Moving up the end date set off a wave of litigation from civil rights advocates who argued that it would cause an undercount of traditionally hard-to-count populations such as minorities, immigrants and low-income people, depriving them of services such as schools, roads, fire stations and hospitals, along with political representation. Businesses and analysts also rely on census data when deciding where to open new locations or make policy recommendations based on population numbers.
The purpose for the rush appears to be political. Meeting the Dec. 31 deadline helps ensure that regardless of November’s election outcome, the president will have an opportunity to fulfill a July memorandum in which he said he wanted to omit the undocumented from apportionment.
A government filing to expedite Supreme Court review of the case said: “Absent some form of relief from the judgment, the Secretary and the President will be forced to make reports by the statutory deadlines that do not reflect the President’s important policy decision concerning the apportionment.”
Removing undocumented immigrants from apportionment would be unprecedented in U.S. history, and critics say it is unconstitutional. It could also affect the makeup of the House and the distribution of electoral college votes, shifting representation from some more diverse states with large immigrant populations to more White ones. A Pew Research Center study this summer found that if the country’s undocumented immigrants were excluded from apportionment, California, Texas and Florida would end up with one less seat and Minnesota, Ohio and Alabama would end up with one more, compared with what they would have gotten with no adjustments.
Trump’s memo sparked a separate spate of lawsuits, including the case that the Supreme Court plans to take up in November. But even as litigation is underway, the government has not explained how it plans to identify and count undocumented immigrants, leaving some worried that reapportionment numbers could be calculated by the White House behind closed doors.
Howard Hogan, a former Census Bureau chief demographer, wrote in a court filing that “traditionally, the Census Bureau takes the Apportionment Counts, that is, state population counts, and applies a complex mathematical algorithm to allocate the 435 Congressional seats to the states.”
But bureau officials “state that the Census Bureau plans to give to the President the components of the Apportionment Counts,” he wrote. “The clear implication is that it would not be the Census Bureau that would compute the actual apportionment, but presumably someone at the White House. The risk to an accurate and fair apportionment [is] enormous.”
The law requires that apportionment numbers be based on an actual enumeration of people living in a state, not estimates. Because no comprehensive list of undocumented immigrants exists, attempts to count them could involve subtracting citizens and legally documented noncitizens from the total population count.
Hints about the process that the White House may use to exclude the undocumented have emerged in court documents. In a filing to the Supreme Court this month, acting solicitor general Jeffrey B. Wall wrote that the bureau planned to “provide the President with information regarding any ‘unlawful aliens in ICE Detention Centers’ whom the President could, consistent with the discretion delegated to him by law, exclude from the apportionment base, thereby partially implementing his Memorandum.”
But the number of people in Immigration and Customs Enforcement detention centers on April 1, the official Census Day, is probably in the tens of thousands and is probably not enough to affect representation in a meaningful way, said Dale Ho, director of the ACLU’s Voting Rights Project, who argued the case. “That’s not going to shift any seats,” he said. “It just seems to me that they must have more in mind.”
In an email to Commerce Secretary Wilbur Ross in late September, Census Bureau Deputy Director Ron Jarmin referred to a plan “to finish the processing of the resident population, federally affiliated overseas and, if requested, unlawful aliens in ICE Detention Centers by 12/31,” adding that other “outputs” related to the presidential memo would be pushed to Jan. 11.
In a statement this month, Al Fontenot, the bureau’s associate director for decennial census programs, said the bureau had postponed “the five days of processing needed to implement the Presidential Memorandum” to the first week of January.
The bureau and the Commerce Department did not respond to questions from The Washington Post about what the “outputs” are.
At a hearing in Maryland on Wednesday in a related suit, a panel of federal judges pressed the government to define the additional “outputs” and processing. Justice Department lawyer John Coghlan said the plan was “a dynamic process” and the bureau was “not in a position to answer that question right now.”
“How can you say we don’t know what the plan is, it’s dynamic, it’s changing, but we know it will take five days?” asked Judge Pamela Harris of the U.S. Court of Appeals for the 4th Circuit. “I am starting to feel that someone is not being honest with us.”
Coghlan told the panel that aside from people in ICE detention centers, “it’s unknown right now what additional categories could . . . potentially be feasibly identified by the secretary.”
Sounding infuriated, U.S. District Judge Paula Xinis noted: “So until the secretary reports the numbers to the president, it is a black box.”
After the government’s attempt to add a citizenship question to the survey was blocked last year by the Supreme Court, Trump issued an executive order directing federal agencies to share information on citizenship with the bureau. It is not clear how much information has been shared or how good the quality is. Only a handful of states complied with a bureau request to share information from state DMVs.
The bureau may be collecting information from other administrative records, including the Social Security Administration, the Internal Revenue Service, U.S. Citizenship and Immigration Services and the Department of Housing and Urban Development.
But that data may not accurately reflect someone’s immigration status or place of residence on Census Day.
“To my knowledge, administrative data sources don’t exist to give you a comprehensive picture for the undocumented populations,” said Amy O’Hara, a former Census Bureau official who is now executive director of the Federal Statistical Research Data Center at Georgetown University’s McCourt School for Public Policy.
Inaccurate or incomplete counts of the undocumented would not affect apportionment if the inaccuracy were spread evenly across states, said Tom Louis, a former chief scientist and former associate director for research and methodology at the Census Bureau. But the addition this year of four political appointees to the traditionally nonpartisan bureau worries him.
“Staff is not going to feel comfortable discussing issues and resolving them with those four around,” said Louis, who, along with Hogan, was on the task force that on Tuesday issued an American Statistical Association report calling for more transparency around the agency’s decennial Census activities. “My concern is that we’ll never know what the values were before the undocumented were removed.”
It is also not clear from the president’s directive who would meet the criteria to be excluded.
“Who is an unauthorized immigrant? It’s not particularly well-defined,” said demographer Jeffrey Passel, who co-authored the Pew analysis. Even after putting aside naturalized citizens, green-card holders, refugees and foreign students, he said, many of the remaining immigrants might have authorization to be in the United States, including people who have DACA or temporary protected status, or who have applied for asylum.
The number of undocumented immigrants is commonly estimated to be between 10.5 million and 12 million people. But Trump’s executive order last year cited a study that estimated the number at between 16.2 million and 29.5 million, sparking concerns that the government may try to exclude people from apportionment who are here legally.
Beyond the battle over apportionment, ongoing problems around the 2020 count have raised questions over whether the data will be accurate enough to rely on for the upcoming decade. Census numbers are also used to determine state redistricting and $1.5 trillion in federal funding each year.
Because of pandemic-related delays, the bureau said in the spring that it would need until Oct. 31 to conduct the count and until April 30 to deliver it. But the administration in August abruptly said the count would end Sept. 30 so that apportionment numbers could be delivered to Trump by Dec. 31. The change compressed six months of post-count analysis into less than half that time. That analysis, which includes verifying the data with other sources, is key to ensuring an accurate count.
The American Statistical Association, whose members include several former bureau officials, recently called for greater transparency “precisely because we believe that there will be resistance . . . to releasing the quality indicators so that an assessment can be made regarding the quality of the 2020 count,” said Robert Santos, vice president and chief methodologist at the Urban Institute and president-elect of the ASA.
“The 2020 counts are stacking up to be the most flawed in history, because of the confluence of everything that has occurred,” he said, including “political appointees at the top levels of the Census Bureau, and schedules that leave no time for any genuine quality assurance. It’s crazy. I’m very disheartened.
“We already know it’s going to be flawed, flawed substantially, to the point that you have to question how flawed it actually is before you can even consider using it,” he said.
Whenever the data from the 2020 Census is released, it will be subject to scrutiny. “A whole group of social scientists and demographers will be ready to grab those data and compare them to independent estimates,” Santos said.
A Census Bureau spokesman said the data for apportionment will be released the same time it is delivered to the president. An FAQ page posted by the bureau on Friday said it is “working hard to process the data in order to deliver complete and accurate state population counts as close to the December 31, 2020, statutory deadline as possible,” raising the possibility that it might deliver them past that date.
Apportionment numbers are scheduled to be presented to the new Congress in early January, and several things could happen next.
The U.S. House could refuse to accept the numbers, said Justin Levitt, a constitutional law professor at Loyola Law School. “You’ve got to produce a full count, legally, and if the president doesn’t, I don’t think that the House of Representatives has to accept anything else,” he said. “Imagine if instead of a full count, the president handed over a Burger King receipt. I don’t think the House of Representatives is required to say, ‘Oh, California gets two Whoppers.’ ”
Whether the House accepts the numbers Trump delivers, individual states might independently assert the right to contest the number of seats they were awarded, creating chaos and sparking new waves of litigation. Or, depending on the outcome of next month’s elections, the new Congress or a new president could ask for revisions, or even request that a new census be conducted.