by Michael R. Strain
The population count has to be accurate. Questions that scare immigrants and minority-group members would make it less so.
There are many ways to illustrate the importance of the U.S. census, and one that should resonate strongly with conservatives is this: It is required by the Constitution. And right up front — Article 1, Section 2.
The Founding Fathers thought it was pretty important, and it is. For one thing, the census — which is designed to be a complete count of the U.S. population conducted every 10 years — determines how many seats in the House of Representatives are awarded to each state. The population counts produced by the census also help determine how hundreds of billions of federal dollars are distributed across states and localities, and influence the decisions of businesses and state and local governments. Problems with the 2020 census will be with us until at least the next count, 10 years later.
It’s critical, then, that the 2020 census counts be accurate. A great way to decrease accuracy is to ask people completing the census survey if they are American citizens. The Justice Department, unwisely, has formally asked the Census Bureau to include a citizenship question on the upcoming census questionnaire. The secretary of commerce should deny this request. If he doesn’t, Congress should step in.
Why? Given the anti-immigrant and anti-minority rhetoric from President Donald Trump and many on the political right, Hispanics, immigrants and members of minority groups probably start by being concerned about answering any survey questions. A citizenship question would only make this problem worse.
It may seem odd that some people would be concerned about answering a simple 10-question form about themselves and the people who live in their houses, especially since the form is confined to basic information about sex, race, birthdate and the like. But experience shows that they are, and were nervous even before Trump took office.
The Census Bureau takes this problem seriously, spending resources on outreach to immigrant and minority communities in an effort to engender trust, and thereby an accurate count. As part of the last decennial, for example, census staff put up signs reading “NO INS. NO FBI. NO CIA. NO IRS” in immigrant communities.
Concern about the 2020 headcount isn’t merely speculative. Census researchers conducting tests noticed in 2017 that respondents were unusually concerned about the confidentiality of their answers.
According to their report, the Census researchers “heard respondents express new concerns about topics like the ‘Muslim ban,’ discomfort ‘registering’ other household members by reporting their demographic characteristics, the dissolution of the ‘DACA’ (Deferred Action for Childhood Arrival) program, repeated references to Immigration and Customs Enforcement (ICE), etc.” Census staff “emphasized facing a ‘new phenomenon’ in the field and reported that respondents’ fears, particularly among immigrant respondents, have increased markedly this year.”
The report continues, “Respondents reported being told by community leaders not to open the door without a warrant signed by a judge.” And Census staff “observed respondents falsifying names, dates of birth, and other information on household rosters.”
One respondent, a legal U.S. resident, told Census staff that “the Latino community will not sign up [i.e., complete the census survey] because they will think that Census will pass their information on and people can come looking for them.”
Census is finding this heightened concern without any mention of a citizenship question. Including a citizenship question will only increase the number of people who refuse to answer the survey or provide false information.
The Justice Department and some others who support including a citizenship question on the 2020 census implicitly argue that these problems are a necessary cost of enforcing the Voting Rights Act. This argument is not compelling. The short decennial questionnaire hasn’t included a question on citizenship in seven decades. And another government survey provides adequate information to the Justice Department on citizenship and the voting-eligible population.
Others support including the citizenship question so that state legislative districts can be drawn using citizen counts, rather than population counts, disadvantaging Democrats in red states. This is likewise a bad reason to sacrifice the accuracy of the decennial census.
There are two reasons why Republicans in particular should oppose a citizenship question.
The first is cost. When someone refuses to complete the survey, the Census Bureau doesn’t just shrug its shoulders and move on to the next house. Instead, the bureau must send in workers with clipboards to encourage the reluctant person to respond. If that doesn’t work, Census sends someone to talk to the person’s neighbors.
So if Republicans want the 2020 decennial to be as cheap as possible, then they should oppose including a citizenship question.
The second reason: By driving down response rates among immigrants (legal and illegal), Hispanics, and minorities, a citizenship question would probably cause those groups to be undercounted. On balance, this undercounting would help red states when it comes time to direct federal spending and apportion seats in the House.
On balance — but not exclusively. Alabama could lose a House seat if a citizenship question is included. Among other red states, Texas and Arizona, each with large Hispanic and immigrant populations, should be concerned as well.
Now would be a good time for leaders from those states to make their voices heard in opposition to this harmful proposal.
Michael R. Strain is a Bloomberg View columnist. He is director of economic policy studies and resident scholar at the American Enterprise Institute.