Census Bureau’s Citizenship Data Collection Is Misguided and Harmful

Steven Dillingham sits behind a desk
Steven Dillingham, Director of the U.S. Census Bureau, testifies before the U.S. Senate Committee on Homeland Security and Governmental Affairs on Capitol Hill in Washington, DC, U.S. on July 16, 2019, discussing the security and accuracy of the 2020 census. Photo by Stefani Reynolds/CNP

As the 2020 Census gets underway, the Census Bureau should be giving its undivided attention to the task of counting every U.S. resident. Instead, the Bureau is trying to juggle that massive undertaking with a separate project that will likely do much more harm than good: gathering records from state governments to help estimate the number of adult U.S. citizens living on every block in the nation.

In a comment submitted to the federal Office of Management and Budget last week, Campaign Legal Center (CLC) argues that the Census Bureau should abandon this effort to collect state administrative records on individuals’ citizenship status.

Contrary to what President Trump’s administration appears to believe, this data collection will not enable the Bureau to produce data reliable enough for states to use in drawing voting districts that equalize citizen voting-age population (CVAP).

The Census Bureau’s effort to collect state records on citizenship follows the failed attempt to add a citizenship question to the 2020 Census.

After the Supreme Court ruled in June that the federal government had illegally “contrived” a rationale for adding the citizenship question, President Trump ordered federal agencies to work together on an alternative method of producing citizenship data. Instead of asking individuals to report their citizenship directly, the Census Bureau would gather citizenship information from administrative records held by the federal and state governments.

According to President Trump’s executive order, a major goal of this project is to provide states with data that they can use to draw voting districts with equal “voter-eligible population,” instead of equal total population.

This is a wrongheaded goal. Equal legislative representation for all U.S. residents—not just adult U.S. citizens—is a matter of fundamental fairness and a tradition rooted in our Constitution.

But even if excluding all children and non-U.S. citizens were acceptable in theory, it would require a set of data reliable enough to make it workable in practice. The Census Bureau is no position to produce CVAP statistics that meet this standard of reliability.

The Bureau’s plan is to combine newly acquired state administrative records with federal administrative data and demographic information to calculate each U.S. resident’s “citizenship probability,” or likelihood of being a U.S. citizen. This method will predictably fail to produce accurate data on the CVAP of each census block.

Simply put, administrative records—especially those held by state governments—are not reliable sources of information on individuals’ current citizenship status. Texas learned this lesson the hard way last year, when it tried to use driver license data to remove registered voters from its voter rolls on the theory that they were non-U.S. citizens.

On behalf of clients including the League of United Latin American Citizens, CLC successfully sued and showed that the vast majority of the individuals Texas targeted were naturalized U.S. citizens whose driver license records had never been updated to reflect their naturalizations.

By relying on faulty administrative data, the Census Bureau will significantly underestimate the numbers of adult U.S. citizens living on some blocks, while overestimating the CVAP of other blocks. The resulting block-level CVAP statistics will be much less reliable than the Bureau’s block-level counts of total population.

As a matter of both constitutional law and common sense, states must use high-quality data for redistricting. Even if a state wanted to base its redistricting plan on CVAP (which it should not), the state could not accomplish that goal with the unreliable estimates of CVAP that the Census Bureau will produce.

Collecting huge amounts of citizenship data from the states cannot change the reality that the Census Bureau’s CVAP estimates will be unusable as a measure of population for redistricting. This data collection will, however, affect the Bureau’s reputation.

The more citizenship information the Bureau tries to gather, the more it will feed existing fears about the Bureau sharing information with immigration enforcement authorities. While these rumors are false, they make it difficult to persuade non-U.S. citizens and their families to participate in the 2020 Census.

The Census Bureau should stop trying to collect new sources of citizenship data from the states, and focus instead on overcoming the many challenges facing the 2020 Census.     

Jeff Zalesin is a Legal Fellow at the Campaign Legal Center.