Visualizing How U.S. County Incomes Have Shifted, 1969–2024

Purpose

These two charts answer a pair of questions that sound identical and are not. What happened to America's places? And what happened to Americans?

Both charts show the same thing on their face: every U.S. county from 1969 through 2024, grouped into income-per-resident brackets, in 2024 dollars, with a play button that runs the 55 years like a film. The difference is what the bars count. In the first chart, every county counts once, whether 48 people live there or 9.7 million. In the second, each county counts by its population, so the bars show how many people live in counties at each income level. Neither view is more correct. They answer different questions, and the gap between them is real information. The first chart is a map of places. The second is a census of people standing on that map.

What this page is

Every chart we publish should be something you can check, question, and rebuild yourself. These two are close to the simplest case: one public federal file, one public price index, and one division. There is no model, no estimation, and no editorial math beyond converting old dollars to 2024 dollars and sorting counties into $10,000 buckets. Where the government published a number, we use that number. This page lays out exactly which file we used, the two places we compute anything, the choices we made about which counties to show and how to draw the top of the distribution, and how the whole thing was checked before it went up.

The data source

Everything comes from one file: the Bureau of Economic Analysis Local Area Personal Income table CAINC4, Personal Income and Employment by Major Component, from the February 2026 release, which covers every U.S. county from 1969 through 2024. CAINC4 is the federal government's primary accounting of income by county. We use two lines of it: total personal income (line 10) and population (line 20).

Personal income is income from all sources: wages and salaries, investment income (dividends, interest, and rent), and government transfers like Social Security. It is a residence-based measure, meaning income is counted where the person lives, not where they work. One thing it does not include, by design, is capital gains. Profits from selling stock or property never appear in these figures.

The price index is the BLS Consumer Price Index for All Urban Consumers, series CUUR0000SA0, U.S. city average, all items, annual averages. We pulled the full 1969 to 2024 series directly from BLS.

Step 1: The two calculations

There are exactly two places we compute anything.

First, income per resident. For each county in each year, we divide the county's total personal income by its population, both straight from the BEA file. BEA publishes its own per capita personal income line, and we verified our division against it: for Teton County, Wyoming in 2024, our figure and BEA's published line 30 agree to the dollar ($532,903), and both match the Wyoming Economic Analysis Division's independent 2024 report.

Second, the inflation adjustment. Every year's figure is converted to 2024 dollars by multiplying it by the ratio of the 2024 CPI-U annual average to that year's annual average. This is the standard constant-dollar conversion. One property of it is worth knowing: because the same deflator applies to every county in a given year, inflation adjustment never changes which bracket a county falls into relative to other counties in that year. It rescales the whole distribution; it cannot reorder it.

That is the entire computational content of both charts. Everything else is sorting.

Step 2: Which counties are shown

Every one of them. All 3,115 counties and county equivalents BEA reports, with no minimum population and no exclusions.

That deserves a note, because our Jackson Hole piece did the opposite. There we dropped counties under 10,000 residents, for a specific reason: that chart was a ranking of the top 15, and in a very small county, income per resident swings wildly from year to year. A handful of unusual tax situations can move the whole county average. Loving County, Texas, population 48, is the standard example. In a ranking, a county like that takes a slot away from a real story.

A histogram has no slots to take. A small county is one count in a bucket, or a few dozen people in the population chart. It cannot crowd anything out. And filtering it would quietly assert that those places do not exist, which is not true and not our call to make. The small counties are also disproportionately the poorest ones, so dropping them would trim the left side of the distribution as well as the right. Everything is in.

This means the two pages use different universes on purpose, and the difference shows up in the numbers. In 2024, 15 counties sit above $150,000 per resident here; the Jackson Hole piece counts 11, because four of the 15 are small counties it excluded.

Step 3: The brackets and the box

The main chart groups counties into $10,000 ranges of income per resident, running from $10,000 to $150,000. The box on the right holds every county above $150,000, split into three tiers: $150,000 to $200,000, $200,000 to $300,000, and $300,000 and up.

The box uses its own magnified scale, and we want to be direct about why. On the main chart's scale, which has to reach 1,815 counties (or 116 million people) to hold the 1969 distribution, the counties above $150,000 would be invisible: there have never been more than 15 of them. The box gives that group its own axis, printed on the chart, so its growth can be seen at all. Both scales are labeled where they apply. Nothing is hidden, but the two panels should not be compared by eye against each other's heights, only against their own printed axes.

The median markers

Each chart carries a marker showing its own version of the middle, and the two markers measure different things.

In the county chart, the marker is the median county: half of all counties sit above it, half below. In 2024 that county sits at $56,066 per resident.

In the population chart, the marker is population-weighted: it shows the income per resident of the county where the middle American lives. Line up all 330-odd million people by their county's income level, walk to the halfway point, and read off that county's figure. In 2024 that is $68,273.

The population-weighted figure runs about $12,000 higher, and that gap is not an error in either number. It says something true: Americans disproportionately live in higher-income metro counties, while thousands of lower-income rural counties hold comparatively few people. The two medians answer the two questions this page opened with.

How we built and checked it

These charts were produced through a human-directed, AI-assisted process. An editor set the questions, the design, and the choices described above, and reviewed the work at each stage. An AI assistant did the data restructuring and built the interactive charts, and ran the validation checks alongside the editor's own review.

The checks were these. The per-resident division was verified against BEA's own published per capita line and against the State of Wyoming's independent report, with exact agreement. The full 56-year CPI-U series used for inflation adjustment was compared value by value against the series as downloaded directly from BLS, all 56 annual averages, with zero discrepancies. And the sum of the income components in the BEA file was checked against the published total for every county and every year in the panel, roughly 174,000 county-year rows, with 100% agreement within rounding.

The BEA and BLS files remain the authoritative references. If you find a number here that disagrees with the source, the source wins and we want to hear about it.

How to use the charts

Press play and the years run forward from 1969; press again to pause. The slider scrubs to any year directly, and the 1x, 2x, and 3x buttons set the playback speed. Labels above each bar show the exact count (of counties or of people) in that bracket, and the box on the right labels its tiers and its total. In the county chart, when the $300,000-and-up tier appears, it is exactly one county, and we label it by name.

Honest notes and limitations

We'd rather tell you the edges of this than have you find them.

Income per resident is an average within each county. A county appearing in a high bracket does not mean its typical household earns that much. Averages can be pulled far upward by a small number of very large incomes, and the most extreme county on these charts is the most extreme example of exactly that, as our Jackson Hole piece lays out in detail.

Small counties are volatile, and they are in here. A county of a few hundred people can jump several brackets in a single year on the strength of one unusual year of income. We include them because they are real, but a bar that twitches at the far right of the county chart is often one tiny county having an eventful year, not a trend. The population chart is immune to this by construction: those counties carry almost no people.

County definitions shift under a 55-year window. The most visible case is Connecticut, where BEA now reports planning regions as county equivalents. We use BEA's geography as published. A few boundary and definition changes over the decades are inherited from the source file the same way.

BEA revises. Local area personal income estimates are updated as better source data arrives, so the February 2026 vintage we used will not match earlier or later vintages exactly. Small differences against numbers you've seen elsewhere are usually this.

Personal income is broader than a paycheck. It includes investment income and government transfers, and some imputed items BEA counts by convention, such as the rental value of owner-occupied homes. It excludes capital gains entirely. It is the federal government's standard measure of income, but it is not the same thing as median household income from the Census, and the two should not be compared directly.