Neil Gorsuch thinks there is nothing unusual about a citizenship question on the 2020 Census. Here’s why he’s wrong.

Last night, the Supreme Court delivered a small but important victory to the Trump administration in the ongoing war over the 2020 Census. In a terse, four-page opinion, Justice Neil Gorsuch temporarily blocked Commerce Secretary Wilbur Ross’s deposition in a case dealing with the Trump administration’s introduction of a citizenship question on the upcoming Census form.

The scope of the ruling was narrow, but the reasoning Gorsuch deployed augurs poorly for those concerned about the politicization of the 2020 count. Specifically, Gorsuch argued that the district court was wrong to level an “extraordinary claim of bad faith” against the Commerce Department:

There’s nothing unusual about a new cabinet secretary coming to office inclined to favor a different  policy direction, solacing support from other agencies to bolster his views, disagreeing with staff, or cutting through red tape. But until now, at least, this much has never been thought enough to justify a claim of bad faith and launch an inquisition into a cabinet secretary’s motives.

But there are some unusual things about Ross’s decision to add a citizenship question to the 2020 Census. For starters, he appears to have been less than truthful with Congress about his motivations for adding the question. More importantly, there hasn’t been a citizenship question on the Census since 1950—prior to the Voting Rights Act and prior to the Supreme Court’s recognition of the “one person, one vote” principle. Indeed, while the Commerce Department contends that there is an “unbroken history” of asking about citizenship on the Census, this is untrue. The citizenship question has only been included on the American Community Survey, a long-form version of the Census sent only to a representative sample of households. Recent Census Bureau research has shown that the addition of such a question could cause an undercount of both documented and undocumented immigrant populations. Further, at the same time that Ross proposed a major change to the Census form, budget shortfalls have limited the Census Bureau’s efforts to lay the groundwork for the 2020 count. In 2017, the Bureau cancelled its field test. A year later, it was forced to drastically scale back its 2018 “End to End” test. Studies by the Government Accountability Office have revealed that inadequate budgets have limited  efforts at enumerating “hard to count” populations by offering non-English language response options and recruiting partner organizations in local communities.

The Census Bureau’s Center for Survey Measurement found fears about responding to citizenship status questions could drive an undercount.

Considering that Article 1 of the Constitution requires Congress to conduct an “actual Enumeration” of the “whole number of persons”, all of the above might be cause for stricter scrutiny of Ross’s actions. One challenge, however, is that the Constitution gives Congress the power to conduct the Census “in such Manner as they shall by Law direct.” Congress, in turn, has delegated authority to the Secretary of the Department of Commerce to take a census “in such form and content as he may determine.” So it might seem that, as Gorsuch argues, there is nothing unusual about a simple change in policy from administration to administration.

Yet the Court has not always given wide latitude to the Commerce Department’s decisions about Census enumeration. Most notably, in the 1999 case of Department of Commerce v. U.S. House of Representatives, the Supreme Court ruled that the use of statistical sampling was inconsistent with the Census Act. In a concurring opinion, the late Justice Antonin Scalia went further to suggest that statistical sampling violated Article I’s requirement for an “actual Enumeration.”

Moreover, the Commerce Department is not bound by the Census Act alone. Congress has designed a number of statutory guardrails that govern the Census. The Information Quality Act, passed by Congress in 2001, requires the Census Bureau to employ data collection standards that minimize respondent burden and achieve the highest possible rate of response. Census Bureau research suggests that adding a citizenship question will do the opposite. Under the Administrative Procedure Act of 1946, agencies may not take arbitrary or capricious actions. There is, however, little evidence to support the Commerce Department’s argument that the citizenship question is necessary for enforcing Section 2 of the Voting Rights Act. The decision also contradicts Census Bureau research and decades of agency decision-making. In short, the courts have more than a few reasons to be concerned about the 2020 Census.


The Trump administration’s actions on the 2020 Census are unusual. Differential undercounts of minority populations, however, are not. Communities of color, low-income households and immigrants are disproportionately likely to be undercounted in the Census.

Differential Undercounts in US Census, 1990–2010

As Margo Anderson highlights in The American Census: A Social History, the issue of the undercount first came to a head in 1960 when––some twenty years after the Bureau began collecting the statistics––it was revealed that one fifth of nonwhite men age 25-29 had been left out of the Census. By 1967, Harvard professor (later Senator) Daniel Patrick Moynihan organized a conference of policymakers and academics to address the issue. As Moynihan put it:

The full enumeration of the American population is not simply an optional public service provided by government for the use of sales managers, sociologists, and regional planners. It is, rather, the constitutionally mandated process whereby political representation in Congress is distributed as between areas of the nation. It is a matter not of convenience but of the highest seriousness, affecting the very foundations of sovereignty.

Moynihan’s conference helped to publicize the undercount issue. Yet a growing array of civil-rights groups, joined by representatives of local governments, kept the issue alive through the 1970s. Executive-branch institutions also cued policymakers’ attention to the undercount. In 1974, the US Commission on Civil Rights released Counting the Forgotten, which accused the Census Bureau of botching its count of Spanish-speaking populations, and urging it to adopt a Spanish-only form.

While the Census Bureau faced a growing number of legal onslaughts surrounding the undercount, fixing the problem was not so simple. One issue was technical capacity. The Census Bureau had been developing a system for correcting undercounts through “dual-system estimation” since 1950. Yet even when the Bureau developed more sophisticated approaches for correcting the undercount, officials remained cautious about employing them. Congressional and judicial skepticism of statistical sampling did not help. In the absence of statistical techniques for correcting the undercount, the Census Bureau has relied heavily on techniques to ensure census mobilization. Census mobilization is a costly business, however, and one that the Trump administration has demonstrated little interest in funding.


Carrying out an accurate Census is a sophisticated technical task and the Census Bureau is a scientific agency with the authority only to collect and analyze data. Yet the origins and development of the Census are the product of American politics. The simple reason for this is that counting the population quite literally constructs the American polity—defining congressional representation and resource allocation. Several political developments have enhanced the zero-sum character of Census politics. First, because Congress has not expanded the size of the House of Representatives since 1910, states whose populations decline or whose inhabitants are undercounted are likely to lose representation in Congress. Second, Congress uses Census Bureau data to determine the allocation of more than $675 billion in grant dollars. Especially when budget cuts are on the table, an undercount or population loss means a diversion of vital resources. Potential losses in resources and representation led to a series of Supreme Court disputes over the 1990 and 2000 counts—not to mention a number of colorful television ads.

Yet the Trump administration’s efforts to introduce a citizenship question are about more than the politics of congressional apportionment and federal grants. Rather, they reflect a deeper legacy of nativist politics that has crossed paths with the Census before. The influx of immigrants into urban areas in the early twentieth century—first revealed in the 1920 Census—linked together rural politicians interested in preserving their power in Congress with nativists bent on immigration restriction. Among those nativists was former Census superintendent Francis Amasa Walker, who declared that Southern and Eastern Europeans were “beaten men from beaten races; representing the worst failures in the struggle for existence.” At the same time, rural congressional delegations recoiled at the evidence—which meant that close to a dozen rural states would lose members of their congressional delegation.

Rural members of Congress rhetorically knitted their regional and racial interests together. As Rep. Edward C. Little (R-KS) warned: “It is not best for America that her councils be dominated by semicivilized foreign colonies in Boston, New York, Chicago.” Others rejected the validity of the Census itself, citing unpaid and inexpert Census employees and anomalies resulting from the timing of the count, which was taken in the aftermath of World War I. Under pressure from a bipartisan coalition of rural members and immigration restrictionists, Congress failed to reapportion itself for a decade–expanding the outsized influence of rural areas in American political life. As Charles Eagles documents in his account of the controversy, Democracy Delayed, rural politicians blocked reapportionment throughout the 1920s by stoking nativist fears, claiming that the enumeration of resident aliens “vitiated the morality of the apportionment.” Urban delegations dug in against such claims, arguing—as did Fiorello La Guardia—that aliens were rapidly naturalizing, except where blocked by the “Ku Klux Klan [members] administering the naturalization department.” Ultimately, Congress did reapportion after the 1930 census, but it denied Americans fair representation for a decade.

Nearly a century has passed since the last apportionment crisis. And much has changed. Following the Great Depression, Congress outfitted the Census Bureau with greater capacities to meet the demands of a complex industrial society. Numerous developments in civil-rights legislation as well as key Supreme Court rulings on apportionment have heightened the demand for a fair and accurate count of the population. Yet controversies over differential undercounts have continued. If anything, the Trump administrations’s brand of nativist politics could catalyze already-existing tensions about the allocation of resources and representation, making the 2020 Census the site of a new political crisis. If that happens, the Supreme Court will assuredly be called upon to intervene. Whether the Court will be able to settle anything is another matter.

Leave a Reply

Your email address will not be published. Required fields are marked *