Research Shows a Citizenship Question Would Suppress Participation among Latinxs and Immigrants in the 2020 Census, Undermining Its Reliability

The U.S. Constitution requires a decennial census of all persons living in the country, and our nation has carried out this duty since 1790. Specifically, the census must count all people living in the U.S. and record where they live. These counts are crucial for determining each state’s number of seats in the U.S. House of Representatives, defining legislative districts, and distributing federal funding for state, local, and tribal governments. The census is also a vital source of population data. Therefore, it is imperative that each decade, the census is methodically planned and carried out.

Under the Trump administration, the U.S. Department of Commerce has proposed to add a citizenship question to the 2020 census. This sudden change would have alarming implications.

Under the Enumeration Clause of the Fourteenth Amendment, the census must count everyone living in the country, regardless of their immigration status. Social scientists, policymakers, advocates, and even former directors of the Census Bureau have argued that introducing a citizenship question — which has not been tested — would have a chilling effect on the census response rate. This would undermine the reliability of census data by undercounting particular populations, especially low-income people and people of color who have already been undercounted in past iterations of the census.

Three federal judges have already found the addition of the citizenship question to be unlawful. Tomorrow, Tuesday, April 23, the U.S. Supreme Court will hear arguments about the question (Dept. of Commerce v. New York; NILC attorneys drafted and filed a friend-of-the-court brief in this case).

Fears of the gravity of an undercount stemming from the proposed citizenship question are empirically supported by new research coming out of the San Joaquin Valley in California. Researchers from the San Joaquin Valley Census Research Project (SJVCRP) conducted 414 in-person surveys with Latinx immigrants and the U.S.-born, adult children of immigrants to determine whether they would respond to the 2020 census with or without the proposed citizenship question.

The San Joaquin Valley is a highly-populated area of the country, with 4.2 million residents. It has a higher-than-average concentration of foreign-born residents and is majority Latinx. It is also a growing community: the population is expected to reach 4.6 million people in 2020, the year the census will be conducted. Surveys for the SJVCRP reached people in locales in the San Joaquin valley ranging from small, rural communities to major cities.

Researchers uncovered a significant and troubling finding from this survey research: Fewer Latinx immigrant households will participate in the 2020 census if the question is implemented, which will result in an undercount. Without the citizenship question, 84 percent of respondents were willing to participate in the census; after including the citizenship question, however, willingness to participate dropped by almost half, to 46 percent. Willingness dropped among individuals across legal status: naturalized citizens, legal residents, and undocumented individuals.

In addition, declines in willingness to participate were especially notable among the “second generation,” that is, U.S.-born citizens who are children of immigrants. Fewer than half of those surveyed were willing to respond when the citizenship question was included. In fact, these U.S.-born citizens were much less likely to answer than naturalized citizens or legal residents.

These results indicate that a census that includes a citizenship question would not only fail to accurately measure the population, with an estimated 4.1 percent undercount, but also would misrepresent population demographics by undercounting first- and second-generation Latinx Americans by nearly 12 percent. Such an undercount is considered by some experts to be a failed census.

What’s at stake if such an undercount occurs in the San Joaquin Valley? Equitable political representation in Congress, for one thing, and at least $198 million in annual federal funding for residents of the valley. Researchers estimate that these results extrapolated to the state of California would cause an undercount of as many as 1.3 million people in the state, resulting in reduced congressional representation for Californians and annual funding losses ranging between $970 million and $1.5 billion.

The researchers conclude that “[p]roceeding with a politicized decennial census — widely understood by Latino first- and second-generation immigrants as compromising a potentially attractive collective endeavor, the process of ‘standing up and being counted’ to assure one’s community gets its fair share of federal funding and equitable political representation — will further erode already-wavering trust in government.”

We must protect the integrity of the census to ensure that all Americans are counted in 2020.

Holly Straut-Eppsteiner is NILC’s Mellon/ACLS Public Fellow and research program manager.