On Monday, the U.S. Supreme Court ruled unanimously in Evenwel v. Abbott that the principle of “one person, one vote” does not require states to use measures of the voting-eligible population in carving up legislative districts. Thus, all 50 states may continue their current practice of drawing legislative districts using counts of the entire population, including children and non-citizens who are not permitted to vote.
At the same time, the court did not forbid states from using alternative measures to draw up legislative districts, such as counts of the adult citizen population. It did not reject alternative measures because plaintiffs in the case had not presented an alternative map for the court to consider. The plaintiffs had simply argued that districts based on population counts were unconstitutional because they gave greater weight to voters living in districts with higher proportions of ineligible voters.
Indeed, Justice Samuel A. Alito Jr., in his concurring opinion, opened the door for states to consider drawing alternative maps in the future, writing:
Whether a State is permitted to use some measure other than total population is an important and sensitive question that we can consider if and when we have before us a state districting plan that, unlike the current Texas plan, uses something other than total population as the basis for equalizing the size of districts.
As a practical matter, it is unlikely that states will be able to produce alternative maps anytime soon. As others have noted, the drawing of equally sized legislative districts requires accurate information at the census block level.
The decennial census is the only dataset that currently has block-level counts, and it does not include any information about the citizenship status of individuals. The census did ask questions about citizenship in its “long form” questionnaire, but this was sent only to about one of every six households. The American Community Survey, which replaced the census “long form” after 2000, is sent to an even smaller proportion of households (about 3 percent) and cannot produce reliable estimates of adult citizens at the block level for purposes of redistricting.
We may therefore see efforts to collect more fine-grained data to produce accurate maps that exclude non-citizens and children. At the same time, the success of these efforts will be tempered by two major factors: expense and the introduction of new data inaccuracies.
It is unlikely that Congress will support increasing the size of the American Community Survey so that it could produce block-level estimates of adult citizens. Republicans have for several years tried to reduce funding for the ACS. They have also objected to the fact that the survey is mandatory. But such a provision is essential to produce accurate estimates of the U.S. population, including its rates of citizenship.
Of course, individual states or cities could decide to produce their own block-level estimates of adult citizens, but these too would be very expensive endeavors.
The biggest gambit might be to mandate the inclusion of a citizenship question in the 2020 Census. In the fall of 2009, amid the final rounds of preparations for the 2010 Census, Sens. David Vitter (R-La.) and Bob Bennett (R-Utah) threatened to withhold funding for the decennial census if it did not include a question on citizenship status.
Vitter and Bennett were not just concerned about redistricting. They wanted to lay the groundwork needed to challenge an even larger principle: apportioning congressional seats based on population count. As Vitter noted in November 2009:
The current plan is to reapportion House seats using that overall number, citizens and non-citizens. … I think that’s wrong. I think that’s contrary to the whole intent of the Constitution and the establishment of Congress as a democratic institution to represent citizens.
Similarly, Bennett noted that the Democratic majority in the Senate “wants to have illegal residents included in the apportionment count so that states like California and New York can gain more congressional seats,” and noted that counting non-citizens was important “so we can fairly determine congressional representation and ensure that legal residents are equally represented.” Ultimately, Bennett and Vitter’s effort to mandate a census question on citizenship failed along a party-line vote on a Commerce Department budget bill that would pay for the 2010 Census.
Given Alito’s reasoning in Evenwel, and a potential Republican Senate majority in 2017 or 2019, the temptation to mandate a question on citizenship in the U.S. census will be greater than ever. However, we should be cautious about the implication of such a change for the integrity of the decennial census, particularly with respect to its constitutional mandate of providing an accurate count of the resident population for purposes of congressional apportionment.
The problem of an undercount based on a compulsory question on citizenship would be a massive one, given that non-citizens account for more than 22 million residents today, about 11 million of whom are unauthorized. Compounding the problem, more than a third of unauthorized immigrants live in mixed status households, which would lead to an even bigger undercount, this time involving U.S. citizens. This undercount would disproportionately hurt states with large cities and many Latino and Asian American communities.
Thus, even if Evenwel leads to new calls for counting U.S. citizens in the decennial census or other data collections, we should be cautious about its implications — not just in terms of financial cost but also in terms of data accuracy and the disproportionate burdens it would place on racial minorities.
Karthick Ramakrishnan is professor of public policy at the University of California, Riverside and author of “The New Immigration Federalism.”