March 6, 2026
I'm sitting here at 8:29 a.m. ET, watching a blinking cursor on a Bloomberg terminal, and I want you to understand the particular psychic weight of the moment before a nonfarm payrolls number drops when crude oil is up 14% in a week, the Dow shed 785 points yesterday, Goldman Sachs is down nearly 4% on the session, and Iran — Iran — reportedly has enough drones to disrupt Persian Gulf shipping for months.
8:30 hits.
February payrolls: 50,000. Unemployment: 4.3%. Steady as she goes.
And somewhere, a macro strategist exhales. An algo buys the S&P. A talking head on financial television calls it "resilient." The chyron reads: JOBS MARKET HOLDS FIRM AMID IRAN CONFLICT.
I stare at this and feel something between exhaustion and dark laughter.
Let's do this properly, because the February payrolls number deserves an honest autopsy before it gets claimed by whatever narrative is most convenient today.
The headline of 50,000 is, in isolation, weak. It's consensus, sure — Dow Jones surveyed economists expected exactly that — but consensus for 50,000 in February was already a downgrade from January's surprise 130,000 print. And January only looked strong because it was measured against December's catastrophic 48,000, which itself was a downward revision from a number that was already bad. Zoom out and the full 2025 picture got quietly annihilated in revisions: the entire year came in at 181,000 total jobs, an average of just 15,000 per month, less than a third of what we thought we were adding. Fifteen thousand jobs a month. That's not a labor market. That's a labor market on a drip.
The sectoral story is worse. Strip out health care and social assistance, and the already-thin monthly averages of 2025 would have essentially gone to zero. We are, functionally, a jobs economy powered by one industry. And now, a Kaiser Permanente strike — 31,000 workers in California and Hawaii, during the BLS survey week — may have further dragged the February health-care numbers, which is the one engine we actually had running.
Bank of America modeled a below-consensus print of 35,000 specifically because of this strike effect. Instead we got 50,000. Which means either the Kaiser disruption was smaller than feared, or something else came in hotter to offset it. We'll find out in next month's revisions. We always find out in next month's revisions.
Meanwhile, in the same 48-hour window that produced this jobs number:
Hundreds of ships remain stuck in the Persian Gulf despite U.S. Navy promises to escort oil tankers through the Strait of Hormuz. The navy is now a shipping concierge service. Remarkable times.
Crude climbed 3% on Thursday to nearly $77 per barrel — up close to 14% from a week ago. The inflation math on that is straightforward and brutal. Oil at these levels doesn't just hit gas prices; it hits freight, it hits plastics, it hits the margins of every company that moves physical goods, which is most companies.
The CME FedWatch Tool is now pricing less than 3% odds of a rate cut this month. Futures are penciling in only a one-in-three chance of any cut by June. The base case has quietly drifted to September. September. Eight months from now. With oil up 14%, manufacturing shedding jobs, and the one sector keeping the payroll number afloat being hospitals.
China, also this week, announced its 2026 GDP growth target at 4.5% to 5% — the lowest target since at least the 1990s. The country that was supposed to be the global demand engine is publicly downgrading its ambitions. Quietly, structurally, with a bureaucratic calm that should alarm anyone paying attention.
And Broadcom. Broadcom posted strong earnings — one of the few tech names that actually delivered — and the stock reaction was telling: investors are rotating away from cyclical industrial and materials plays toward anything with a durable earnings story, as Caterpillar fell 3.6% and GE Aerospace dropped 3.4%. The market is quietly repricing who survives a stagflationary supply shock and who doesn't. Hint: it's not the companies that need cheap energy and unconstrained supply chains.
The Block news deserves more attention than it got.
Jack Dorsey announced this week that Block would slash roughly 40% of its payroll in response to AI adoption. Forty percent. Not a restructuring. Not a "right-sizing." Forty percent of humans, replaced or rendered redundant by software. And the number that showed up in payrolls data this morning doesn't fully capture it yet, because layoffs take time to process, and the BLS survey week is a snapshot, not a film.
This is the part of the jobs conversation that nobody in financial media wants to sit with for very long, because it complicates the narrative. The "stable" labor market — the low-hire, low-fire equilibrium that economists keep describing with a kind of relieved shrug — is stable the way a patient is stable when they're not getting worse. Stability is not health. And AI-driven displacement, at scale, in white-collar sectors, arriving quietly beneath a headline number of 50,000, is not a jobs market trend. It's a structural shift in how labor works.
The irony, of course, is that the Fed is on hold precisely because the labor market looks "okay." And the labor market looks okay precisely because the measurement tools are lagged, sector-narrow, and revision-prone. By the time the data catches up with reality, we'll be in a different movie entirely.
The market closed the week in a peculiar kind of limbo.
The S&P 500 is still up nearly 19% year-over-year. Oil is pricing in a war premium that hasn't fully passed through to CPI yet. The Fed is frozen, not because it lacks tools, but because every tool it has is wrong for the current environment — cutting would pour fuel on an oil-driven inflation fire, holding risks breaking a labor market that is already quietly hollowing out.
The Kospi rebounded Thursday after its circuit-breaker drama earlier in the week. Gold is above $5,300. The 10-year Treasury yield is elevated. The dollar isn't panicking but it's not celebrating either.
And tomorrow, or next week, or the week after, there will be another data point, another war update, another Fed speaker threading a needle with language carefully calibrated to mean as little as possible.
Here is what I actually think, for whatever that's worth on a Friday:
The jobs market is not about to collapse. But it is being hollowed out from underneath in ways that aggregate payroll numbers structurally cannot capture. The Fed knows this and cannot say it. The Treasury knows this and won't say it. The sell-side knows this and will publish a note saying "robust underlying demand" because their compliance team won't clear the alternative.
Oil at $77 on the back of a shooting war in the Gulf, with a Fed that can't cut, a labor market powered by hospitals and not much else, and a technology displacement cycle that is barely visible in the data yet — this is the actual picture. Not a crisis. Not a melt-up. A slow compression, playing out across a dozen data sets that each look fine individually and alarming in aggregate.
The payrolls number today was 50,000. Fine. The doctor said your blood pressure is normal.
The building is on fire, but your blood pressure is normal.