Center for Immigration Studies Says Census Survey Shows Foreign-Born Population Down 2.9 Million From January 2025 to July 2026
The private research group's reading of the Current Population Survey also estimates the unauthorized population fell 2.3 million; economists at Brookings and the Peterson Institute say the survey's weighting makes the size of the drop uncertain.
A Number and Its Shadow Number
On September 3, 2026, the Center for Immigration Studies put out a headline figure: the foreign-born population of the United States had fallen by 2.9 million since January 2025[1]. The group, which argues for lower immigration levels, said the drop meant net migration had gone negative for the first time since the 1930s[1]. It also offered a preliminary estimate that the unauthorized immigrant population had fallen by 2.3 million over the same 18 months, from roughly 15.8 million to about 13.5 million[1].
But the government's own most recent full-year count tells a different story. The Census Bureau reported that net international migration was 1.3 million for the year ending June 2025 — a historic drop from 2.7 million the year before, but still a positive number, not a decline[3][4].
Both figures are real. Neither is fake. They just aren't measuring the same thing, over the same window, with the same tool[1][3]. That gap is the whole story.
Why a Jobs Survey Can't Easily Count People Leaving
The CIS number comes from the Current Population Survey, the same monthly poll that produces the unemployment rate. It's the CPS — a rolling household survey of about 60,000 homes that asks where each person was born. It does not ask about immigration status[1][7].
That second point matters a lot. The unauthorized-population estimate of 2.3 million isn't a count. It's an estimate built on top of another estimate, using a statistical method called residual estimation that infers who's undocumented by subtracting known legal populations from the total foreign-born count[18]. Small errors in the first number can widen fast in the second.
There's also a design problem economists keep pointing to. The CPS doesn't just tally who answers the survey — it scales that small sample up to match preset population totals, called population controls, that the Census Bureau calculates in advance[7]. For 2025, those totals were set in late 2024, before anyone knew immigration would fall sharply[5][6].
Here's why that's a bigger deal than it sounds. If the preset totals assumed more immigrants than actually showed up, the survey doesn't just show fewer foreign-born people — it has to shove those "missing" people somewhere else in the count to make the totals balance. Economists at the Peterson Institute for International Economics and the Brookings Institution say that's exactly what's happening: the same data that show 2 million fewer foreign-born residents also show an implausible jump of more than 2 million in the U.S.-born population, a group that simply can't grow that fast[5][6]. On top of that, immigrants under enforcement pressure may just stop answering surveys — someone who's gone quiet and someone who's actually left the country look identical in the data[6][7].
The Case for Reading the Number as Vindication
CIS and the Trump administration have a straightforward argument, and it isn't a crazy one: if people are leaving on their own because enforcement went up and incentives to stay went down, that's the policy working without mass roundups. The White House said in March 2026 that net migration had turned negative across every metro area it tracked[15]. Record-low Border Patrol apprehensions offer a second, independent signal that doesn't depend on how the CPS is weighted[1].
CIS also argues the labor-market alarm is overstated by its own numbers. It estimates only 44% of the foreign-born decline is workers — the rest are children, retirees and other non-workers[1]. And the group isn't hiding the uncertainty: its own report states the margin of error is large, and its lead author, Steven Camarota, wrote back in March 2026 that newly reweighted data actually showed a smaller decline than earlier monthly numbers had suggested[1][8].
For the administration, there's a political incentive here too, heading into the midterms: a concrete number to point to. But that same number cuts both ways. A smaller labor force tends to mean slower growth in jobs, spending and GDP — a cost the administration also has to account for[5].
What the Number Leaves Out
Turn the number over and it describes something else: 2.9 million people not arriving, or leaving, or going quiet — some of them here entirely legally. CIS puts the legally present share of the decline at around 600,000, covering green-card holders, students and refugees alongside people without legal status[1]. Advocates for immigrants say that's the part a population count can't capture: it treats a departure as a data point rather than a disruption.
Employers in construction, farming, meatpacking and elder care say those workers aren't easy to replace at current wages. One projection, from the National Foundation for American Policy, puts the potential loss at roughly 19 million worker-years by 2028[16]. Brookings has separately tied the immigration slowdown to slower population growth across major U.S. metro areas that had been growing mainly through immigration[17].
None of that disputes the direction of the trend. Immigration fell after January 2025 — every source in this dispute agrees on that[1][3][5]. The disagreement is about how much, how fast, and whether an 18-month survey can measure a shift like this cleanly at all.
How the Same Number Traveled Differently
The CIS report's headline — "Census Bureau Survey Shows Total Foreign-Born Down 2.9 Million" — leads with the Census Bureau's name, which lends a federal-sounding authority to what is a private group's analysis of federal data[1]. The press release version, distributed over GlobeNewswire the same day, carried that framing with none of the caveats CIS's own report text included[1][2]. The Manila Times republished that release almost verbatim, with no independent reporting attached[13].
On the right, RedState ran the figure as "Stunning" and framed it as happening "Under Trump," converting an 18-month trend into a scoreboard, again without the margin-of-error language CIS itself had published[11]. Breitbart used the same household survey for a related labor-force claim, in the direction that supported its point, without mentioning the weighting problem economists say cuts the other way too[12]. Fox News was more careful, attributing the claim to "economists" — though the economists who question the CPS's weighting got less space than the ones who don't[10].
Coverage that leaned left ran differently: not as competing news stories, but through economists at Brookings and the Peterson Institute, arguing the method rather than the politics[5][6]. Left-of-center outlets appear to have mostly stayed out of it as a news story altogether.
What neither side is advertising: CIS's own reweighted data in March 2026 showed a smaller decline than its earlier estimates had implied, and the Census Bureau's last official full-year figure — 1.3 million — is still a positive number, not a drop[3][8]. The more precise size of the fall in immigration won't be settled until revised Census estimates and reweighted survey data come out, sometime next year.
Summary
On September 3, 2026, the Center for Immigration Studies published an analysis of the Census Bureau's monthly household survey. It reported that the U.S. foreign-born population fell about 2.9 million between January 2025 and July 2026[1][2]. The group also gave a preliminary estimate that the unauthorized immigrant population dropped 2.3 million over the same 18 months, from about 15.8 million to about 13.5 million[1]. CIS said the decline means net migration to the United States was negative for the first time since the 1930s. It noted this follows an 8.3 million rise in the foreign-born population over the prior four years[1].
CIS is a private research group in Washington that argues for lower immigration levels. It grew out of the network founded by John Tanton and has received millions of dollars from the Colcom Foundation, which also funds restriction advocacy[9]. Its numbers are not government findings. They are one group's reading of a government survey.
The survey in question is the Current Population Survey, or CPS. It is the same monthly survey that produces the unemployment rate. It asks about 60,000 households where each person was born. It does not ask about legal status[1][7]. That matters for the second number: the 2.3 million unauthorized figure is an estimate built on top of an estimate[18].
The genuine dispute is not whether immigration fell. Everyone agrees it did. The dispute is how much, and whether the survey can measure it this precisely. Economists at Brookings and the Peterson Institute for International Economics argue the CPS is weighted against population targets that were projected in late 2024 — before anyone knew what would happen — and that the same data imply an implausible jump of more than 2 million in the U.S.-born population[5][6]. The Census Bureau's own official estimate for the year ending June 2025 still showed positive net international migration of 1.3 million, though sharply down from 2.7 million[3][4]. CIS acknowledges the margin of error is large, and its own author wrote in March 2026 that new statistical weights made the decline look smaller[1][8].
The Event
The Center for Immigration Studies released a report on September 3, 2026, titled "Census Bureau Survey Shows Total Foreign-Born Down 2.9 Million"[1]. The report analyzes Current Population Survey data from January 2025 through July 2026 and states the foreign-born population fell about 2.9 million over that period[1]. It gives a preliminary estimate that the unauthorized immigrant population fell about 2.3 million, to roughly 13.5 million in July 2026 from roughly 15.8 million in January 2025[1]. The report was distributed the same day as a press release over the GlobeNewswire wire and republished by outlets including The Manila Times[2][13].
Undisputed Facts
- The Center for Immigration Studies published the analysis on September 3, 2026, using Current Population Survey data through July 2026[1][2].
- The Current Population Survey records where a respondent was born, but does not ask immigration status; unauthorized-population figures are derived estimates, not counts[1][18].
- The Census Bureau reported net international migration of 1.3 million between July 1, 2024 and June 30, 2025, down from 2.7 million the previous year, which it called a historic decline[3][4].
- U.S. population growth between July 1, 2024 and July 1, 2025 was 1.8 million, or 0.5 percent, and the Census Bureau attributed the slowdown mainly to lower net international migration[4].
- CIS states in the report itself that the margin of error around its figures is large and that the point estimates have fluctuated[1].
- CIS estimates that 44 percent of the foreign-born decline consists of workers; the rest are children, the elderly and other non-workers[1].
- CIS author Steven Camarota wrote on March 6, 2026 that reweighted January and February 2026 CPS data showed a smaller decline since January 2025 than the 2025 monthly data had indicated[8].
- The Brookings Institution has published estimates that U.S. net migration in 2025 was between -295,000 and -10,000, and could reach -925,000 in 2026 — figures that are estimates, not official counts[5].
- The White House issued a March 2026 release stating the U.S. recorded net negative migration across every metro area[15].
The Pressure
Strip away the moralizing and blame. What structural realities persist regardless of which narrative wins?
- A number arrives before the data settles
- The CPS is monthly and fast. The Census Bureau's fuller estimates come once a year and get revised[3][4]. Whoever publishes first sets the frame, and the revision lands months later with a fraction of the attention. That timing advantage is structural, and both sides know it.
- The survey was not built for this
- The CPS exists to produce the unemployment rate. It weights a 60,000-household sample up to population totals set in advance[7]. Measuring a fast demographic shock is outside its design. That is a real limit, not an excuse invented for this report.
- Falling response looks like leaving
- When enforcement rises, immigrants answer government surveys less[6][7]. Someone who stops responding and someone who has left the country produce the same missing row. No survey can tell them apart in the moment.
- Fewer workers is a cost the policy's supporters also carry
- A smaller labor force slows growth in employment, spending and GDP[5]. Metro areas that grew only through immigration stop growing[17]. This constraint applies regardless of who is right about the exact figure.
Material realityImmigration to the United States fell sharply after January 2025. Every source in this dispute agrees on that direction. The Census Bureau's own official figure — 1.3 million in net international migration for the year ending June 2025, down from 2.7 million — is a historic drop, and it is still positive[3][4]. CIS's 2.9 million decline covers a longer and later window, January 2025 to July 2026, and comes from a different, faster, noisier instrument[1]. The two are not directly comparable, and readers should not treat them as competing versions of one figure. What is on the ground either way: fewer arrivals, some number of departures, shrinking populations in metro areas that had been growing only through immigration, and employers in low-wage, physically demanding sectors reporting they cannot fill shifts[16][17]. The precise size of the drop will not be known for at least another year, when revised Census estimates and reweighted survey data are published.
Narrative as a weaponTwo campaigns are running on the same dataset. The Center for Immigration Studies wants you to read a survey estimate as a headcount, and its press release does the work its report text is more careful about — the phrase "Census Bureau Survey Shows" borrows federal authority for a private analysis[1][2]. Downstream, conservative outlets dropped the caveats CIS itself kept[11][12]. On the other side, economists at Brookings and the Peterson Institute want you to focus on the weighting flaw, and their arithmetic point is strong — the same data imply an impossible surge in U.S.-born Americans[5][6]. But that critique attacks the magnitude, not the direction, and some of its amplifiers use it as if it disproved the trend. Note what neither camp advertises: CIS's own author revised the decline downward in March 2026 after new weights arrived[8], and the Census Bureau's official number for the most recent full year it has measured is still a positive 1.3 million[3]. Neither of those facts fits either campaign, which is a decent reason to keep them in view.
How Each Side Sees It
Each major actor’s view — how it frames things, its underlying incentive, and how it’s materially affected. Tap a side to read it.
Frames it asTheir strongest case is that the policy worked as advertised, and worked without mass forced removals. They argue the post-2020 increase of 8.3 million foreign-born residents was itself the anomaly, absorbed by schools, hospitals and housing markets that never voted on it[1]. If enforcement plus reduced incentives causes people to leave on their own, they say, that is the humane version of the policy — no camps, no raids at that scale. They also argue the labor-market alarm is overstated on their own numbers: fewer than half the departures were workers[1]. On the data fight, they note they published the margin-of-error caveat themselves and revised downward in March when new weights arrived[1][8].
WhyCIS exists to argue for lower immigration levels and has done so since 1985[9]. A large, measurable drop is the clearest possible proof of concept for its policy program, and arrives while that program is being contested in Congress and the courts.
Impact on themThe group's influence runs on being first with a number that reporters use. Its figures were picked up within hours by wire services and conservative outlets[2][11][13]. The reputational risk is symmetric: if revisions cut the 2.9 million figure substantially, the citation trail already exists.
Frames it asOfficials argue enforcement plus removal of pull factors produced a result decades of promised reform never did. The administration's own release says net migration turned negative in every metro area[15]. The underlying principle they invoke is that a country decides who enters, and that a border which is enforced is not a cruelty but a precondition for any legal system of admission. They point to record-low Border Patrol apprehensions as an independent check that does not depend on survey weighting[1].
WhyA concrete, countable result ahead of the midterms, and a rebuttal to the charge that the enforcement push produced disruption without measurable change.
Impact on themPolitically, the number is an asset. Economically, it cuts both ways: a smaller labor force means slower growth in employment, consumer spending and GDP, which the administration also has to answer for[5].
Frames it asTheir argument is technical, not political, and it is arithmetic. The CPS surveys a sample and then scales it up to match independent population totals — the "population controls." Those controls for 2025 were projected in late 2024, before the policy shift. If the control totals are wrong about immigration, the survey redistributes people rather than losing them. That is why, they say, the same data that show 2 million fewer foreign-born residents also show an implausible increase of more than 2 million U.S.-born residents — a population that cannot grow that fast[5][6]. Separately, immigrants under enforcement pressure may simply stop answering the door, which looks identical in the data to leaving[6][7]. Their point is not that immigration did not fall. It is that a survey built for the unemployment rate cannot cleanly measure a demographic shock in real time.
WhyProfessional stake in method. Their credibility rests on estimates surviving revision, and they were publicly skeptical of a smaller version of this claim in 2025[14].
Impact on themIf they are right, downstream forecasts of labor supply, Social Security and GDP built on the larger figure are wrong. Brookings has already tied the slowdown to falling population growth across major metro areas[17].
Frames it asFor them the crux is not the size of the number but what it describes. A 2.9 million decline is 2.9 million individual departures, removals or absences — including lawful permanent residents, students and refugees, since CIS puts the legal share of the drop at about 600,000. Advocates argue people who are legally present are being pushed out alongside those who are not. Employers in construction, agriculture, meatpacking and elder care say the workers are not replaceable at current wages; the National Foundation for American Policy projects a loss of roughly 19 million worker-years by 2028[16]. Their strongest principled claim is that a policy measured only by how many people left has no way to register what those people were doing.
WhyAdvocacy groups need the human and economic cost made legible before the population figure hardens into a settled success story. Employers need workers.
Impact on themDirect and material: separated families, unfilled shifts, and shrinking populations in metro areas that had been growing only through immigration[17].
Like this article?
The Bias Ledger average rating 5.9
The same story, as framed by outlets across the spectrum, ordered least to most biased. The bias score (1 = straight, 10 = heavily spun) is an AI assessment of that framing — click an outlet to see its track record. The tell is the word choice or omission that reveals the angle.
| Outlet | Vantage | Bias | How they frame it | The tell |
|---|---|---|---|---|
| PolitiFact | U.S. center-left fact-checking nonprofit, Poynter Institute | 3 | Checked a 2025 DHS claim of a 1.6 million immigrant-population drop and focused on survey limits rather than the politics. | Its skepticism is aimed at the claim rather than at the underlying trend, which it does not dispute. Critics on the right note that similar scrutiny was applied less often to immigration estimates moving the other way. |
| The Manila Times | Philippine, center-right daily | 5 | Republished the GlobeNewswire item under the CIS headline, in a newswire section. | Overseas republication with no local reporting and no counterpoint. The omission is not spin but distance — readers abroad get the framing with none of the U.S. methodological dispute attached. |
| Fox News | U.S. right | 5 | "US sees net negative migration for first time in decades amid Trump admin enforcement, economists say" | More careful than its ideological neighbors — it attributes to "economists" and uses "amid" rather than claiming cause. But the economists cited are the ones who agree; those disputing the survey's weighting are not given comparable space. |
| Center for Immigration Studies | U.S. right, immigration-restrictionist advocacy | 6 | "Census Bureau Survey Shows Total Foreign-Born Down 2.9 Million" — leads with the Census Bureau's name, which lends federal authority to a private analysis. | The report does carry real caveats about margin of error. But the headline and press release do not, and the phrase "Census Bureau Survey Shows" invites readers to hear this as a government finding rather than one group's reading of a government dataset. |
| GlobeNewswire | Paid press-release wire, no editorial screen | 7 | "Total Foreign-Born Down 2.9 Million" — the advocacy group's own words carried unchanged onto a newswire. | A press release formatted like a wire story. Nothing is false, but nothing is checked either, and the distribution gives the claim the visual authority of reported news. |
| Breitbart | U.S. right, pro-restriction | 7 | "Jobs Report Shows Nearly One Million Fewer Foreign Born Men in U.S. Labor Force" | Uses the same CPS household survey, but only where the figure runs one direction. The known weighting problem in that survey — which economists say also inflates the U.S.-born count — is not mentioned. |
| RedState | U.S. right, opinion-driven | 8 | "Foreign-Born Population Drops by Nearly 3 Million Under Trump: Stunning New Report" | "Stunning" is the writer's verdict, not a finding. "Under Trump" converts a correlation over 18 months into a scoreboard, and CIS's own margin-of-error language does not appear. |