What the Unemployment Rate Is, and Who It Doesn't Count
A headcount of people looking for work, taken from 60,000 households. It is the most watched number in the economy, and it leaves out everyone who stopped looking.
October 28, 2026 · updated October 29, 2026
In one sentence
The unemployment rate is the share of people who want a job, looked for one in the past four weeks and don't have one, out of everyone who either has a job or is looking.
It comes from a monthly survey of about 60,000 households, not from unemployment-benefit claims. You can be unemployed without ever filing for benefits, and you can be ineligible for benefits and still count. The two numbers move together, but they are different counts of different people.
New articles and charts arrive in one short email on Monday mornings. Get the weekly digest
The analogy
Picture a party, and you want to know how it's going, so you count the people at the door waiting to get in. That is the unemployment rate: not the people inside (employed), not the people who went home (out of the labor force), only the ones at the door, still trying.
The count is useful because the door is where the pressure is. But it only counts people in the line. If the wait is long enough that some give up and go home, the line gets shorter and the number improves, though nothing about the party did. If word gets out that the party is good, more people show up, the line grows, and the number looks worse. That is the whole trick to reading it: a change can mean the party changed, or that the line did.
How it reaches your kitchen table
Your raise: within a year. When the line is short, employers compete for workers and pay climbs, fastest for people who switch jobs. In 2022, with the rate at 3.5%, typical hourly pay rose about 5% a year. When the line is long, raises stall. This is the channel the Fed watches, because wage growth feeds prices.
Your hours and your job: now, if it's you. The rate is an average; for the person laid off it is 100%. The useful signal for everyone else is direction. The rate rarely rises a little: once its three-month average is half a point above its low of the past year, a recession has usually already begun. That happened within four months of the start of every recession since 1970, and twice without one, for a month in late 1976 and in 2024 (see What a recession is).
Benefits: by state formula. Extended unemployment benefits switch on when a state's own rate crosses set thresholds, so the state figure, not the national one, decides how long a check lasts.
Interest rates: through the Fed. A rising rate is the strongest argument for cuts; a falling one with inflation above target is the argument for hikes. The committee publishes the rate it considers normal (4.2% as of September 2026) and reads the gap.
Your kids and your neighbors: unevenly. The national rate hides a spread. Teenagers run about three times the national figure, workers without a high school diploma well above it, Black workers roughly twice the white rate in every cycle on record. A 4% headline means something different in each of those lines.
What moves it
Hiring minus separations, mostly. Layoffs raise it quickly; hiring lowers it slowly. Then the size of the line: graduates and immigrants arriving, retirees and the discouraged leaving. When the working-age population grows, the economy has to add jobs just to hold the rate steady. When it stops growing, as immigration slowed in 2025–26, the rate can hold near 4% with very little hiring, which is roughly where 2026 has been. Recessions do the rest.
How to read this month's number
Tenths are noise. The survey's margin of error is about two-tenths of a point. A move from 4.1 to 4.2 by itself means nothing; three such moves in the same direction mean something. Watch the three-month average.
Direction over level. Where it stands matters less than where it's going, and recessions start with the rate near its low. The Sahm indicator (the three-month average minus its twelve-month low) makes the direction visible; 0.5 has been the line, and it is on the chart as a companion.
Check the other gauges. U-6 adds people working part time who want full time and people who looked in the past year but not the past month; it runs three to four points above the headline. Prime-age employment-to-population, the share of 25-to-54-year-olds with a job, skips the question of who is "looking" entirely. If the rate falls and the prime-age share doesn't rise, people left the line. If both improve, the party got better. In October 2026: rate 4.1%, U-6 7.9%, prime-age employment 80.7%, a little over a point below its 2000 record.
Where the analogy breaks
A party has a door. The labor force doesn't. People drift between looking and not looking by the week, so the line is counted by asking a sample of households what they did lately, and the boundary is blurrier than a headcount suggests. One hour of paid work last week puts you inside the party. The rate also says nothing about the jobs inside, what they pay or whether they are full time, which is what U-6 and the wage lines are for.
What people argue about
How low is too low. Economists estimate a "natural" rate below which inflation accelerates; the CBO puts it near 4.4%, the Fed near 4.2%. In 2019 the rate sat at 3.5% for a year with inflation under 2%, which taught everyone the estimates are soft. One side says the concept is still essential, because the Fed has to aim at something. The other says a number that gets revised after every miss can't be aimed at, and aiming at it has cooled job markets that weren't overheating. Data can show where inflation accelerated last time; it can't show the number in advance.
Participation and immigration. The share of adults in the labor force has fallen since 2000, mostly from aging and partly from prime-age men leaving. One side reads that as hidden unemployment and prefers the prime-age employment ratio; the other says retirement and school are not unemployment. Immigration is the other lever: more workers raise the number of jobs needed to hold the rate steady, and also fill them. Both sides agree on the arithmetic and disagree on the policy, which the chart can't settle.
"The real rate is much higher." Claims that true unemployment is 15% or 25% usually count everyone without a job, including retirees, students and parents at home. The broadest official measure, U-6, is published every month and sits on the chart. It is higher than the headline and it moves with it.
See also
- The Fed funds rate: what the Fed does when this number moves, and how long it takes.
- Inflation: the other half of the Fed's mandate, and why the two pull against each other.
- The almanac's unemployment chart with every overlay, the Sahm view, and jobs added.
Sources
- BLS, How the Government Measures Unemployment, the definitions and the survey.
- BLS, Alternative measures of labor underutilization (Table A-15), U-1 through U-6.
- FRED: Unemployment rate (UNRATE), U-6 (U6RATE), Prime-age employment-population ratio (LNS12300060), Sahm rule (SAHMREALTIME), CBO natural rate (NROU).
- FOMC Summary of Economic Projections, September 2026, the longer-run estimate.
- Sahm, "Direct Stimulus Payments to Individuals" (Hamilton Project, 2019), where the rule was proposed; FRED: Real-time Sahm rule indicator (SAHMREALTIME), its record since 1959, including late 1976 and 2024.
Corrections
October 29, 2026: An earlier version said a half-point rise in the three-month average had meant a recession had begun every time since 1970. It also happened without a recession in late 1976 and in 2024.