4.3% and other beautiful lies

Once a month, on the first Friday, a number falls out of the sky and lands on every newspaper in America. In May of 2026 the number was 4.3%. The economy had made 172,000 new jobs. A man who worked at a bank told a television station it was a strong report from every angle. The Federal Reserve nodded its enormous head. Bond yields went up. The number had done what it was built to do, which was to tell everybody a story before they’d finished their coffee.

If you happen to live inside this economy, and if you’ve been laid off, or you drive for an app to keep the lights on, or you watched the rent swallow whatever was left, then 4.3% reads like a postcard from a country you have never visited. You aren’t crazy. The number is real. It’s just measuring something much smaller than you think it is, and the something it measures has been quietly drifting away from your actual life for about eighty years now.

Let me explain.

What the government means when it says “unemployed”

The famous number, the one in the headlines, is called U-3. To be counted in it you have to be out of work, ready to work, and you have to have gone looking for work in the last four weeks. That’s all. That definition has barely budged since 1940, and the Bureau of Labor Statistics will tell you so proudly. They say the official idea of unemployment has been thoroughly reviewed and validated since the Current Population Survey began in 1940.

Think about what that means. The machine for counting joblessness was built for a world of factory whistles and one company you worked at until you died. It was built before delivery apps, before everybody became an independent contractor, before “I’ve got three part-time things and not one of them comes with health insurance” turned into an ordinary sentence a person could say out loud. The definition isn’t broken, exactly. It just keeps answering a 1940 question, and we keep asking it a 2026 one.

Fun facts: a couple jobs that existed in 1940 that don’t exist today include a “Pinsetter.” This is a job held by a child at a bowling alley and they manually set the pins up after they fell over. Another is a “knocker up,” and these people acted as human alarm clocks and would go wake people up. No joke.

And here’s the part that ought to make you laugh, or maybe sit down. The government already knows U-3 is too small. That’s why it quietly publishes five other unemployment rates, U-1 through U-6, each one a little bigger, each one catching a little more of the misery. The biggest of them, U-6, scoops up the people who want a job and looked this past year but not this past month, and the people who’ve given up because they figure there’s nothing out there, and the people stuck working part-time who’d give anything for full-time. In May of 2026, U-6 was 8.1%. Nearly double the headline. The government makes both numbers. It just hands the microphone to the little one.

Which jobs, not how many

Edited section: Thank you for the idea!

Now, suppose you ignore all that and take the latest number at face value. May of 2026: 172,000 jobs, beating what the experts expected, and one fellow at an investment bank actually called it a “Payroll Blowout!” with an exclamation point, which is the kind of thing people say right before they’re wrong.

But, look at where the jobs came from, because that’s the whole story. Of those 172,000, about 70,000 were in leisure and hospitality, roughly 5X the usual monthly haul, almost certainly a one-time blip from hiring for the World Cup, which is a thing that happens to a country once and then leaves. Another 55,000 were local government, and some 35,000 were health care. Twelve thousand more were social assistance, which is to say people hired to help other people get by. Add those up and you’ve found nearly all of it.

And, now look at the other ledger, the one the headline skips. Financial activities lost 22,000 jobs in May, and they’re down more than 100,000 since this time last year. Transportation and warehousing have shed 92,000 since early 2025. Construction, manufacturing, retail, information, professional and business services, the entire white-collar, market-driven middle of the economy, “showed little change,” which is the government’s gentle way of saying nothing happened.

So, here is the actual shape of it: if you’re a home health aide, or you work for the county, or you got hired to pour beer at a soccer match, the economy is hiring. If you work in an office, or finance, or you move things around the country for a living, the economy has quietly closed for business and forgotten to flip the sign on the door. These are two different countries living inside one decimal point. The 4.3% is the average of a person who’s drowning and a person who’s fine, and the average, as always, is a man with his head in the oven and his feet in the freezer who reports feeling roughly comfortable.

The people who fall through the floor

The ones who’ve been out a long time

If you lose your job and you still can’t find one half a year later, the statistics start to lose interest in you. You stay in the headline number only as long as you keep actively looking. The longer you’re out, the more likely you slip into “discouraged,” or “marginally attached,” and slide right off the edge of the famous number.

This is not some tiny pocket of people. In May of 2026, two million Americans had been out of work for 27 weeks or longer, and that crowd had grown by 524,000 in a single year. They were more than one in four of all the unemployed. One in four, in an economy the headline keeps calling strong. When their job hunts run out of gas, the data doesn’t hire them back. They just stop existing, statistically speaking, which is a strange and terrible way to stop existing.

The gig workers

This is where the little gap becomes a canyon. The surveys that cough up the jobs numbers were designed to count payroll jobs, and the gig workers keep falling out the bottom: the independent contractors, the 1099 people, the ones who work for an app. Researchers have found that leaning on the household survey shoves a lot of working gig workers into the wrong box entirely, counting them as unemployed or as not in the labor force at all.

And the size of it is enough to make you dizzy. The Federal Reserve has figures suggesting more than 7 million workers might be missing from the monthly count, and by some estimates the real number could be off by as many as 13.2 million human beings. ADP looked at the payroll data of more than 1.1 million employers and found that people getting a short-term W-2 or a 1099 made up 27% of all the jobs held in 2024, and that by the end of the year, fully one in four workers had done some kind of gig work.

Now, whatever you make of that kind of work, the headline rate will never tell you the cruelest part: most of these folks have no net under them at all. When the gig money dries up, the typical gig worker can’t even file for unemployment, because as an independent contractor he was written out of the system on purpose. No unemployment check. No workers’ comp. Usually no health insurance because only about 40% of gig workers have any, against 82% of full-time employees. The headline counts a flexible, modern labor market. It doesn’t count the floor quietly disappearing underneath it.

And if you do qualify? The check won’t save you.

Say you’re one of the lucky ones the system actually sees. A regular employee, laid off, eligible. Here is what’s waiting for you.

Across the country, unemployment insurance replaces less than 40% of what you used to make. In the spring of 2025 the average weekly check was $451 and as low as $221 in Mississippi, and as high as $738 in Washington. Most states are aiming to give you back somewhere between 30 and 50 cents on your old dollar.

Now, set that next to the plain cost of being alive. The average American household spends around $6,545 a month, about $78,535 a year, just to live, according to the government’s own spending data. The average one-bedroom apartment rents for around $1,506 a month. A $451 weekly check comes out to roughly $1,950 a month, before one single other bill. The arithmetic doesn’t work, and it was never meant to. The Economic Policy Institute says it flat out: in no state are unemployment benefits enough to cover a worker’s basic needs. The check was never built to let you live. It was built to slow the bleeding while you run around looking for work. Most people collecting it already know this in their bones. The statistics just won’t say it out loud.

And it gets thinner. Only about 29% of unemployed people actually got benefits in 2023. So the famous safety net reaches a minority of the people who lose their jobs, hands them less than half their income when it reaches them at all, and pays out a sum that covers a sliver of what it actually costs to exist. Three filters, each one shrinking the help.

The crisis the number can’t see

And here is the reason the number and the mood have drifted so far apart: you can be employed, completely, officially, headline-certified employed, and still be going under. The unemployment rate was never built to notice that, and these days, that’s where most of the suffering lives.

Heading into 2026, roughly two out of three Americans were living paycheck to paycheck. For the first time in three years, prices were climbing faster than paychecks: between April of 2025 and April of 2026, wages rose 3.6% and prices rose 3.8%. Groceries are up about 30% since January of 2020. In surveys, 92% of Americans said they cut back their spending in 2025, including on food and medicine, and nearly half said they’d dipped into savings just to get through. The share of people who’d skipped a meal in the past year climbed from one in four to more than one in three, and it did that in a matter of months.

One analysis in TIME put it about as plainly as it can be put: America’s cost-of-living crisis is really a pay crisis. For something like 45 years, wages for most people have failed to keep up with the country getting richer. A “4.3% unemployment rate” has nothing whatsoever to say about any of that. You can have a job and still not afford the life the job was supposed to buy you. The number measures whether you’re hooked up to employment. It was never built to ask whether the employment is enough.

What to do with the number

None of this means anybody’s cooking the books. The truth is duller and meaner than that. It’s an honest, accurate answer to a 1940 question, read aloud as though it answered a 2026 one. The official rate is exact about a definition that has quietly stopped describing how Americans work, earn, and go without.

So the next time that one number flashes across your screen with its tidy little verdict already stapled to it, treat it as the start of a question instead of the end of one. Ask about U-6, sitting there at 8.1%. Ask about the two million people who’ve been out of work so long the count gives up on them. Ask about the millions of gig workers the surveys can’t find and the net won’t catch. Ask whether the check, for the few who get one, pays for a single month of a real human life.

The headline says the economy is strong. An awful lot of people, doing the actual math at their actual kitchen tables, have decided the headline isn’t about them. And the funny thing, the thing that ought to make you set down your coffee: if you read past that first number, the government’s own data agrees with them.

That’s all.

Sources: U.S. Bureau of Labor Statistics (Employment Situation, May 2026; Alternative Measures of Labor Underutilization); Center for American Progress; Federal Reserve Banks of Boston and Minneapolis; The Hill; ADP Research; Economic Policy Institute; National Employment Law Project; Peter G. Peterson Foundation; Motley Fool Money / BLS Consumer Expenditure Survey; TIME; The Century Foundation; Bloomberg. Figures reflect the most recent data available as of June 2026.