The headline number
We matched all 3,221 counties across the two most recent same-month readings a year apart: July 2025 and July 2026. In both months, 1,610 counties sat in our lowest band, 806 in the next, 483 in HIGH, and 322 in CRITICAL — 805 counties, exactly a quarter, in the top two bands both times.
But the counties inside that quarter are not the same counties. Over the year, 199 counties that were LOW or MODERATE in July 2025 climbed into HIGH or CRITICAL by July 2026. Separately, 199 counties that were HIGH or CRITICAL a year earlier fell back to LOW or MODERATE. The topline number is frozen by how the score is built. The membership underneath it is not.
Why that is exactly symmetric, and what it means
Our score ranks every county against every other county in the same month, so the size of each band is fixed by construction — the same point we make whenever we discuss the score at all. A county can only enter the top quarter by displacing one that leaves it. That is why 199 in and 199 out is not a coincidence; it is close to a mathematical certainty for a ranking built this way.
The two pools do not churn at the same rate, though. Of the 2,416 counties outside the top two bands in July 2025, 199 — about 8 percent — crossed up within the year. Of the 805 counties inside the top two bands, 199 — nearly 25 percent — dropped out. A county already flagged HIGH or CRITICAL is roughly three times more likely to move off that list within a year than an unflagged county is to move onto it. The flagged group is the less stable one, not the more stable one.
Two counties, moving in opposite directions
Greeley County, Nebraska, was LOW in July 2025 at the 42nd percentile and CRITICAL in July 2026 at the 93rd — even though its unemployment rate fell, from 3.1 percent to 2.4 percent, over that same year. Treasure County, Montana, ran the other way: CRITICAL in July 2025 at the 91st percentile, LOW in July 2026 at the 15th, while its unemployment rate rose slightly, from 1.7 percent to 2.1 percent. Neither county carries unusual seasonal swings in our data, so neither move is a seasonal artifact — both are real changes in how each county's trend, not its level, compares to everyone else's that month. We also checked that this specific comparison, July to July, does not straddle an earlier point this year where our own scoring configuration changed; it does not, and the two months rank consistently with each other.
What this is useful for
If you track a defined list of counties — a loan book, a service territory, a set of client sites — the useful question is not "how many counties are HIGH or CRITICAL nationally," which never changes much. It is "did any of my counties cross that line in the last year, in either direction." That is a real signal about your specific exposure, built from the same table, and it survives the fact that the national band sizes are fixed.
What it is still not
This is still a relative ranking of one month against counties scored the same way, and the newest month it can describe is July 2026 — our county data runs on the usual unemployment-reporting lag, so by the time you read this the gap is roughly three and a half months. A county that crossed into CRITICAL by July has had that long to keep moving, for better or worse, before this number updates again.