The Washington PostDemocracy Dies in Darkness

Opinion The Trump administration’s census question degrades our data — and our democracy

Commerce Secretary Wilbur Ross, who added a citizenship question to the 2020 Census, testifies before a House committee last month. (Jose Luis Magana/AP)

It’s not enough that President Trump and his advisers have been arguing for years that official government data is bad, untrustworthy, phony, manipulated for political again. Now they are working to lend credence to these smears and conspiracy theories — by making them true.

Unless, that is, the Supreme Court intervenes.

During the Obama administration, Trump repeatedly claimed that official numbers released by our independent federal statistical agencies — such as the unemployment rate — were fake. Legions of career civil servants were all cooking the books to make Democrats look better, he claimed. Trump’s economic advisers and boosters (including Stephen Moore and Herman Cain, whom Trump now plans to nominate for the Federal Reserve Board) joined in the baseless conspiracy theorizing. As did some other high-profile Obama critics who should have known better.

Troublingly, it turns out a lot of other Americans are on board with this numerical nihilism. In a poll last fall from Marketplace and Edison Research, about 4 in 10 Americans said they either completely or somewhat distrust data about the economy reported by the federal government.

And since Trump has taken office, he has worked to justify such distrust by actively degrading the quality of data — specifically, by seeking to make the 2020 Census less accurate.

The Trump administration wants to add, at the last minute, a new question to the census. I say “last minute” because usually new survey questions go through years of research, field-testing and public comment, as required by law and federal regulations.

This is to make sure that, among other things, any changes will not disrupt the accuracy of an enumeration mandated by the Constitution.

“It’s pretty well known that when you change the context of a data-collection instrument, unexpected things can happen,” said John Thompson, a former director of the U.S. Census Bureau. “The only way to understand what’s going on is to test it.”

The question the administration wants to shoehorn in without this process turns out to be particularly disruptive: It asks about citizenship. Given rising levels of government distrust among immigrant and ethnic minority populations, the question could be reasonably expected to depress response rates among these groups and lead to significant undercounts or otherwise inaccurate data.

In fact, in unrelated survey testing in 2017, respondents told census workers that they fear how their data might be used against them or their loved ones. They expressed concerns about the “Muslim ban,” anxiety over “registering” household members and the dissolution of the Deferred Action for Childhood Arrivals program. Some falsified their names, birth dates and other demographic information.

Despite such warning signs, as well as an explicit recommendation against inclusion of a citizenship question by experts within the Census Bureau and six former bureau directors  appointed by Republicans and Democrats alike, the Trump administration barreled ahead.

Three federal courts have so far blocked the question, finding that the administration violated administrative law. Two of those courts also found it violated the Constitution. Next week, the issue heads to the Supreme Court.

So what happens if the Supreme Court sides with the Trump administration instead?

In the near term, the consequences could be severe. Hundreds of billions of dollars are allocated annually based on the decennial census. Congressional seats are apportioned, and districts are redrawn. Perhaps not coincidentally, blue states are likely to be the biggest losers in both dollar terms and political representation if, as expected, this new survey question results in significant undercounts of immigrant and Latino populations.

The decennial census data is also the baseline against which virtually all other surveys are calibrated. Which means that whatever its motives, the administration’s innumeracy is likely to skew all sorts of other critical information that government agencies use to evaluate economic trends and health epidemics; that businesses rely on to decide how much to invest and hire and where; and that workers and families use to determine where to live, what to study, how much to spend on a home. 

Even giving the public reason to believe the numbers have been either manipulated or mismanaged will cause people and businesses to make worse choices. Just like the Fed — and other apolitical institutions Trump has been working to sabotage — our statistical agencies must be free of political influence both in practice and perception to be useful.

And that’s the longer-term risk here. One basis of a democracy — not to mention a healthy economy — is good official statistics so that the people and their representatives can make informed decisions. By throwing the numbers into doubt, the administration jeopardizes our democratic and economic health, not only today, but for many years to come.

Read more:

Ronald L. Wasserstein: The Trump administration’s statistical malpractice on the census

The Post’s View: This should be the final answer to the census question

Carolyn B. Maloney: How the Trump administration tries to skew democracy by attacking the census

The Post’s View: The Trump administration seems to be doing everything it can to bias the 2020 Census

Karen Tumulty: The Trump administration’s plan to make people disappear