2020 Census: History, Justice, Representation

The citizenship question will not appear in the 2020 Census. But has the very partisan debate surrounding the question fundamentally changed the way the census will be conducted?

By Robert Nolan July 23, 2019

Carnegie Reporter Fall 2019 Census

From the days of the Founding Fathers, the American census has provoked questions about who should — and who shouldn’t — be counted. But turmoil over a question about citizenship has turned the exercise, constitutionally mandated to take place every 10 years, into a political football like never before. Kenneth Prewitt, former United States Census Bureau director and Carnegie Professor of Public Affairs at Columbia University, talks with Carnegie Corporation of New York’s Robert Nolan.

Many people had no idea about what the census was until it began popping up in headlines over the past six months, mostly because of a controversial question the Trump administration wanted to add about citizenship. What do you tell people at cocktail parties when they ask you about why the census is so important?

Article I of the Constitution requires a census every ten years; the first was conducted in 1790. Indeed, the census was one of the first things the country’s founding fathers mandated. It differed sharply from the Colonial censuses, with which of course, they were familiar. Those censuses were tools for the powerful — control and tax the population, exploit natural resources, and so forth. The American census flipped this 180 degrees, building on the Boston Tea Party cry: “No taxation without representation.” In fact, for the founders, the right to be represented is as fundamental as the right to vote, the latter restricted to adult white male property owners but the former for all women and children, and indeed all people residing in the initial 13 states. This “right to be represented” soon moved westward, adding new states — Kentucky and Tennessee — in the early 1790s, securing the Jeffersonian vision of a free and independent Republic.

How does this right to be represented manifest in the real world today?

Two ways. First, of course, seats in the House of Representatives are allocated proportional to population — the more populous the state the more representatives it sends to the capital. The second, a 20th-century development, is in redistributing federal taxes, now sustaining social services, back to the states. The 2020 Census counts will be used over the next decade to return to the States nearly a trillion dollars allocated to health care, disaster preparedness, transportation projects, benefits for veterans, and much more. The funds, like political representation, will be allocated proportionate to the respective size of each state, as counted by the census. The right to be represented is manifested, then, in power and money.

Incidentally, census numbers have a life across the decade. If a state is shortchanged in the apportionment of congressional representation or a school system is shortchanged in new construction funds, it’s not a one-year problem but a ten-year problem.

A state, city, or even neighborhood that is undercounted does not get its fair share of federal health or education programs. The children who are left out of the census are of course still there, and need a seat in the classroom and a teacher.

We know that the census began recording whether those surveyed were native-born or foreign-born in 1850. How is the addition of a question about citizenship status so different from that? Wouldn’t we want to have an accurate head count of American citizens?

Oh, absolutely. The census form that is getting so much attention right now only has eight questions. A large number of critical characteristics are not on the decennial form, because the census bureau has less expensive and more timely sources — especially the American Community Survey and, increasingly, federal and state administrative records. Members of Congress who are asking “Why would we not want to know how many American citizens there are?” — they are either deeply misinformed or just making political points. They could go to their computers and in seconds get very reliable statistics on the number of citizens and noncitizens, and among the latter, the number of documented and undocumented.

We have low-turnout elections, but they still send candidates to political office, from city councils to the White House. There is no such thing as a low-turnout census. If there is a low turnout, it's not a census anymore.

I would like to make a more general point. There is extensive information about America’s population recorded in administrative records. The federal statistical system — 15 agencies and nearly 100 smaller programs — is steadily improving its use of those data, largely through advances in technology and the number of characteristics electronically available — health, employment, living conditions. There are challenges. One is ensuring the quality and coverage of administrative data. The other is public concern with privacy. However, every census is used to improve the census that follows it. The 2020 census will help us learn the strengths and weaknesses of administrative data, and how to apply sophisticated statistical tools in privacy protection. I am confident that 2030 will apply lessons learned in 2020, leading to increased use of administrative data.

The Trump administration has said it wants to include the question on citizenship to help better enforce the Civil Rights Act and other laws that protect minorities. While the Supreme Court rejected this justification, there is a real relationship between the census and civil rights. Can you help to explain that?

The key issue regarding the census and civil rights is the Voting Rights Act, as well as the administration of all the antidiscrimination laws that were created in the 1960s and then laws, especially gender equality, that have been gradually enacted over the years since. Civil rights and social justice laws very often require a denominator. [Editor’s note: a denominator refers to a specific set of data about the population.] I oversimplify to make the basic point. Consider the types of questions asked that launched the civil rights era a half-century ago: if the African American population is 12% but they comprise only a few percent of university student populations, we have an unfairness problem; if women are half the population but nowhere near that proportion of elected officials, we have an unfairness problem.

Justice is more than proportionate representation, but certainly such representation is relevant to social justice. Proportionality requires a denominator, and the census is the most powerful denominator we have. The social justice advances in the last half-century draw from data provided by the federal statistical system — on employment, wages, access to health, education opportunities, livable housing. All of these survey-based statistics are calibrated to the basic count provided by the census.

What are some of the potential outcomes in the run-up to the census, and what if it basically fails?

We’ve never had a census that did not start with the general good will of the public and the full backing of the political leadership. Neither of those conditions are a certainty for 2020. Whether the country’s going to pull itself together and have a good census in eight months depends on improvement on both counts.

Consider this. In preparing for 2020, the Census Bureau (2018) took a survey, with these results:

• 60% of the public mistrusts the federal government
• 53% percent believe the census is used to locate the undocumented
• 28% doubt that the Census Bureau will keep their answers confidential
• One in five Americans believe census answers can be used against them

I put it this way: we have low-turnout elections, but they still send candidates to political office, from city councils to the White House. There is no such thing as a low-turnout census. If there is low turnout, it’s not a census anymore. While I don’t expect the 2020 census to “fail,” I do think it requires and deserves stronger support than it is presently getting from the administration and from the general public. The Census Bureau is not the problem — the support system is. Can this turn around? Absolutely, and hundreds of civic organizations, state and local governments, businesses, and advocacy groups will step up — as so many already did, successfully, in the wasteful and needless battle over the citizenship question. I have no crystal ball. But if the political rhetoric can be dialed down, there is a civic mobilization effort poised to protect the census from toxic partisanship, and redefine it in terms the Founding Fathers would recognize.

In the 2000 and 2010 censuses, one of the big issues was ethnic diversity. I read that in the last census, 20 million Americans chose “other” as their race. How important is it that the census evolve with the rapid pace of immigration and ethnic identity? And what about sexual identity? Is that the next challenge?

The Census Bureau expected to introduce an extensively tested and much stronger race/ethnicity question in 2020 than the version used in 2000 and 2010. It was designed with many more options for national origin, ethnicity, and race. It made the “other” line redundant. I am hopeful that we will not have to wait until 2030 to introduce this much improved question. Instead, replace an old and flawed question with the improved version in the American Community Survey, perhaps as early as 2021. It will migrate across the entire federal statistical system and gradually make its way into the private sector. Sexual identity has a different set of challenges, but I expect major improvement in the coming decade. In brief, the nation, unfortunately, needs to complete the task of creating racial and gender justice. Each requires a more robust denominator in official statistics.

Finally, what about public education? What role have civil rights groups and litigators played in the fight against the citizenship question being included?

A very important role, as I indicated previously. And these players have more to do. The 2020 census needs multiple “trusted voices.” Americans hesitant about cooperating with the census can best be persuaded by their minister, school teacher, best friend, or the person they go bowling with. Trusted voices who will say, “Oh, come on, you should do this.” Of course some of the traditional “trusted voices” are themselves hesitant, especially where the undocumented are involved.

Another important initiative takes place in the public school system, where the children bring back home the message that the census matters, based on materials provided and experience as mock enumerators and census-takers. Also, we have just witnessed extensive media coverage of the legal back and forth over the citizenship question. This was useful preparation for the much greater task of encouraging census cooperation, and many journalists have now graduated from “Census 101” and will undoubtedly make a positive contribution in 2020.

I summarize: there is lingering damage from the very partisan debate about the citizenship question. If the toxic partisanship persists — especially if sizeable sections of the population remain fearful that the census will be used against them — the census will be fighting strong headwinds.

Simultaneously, however, in the Census Bureau and across a broad array of civic actors, there is preparation for a census with a positive message: if you care about your country and want to help your community, cooperate with the census that our Founding Fathers gave to us.

For now, we must cope with some uncertainty about whether there will be more tailwinds than headwinds come April 1, 2020.

2020 Census: History, Justice, Representation

Read more stories like this in the Carnegie Reporter