Children Are Already Undercounted in the Census. Proposed Changes Could Make It Worse.
Newly proposed rules for who is counted and where could make millions of children harder to count in the 2030 Census.
More than 1.5 million children were missed in the last U.S. census. In 2020, the census undercounted children under age 18 by an estimated 2.1%, according to the Census Bureau’s Demographic Analysis, an independent method for evaluating census coverage. The undercount was even higher for children under age 5, at 5.4%—the highest rate of any age group.
Now the Census Bureau is proposing major changes to 2030 Census rules that could make some children even harder to count.
The proposal, released Sept. 10, would count only U.S. citizens and lawful permanent residents for congressional apportionment, excluding other foreign citizens who live here— including people on temporary student or work visas, asylum seekers, DACA or TPS recipients who are not lawful permanent residents, and undocumented immigrants. It would also change how the Bureau determines where people live, including using tax records as evidence of usual residence, and remove race and ethnicity questions from the census form. The Bureau is also considering adding a question about legal status.
Which Children Are Most Likely to Be Affected?
Two of these changes would fall hardest on particular groups of children: the possible legal-status question, and new rules for determining where people are counted.
Children in immigrant families are the clearest case. In 2024, 18.3 million U.S. children under age 18—more than one in four—lived with at least one immigrant parent, and 15.5 million of them were born in the United States.
Those U.S.-born children are citizens and would remain in the apportionment count. The concern is that a legal-status question could discourage their families from responding at all, a so-called “chilling effect.” A recent analysis of the 2019 Census Test found that a citizenship question reduced the response rate among households with noncitizens, and households that did respond were more likely to leave people off the form.
The second change, to the residence rules, could make it harder for the Census Bureau to count children “once, only once, and in the right place.” A person’s usual residence would be where they spent the most days between January 3 and April 1, with at least one hour at a residence counting as a full day. Tax records would serve as evidence of where someone usually lives.
The proposal has a separate provision for dependents, although it does not define either “dependent” or “provider.” Dependents without their own usual residence would be counted at the home of the provider they spend the most days with.
Those rules could be especially hard to apply to children who split time between parents, move between households, or live with grandparents or other relatives. Census data show that about 9 million children—roughly 1 in 8—live in households headed by someone other than a parent (figure). Most are the householder’s grandchildren; others are more distant relatives, foster children, or unrelated children. Another 147,000 children live in group quarters such as college housing, juvenile facilities, residential treatment centers, group homes, or shelters.
Figure. One in Eight U.S. Children Live in a Household Headed by Someone Other Than a Parent
Children Under Age 18 by Living Arrangement, 2024 (in Millions)
Source: PRB analysis of U.S. Census Bureau, 2024 American Community Survey 1-year data, Table B09018.
These living arrangements are common, and they are where children are often missed or counted in the wrong place. Research on the 2010 Census found that coverage problems rose steadily with a child’s relational “distance” from the householder. The rate was about 10% for young children who were the householder’s biological child, 16% for grandchildren, 22% for other relatives, and 25% for unrelated children.
2020 Census residence rules provided instructions for children with complex living arrangements, including kids in shared custody, at college, living in shelters, or whose families were in the process of moving. The proposed rule drops that guidance.
Take a child who alternates between two parents’ homes. Under the proposed rule, the child would be counted at whichever home they spend the most days in. But an hour counts as a day, and the same calendar day can count at more than one residence. So a child who spends an hour or more at one parent’s home and sleeps at the other’s would log a day at both. The proposal doesn’t say how it would resolve these cases.
It offers tax records as evidence of where someone lives. But children can be missing from those records entirely. A child appears in IRS Form 1040 records only if someone lists them as a dependent, generally with a Social Security number or other taxpayer identification number. A W-2 identifies the worker, not the worker’s children. And even when administrative records include a child, they may not reliably identify where that child lives. Census Bureau research has found that these records can contain outdated addresses, making it harder to place people accurately at their Census Day residence. Tax records also miss some children more than others. A 2024 Census Bureau study using California birth records found that kids in lower-income families and those whose parents had lower English proficiency were less likely to appear in IRS dependent records. The researchers concluded that IRS records fall short for some of the same populations that are already at higher risk of being missed.
Removing Race and Ethnicity Questions
Removing race and ethnicity questions from the census form would affect all children, not just particular groups. These data are used to plan and fund government programs, measure disparities, evaluate whether programs are reaching different populations, and enforce civil rights laws. Losing them would matter especially for kids. The 2020 Census showed that children were more racially and ethnically diverse than adults, and children under age 5 were the nation’s most racially and ethnically diverse age group.
These data also reveal who the census misses. According to a separate Census Bureau analysis based on the Post-Enumeration Survey, the 2020 Census undercounted Black children ages 0 to 4 by an estimated 8.2% and Hispanic/Latino children of the same age by 7.3%, while non-Hispanic white young children saw no significant undercount. Without race and ethnicity questions, it would be much harder to tell which groups of children are missed most, where, and whether those gaps are changing.
Why It Matters
The decennial census determines how seats in the U.S. House of Representatives are divided among states and provides the population counts used for redistricting. Its influence also lasts through the decade. Census results help form the base for the Census Bureau’s annual population estimates, which track population change by age, sex, race, and Hispanic origin for states and counties. Those estimates, in turn, are used as population controls for the American Community Survey. An inaccurate count of children can therefore affect other statistics that policymakers and communities rely on for years.
Money is also at stake. In fiscal year 2021, data from the decennial census and related programs helped distribute more than $2.8 trillion through 353 federal assistance programs, including roughly $700 billion in one-time pandemic programs. Those programs included Medicaid and CHIP, nutrition assistance, child care, school lunches, housing, and other services that children and families rely on.
The Census Bureau is accepting comments on the proposed rules through October 13, 2026. The central questions extend well beyond who gets included in a single 2030 Census count: How accurately would the proposed rules count children, where would they be counted, and what information would remain available to identify and address gaps? For our kids, the consequences of getting those answers wrong could last a decade.
Related Posts
Understanding and Comparing Population Projections in Sub-Saharan Africa
The release of a new set of population projections by the Institute for Health Metrics and Evaluation (IHME) in July 2020 was greeted with extensive coverage in the media and stimulated fresh debate among demographers about calculation and communication of projections.
Population Age Structure and Pathways to Inclusive, Effective Governance
Among countries with a youthful population, a window of opportunity to achieve key development goals opens across four sectors―health, education, the economy, and governance―as fertility declines and the age structure of the population gets older.
