Jurisprudence

The Census Case Is a Test of Whether the Truth Still Matters at the Supreme Court

U.S. Commerce Secretary Wilbur Ross testifies on Capitol Hill on March 14 in Washington.
U.S. Commerce Secretary Wilbur Ross testifies on Capitol Hill on March 14 in Washington. Mark Wilson/Getty Images

On Tuesday, the Supreme Court will hear arguments in Department of Commerce v. New York, a challenge to the Trump administration’s addition of a citizenship question to the 2020 census. At the heart of the case lies a simple inquiry: Can unelected bureaucrats violate the law, lie about this misconduct, and then get away with it? The answers will depend on the justices’ willingness to favor truth over falsehoods. Indeed, while Department of Commerce involves a number of complex legal disputes, they all boil down to the question of whether facts still matter in this Supreme Court.

The fundamental purpose of the decennial census, as mandated by the Constitution, is the “actual enumeration of the people”—not just citizens, but everybody living in the United States. Secretary of Commerce Wilbur Ross wants the 2020 census to ask every individual whether they’re an American citizen for the first time since 1950. Whether the reinsertion of this question would so corrupt the “actual enumeration of the people” as to violate the Constitution is an interesting conundrum. But the court need not decide it to rule against Ross. That’s because the Administrative Procedures Act prohibits “arbitrary and capricious” agency actions, and the conduct of Ross and his associates is a quintessential example of arbitrary and unscrupulous decision-making.

Consider the chain of events that led to the citizenship question’s inclusion, laid out by federal Judge Jesse Furman in his 277-page opinion blocking its inclusion on the census. On March 26, 2018, Ross issued a memo asserting that he was going to add a citizenship question to the 2020 decennial census. In his memo, and again in testimony before Congress, Ross falsely stated that his decision was motivated only by a request from the Department of Justice for better data with which to enforce the Voting Rights Act. But Ross lied. He had made up his mind to add a citizenship question almost as soon as he took office, and his aides solicited DOJ to request the citizenship question to reverse-engineer a rationale and a record that they hoped would withstand scrutiny from the Supreme Court. (In fact, DOJ has been able to enforce the Voting Rights Act successfully since its passage in 1965 without the aid of a citizenship question.)

Career data scientists at the Census Bureau were asked to evaluate whether a citizenship question was a good way to get the data that DOJ had supposedly requested for VRA enforcement. They unanimously said no. Consistent with the Census Bureau’s position for decades, under directors appointed during Republican and Democratic administrations, the bureau’s data scientists told Ross that adding the citizenship question was a bad idea. The decennial census’ most important function is the “actual enumeration” of the population. That enumeration is used to apportion congressional districts among the states and also to allocate roughly $900 billion in federal funding. The evidence was clear: A citizenship question would disproportionately harm response rates among Hispanics and noncitizens. The result would be an undercount of the population in states with large populations of those groups—depriving the populations of these states of political power and vital federal resources.

Census Bureau data scientists told the secretary that substantially more accurate data on citizenship could be gathered at lower cost through the use of administrative records from the Social Security Administration and U.S. Citizenship and Immigration Services. Because the accuracy of enumeration is so important, and adding questions to the Census diminishes response rates, the Census Act has required the use of administrative records to the maximum extent possible instead of adding questions since 1976. Moreover, the citizenship question hadn’t been pretested in accordance with well-established bureau standards. In other words, all of the evidence and all of the policies weighed against asking a citizenship question. But Ross dismissed all of it and pushed ahead with the citizenship question anyway.

This is a textbook APA violation—and that should matter, particularly to conservative justices. The Roberts court has generally been strict about ensuring that executive agencies comply with the APA. The Department of Justice has argued that the court shouldn’t even review Ross’ decision, which is absurd: There is a strong presumption that courts can review agency decisions. As Chief Justice John Roberts reminded the government in Weyerhaeuser v. Fish and Wildlife Service, “legal lapses and violations occur, and especially so when they have no consequence. That is why this Court has so long applied a strong presumption favoring judicial review of administrative action.”

Yet the DOJ insists that Ross has untrammeled discretion to manipulate the census—powers the department describes as “at least” as vast as the CIA’s director authority to fire secret agents. This is absurd. As Justice John Paul Stevens wrote in Franklin v. Massachusetts: “While the operations of a secret intelligence agency may provide an exception to the norm of reviewability, the taking of the census does not. … The reviewability of decisions relating to the conduct of the census bolsters public confidence in the integrity of the process and helps strengthen this mainstay of our democracy.” The DOJ’s position would immunize the census from judicial scrutiny, creating a back door for partisan manipulation of the apportionment of political power by an administration—which is exactly what seems to be going on here. At oral argument before Furman, Ross’ lawyers took the position that a court could not review a decision to add questions such as “whether and how many guns people owned,” or “who you voted for in the last presidential election.” The Supreme Court should not let Ross turn the Census Bureau into a pseudo-CIA that can design and deploy the census as a political weapon to undercount an administration’s disfavored classes.

Finally, the Justice Department suggested that Ross’ decision should be upheld because the evidence against the citizenship question was not “definitive.” That’s silly. All of the quantitative evidence weighed against using a citizenship question. True, there wasn’t a formal test of a citizenship question to give a “definitive” answer on whether it will provoke an undercount. But that’s Ross’ own fault: He chose not to test the question, presumably because all available evidence suggested that it would validate the consensus against the question long-held by data scientists and demographers. The Supreme Court should not hand Ross a victory because he managed to suppress further research proving the dangers of a citizenship question.

And that research is damning. A study by the Harvard Kennedy School’s Shorenstein Center, which surveyed 9,000 people, estimated an even more dramatic effect: It found that the citizenship question could lead to an undercount of more than 4 million Hispanics. The Justice Department stipulated a drop in self-response of 6.5 million people. And the Census Bureau itself said it would lead to a 5.8 percent decline in response rates among noncitizen households at a minimum.

To the Trump administration, this undercount wouldn’t be an unfortunate side effect—it appears to be the entire point of the citizenship question. As the Harvard study illustrated, many Hispanics and immigrants are fearful of this question, afraid that it might help the Trump administration identify unauthorized immigrants, their families, and households. An undercount could cause states with large populations of Hispanics and noncitizens—including Arizona, California, Florida, Illinois, and New York—to lose seats in Congress, votes in the Electoral College, and billions in federal funding.

The Supreme Court, however, need not wade into these political considerations to resolve this case. It need only decide whether Ross’ actions were “arbitrary and capricious.” Every fact on the record, every iota of evidence, proves that they were. If truth still holds sway at the Supreme Court, this case should be an easy one. Ross lied over and over again. Those lies were meant to conceal the arbitrary, insidious nature of his crusade for a citizenship question. And if the Supreme Court pretends to believe them, it will become complicit in the Trump administration’s most flagrant and bungled cover-ups.

Disclosure: Perry Grossman helped litigate this case with the New York Civil Liberties Union.