When people search "graph us unemployment," they're usually looking for one of two things: a visual picture of how unemployment has moved over time, or a way to make sense of a number they just heard in the news. This article explains what those graphs show, where the data comes from, how to read it, and what the numbers actually mean.
The unemployment rate you see in headlines and on charts is produced by the Bureau of Labor Statistics (BLS), a federal agency within the U.S. Department of Labor. Each month, the BLS conducts the Current Population Survey (CPS) — a household survey of roughly 60,000 households — to estimate how many Americans are unemployed.
That number is then divided by the civilian labor force (everyone either working or actively looking for work) to produce the unemployment rate. The result is published as a percentage, updated monthly, and revised as more data becomes available.
The BLS also publishes historical unemployment data going back to 1948, which is what most graphs draw from.
A standard historical unemployment graph plots the monthly rate on the vertical axis and time on the horizontal axis. A few patterns stand out immediately:
Some of the most prominent spikes visible in historical graphs include:
| Time Period | Approximate Peak Rate | Primary Cause |
|---|---|---|
| Early 1980s | ~10.8% | Federal Reserve rate hikes / recession |
| 2008–2010 | ~10.0% | Financial crisis / Great Recession |
| April 2020 | ~14.7% | COVID-19 pandemic |
| Early 1990s | ~7.8% | Gulf War-era recession |
| Early 1970s | ~9.0% | Oil embargo / stagflation |
These figures come from BLS historical records and reflect the U-3 unemployment rate, which is the most commonly cited measure.
The rate most often graphed is called U-3, and it counts people who are jobless, available for work, and have actively looked for a job in the past four weeks. But the BLS publishes six different measures, labeled U-1 through U-6:
U-6 is always higher than U-3. During the Great Recession, for example, U-6 peaked near 17%, even while the official U-3 rate sat around 10%. Graphs that use U-6 will look significantly worse than those using U-3, even covering the same time period. When comparing graphs from different sources, checking which measure is being displayed matters.
The BLS publishes interactive charts at bls.gov, including the FRED database (Federal Reserve Bank of St. Louis), which is one of the most widely used tools for graphing unemployment data. FRED allows users to:
News outlets, academic institutions, and policy organizations also produce their own charts drawing from BLS data. The underlying numbers are the same — the presentation and framing vary.
Raw unemployment rate graphs have real limitations. They don't show:
State unemployment rates are published separately by the BLS through the Local Area Unemployment Statistics (LAUS) program. A national graph may show a rate of 4%, while individual states sit anywhere from 2% to 7% or higher.
The unemployment rate and unemployment insurance (UI) claims are related but distinct measurements. The unemployment rate comes from a household survey. UI claims — initial claims filed each week — come from state agency administrative data and are reported separately by the Department of Labor.
Initial claims (people filing for benefits for the first time) and continuing claims (people still receiving benefits week over week) are tracked as leading economic indicators. During the COVID-19 spike in spring 2020, initial claims reached levels never previously recorded — over 6 million in a single week. That context is visible in graphs of UI claims but not always in the headline unemployment rate, which lags by weeks.
Unemployment graphs are tools for understanding economic conditions broadly. What they reflect about any individual's situation — whether someone qualifies for benefits, how much they'd receive, how long coverage might last — depends entirely on the state where they worked, their earnings history, and the specific circumstances of their job separation. National averages don't translate directly into individual eligibility, and historical trends don't predict individual outcomes.
The data is publicly available, consistent, and worth understanding. What it means for a specific person's claim is a separate question entirely.