The Washington PostDemocracy Dies in Darkness

In Md. census trial, Trump administration tries to show citizenship question would not harm the 2020 count

January 31, 2019 at 7:04 p.m. EST
(Michelle R. Smith/AP)

Testimony wrapped up Thursday over the Trump administration’s addition of a citizenship question to the 2020 Census as government attorneys sought to show it would not harm the accuracy of the count.

In the second week of trial at U.S. District Court for the District of Maryland in Greenbelt, the Census Bureau’s chief scientist, John Abowd, was called to testify by both sides.

Abowd, who has testified in similar trials in New York and California, told government lawyers that although the citizenship question would be likely to produce a drop in the initial ­self-response rate and make the count more costly, the undercount could ultimately be mitigated by census enumerators doing a Nonresponse Followup Operation (NRFU).

But when questioned by plaintiffs’ lawyers, Abowd said that even if the households that failed to initially respond could ultimately be counted by the NRFU, adding the question would irreparably harm the accuracy of many of those responses. “The increase in cost and the degradation of the data cannot be remediated by NRFU,” he said.

During two days of testimony, his opinion on adding the citizenship question was consistent — and stark.

“As the chief scientist of the Census Bureau, you do not think that adding a citizenship question to the 2020 Census is a good idea, correct?” asked Karun Tilak, an attorney at Covington and Burling, which represents some of the plaintiffs.

“That’s correct,” Abowd said.

At the core of the challenge is the premise that asking about citizenship on the constitutionally mandated decennial survey of every household in the United States will scare people who are not citizens or who live with noncitizens, resulting in an inaccurate count. Census data is used to dole out federal funds, apportion congressional seats and draw districts. It is also used by private and public organizations for planning and analysis.

The Maryland trial, which addresses two lawsuits, is the third seeking to block the question. A federal judge in New York ruled Jan. 15 that the government must stop its plans to add the question. In that case, the government appealed to the Supreme Court to quickly bypass its normal procedures and rule before the forms go to the printer this summer. A similar trial challenging the question is underway in California. The government has sought unsuccessfully to have all three dismissed.

In each case, the government is accused of failing to follow the Administrative Procedure Act, which governs the process by which federal agencies develop and issue regulations.

But one of the Maryland lawsuits also accuses the government of conspiracy “to depress the count of immigrant communities of color, thereby decreasing this population’s impact on and benefit from apportioned political power,” and charges that Commerce Secretary Wilbur Ross “engineered the Voting-Rights-Act rationale with the assistance of the Department of Justice to cloak Defendants’ true purpose.”

In court Thursday, Burth Lopez, an attorney with the Mexican American Legal Defense and Education Fund, which represents several plaintiffs, argued that the decision to add the question “was made not by Secretary Ross by himself, but by a group of decision-makers near the beginning or at the beginning of the Trump presidency” and that it “was moved forward over a year later by an announcement by Secretary Ross.” Government records indicate that former White House adviser Stephen K. Bannon and former Kansas secretary of state Kris Kobach consulted with Ross on the question.

In announcing the addition of the question in March, Ross said it would help enforce the Voting Rights Act. A Justice Department spokesman declined to comment on the trial Thursday.

The other Maryland case says the question would harm people — including U.S. citizens — living in areas such as Prince George’s County that have a high proportion of immigrants and minorities and are vulnerable to being undercounted.

Last week, attorneys for the plaintiffs called upon public policy, statistics and survey experts, as well as immigrant community leaders, who said adding the question is likely to depress response rates and harm the count.

In addition to Abowd, the government called Stuart Gurrea, senior vice president of Economists Inc., a private consulting firm, who said analyses that the plaintiffs’ witnesses had performed or incorporated into their testimony was “unreliable” and that it overstated the perils of adding the citizenship question.

Plaintiffs’ attorneys grilled Gurrea on his credentials, noting that he had no experience working with census issues, demography, reapportionment or redistricting, or survey methodology, and no experience working in government.

Trump administration lawyers also sought to block evidence including tweets by the president that negatively characterized immigrants and claimed that noncitizens were voting in U.S. elections.

The attorneys said the tweets were irrelevant, but Judge George J. Hazel said he would allow them as they “could show that to the extent that Secretary Ross is doing something, he’s doing something because his boss holds this view.” But Hazel added that their relevance was “marginal at best, because I’ve seen no evidence that Secretary Ross saw or was motivated by these tweets.”

Closing arguments in the Maryland citizenship question trial are scheduled for Feb. 21, and in the California trial for Feb. 15.

Abowd said the current political climate around immigration would be likely to increase the number of people who do not respond to the survey.

Abowd said a projected 5.1 percent difference between the effect of the question on all-citizen households vs. households with some noncitizens was sufficient for bureau senior executive staff to recommend against it. In addition, a report by the bureau last week found many Hispanics to be wary of responding to the survey, even if they are citizens, when people in their households are not.

A poll released Thursday by Quinnipiac University found that Hispanics and immigrants in New York City said they would be less likely than others to fill out the census.