Click a state to focus on it; click it again (or press Esc) to return to all states. Puerto Rico is shown as the square near Florida. What do the ± numbers mean?
Latest net migration: Census population estimates
Hover a state for its components of change; click it to show its state-to-state flows on the map above.
About the sentiment view
Net migration mostly tracks size: a big state sends more people everywhere simply because it has more people. The Sentiment view takes size out by comparing rates: the share of one state's residents who moved to the other state, against the share who moved the other way.
In 2024, 77,161 Californians moved to Texas, 197 of every 100,000 California residents. Going the other way, 45,447 Texans moved to California, 148 of every 100,000 Texans. So people in California were 1.33× as likely to move to Texas as people in Texas were to move to California. On the map, that's an arrow pointing to Texas labeled 1.3×.
Sentiment can point the opposite way from net migration. More people moved from California to Nevada (53,289) than from Nevada to California (21,133), so Nevada gained. But Nevada has a twelfth of California's population: 658 of every 100,000 Nevadans moved to California, against 136 of every 100,000 Californians moving to Nevada. By sentiment, Nevadans were 4.8× as likely to make the move, so the arrow points to California.
How to read it
- With a state selected, partners are green when their residents moved to it at a higher rate than its residents moved to them, and red when the reverse. Arrows point toward the favored state and are labeled with the ratio of the two rates.
- With all states shown, states aren't shaded. The arrows and the table show the most lopsided pairs among those with at least 1,000 movers each way (per survey).
- Rates use each state's population on July 1 of the year before the survey, when the movers still lived there, from the Census Bureau's population estimates.
- Each rate has its own margin of error, and so does the ratio. Washington and Arizona in 2024: 225 ± 42 of every 100,000 Washingtonians moved to Arizona, and 116 ± 35 of every 100,000 Arizonans moved to Washington. The ratio is 225 ÷ 116 = 1.94×, but with both rates uncertain, its 90% range is about 1.25× to 2.62×. Tooltips and the table show that range, and a ratio is gray when its range includes 1× (see About the ± numbers).
- Sentiment strongly favors large states. A big state offers far more jobs and homes to move to, so a small state's residents move to it at a much higher rate than the reverse, even without any special preference. Nearly every state favors California this way, many by 20× or more, and Alaska by 83× in 2024, largely because California has about 54 times Alaska's population. Sentiment is most telling between states of similar size, like California and Texas, or when comparing a state's partners with each other.
About the ± numbers
The state-to-state flows come from the American Community Survey (ACS), which asks a sample of about 1 in 40 homes each year where everyone lived a year earlier. The Census Bureau scales the answers up to the whole population, so every number is an estimate, and it publishes a margin of error with each one. “2,963 ± 1,488” means the Bureau is 90% confident the true number is between 1,475 and 4,451.
Why the margins can be so large
- Few people in the sample made that move. Only about 2% of people move to another state in a year, and those moves are spread over 2,652 pairs of states. A flow of a few thousand people rests on a few dozen survey answers, so it can't be pinned down closely. Big flows are much more precise: California → Texas in 2024 was 77,161 ± 9,059 (±12%), while Massachusetts → Washington was 2,963 ± 1,488 (±50%). In 2024, more than a quarter of the state-to-state flows have a margin as large as the estimate itself.
- Zero doesn't mean nobody. When no one in the sample made a move, the estimate is 0, but the margin (for example, Wyoming → Vermont: 0 ± 170) says it could still be a few hundred people.
- Net numbers are the least certain. A net is the difference between two flows that are often similar in size. The flows mostly cancel, but their uncertainties don't: they combine. Washington's exchange with Massachusetts in 2024 was 2,963 in and 2,252 out, a net of +711, but the margin on the net is ±1,755 — bigger than the net itself. The true net could be a loss of 1,044 or a gain of 2,466, so the map shows it in gray.
How to read them here
- When the margin is larger than the estimate, the number is not statistically significant: the survey can't rule out zero. On the map those arrows are dashed and unlabeled, uncertain nets are gray, and in the tables the values are gray italic. Hover for the full range.
- Totals and longer periods are more precise. A state's total moves combine dozens of flows, and the combined 5- and 10-survey datasets add up independent samples, so their margins are smaller relative to the estimates.
- The population estimates in the bottom section show no margins. They aren't a survey: the Census Bureau builds them from tax returns, Medicare enrollment and other records, and doesn't publish margins of error for them. They're not exact either.