The output is counted where it happens. The people are counted where they sleep

Four numbers describe the District of Columbia in 2025. Its gross domestic product per resident is $277,689. Its real GDP per resident, in chained 2017 dollars, is $210,045. Its per capita personal income is $116,121, and its midyear population is 693,645.

Every one of those comes from the Bureau of Economic Analysis. Every one is correct.

That is the difficulty. Search for GDP per capita by state and the first page of results carries three separate leaderboards, two different jurisdictions at the top, and figures for the same state more than thirty thousand dollars apart. An encyclopaedia table gives New York $123,369. A financial advisory page ranking on the same query gives New York $90,730. A federally funded data service ranking what a reader would take to be the same thing puts the District of Columbia first at $116,121 and Connecticut second at $98,879. A widely downloaded public dataset is denominated in chained 2009 dollars, a price base BEA has long since retired.

Checked against the underlying files, most of those pages are right to the dollar. The encyclopaedia table matches. So does the data service. They are answering different questions, in different price bases, from different vintages of the same federal source, and most of them do not say which.

The reconciliation is published. It sits in a line item BEA has carried for decades, and once you have read it the whole spread of numbers stops being confusing and starts being informative.

Which state has the highest GDP per capita?

Answer: New York, at $123,369 in 2025. The District of Columbia is higher at $277,689 but it is a federal district and not a state. Massachusetts follows New York at $114,635, then Washington at $111,860, Delaware at $110,588 and California at $108,012. The national figure is $90,004 per person.

The distinction between a state and a federal district sounds like pedantry until you see what it does to the answer. Google's own related-questions panel on this query asks which US state has the highest GDP per capita, and the pages it draws on mostly lead with the District of Columbia. One is a question about states. The other is a ranking headed by something that is not a state, and which, for reasons set out below, is the most distorted entry in the entire table.

At the other end, Mississippi produced $55,877 per resident, West Virginia $61,873 and Arkansas $63,703. The gap between the top state and the bottom state is a factor of 2.21. Put the District of Columbia back in and the spread across the fifty-one jurisdictions widens to a factor of 4.97.

Why does the District of Columbia read $277,689 per person?

Answer: Because gross domestic product is credited to the place where output is produced, and population is counted where people sleep. In the District those two places differ for a very large share of the workforce. BEA measures the gap directly: its adjustment for residence subtracted $63.7bn from $132.4bn of earnings by place of work in 2025, which is 48.1 percent.

The adjustment for residence is line 42 of BEA's state personal income table, and it is the most useful line in the regional accounts that almost nobody quotes. It exists for a plain accounting reason. Personal income is a resident measure and earnings are a workplace measure, so BEA has to move labour income across state borders to get from one to the other, and the sign of the adjustment tells you which way the commuters run.

Nearly half of the labour earnings generated inside the District of Columbia leaves it every evening. That money lands as income in Maryland and Virginia, both of which carry large positive adjustments, and it never touches the denominator of the District's per capita GDP. What is left is a statistic that describes an office district rather than a population.

The same mechanism, in smaller doses, runs up and down the table.

New York carries the largest negative adjustment in dollars anywhere in the country. New Jersey carries the largest positive one, and the two are substantially the same commute seen from opposite ends. A New Jersey resident working in Manhattan adds to New York's gross domestic product, adds to New Jersey's personal income, and appears in New Jersey's population.

None of that is a flaw in the statistic. GDP by state does exactly what it says, which is to measure production inside a border. The flaw is in reading a production measure as a description of how well off the people living there are.

Why does the ranking change when the measure is personal income?

Answer: Because personal income follows the resident and GDP follows the workplace, and the two diverge by more than most rankings admit. Connecticut is eighth on GDP per capita and second on per capita personal income. Delaware is fifth and twenty-ninth. Georgia is twenty-fifth and forty-second. Florida is thirty-fourth and eighteenth.

Commuting explains Connecticut, and it explains New York in reverse. It does not explain Delaware, whose residence adjustment is a modest negative 5.9 percent. Something else is going on there, and the industry table says what.

Finance and insurance produced 25.8 percent of Delaware's gross domestic product in 2025, against 7.9 percent nationally. Add real estate, rental and leasing and the combined group reaches 45.7 percent of a $117.2bn state economy. Value added booked inside a state does not have to become income of that state's residents. In Delaware a large share of it does not.

North Dakota reaches the same divergence from an unrelated direction. Mining, quarrying and oil and gas extraction produced 14.0 percent of its output against 1.2 percent nationally, and capital-intensive extraction generates a great deal of value added per worker with comparatively few workers on the ground. Wyoming runs the pattern backwards, ranking eighteenth on production and fifth on income, because dividends, interest and rent are counted where the owner lives rather than where the asset sits.

The District of Columbia is the clearest case of all. Government and government enterprises produced 30.1 percent of its output in 2025, against 11.2 percent nationally, and professional, scientific and technical services produced another 20.8 percent against 8.1 percent nationally. Half the economy of the District is the federal government and the firms that sell to it.

Why do published figures for the same state differ by thirty thousand dollars?

Answer: Three reasons, and they compound. Current-dollar GDP and real GDP in chained 2017 dollars are different series, and for New York in 2025 they read $123,369 and $94,702. Older pages use a retired price base. And BEA revises the whole history annually, most recently on 9 April 2026.

The price-base problem is worth understanding rather than memorising, because it changes what a rising number means.

Between 2021 and 2025, United States GDP per capita in current dollars rose from $71,441 to $90,004, an increase of 26.0 percent. Over exactly the same period, measured in chained 2017 dollars, it rose from $64,837 to $69,782, an increase of 7.6 percent. The economy did not grow twice. Roughly two thirds of the headline gain is the price level.

For the District of Columbia the divergence is starker. Its current-dollar figure went from $233,830 in 2021 to $277,689 in 2025, a rise of 18.8 percent. Its real figure went from $214,054 to $210,045, a fall of 1.9 percent. A reader looking only at the current-dollar series would conclude the District had a strong four years, while a reader looking at the volume series would conclude it produced slightly less per resident in 2025 than it did in 2021. The second reader is closer to right.

BEA attaches a further caution to the chained-dollar series which is easy to skip and worth restating. Because the chain formula uses weights from more than one period, chained-dollar estimates are usually not additive. Summing states in chained dollars does not give the national total.

Summing them in current dollars does not quite give it either. Adding all fifty states and the District of Columbia produces $30,594.3bn for 2025 against a published national total of $30,762.1bn. The $167.8bn difference, 0.55 percent, is federal military and civilian activity located overseas plus differences in the source data behind the industry estimates and the expenditure measure of GDP. BEA says so in the footnote to the table, which is the sort of disclosure that makes the number usable rather than merely impressive.

What does the full 2025 ranking look like?

Answer: The table below carries all fifty states and the District of Columbia for 2025, ranked by current-dollar GDP per capita, with real GDP per capita in chained 2017 dollars and per capita personal income alongside. The final column is where each jurisdiction sits when residents rather than workplaces are counted.

Read the last two columns together and the divergences jump out. Wyoming sits eighteenth on production and fifth on income. Florida sits thirty-fourth and eighteenth, which is what a state with a large retired and investor population looks like when you measure it two ways. Georgia sits twenty-fifth and forty-second, the widest downward gap after Delaware.

How many states are above the national average?

Answer: Fifteen states and the District of Columbia, out of fifty-one jurisdictions. The national figure is $90,004. Thirty-five states sit below it, which is what happens when a small number of very large economies pull an average upward and most of the country does not.

The list of jurisdictions above the line is short and it is not intuitive: Alaska, California, Colorado, Connecticut, Delaware, the District of Columbia, Illinois, Maryland, Massachusetts, Minnesota, Nebraska, New Jersey, New York, North Dakota, Texas and Washington.

Concentration is part of the reason. California alone produced $4,251bn of output in 2025, 13.8 percent of the national total. Add Texas at $2,904bn, New York at $2,468bn, Florida at $1,835bn and Illinois at $1,202bn and those five states account for 41.2 percent of United States gross domestic product. The top ten reach 56.6 percent.

Two of those five, Florida and Texas, sit below the national average on a per capita basis. Size and intensity are separate things, and the ranking most people are looking for when they type this query is the second one.

What the number will not tell you

GDP per capita is a production measure divided by a headcount. It is not income, it is not wages, it is not adjusted for the cost of living, and it says nothing about distribution. A state can lead the table because a small number of very large firms book value added there. A state can trail it while its residents draw substantial income from assets held elsewhere.

There is also a hard limit on how far back the series can be read. BEA cautions users against appending the pre-1997 statistics to the current ones, because the data change from the older industry classification to the current one at that point, the older statistics are consistent with gross domestic income rather than gross domestic product, and improvements made in the 2014 comprehensive revision, including the capitalisation of research and development, were never carried back. The caution is explicit and it applies to both the levels and the growth rates.

Within those limits the current statistics are recent. State GDP is published alongside the third estimate of national GDP each quarter, so the latest reading covers the first quarter of 2026 and was released on 25 June 2026. Real GDP rose in 46 states and the District of Columbia in that quarter, ranging from 4.5 percent at an annual rate in Washington state to a decrease of 1.6 percent in South Dakota, and it was unchanged in Delaware. Information was the leading contributor in Washington. Agriculture, forestry, fishing and hunting was the leading contributor to the decline in South Dakota. The next state release is 30 September 2026.

Which leaves the question of what to do with the figure. Where a state's workers and a state's residents are largely the same people, GDP per capita is a fair reading of how intensively that economy produces. Where they are not, it is a reading of the commute. The adjustment for residence is the line that tells you which case you are in, it is published for every state every year, and it takes about a minute to look up.

Related: how accurate economic forecasts turn out to be, from the Federal Reserve's own error table, and why the trade deficit by country depends on which flow you count.