The Data of Demographic Change - Part 2
New data, new concerns
In December 2024, the Census Bureau’s Population Estimates Program announced a methodological adjustment to account for the surge in refugees, Border Patrol releases, and CBP paroles between 2021 and 2024. The revision nearly doubled estimated humanitarian migration nationally, then distributed those figures to states and counties based on immigration court caseloads from the Executive Office for Immigration Review. This sudden change is perhaps a symptom of a growing and worrisome trend: a depleted and increasingly isolated Census Bureau stretching its most vital data to their limit during a period of unprecedented turmoil and demographic shift.
Source: US Census Population Estimates Program.
The Census made this methodological change at the end of the Biden administration in 2024, but when President Trump took office he put in place policies that led to a significant drop in international travel and migration to the United States. Now the gap between the Census’s estimates and what is truly happening on the ground is growing, leading analysts and planners to require caution in applying this data.
A closer look at Oregon
The practical effect of the 2024 change: communities across the country received a retroactive boost to their population growth numbers. In the Portland, OR metro area, for example, 2023’s overall population growth rate was revised from effectively zero to 0.5%.
In the Portland, Oregon metro area, for example, 2023’s growth rate was revised from effectively zero…
to 0.5% of population.
Source: US Census Population Estimates Program.
Methodological Concerns
Doubling national humanitarian migration over a four-year window and allocating it by court caseload produced estimates that may not reflect actual settlement patterns, particularly as ICE operations take on the nature of sudden, highly localized, politically motivated raids in communities like Minneapolis, Chicago, and Los Angeles.
Net international migration fell in 2025, and should be expected to fall further as the country’s reputation for hostility toward migrants spreads. But our capacity to measure that decline accurately is eroding in parallel. A federal hiring freeze and a wave of early retirements driven by workforce reductions at federal agencies in early 2025 have introduced new friction into the data pipeline. Now, a short-staffed bureau must stretch statistical methodologies to make up for new gaps introduced by federal policy changes.
One concrete example: delivery delays meant the most recent PEP vintage could not incorporate the full-year Medicare and IRS records it has historically relied on, and therefore effectively carried forward rates from 2024.
Data like the PEP estimates remain essential, and the bureau staff are not to blame for these new added pressures. Regardless, the data demand a more critical eye. The priority now is to understand the direction and implications of these trends, not to fixate on the precision of any single year’s estimate.
Imperfect Timing
All of these factors combine to form a structural inflection point: as natural population change declines in more communities, growth depends increasingly on migration. Yet, at the same time international migration is weakening, statistical uncertainty is rising.
For states and localities, the stakes are high; labor force projections, housing construction, school enrollment forecasts, and revenue capacity all flow from these vital data. This situation raises the unwelcome question of whether we are equipped with the data to diagnose and plan for a future in which slow growth is the norm and outright population decline is a real possibility.
Hiring Freezes and Declining ACS Response Rates
The American Community Survey faces parallel threats. The ACS, the backbone of research on household composition, housing, public health, educational attainment, and more, has surveyed roughly 1% of the U.S. population annually since 2005. The mandatory survey requires sustained funding, staffing, and institutional capacity, all now under pressure from federal hiring freezes and legislative efforts to constrain follow-ups or make the ACS voluntary.
Even before these changes, ACS response rates were in noticeable decline, a trend that accelerated during the pandemic and has not recovered. Lower response rates mean larger margins of error, particularly for small geographies and subpopulations where local planners need the most reliable data. When fewer people respond, the Bureau leans more on imputation and weighting adjustments, compounding uncertainty in estimates that already carry significant caveats at the county and tract level. For practitioners who depend on ACS data to calibrate models, allocate resources, or justify investment, this erosion in reliability is not hypothetical.