond portfolio cap per GSE ran from $40 billion before the directive, on January 8, to $225 billion after it, on January 12 — a $170 billion jump in purchasing capacity, beyond what the $200 billion directive alone required. The mortgage-to-Treasury spread ran 200 basis points in summer 2025, compressed to a GSE-buying low of 189 basis points, then reverted to 219 basis points by September — back near where it started. The Fed's own balance sheet moved from $0.9 trillion pre-2008, to $5.3 trillion by the mid-2010s, to a 2022 peak of $9.7 trillion, to $6.7 trillion now — more than 7 times its pre-2008 level, 17 years after the 'temporary' emergency that supposedly required it. And on who actually drove the inflation that followed: corporate profits accounted for 126% of the increase in year one, 2021; cost increases drove most of year two; the two-year average came in at 41%, below the 59% historical post-recession norm.

A week after my cousin Becca passed from cancer, the fare from Hartford to Atlanta showed $1,097 — roughly four times what the same route cost two years ago. I do what I always do with a number like that: I look at how full the plane actually is, because eventually supply and demand catches up to a price that's gotten ahead of itself. Two days before the service, the airline dropped it to about $890. The plane was still half empty.

Put that against the number everyone's told to trust. The airfare component of the Consumer Price Index — already one of the most elevated line items in the entire index, elevated enough that it gets covered as its own story — is up a cumulative 27.5% over the same two-year stretch. My own fare on this exact route didn't rise 27.5%. It roughly quadrupled, easily tripling at minimum on any month I've priced it out. And that gap doesn't even count the part no index measures at all: two years ago this route had multiple daily nonstops to choose from. Today a nonstop is the exception rather than the rule, connections through JFK or Chicago have become the norm, and the flights that do run are doing it with fewer seats filled at a higher price. One routing option actually offered on this trip makes the absurdity concrete: fly from Atlanta to Hartford by way of Chicago — geographically backward, hundreds of miles out of the way — with a two-hour layover that, by itself, runs longer than the direct flight it's replacing. An index built to track the price of a ticket was never built to track the disappearance of the ticket you actually wanted.

The return trip made the point more clearly, because it came with a natural experiment attached. The original flight got canceled by a nor'easter. Every experienced traveler knows what's supposed to happen next: the flights that do get out after a weather event are packed, because everyone who got bumped is now competing for the same seats, and I heard other passengers saying exactly that on the way to the gate. Instead, the Atlanta-to-JFK leg — a day and a half after the storm, exactly when it should have been overbooked — was half empty. JFK to Hartford was one-fifth full.

I'm not going to tell you what that means. I'll just tell you what it isn't: it isn't a price finding its market-clearing level, which is what a mostly-empty plane at a record fare is supposed to trigger. Something is holding that price up against exactly the pressure that should be pushing it down.

It turns out the airlines have said, on the record, why. Jet fuel hit $4.51 a gallon the week of September 18 — up nearly 80% year over year — and fuel typically runs a quarter to a third of an airline's operating cost. United's CFO put the strategy plainly this quarter: "We are not flying to maximize market share. We're flying to maximize profitability and free cash generation." American has already dropped routes out of Cleveland and Pittsburgh. Southwest's CFO said the same thing in different words: if fuel stays high, capacity gets trimmed. Domestic airfares were up 26.5% year over year in June and 25.5% in July. None of that is a coincidence sitting next to a half-empty plane at a four-figure fare — it's the strategy working as designed. When the marginal seat costs more to fly than it used to, airlines aren't chasing that seat anymore. They'd rather fly fewer planes, fuller with people who have the ability to pay, than cut the price and fill every row with people who don't. That distinction matters — willingness assumes everyone's simply choosing not to pay, when the actual constraint for most people is that they can't. Corporate and premium travel is carrying the math: American's premium seats brought in half its revenue this year from just 30% of its capacity.

That's a rational response to costs, not a conspiracy. But look at what it does to two different versions of the same trip.

I found out about the death in the middle of a business conference, on schedule, in front of people I couldn't just walk away from — so I did what you do: held it together, and spent the next four hours working the phone and the apps trying to find a flight that would get me there and back without costing me other business conference commitments I had on the calendar. I found one. I paid for it. That's the part where the fare mattered less to me than it does to most people, and I'll say plainly that I benefited from having the means to solve the problem with money instead of losing the trip entirely.

But solving it with money didn't mean it cost nothing. The return leg alone — booked separately, after the original flight was canceled — ran roughly half of that discounted $890 fare, call it $450, and even at that price the trip took eleven hours, routed through JFK, for what used to be direct in three and a half: a delay out of JFK on a layover that didn't exist three years ago, then close to another hour sitting on the tarmac in Atlanta, then my reserved rideshare never showing after I waited close to an hour outside, then a cab that cost 30% more than what I'd already paid for a ride that didn't come. None of that stopped me from getting there. All of it made getting there cost more — in time, in money, in the accumulated friction of a system with less slack in it than it used to have.

Here's the part that's less visible than the delay, and probably more important: once I saw the flight was going to be roughly half empty, I did the math the way I do with any position — what am I actually getting for this price. Two years ago, a business-class seat on this route — one tier up from coach, priority boarding, a free checked bag, room to actually rest — ran about $800. This time, at a coach fare, I effectively got that same seat for free: nobody next to me, nobody in front of me or behind me, room to stretch out and sleep on a flight that would otherwise have been a bad way to end a bad week. I didn't buy business class. The empty plane gave it to me anyway.

That's before the part that compounds. Most airline loyalty programs now award miles and status based on dollars spent, not just distance flown, so a higher fare — even at a discount off a still-elevated sticker price — earns disproportionately more toward status than the same seat did at yesterday's lower price. I already hold a top-tier status on this airline, and status like that isn't incidental to how I fly, it's a strategy: every January, I buy two full-price, first-class, transcontinental tickets. That alone jumps my status to one tier below the top for the rest of the year. Every flight after those two doesn't need to be first class, and doesn't need points or miles spent on it, because the status itself triggers a free upgrade automatically. By the end of a given year, my average cost per flight comes out to around $300, and I'm sitting in first class or one seat back from it almost every time I fly. A coach flyer without that status is paying something closer to $600 a flight before a checked bag, plus another $100 or so in bag fees on top of that — for a worse seat, on a more crowded plane, with none of it compounding into anything for the next trip.

I'll let the individual determine for themselves whether they think that's fair or not. One thing is certain: it's our reality, running quietly in the background of every fare on that half-empty plane, whether anyone in coach ever sees the math behind it or not.

Think of the empty rows as a storm surge running in reverse — the water pulled back out to sea before the wave that's still coming in. That same undertow — protect the margin, let the volume go, and let whoever can't absorb the price absorb the disruption instead — doesn't stay in the sky. It touches ground on American farmland this fall.

In numbers: domestic airfare ran 26.5% year-over-year in June 2026 and 25.5% in July — rising even as this issue's own flights flew half to one-fifth full.

Four pressures are landing on the same group of people at the same time this fall, and none of them are coincidental — each one traces to a specific policy choice made by this administration.

Start with tariffs. Soybean farmers lost roughly $75 for every harvested acre in the last crop year, even after a $12 billion federal aid package arrived in December to offset the damage. Soybean exports are running 15% to 20% behind normal. In a late-March survey of 400 farmers by Purdue researchers, almost half said their operation is worse off than it was a year ago. This isn't a novel mechanism — USDA's own retrospective on the 2018–19 trade conflict put the agricultural export loss at more than $27 billion, and this round is repeating the pattern: a tariff is a tax on a foreign buyer that a domestic seller ends up eating when that buyer simply buys from someone else.

Layer fertilizer on top of that. Anhydrous ammonia has risen 35.6% since the war with Iran began — from $828 a ton before the conflict to $1,123 a ton by mid-April — and nitrogen solution is up 25%. For a corn farmer, that alone adds $20 to $27 an acre in fertilizer cost, arriving at the exact moment this issue's own Insurance Mechanism section is tracking how that same conflict has driven up shipping and insurance costs everywhere else in the supply chain. This administration is a direct party to that conflict, not a bystander to it, which puts fertilizer cost squarely alongside tariffs as a pressure with a policy address.

Then add labor, at harvest, when it matters most. In parts of the Rio Grande Valley, up to 75% of farm workers stopped reporting after immigration enforcement actions swept through; in California, Texas, and Pennsylvania combined, the figure runs as high as 70%. Nearly half the national agricultural workforce lacks legal status, according to USDA's own estimate, which means enforcement policy and harvest labor aren't two separate stories — they're the same story. Reporting from Texas describes entire farms sitting empty for days during the exact window when a crop has to come in or it's lost. A farm bureau representing more than 25,000 members put it as plainly as an industry group ever does: this threatens both food affordability and availability nationwide.

The bottom line shows up in bankruptcy filings, not sentiment surveys. Family-farm Chapter 12 bankruptcies rose 46% in 2025 to 315 filings, up from 216 the year before and a 2023 low of 139. April 2026 alone saw 62 filings — the highest single month since February 2020. Corn and soybean growers are losing more than $100 an acre before any subsidy arrives. Bankruptcy rates are up more than 69% across the Midwest, Southeast, and Southwest. Three separate policy pressures — trade, a war, and enforcement — landing on the same balance sheet in the same season is not something a farmer's own decisions can hedge against.

And on the other side of that same distress, capital is moving in. Farmland prices still rose 4.7% nationally in 2025, to $4,350 an acre, even as the people working that land lost money on every acre they planted. Institutional buyers aren't waiting on the sidelines — Nuveen Natural Capital, Farmland Partners with 71,000 acres and growing, AcreTrader before its recent acquisition by Proterra — are actively expanding into the same land distressed operators are being forced to exit, describing it openly as an inflation hedge and a buying opportunity. Someone is losing $100 an acre. Someone else is buying that acre at a rising price because they can afford to wait out the very distress the first person can't survive. That is the K-shape in its most literal form: two people, standing on the same soil, living two different outcomes from the same policy environment.

There's an irony underneath all of it worth stating plainly, because this newsletter doesn't pretend not to see it: a meaningful share of the people now filing Chapter 12, watching their labor disappear mid-harvest, and eating a fertilizer bill that jumped 35% inside a year are in the bottom half of the K — and many of them supported the administration whose tariff policy, war posture, and enforcement priorities are the direct, traceable source of all three pressures. That's not offered as a judgment on anyone's vote. It's offered as the most literal test case this newsletter has seen yet of its own core thesis: policy is never neutral, someone always absorbs the cost of a choice, and the people absorbing this one are disproportionately the people the policy was sold to.

We don't have a mental model to offer that makes that easier. We have the arithmetic, and the arithmetic is the point.

Tariffs, a war, and an enforcement sweep are three fronts of the same weather system, not three separate storms — three swells building in the same direction, stacking on top of each other until the surge is bigger than any one of them could have produced alone. Pull the camera back far enough and they're all one pattern: a cost gets imposed somewhere, and it comes down hardest on whoever has the least shelter built.

In numbers: family-farm Chapter 12 filings ran 139 in 2023, 216 in 2024, and 315 in 2025 — a 46% jump in 2025 alone, with April 2026 alone seeing 62 filings, the highest single month since February 2020.

Inflation is not a weather event that falls evenly on everyone standing outside in it. It is regressive by construction, and the mechanism is simple enough to fit on an index card: the bottom of the economy holds cash and fixed wages, which lose real value every year prices rise. The top of the economy holds appreciating assets — stocks, real estate, businesses — which inflation tends to lift rather than erode. The same event that quietly taxes one household enriches the other, and it does this by design, not by accident.

The numbers make the mechanism impossible to wave away. Only 34% of the bottom half of the income distribution owns any stock at all, versus 95% of the top decile. Among people who do own stock, the bottom half's median holding is $12,600; the top decile's median holding is $608,000 — a roughly 48-fold gap. Run both through an ordinary 7% year: the bottom half gains about $880. The top decile gains about $42,560. That's not two people experiencing the same market from different starting points. That's one household riding the swell and another standing in the flood plain beneath it, in the same economy, in the same year, watching the same headlines about "the market" being up. Between 2019 and 2022 — a stretch that included the largest three-year jump in real median net worth this survey has ever recorded — the bottom income quintile's own net worth actually fell, from $132,200 to $129,700. The gains were real. They just didn't go to the people this newsletter is written for by default.

This is the argument for taking inflation seriously as a rate problem, not managing around it. A 100 to 150 basis point hike, delivered with real conviction, causes near-certain pain — homeowners refinancing into a higher rate, car buyers priced out of a loan, households already carrying maxed-out debt watching their minimum payments climb. That pain isn't a side effect to be weighed against the alternative of doing less. It is the cost of the fix, paid up front, by people who can least afford to pay it — which is precisely why it's tempting for anyone in charge to avoid paying it.

The alternative isn't painless. It's just slower, and it compounds. Under-treating inflation doesn't make the bill disappear — it turns a sharp, visible cost into a long, quiet one: continued erosion of real wages, continued softening of the currency's purchasing power, continued drift toward what economists call a Sovereign Debt Crisis — a national insolvency problem instead of a household one. The bottom rung absorbs the cost either way. The only real choice is whether it's paid all at once, where everyone can see it and hold someone accountable for it, or in small increments nobody quite notices until they add up — the equivalent of ripping a bandage off in one motion versus pulling the hairs out from underneath it, one at a time, for years.

I have a personal reason for holding this position as firmly as I do. I was a child when Paul Volcker took the federal funds rate to nearly 20%, and I remember what that did to my own family — not in the abstract, in the actual poverty it caused, before I was old enough to understand why it was happening. I learned the reasons later, in school and after — and as I got older, I came to understand something else, too: the reckless spending and the wars run up by irresponsible politicians in the years before, the deficits that got spent into an economy with no plan for how they'd be paid down, were what created the inflation Volcker was then forced to burn out of the system at such a brutal cost. The pain didn't start with the fix. It started with the decisions that made the fix necessary, made by people who were never the ones who had to live through it. What Volcker did was necessary. Inflation had run at 11% in 1979; by the time the fix took hold in late 1982, it was down to 5%, at a real cost of nearly 11% unemployment — the highest of the postwar era, with residential construction and auto manufacturing absorbing unemployment rates north of 20%. That was the price for actually solving the problem, paid honestly, in public, by a country that then moved forward.

I'd be remiss if I didn't name the other government decisions from that same era that made the poverty worse, and that eventually reached the block I grew up on. The deficit spending and the wars of the 1960s into the mid-1970s didn't just create the inflation Volcker was later forced to burn out of the system — they also set the stage for what came next. In the years right after Vietnam, the U.S. government's own Contra war in Nicaragua became entangled with cocaine trafficking into American cities, largely through Contra-connected pilots, suppliers, and financiers. The 1988 Kerry Committee found "substantial evidence of drug smuggling" on the part of individual Contras, Contra suppliers, and their supporters, and documented State Department humanitarian-aid payments made to people who had already been indicted on drug charges. The CIA's own inspector general, in a 1998 review, concluded the agency did not deliberately orchestrate drug distribution in American cities, but acknowledged it had, in his words, exempted its own officers from a legal duty to report narcotics trafficking by its assets and repeatedly "didn't bother to look into" credible allegations against people it was working with — negligence at best, a blind eye at worst, by the inspector general's own description. What isn't in dispute is the timing: cheap cocaine, and then crack, moved into American communities already hollowed out by deficit-driven inflation and a manufacturing base that hadn't recovered, at the exact moment people out of work were most susceptible to an illegal economy offering a way in. I see this as the first imperfect storm of my own life — deficit spending driving inflation, inflation driving the country off the gold standard, that shock deepening the poverty already spreading through communities like mine, and drugs and arms arriving in it at the same time, all compounding into the conditions I was raised in, in South Central Los Angeles and Compton, from the early 1980s through the mid-1990s. That storm didn't pass quietly. It culminated in mass incarceration that irreparably altered two generations, Black and Latino communities most of all, in cities from Harlem to Oakland to the one I called home. And prison wasn't the only thing that storm took. It cost lives outright, not just years: the national homicide rate, after a first peak of 10.2 per 100,000 in 1980, fell through the middle of the decade before climbing back to a second peak of 9.8 per 100,000 in 1991, as the crack years crested — effectively matching the rate that had already alarmed the country a decade earlier. Behind that number were people, disproportionately young, Black, and male, who didn't get the chance to grow past those two decades at all. Any honest accounting of that era has to hold both costs at once: the generations rerouted into prison, and the ones who never made it out of the storm to begin with.

There's a harder truth underneath all of it that I've come to believe the older I get: hurt people hurt people. When an economy strips enough people of a legitimate way to earn a living, it doesn't just create desperate buyers. It creates desperate suppliers, too. Somebody struggling to feed a family will eventually find a way to provide the escape that somebody else in pain is already looking for — that's not a moral failing unique to any one neighborhood, it's what happens anywhere hope and opportunity get hollowed out at the same time, in the same place. The person selling and the person buying were, more often than not, failed by the same storm, standing on opposite ends of the same transaction.

I don't think that storm ever fully passed. I think it just found new vehicles. The drug trade didn't go away when the crack years ended — it professionalized, and cryptocurrency became one of its preferred instruments. TRM Labs' 2026 Crypto Crime Report puts the online illicit drug trade at over $3.4 billion in incoming cryptocurrency volume in 2025 alone, with Chinese money-laundering networks now moving billions of dollars in proceeds between cartels and the precursor-chemical suppliers who feed the fentanyl pipeline — laundering that barely existed at scale a decade ago and now runs in the billions, precisely because crypto lets that volume of money cross borders without ever touching a bank that has to ask where it came from. Artificial intelligence is what makes the next layer of that laundering possible. TRM Labs' 2026 AI-in-Crime Adoption Index documents synthetic-identity kits and real-time deepfake face-swap and voice-cloning packages, some sold for as little as $500, that let traffickers open cryptocurrency exchange accounts under fabricated identities — defeating the "know your customer" checks that exist specifically to catch this kind of money before it moves. History doesn't repeat itself, but it rhymes. The uprising and political turmoil of the 1960s gave way to the deficit spending, poverty, and drug-and-arms flood of the 1970s and '80s traced earlier in this section. I believe some version of that cycle is due again — compressed, not distant, because the money, the drugs, and now the disguise all move faster than they did then. The same out-of-work opportunism that turned to dealing when the last storm hit will turn to cryptocurrency to hide the proceeds, and artificial intelligence will be the tool that helps that money, and the identity behind it, disappear. Neither technology created these criminal networks. Both of them just made the same old exploitation of desperate people faster, harder to trace, and larger in scale than the analog version I grew up watching play out on the block. Layer that onto the storm this issue has spent its pages tracking, and the surge isn't the ordinary kind anymore. Thirty-foot swells. Flooding that doesn't recede between systems. Lightning and thunder behaving less like a storm and more like a tornado touching down without warning. Every instrument this issue reads this season is telling the same story: this isn't a drizzle to wait out. It's a storm that keeps restocking itself with a new hazard every time the last one starts to clear.

Set against that same backdrop, it's worth asking what that 11% actually looks like next to today's headline number, since the way unemployment gets counted has itself changed since then. The official rate ticked up to 4.2% in September 2026, from 4.1% in August — but that figure excludes long-term discouraged workers, people who've stopped actively looking for work, a category the government defined out of the official count in 1994. The September print itself was weak on every axis that matters: the economy added just 29,000 jobs against the roughly 90,000 economists expected, August's already-soft 162,000 was revised down to 133,000, and average hourly wage growth came in at 0.1% for the month, a third of the 0.3% forecast. The Bureau of Labor Statistics' own broader measure, U-6, which counts people working part-time who want full-time hours along with those only loosely attached to the labor force, ran to 7.7% as of the last reading. Add back the long-term discouraged workers the pre-1994 methodology used to count — a group independent analysts, Shadow Government Statistics among them, estimate would push the fuller rate into the mid-20s — and the honest comparison to 1982 looks considerably closer than the headline number alone would suggest.

No one today would dare do even half of that. And that reluctance is not compassion. It's the reason this newsletter keeps returning to the same image: a slow, persistent drain, tedious enough that no single week of it feels like a crisis, closer to a disease eroding the body's vitality than an injury anyone would notice and treat. A country that won't take the fast, visible cost has chosen the slow one instead — and the slow one is still a cost. It's just been quietly reassigned to the people with the least ability to negotiate the price.

Before this issue returns to spreadsheets and scenario tables, a different kind of forecasting — the kind that gets handed down in a family rather than published by a data desk.

In numbers: 34% of households in the bottom half own any stock at all, versus 95% in the top decile. And among those who do own stock, the median holding is $12,600 in the bottom half versus $608,000 in the top decile — roughly a 48-times gap. At an ordinary 7% year, that's about $880 in gains against about $42,560 on the same market move. The homicide rate over that same stretch moved in two waves: 10.2 per 100,000 in 1980, down to 7.9 by 1984, back up to 9.8 in 1991 as the crack years crested, then down to 8.0 by 1995 and 5.5 by 2000.

My uncle had a way of reading weather that astounds me to this day , even having seen it work more than once. The clearest example was a family reunion, late July 2023. We were all together, the whole thing in full swing, when he said it: "A storm's coming."

You have to understand how that phrase gets used where our grandparents are from , the South, the Mississippi Delta specifically, not where he actually grew up. My uncle was raised in Los Angeles. But he spent summers there as a kid with his grandparents — my great-grandparents — and other family, and I have to assume this got handed to him the same way it eventually got handed to me: not taught so much as absorbed, standing next to someone who already had it. Down there, "a storm's coming" doesn't mean rain. Rain happens constantly and nobody clears a room for it. It means something you need to be prepared for, in the true sense of the word.

Picture the day itself the way I still carry it, because I can close my eyes and walk back into it whole. The grill had been going since before most of us were out of bed, smoke curling up lazy over the yard, ribs and chicken laid out in neat rows, that low-and-slow smell working its way into everybody's clothes for the rest of the weekend. The potato salad sat under foil on the folding table, mustard-yellow and studded with egg, and there was a rule nobody had to say out loud: you waited for the blessing before a single plate got made, and you did not, under any circumstances, ask what was in it. Somebody had the ice cream maker going in the shade — the electric kind, rock salt and cream and that steady motor hum running under everything else — because the payoff, vanilla, still soft, eaten before it had any business being eaten, was worth the wait. Off by the carport, the uncles and the older cousins had a card table given over to dominoes, and that table ran louder than anything else on the property: bones slapping the wood hard enough to jump, somebody hollering about a count nobody had agreed to, laughter with nothing polite in it. Kids ran the whole yard in a pack, barefoot by early afternoon, playing some game whose rules changed depending on who was winning. And underneath all of it, somebody's speaker was doing the real work of holding the day together — the O'Jays, the Isley Brothers, Frankie Beverly and Maze, the kind of music that doesn't play in the background so much as become the background, so that even now I can't separate the smell of barbecue smoke from the sound of a Sunday afternoon refusing to end.

There were maybe a few clouds in the sky that day. That was it. Nothing that would have registered to anyone else as a warning. But it took him only a few seconds to decide, and within twenty minutes he had packed the car and was ready to go , in the middle of a family reunion, walking away from everyone else still enjoying the afternoon. We got in the car and started driving. Within the hour, the sky had opened into something close to a monsoon , flood warnings, lightning like I'd genuinely never seen. Much of the family got soaked just making it back to the hotel. This wasn't the first time he'd called it this precisely. It was just the one that stayed with me.

My grandmother had the same gift, and with over thirty grandchildren to keep track of, hers usually arrived the same way: "Uh ruh, baby" , not because that was your name, but because at that number of grandkids, and at her age, the specific name could take a second to arrive, and "baby," or whatever the equivalent was, got there faster. Then: "A storm is coming. Bring those clothes in off the line." We'd look up. A few scattered clouds, nothing that looked like anything. Nobody argued with her, though , not because we were naturally obedient, but because talking back, what that generation called sassing, got you the switch, and you went and cut your own if it came to that. That was simply how things were, and whether it was right or wrong by today's standard isn't really the point I'm making here. Every single time, minutes after the last shirt came off the line, the sky would open up like it had been waiting on us.

I've come to understand that neither gift came from nowhere, and I don't think the discipline did either. Reading the sky wasn't a hobby for the people who passed it down , both the African American and Native American sides of my family. Generations back, in the fields of the Delta, during enslavement, knowing what the weather was about to do wasn't a curiosity. It was survival. The same is likely true of the switch and the silence that went with it , a household run on strict, immediate obedience wasn't cruelty for its own sake in a world where a child talking back to the wrong person, at the wrong time, in that era, could carry real danger. That danger wasn't abstract or distant from where this story is set. This is the same Mississippi where Emmett Till was lynched in 1955, in Money, for the accusation of saying the wrong thing to the wrong person, and where Medgar Evers was gunned down in his own driveway eight years later for refusing to stay silent. Neither was far removed, in distance or in years, from the grandparents this story is about. The lesson a grandmother drilled into a child wasn't paranoia. It was accurate. Both things got carried forward hand to hand long after the reasons that first made them necessary had changed. I don't know if my grandmother or my uncle ever thought about either one in those terms. I hadn't either, until sitting down to write this. But I can't tell the story honestly without saying it.

Some people in my life think I inherited a version of it, just pointed at a different sky. In my investment group, and among a few of my fraternity brothers, I've picked up a couple of nicknames over the years , Q Stradamus, a play on my name and Nostradamus, and, depending on who's talking, a nod to Paul Mooney's Negrodamus character from the Chappelle Show, both said with real affection and a lot of laughing at my expense. It's not weather I've learned to read. It's macro. Sector rotations, credit cycles, the places the data quietly disagrees with the headline , models I've built and refined over twenty years, well enough to have beaten the SandP 500 by somewhere between 50 and 100 percent over that stretch. I don't think that makes me special. I think it proves something I actually believe: people make good decisions for their own circumstances when they have real information. Most of the time, the problem was never the person. It was what they'd been handed.

So here's an example, offered the same way my family would have offered it , not a prediction, just: I saw a few clouds, so let me count.

The clouds, in this case: a crypto market-structure bill collapsed in the Senate this month on an ethics dispute nobody in either party wants to own before an election. If the midterms go the way a lot of forecasters currently expect , Democrats taking the House, and the House matters more here than the Senate , that fight doesn't quiet down. It moves into committee rooms with subpoena power. Bitcoin's already down close to 40% from its October peak, and the corporate structure most levered to it, Strategy, is running a financing loop that only works when its stock trades at a premium to the coin it holds , a premium that's already compressed close to flat this year. A serious political overhang landing on top of a structure that thin isn't a guarantee of anything. Whether it turns into an actual storm , another 10 to 20% down, which isn't even that unusual a move for an asset this volatile , is the part I can't count yet, because I don't know how far away it is.

A few clouds don't mean rain. But when the people who are rarely wrong say a storm is coming, you bring the clothes in off the line and you pack the car.

Some storms show up on radar days out, wind speed climbing on schedule, everyone able to watch the pressure drop in real time. This one showed up in an insurance filing, with almost nobody watching.

The physical chokepoints get the headlines. The financial one underneath them gets almost none.

Global shipping doesn't run on the absence of danger , it runs on the ability to price it. Every tanker that transits a war zone does so under a separate war-risk policy, layered on top of standard hull and cargo coverage, and that policy is what makes the voyage possible at all. Take the policy away and the ship doesn't sail, regardless of whether it would have made it through safely.

That's what happened at the Strait of Hormuz this year. War-risk premiums there rose from roughly a quarter of one percent of a vessel's hull value before the war to as much as 10% at the peak , turning a routine $250,000 line item into a $7-10 million one for a single transit on a $100 million tanker. Traffic through the strait fell by an estimated 95% from its prewar average of about 178 ships a day, according to World Economic Forum reporting on the episode. Not because 95% of ships were being hit. Because the insurance math stopped working.

Bab al-Mandeb premiums had been climbing on the same pattern , a few tenths of a percent, still underwritable , before this month's Houthi takeover of the strait itself. Whether that repeats Hormuz's trajectory is the open question, not a foregone conclusion.

There's a name circulating for the broader strategy this describes: rather than concentrating force on a single chokepoint, spread unpredictable, low-cost attacks across several at once , Hormuz, the Red Sea, the Suez approach , so that insurers can no longer model the risk region by region. Professor Jiang, discussing the dynamic on Steven Bartlett's Diary of a CEO podcast in mid-September, was careful to frame it as a possible pattern rather than a forecast , a caution worth keeping here too. The mechanism itself, though, isn't theoretical. It already shut down 95% of one strait this year.

If shipping insurance is one quiet pressure gauge, the bond market is the loudest instrument this issue has , and its barometric pressure has been falling, steadily, for weeks.

In numbers: the war-risk premium on a Hormuz transit ran 0.25% of hull value before the crisis, and peaked at 10%. Vessel transits through the strait ran about 178 a day before the war, and fell to about 9 a day during the crisis — a roughly 95% decline, not because ships were being hit, but because the insurance math stopped working.

On September 16, the Federal Reserve raised the federal funds rate 25 basis points, to a range of 3.75%–4.00%. The vote was unanimous, 12–0. It was the first hike since July 2023, ending a cutting cycle that had run through most of the previous two years. This is no longer a forecast this newsletter has to hedge — it happened, on the record, and the Fed's own dot plot shows 16 of 18 officials projecting at least one more hike before the year is out.

The next live decision point was the FOMC's October 27–28 meeting. In the final week of positioning before it, the market's read moved twice. CME FedWatch had better than 75% odds of a second straight hike as recently as September 23, after hawkish comments from Fed Governor Barr and a hot inflation reading. Then, on September 30, the Fed's preferred inflation gauge — the August PCE index — came in cooler than forecast (headline 3.4%, core 3.0%, both below expectations), and New York Fed President John Williams said the Fed "has time." October odds collapsed to essentially a coin flip heading into the meeting: 52.9% hold, 47.1% hike, per CME FedWatch. The committee wasn't of one mind, either — Williams read as the dove in the room, while regional presidents Hammack, Schmid, and Paulson had each delivered firm anti-inflation messaging in the same window. What hadn't moved was the full-year call: Polymarket still priced 81% odds of at least one more hike before year-end, with December the window traders considered more likely than October. This issue closed before the meeting's outcome was confirmed; readers will know by now which way the committee went.

One more data point landed just as this issue was closing, and it points the same dovish direction as Williams: September's jobs report, released October 2, showed the economy added just 29,000 jobs against the roughly 90,000 economists expected, while August's already-disappointing 162,000 was revised down to 133,000 and the unemployment rate ticked up to 4.2%. Wage growth came in at 0.1% for the month, a third of the 0.3% forecast. A labor market cooling that fast gave the hold camp a fresh, concrete argument heading into October 27–28 — though it did nothing to cool the inflation data driving the hike camp's case, which is exactly the tension this newsletter expects to keep yields elevated regardless of which way the committee ultimately leaned.

What changed the market's mind in the first place is largely the bond market itself — the barometric pressure this newsletter watches most closely, and it has kept rising even as the October odds swung. The 10-year Treasury reached 5.338% intraday on October 1 — its highest level since April 2002, a 24-year high — up from 5.223% just a week earlier, while the 30-year hit 5.620% intraday on September 29, its own highest level since 2002, before settling near 5.565%. Even the weak September jobs report two days later didn't cool things off: the 10-year ticked higher again on October 2 despite a payrolls miss that would normally argue for lower yields, a sign that fiscal-deficit concerns and term premium are now doing more to set the long end's price than the next rate decision is. A 5-year Treasury auction on September 23 cleared at 5.033%, up 64 basis points from the same auction in August, with a weak bid-to-cover ratio of 2.21 signaling softer demand than usual. None of that happens in a vacuum: oil back above $103 a barrel, September input costs rising at their steepest pace in four years per S&P Global's own PMI reading, and University of Michigan consumer inflation expectations jumping to 4.6% for the year ahead, up from 4% in August. The Fed hiked in September into an economy that was already telling it, through the bond market, that one hike wasn't going to be enough to convince anyone rates were under control — and the long end kept climbing even as the next-meeting odds swung the other way.

There's a structural piece worth naming plainly, because it's not just the deficit doing this anymore. Vanguard estimates that five major hyperscalers — Alphabet, Amazon, Meta, Microsoft, and Oracle — issued roughly $132 billion in debt through July alone to fund AI data-center buildouts, up sharply from a roughly $35 billion annual average between 2020 and 2024. Broader AI-related debt issuance across the data-center, semiconductor, and utility ecosystem could reach $300 to $570 billion this year. That's real new supply competing with Treasuries for the same pool of buyers, at the same moment those buyers are already demanding a higher premium for long-duration government debt. The AI buildout isn't just a story about earnings and capex anymore — it's now a bond-market story too, showing up directly in the yield this newsletter's own GSE and Cost of the Fix sections are tracking.

This is also a fair moment to check this newsletter's own work, since a forecast that's never revisited isn't worth much. Two calls this newsletter made in prior issues have now played out plainly in the data above, not as a coincidence but as the same mechanism continuing to run. In "To Serve Man," this newsletter called the long end "hostile" and expected the 30-year to hold above 5% and grind higher through the back half of the year — it's now at 5.565%, having touched 5.620%, a level last seen in 2002. In "The Fifth Economy," after Fed Governor Warsh's hawkish Jackson Hole remarks in late August, this newsletter revised its own call from a September hold to a September hike as the more likely outcome — the Fed delivered exactly that hike on September 16. Both calls were opinion when they were made, with a stated "wrong if." Neither was wrong. My grandmother would have put it more simply: when the people who are rarely wrong say a storm is coming, you bring the clothes in off the line.

Here's the scenario table, updated to reflect both the hike that's already happened and how quickly the next-meeting odds can move in nine days.

Here is the scenario table, in words. An October hike of 25 basis points, to a range of 4.00%–4.25%, is priced at around 47% — consistent with the hawkish wing of the committee; odds were above 75% just a week earlier, before the cooler August PCE print. An October hold, rate unchanged, is now the current favorite at around 53%, per CME FedWatch, after Williams said the Fed "has time" — that shifts the market's attention to the December meeting instead. An October cut is priced in the low single digits, and would require a sharp deterioration in employment or financial stability — not the base case given current data.

None of this means equities are next in line. Here's the mechanism, in plain terms, because it explains why a 24-year-high Treasury yield matters to a stock portfolio at all. A share of stock is really just today's value of all the cash a company is expected to hand its owners over many years to come — and that "today's value" shrinks the moment the safe, risk-free rate investors could earn instead goes up. Say a relative owes you $1,000, due in exactly one year — a stated, illustrative example, not a claim about any specific company. If a safe one-year Treasury pays 2%, you'd take about $980 today instead of waiting, since $980 earning 2% gets you to roughly $1,000 in a year anyway. But if that same safe Treasury now pays 5.3% — close to where the 10-year sits today — you'd only take about $950 today for that same $1,000 IOU, because $950 at 5.3% gets you there just as fast. Nothing about your relative's ability to pay changed. The rate going up is what made the future promise worth less, right now, in your pocket. A share of stock is that same IOU, just stretched across many years of a company's future profit instead of one payment a year out — so when the 10-year jumps to a 24-year high, every dollar of profit a company is expected to earn five, ten, twenty years from now gets marked down today, before the company's actual business has changed at all. That's the first hit. The second, slower one is higher borrowing costs actually dragging on the economy and denting the profits themselves, months or years later. The two don't have to arrive together: the S&P 500 kept climbing through 1999 while the yield curve inverted, and again through most of 2007, in both cases well before the second hit eventually landed. The St. Louis Fed's own research on Fed tightening cycles since the 1980s found that in the cycles where a recession did follow, the average gap between the last hike and the subsequent business-cycle peak was about 15 months. The fork that decides which kind of rising-yield story this is: yields climbing because the economy is genuinely growing faster are tolerable; yields climbing because of inflation, fiscal deficits, and the term premium investors demand to hold long-term government debt are not, since the borrower pays more without the economy producing more in return. This issue has already made its case for which kind this is. The S&P sitting at roughly +13% for the year, against a 24-year-high 10-year and a 7.49% 30-year mortgage rate, isn't evidence the bond market has it wrong — historically, equities are often just the last to know. The index's advance has also grown narrower underneath as it climbed. The Magnificent Seven alone now account for roughly a third of the S&P 500's total market capitalization — Palantir, increasingly mentioned in the same breath this year, adds further concentration on top of that — while the median stock in the index sits about 16% below its own 52-week high even as the index trades near records. On a typical day this past month, only 2–3% of S&P 500 members registered a new 52-week high against 5–6% that hit a new 52-week low, and fewer than half the index's constituents remain above their 200-day moving average, down from 73% earlier this year. A handful of names are carrying the tape; most of the index is not confirming the move.

One more number deserves the same scrutiny. The 7.49% rate quoted on CNBC or Fox Business isn't what most buyers are actually paying. It's the Freddie Mac survey average for a conventional loan, 20% down, good-to-excellent credit — a specific, narrow box. Step outside it — a smaller down payment, an FHA loan, a lower credit tier, a loan above the conforming limit — and the real rate climbs past 8% for a significant portion of today's buyers. The quoted average describes the best-qualified borrower. It doesn't describe the market.

In numbers: the Treasury curve as of October 1 ran 5.03% on the 5-year, 5.34% on the 10-year — its highest since April 2002 — and 5.57% on the 30-year. And the market's own odds for the October 27 to 28 meeting, as of September 30: 47.1% for a quarter-point hike, 52.9% for a hold, and about 1% pricing a cut.

Twenty-nine percent of BNPL users say they've used a buy-now-pay-later loan for groceries — up from 25% a year ago, and 14% two years before that. That's not a slow drift. It's the use case roughly doubling in two years, in a product category that was never built for it: BNPL was designed to spread a couch or a laptop across four payments, not a Tuesday grocery run.

The late-payment side of it is worsening at the same pace. 41% of BNPL users paid at least one installment late in the past year, up from 34% the year before — a seven-point jump in twelve months. And the grocery-financing habit isn't confined to the bottom of the income ladder the way the easy assumption goes: 38% of Gen Z BNPL users are financing groceries, 34% of parents with kids under 18, and 33% of users earning $100,000 or more. This has moved well past a poverty-only story. It's a household-cash-flow story, cutting into income brackets that aren't supposed to need this.

The Federal Reserve Bank of Richmond puts total BNPL purchase volume at roughly $70 billion in 2025 — about 1.1% of total U.S. credit card spending — and finds the category has grown at a real rate of roughly 20% a year since 2021. That's a financial product built for discretionary purchases now absorbing a meaningfully larger share of non-discretionary ones, at a pace that's accelerating rather than leveling off.

None of this shows up in the inflation prints this newsletter has already covered. A grocery bill split into four payments still counts as one purchase in the CPI basket — the strain of paying for it shows up nowhere except in a lender's late-payment data, months later. It's the same "imperfect data" problem as the airline seats and the empty rows: the number that's supposed to tell you whether people are financially fine is quietly missing the part where they aren't.

Grocery bills financed in installments are one hidden pressure moving through this economy — a swell nobody's measuring because it never breaks the surface. The trucks that deliver those groceries are absorbing a bigger one, and this one does show up in the data — if you know where to look.

In numbers: groceries financed through buy-now-pay-later roughly doubled in two years, reaching 29% of users now. And among BNPL users, those paying at least one installment late ran 34% a year ago, and 41% now — a seven-point jump in twelve months.

On-highway diesel opened 2026 at $3.477 a gallon. By the first week of August it was $5.348 — a 53.8% rise in seven months, the largest January-to-August increase in the full 32-year history of the EIA's data series. It kept climbing. On September 7, diesel hit $5.967 a gallon, breaking the previous weekly high of $5.810 set in June 2022. It did not stop there: by September 22, the national average reached $6.527 a gallon, a fresh record set just two weeks after the last one. The year-to-date average is now tracking well above the 2022 annual record of $4.989, and the gap is still widening, not closing.

Retail gasoline, by contrast, is still below its 2022 peak. That divergence is the story. Gasoline is the price everyone sees and everyone budgets around. Diesel is the price that moves freight, agriculture, and shipping — the input cost baked into everything that arrives on a truck, which is to say nearly everything — and it's the one running hotter than the last genuine energy crisis, quietly, without a headline of its own. You already saw its fingerprints in two other sections of this issue: it's a direct line item behind why airlines are protecting margin over filling seats, and it's a direct line item behind the fertilizer and input costs squeezing farmers at harvest.

The trucking industry is where the price stops being an abstraction and starts being a bankruptcy filing. At $6.39 a gallon — up roughly 70% from $3.71 a year ago, after touching $6.53 at the recent peak — fuel now runs an estimated $3,000 per truck on a single cross-country haul. Eight large trucking firms filed for Chapter 11 in September alone, including Xoco Transport, which posted a $609,000 net loss against just $2.2 million in assets and $3.3 million in liabilities over the first seven months of the year — a debt load no freight rate at today's diesel price can service. Layoffs are stacking on top of the bankruptcies: roughly 2,000 jobs cut in recent weeks across delivery, manufacturing, and packaging firms tied to freight, including 230 positions at 4XH Logistics, an Amazon delivery partner in Texas, and 106 more at California-based Capstone Delivery, effective November 21. As Harvard Business School's Willy Shih put it, plainly: "If you think that isn't going to get passed on to consumers, think again, because nobody else along the chain can afford to eat that increase in costs." That's the same input cost already showing up in airfares and farm input bills elsewhere in this issue, now showing up as the freight sector's own layoffs and bankruptcy filings.

The Baltic Dry Index tells a related but more complicated story. It's trading north of 3,500 — the strongest sustained run since before COVID. Some of that is disruption: the Strait of Hormuz situation this issue has already covered, plus congestion at the Panama Canal. But unlike earlier this year, the rally isn't purely a disruption artifact anymore — genuine demand is underneath it too: West African iron ore shipments now exceeding South African volumes, Chinese port inventory building, and North American grain loadings up 23% year-over-year as harvest season ramps. That's a real signal mixed with a geopolitical one, which is exactly the kind of muddied bellwether this issue's own thesis warns about: imperfect data, still worth reading, as long as you're honest about which part of it you can trust. The swells in the shipping lanes are getting larger either way — the only open question is how much of that is wind, and how much is current.

Every number in this issue eventually reduces to one question: what does a dollar actually buy, right now, compared to what it used to. Here's the formula for answering that honestly.

In numbers: on-highway diesel set two weekly records in two weeks this September — $5.967 a gallon, then $6.527 — both breaking the prior high of $5.810 set in June 2022.

Every rate you see quoted — a savings account APY, a bond yield, a CD rate — is a nominal rate. It tells you how many more dollars you'll have. It says nothing about how many more groceries, gallons of gas, or square feet of housing those dollars will actually buy. The gap between those two numbers has a name, and a formula simple enough to do in your head.

The Fisher equation, named for economist Irving Fisher, says it plainly: real rate approximately nominal rate minus inflation rate. A savings account paying 4% nominal interest, sitting in an economy running 5% inflation, isn't growing your money. It's losing you about 1% of purchasing power a year, even as the account balance ticks up every month and looks, on paper, like progress.

Here's the grocery-cart version, purely illustrative: say a cart of groceries costs $100 today, and grocery-specific inflation runs 6% this year — a stated assumption for this example, not a claim about the actual current rate. That same cart costs $106 in twelve months. If your savings account paid 4% nominal interest over that same year, $100 in the account grew to $104. You have more dollars. You can afford less cart. The account isn't broken and the bank didn't lie to you — the nominal number was always honest about dollars. It just was never designed to tell you about groceries.

Now apply the same math to where the bond market actually sits today, because this isn't hypothetical anymore. The 10-year Treasury yield reached 5.34% on October 1 — a 24-year high, covered earlier in this issue. The University of Michigan's survey of year-ahead inflation expectations came in at 4.6% in September, up from 4% in August. Run the Fisher equation on those two real, current numbers: 5.34% minus 4.6% approximately 0.7%. That's the real return an investor is being offered for locking up money in the single safest, most liquid asset in the world for ten years — a genuinely thin margin, and one that shrinks further the moment inflation expectations tick up again, which is exactly the scenario the Fed is trying to prevent with the hike this issue already covered.

This is why "the Fed is raising rates" and "your money is finally earning something" aren't the same sentence, no matter how often they get said together — a rising tide doesn't lift a boat that's taking on water faster than the tide is coming in. A higher nominal rate only helps if it's rising faster than the prices it has to outrun. When it isn't — when a savings account, a CD, or a Treasury note pays a rate that barely clears expected inflation, or doesn't clear it at all — you can be earning interest and still losing ground, and the monthly statement will never once tell you so. The number to ask about isn't "what's the rate." It's "what's the rate, minus what things are actually going to cost."

That's the last instrument this issue has to offer. What's left is what you do with the reading.

In numbers: a 5.34% nominal 10-year yield, minus 4.60% expected inflation, leaves a real return of about 0.70%.

There used to be a commercial , Ames Home Loans, as it turns out, memory confirmed , that I must have seen a dozen times growing up. A man lying back in an inflatable inner tube, adrift on open water, a shark circling him in slow, unhurried laps while some version of the Jaws theme played underneath. Storm clouds stacking up on the horizon behind him, dark and moving in. And in the middle of all of it, he calmly pulls out the Yellow Pages and looks up Ames Home Loans , at which point a ship arrives to rescue him. Even as a kid, before I understood a word of what the ad was actually selling, I remember the feeling of it: something is coming, he is already out past where it's safe, and somehow he is still calm enough to go looking for help in a phone book.

Watching it back now, it's a genuinely strange commercial to have lodged itself in a kid's memory this precisely , a shark, a storm, and a phone book, all in service of a home loan , not an insurance policy, a loan against the equity already sitting in someone's house. But the strangeness was never really the point of watching him out there. The point was always what he'd chosen to do, or not do, before he ended up in the water with a shark in the first place, and that he didn't panic once he was in it.

That's the distinction this issue has tried to hold onto. None of the data in these pages is clean. The jobs numbers get revised. The Fed's next move is a probability, not a certainty. The Red Sea and the Strait of Hormuz are being fought over in real time as this goes to print, and nobody , us included , knows exactly how that resolves. A hundred-basis-point hike would hurt in ways that are almost certain, and the alternative isn't painless either, just slower. None of that is a forecast anyone can be fully confident in.

But the man in the inner tube didn't need a perfect forecast. He needed to notice the barometer was already pinned to "storm" long before he ever got in the water, and he needed to know who to call once he was in it. Imperfect data is still actionable data , that was the whole thesis of this issue, stated at the top and worth repeating at the bottom. You don't get to wait for certainty. You get to decide what you do with what you can already see, and when.

The clouds are visible. Go find your umbrella.