Symposium: Facts, not fears, should control outcome of census citizenship question case

Mithun Mansinghani serves as solicitor general for the state of Oklahoma, which filed an amicus brief joined by 16 other states in support of the petitioners in Department of Commerce v. New York.

Pundits, politicians and the media have warned the public for years about the dangers of fear-based politics. In one heavily publicized moment in 2010, comedian-commentators Jon Stewart and Stephen Colbert went so far as to hold a half-satirical rally calling for a restoration of sanity in government and an end to a culture of fear in politics. Nearly a decade later, it’s clear that the use of fear as a partisan tool is here to stay.

The latest attempt to leverage trepidation for political purposes — the challenge to reinserting a question about citizenship in the 2020 census — has made its way to the U.S. Supreme Court. But if the court focuses on the facts of the case, rather than the fear-based myths that have arisen around the citizenship question, upholding the restoration of this census question should be an easy task.

So let’s tackle some myths about the citizenship question.

MYTH: The government can use the citizenship question to find and deport illegal immigrants.

FACT: By law, the Census Bureau cannot disclose anyone’s responses or share data from which an individual can be identified. Violation of this law comes with stiff criminal penalties, including five years in prison. Not even federal and state law-enforcement agencies can access individual census responses. On a more practical level, a person answering that they are not a citizen does not give any reliable information as to whether they are lawfully present. Fewer than 12 million of the more than 44 million immigrants in the United States are undocumented. So even if a person identified as a noncitizen on the census, and federal agents risked being thrown in jail by disclosing that response, law enforcement could not reliably use the census to determine that person’s legal status.

This pernicious myth, spread by those who benefit from the fears it generates, appears to undergird the primary concerns with the citizenship question. We would all do well by actively informing our local communities that there can be no legal repercussions for truthfully answering any census question. Instead, it is failing to respond to the census, or responding falsely, that carries with it legal consequences.

MYTH: Asking about citizenship on the census is unusual.

FACT: Most censuses throughout U.S. history have asked about citizenship in one form or another. The practice goes back millennia, to ancient Athens and beyond, and continues in countries across the globe. The citizenship question began in the United States with the 1820 census and was asked consistently through the 1950 census (with the exception of 1840). The 1960 census did not ask about citizenship, but did ask place of birth. From 1970 to 2000, a citizenship question was asked on the “long-form” census, which was the census sent to about one in six households instead of the shorter census. The “long-form” was eliminated on the 2010 census, but the citizenship question continues to be asked on a yearly Census Bureau questionnaire called the American Community Survey (ACS). Tallying all of these up, the Census Bureau has asked U.S. residents about citizenship over a billion times.

MYTH: The citizenship question is untested.

FACT: It is strange to claim that a survey question used over a billion times is untested. In fact, the Census Bureau’s chief scientist, who has served in that role since the Obama administration, informed Secretary of Commerce Wilbur Ross that “[s]ince the question is already asked on the American Community Survey, we would accept the cognitive research and questionnaire testing from the ACS instead of independently retesting the citizenship question.”

MYTH: Having better data on citizenship serves no useful purpose.

FACT: Citizenship data is critical to ensuring compliance with Supreme Court precedent on the Voting Rights Act. Without such information, it may be impossible to create districts designed to provide minority communities the opportunity to elect the candidate of their choice. Because Hispanic communities, for example, often contain many noncitizens who cannot vote, it is not enough that Hispanics hold a bare majority to elect their preferred candidate — they must comprise a majority of voting-aged citizens in the district. Lack of citizenship data could actually hurt minority communities.

Although the ACS provides some citizenship data, there is widespread agreement that this data is highly flawed. Even some of the states suing the federal government in this case recently complained about the ACS to the Supreme Court. The ACS is sent only to about 2.5 percent of households on a rolling basis. It is therefore less accurate than census data, it provides estimates with high margins of error for only very large slices of the population (and thus is difficult to use in redistricting smaller communities), and it is often incompatible for use with other census data. Moreover, these inaccuracies mean that ACS data is subject to interpretation — and manipulation — that mires the states in litigation over whether minority-opportunity districts were appropriately drawn. Combined with other methods that the Secretary is proposing, the census citizenship question will provide the best citizenship data in decades for use by states, the Justice Department and voting rights advocates alike.

MYTH: The empirical data presented to the Secretary of Commerce proved that the citizenship question will cause millions to go uncounted in the 2020 census.

FACT: Asking about citizenship on the ACS, as well as on private-sector surveys, has had few significant effects on minority community response rates. Yet the challengers in this lawsuit claim that a citizenship question will cause 5.1 percent of noncitizens to refuse to respond to the census, based on data presented in internal Census Bureau memos. But a closer look at these memos reveals the flaws in this interpretation. This requires diving into some numbers, so bear with me.

The 5.1 percent figure results from a comparison between the response rate to the 2010 census (which did not include a citizenship question) and the 2010 ACS (which included a citizenship question). Although the nonresponse rate to the 2010 ACS was greater than that for the 2010 census with all groups, the drop in the noncitizen response rate was 5.1 percentage points larger than the drop in the citizen response rate. So, the assumption goes, the citizenship question will cause 5.1 percent of noncitizens to abandon the census.

But it’s just an assumption — and a wildly irresponsible one at that. This conjecture ignores many other reasons why noncitizen responses decreased at a higher rate in the 2010 ACS. The ACS contains more than five times as many questions as the 2010 census, which undoubtedly affects response rates, especially in households that lack the resources to respond to long surveys. And any one of those many extra questions, aside from the citizenship question, could have caused nonresponse. The census also involves greater outreach and follow-up, which may not be as effective in noncitizen communities, leading to greater nonresponse to the ACS. These are just a small sampling of the many alternative explanations that demonstrate how reckless it is to assume that the entire 5.1 percent differential decrease in response among noncitizens is attributable to the citizenship question alone.

Meanwhile, other data presented to the secretary — such as data on how many people taking the 2016 ACS survey online stopped when they reached the citizenship question — shows a far smaller disparity in response rates among Hispanics. After accounting for normal follow-up procedures the bureau uses to count people who initially fail to respond to the census, the undercount caused by the citizenship question may be as low as 0.001 percent of the population.

MYTH: The Constitution prohibits any practice that might reduce response rates to the census.

FACT: The challengers in this case argue that, because the Constitution requires an “actual Enumeration” of the population every 10 years, any census that leads to an undercount is unconstitutional. But the census since the beginning has always gathered more data than a simple enumeration, and every census has suffered from an undercount. The 2010 census disproportionately undercounted Hispanics, all while the Obama administration was deporting a record number of illegal immigrants — yet no one sought to declare as unconstitutional the 2010 census question singling out Hispanic origin.

Because no census will ever be perfect, and because Congress has the well-recognized power to conduct the “Enumeration … in such Manner as they shall by Law direct,” courts for centuries have rejected this constitutional challenge to a citizenship question. And although the California court in this case accepted this “enumeration clause” claim, the New York court appropriately threw it out.

In the end, better census data is better for everyone. We shouldn’t be trying to suppress it. Instead, we should be actively encouraging those around us to respond accurately to all census questions, including those about citizenship. Any less from those challenging the citizenship question, who are spreading panic about its inclusion, would be a self-fulling prophecy — creating the very undercount they seek to avoid. The Supreme Court should reject this path of fear and allow the census citizenship question to go forward as it has for almost 200 years.