One country, many job markets

Four point two was nobody’s rate

The United States unemployment rate was 4.2 percent in June 2026. South Dakota’s was 2.0 percent. The District of Columbia’s was 6.0 percent. Same month, same release, three times the distance between the two ends.

Those figures come from the Bureau of Labor Statistics State Employment and Unemployment release published on July 21. Seventeen states sat below the national rate, eight states and the District sat above it, and 25 fell in between. The job market by state is not one market reported at different volumes. It is a set of separate markets that share a headline.

The year pulled states apart

Payrolls barely moved in the month. Nonfarm employment rose in 3 states in June, fell in 1, and was essentially unchanged across the other 46 and the District.

The twelve months behind it tell a different story. Thirteen states carried jobless rates higher than a year earlier, the largest increases in Connecticut at 1.3 percentage points and Oklahoma at 1.0. Seven states improved, Ohio by a full point. For 30 states and the District, the rate was not notably different from where it started. One average laid over that mix is a number almost nobody lives in.

Metro lines cut finer than state lines

Go one level down and the spread widens again. Of the 387 metropolitan areas the Bureau tracks, 226 had June jobless rates below the national figure of 4.4 percent and 145 had rates above it. That 4.4 is the unadjusted rate, which is the only version metro data can be read against.

El Centro, California ran 17.6 percent. Among the 56 largest metro areas, Raleigh-Cary, North Carolina and Urban Honolulu, Hawaii tied for the lowest at 3.1 percent. Over the year, rates fell in 27 of those large areas and rose in 26, close to an even split. Hartford-West Hartford-East Hartford, Connecticut posted the largest increase at 1.3 points.

What a rate cannot tell you

An unemployment rate counts people, not openings. It can fall because hiring picked up or because the labor force shrank, and from outside the two look identical. Small-area estimates are modelled rather than counted directly, so a single metro reading carries more uncertainty than a state or national one.

The national backdrop is thin either way. Payrolls fell by 23,000 in July, and revisions took a combined 103,000 off May and June, in the Employment Situation released August 7.

Where the search actually runs

Where a search runs matters as much as what it is for, and nobody reads 387 markets by hand. CoBlack sources openings from validated employer career pages and ATS feeds across the United States and Canada, matches them against what a person has demonstrably done, and submits server-side.

The rate on the news belongs to the country. The one that decides a search belongs to a much smaller map.