Today's Friday Study shares information from the The Urban Institute about a projected 4 million people that could be missing from the 2020 Census. A net national undercount of 4 million people would reflect a census that missed far more millions of people of color and people with low incomes and double counted millions of higher income people and non-Hispanic whites.
At the local level, these miscounts don't cancel each other out; they increase the inequities flowing from a census that misses people of color and low-income households at much higher rates than other communities. An inaccurate census count could exacerbate systemic inequities for people of color and people with low incomes.
The Urban Institute's analysis concludes that the 2020 census could lead to the worst undercount of black and Latinx people in the U.S. since 1990. This follows reports from the Census Bureau that the addition of the citizenship question could cause 6.5 million people to not respond in the self-response period, leading to a more costly and time-intensive "non-response-follow-up" period. See below for additional information about the status of the citizenship question and a link to a communications toolkit.
Challenges threatening the upcoming 2020 census could put more than 4 million people at risk of being undercounted in next year's national head count, according to new projections by the Urban Institute.
The nonpartisan think tank found that the danger of an inaccurate census could hit some of the country's most difficult to count populations the hardest. Based on the institute's analysis, the 2020 census could lead to the worst undercount of black and Latino and Latina people in the U.S. since 1990.
"Miscounts of this magnitude will have real consequences for the next decade, including how we fund programs for children and invest in our infrastructure," says Diana Elliott, a senior research associate at the Urban Institute who co-wrote the report released Tuesday.
Nationally, black residents could be undercounted by as much as 3.68%.
"That doesn't sound terribly high, but when you realize that that's 1.7 million people, that's a lot of people to be missed in the overall count," Elliott explains.
The institute also projects as many as 2.2 million (3.57%) Latinos and Latinas around the U.S. could be undercounted in the 2020 census.
Children under the age of 5 — another hard-to-count group — also face an undercount as high as 6.31%, or about 1.3 million young children.
All of these projections are based upon what the Urban Institute considers a "high-risk" scenario. Still, John Thompson, a former Census Bureau director who reviewed the report, says that these estimates "may be a little bit on the conservative side."
"It could be as bad as 1990. It could be worse," Thompson says, raising concerns about the bureau's ability to encourage all households to take part in next year's constitutionally mandated head count of every person living in the U.S.
A Range of Hurdles
To produce these projections, researchers factored in a range of hurdles that could undermine the accuracy of the census.
They include the controversial question the Trump administration wants to add to forms for the 2020 census: "Is this person a citizen of the United States?"
Census Bureau researchers have warned that including the citizenship question would very likely scare households with noncitizens into not responding to the census. In a separate study, the bureau concluded the question was a "major barrier" to full participation in the head count, especially at a time of increased immigration enforcement and rising anti-immigrant rhetoric around the U.S.
The Supreme Court is expected to rule by the end of June on whether the Trump administration can include the citizenship question. Newly disclosed documents belonging to a major GOP redistricting strategist involved in the administration's push for the question are complicating the legal battle.
Regardless of how the court rules, the Urban Institute researchers say all of the public attention on the question has created a chilling effect on census participation among Latinx and immigrant groups — a factor they included in their projections for a "high-risk" scenario.
The report also points out new ways of conducting the U.S. census that have not been thoroughly tested and could pose another risk to the count's accuracy. These methods include allowing all households to complete an online form and expanding the use of existing government records to help complete questionnaires for households that don't respond themselves. Uncertainty in funding in recent years has led the Census Bureau to cancel field tests for the 2020 census, including test runs designed for rural and Spanish-speaking areas.
"Not only are these new additions insufficiently tested in a decennial census environment," write the report's authors, "but the best evidence suggests they will disproportionately improve the count of those who are already easiest to count, leaving the hard-to-count population a lingering challenge."
"An Incorrect Vision"
Among racial and ethnic groups, only white people are projected to be overcounted, while other groups are expected to see undercounts — including as high as 1.36% (or about 306,000 people) for Asians and Pacific Islanders and 2.12% (102,000 people) for American Indians and Alaska Natives nationally.
At the state level, these trends mean that states with more historically undercounted groups — including people of color and renters — are more likely to have inaccurate population counts in 2020. While California, Texas and Nevada face high undercount risks, states with older populations that are more likely to be white and owning homes — including Maine, New Hampshire, Vermont and West Virginia – have the greatest potential for being overcounted, according to the institute's analysis. In the 2010 census, for example, white homeowners were overcounted because some with multiple homes were counted incorrectly at multiple addresses.
Whether it's an overcount or undercount, the concern is that political representation and federal funding will not be fairly shared after the 2020 census. The new population numbers will determine how many congressional seats and Electoral College votes each state gets, as well as guide the distribution of around $880 billion a year in federal tax dollars for schools, roads and other public services.
"When it comes time to allocate resources," says Robert Santos, the Urban Institute's vice president and chief methodologist, about the consequences of an inaccurate census, "you end up with an incorrect vision of where the population is and where the funding should go."
Despite their report's dire warning about potential undercounts, the Urban Institute's researchers emphasize there is still an opportunity to overcome these challenges by driving up public interest and participation in next year's count.
"This is by no means a critique of the Census Bureau," says Elliott, who once worked at the bureau as a demographer. "The Census Bureau is increasingly asked to do more with less, and as we see, it's more and more of a challenge to count the nation's population with every passing decade."
The institute provided the bureau an advance copy of the report before its release. In a written statement, spokesman Michael Cook said the bureau is "laser focused" on working with national organizations and local community groups to help promote the census.
"We hope that the known challenges encourage individuals across the country to apply for 2020 Census jobs and partner with us to help ensure an accurate count," Cook said.