Who Counts in America?

Anne Kim: Let’s start with some basics. Because I think a lot of people just take the census for granted and don’t really know why it exists. But this is something that’s in the U.S. Constitution, right? Can you give us a little bit of the history of why we have it?

Margo Anderson: Yes, in fact I carry around my Constitution at all times. In the 1780s, just after the Revolutionary War, a bunch of men got together in Philadelphia and wrote the document that we still use now. They put a census every ten years into that document for two purposes at the time. One was to allocate seats among the states according to their respective numbers in the House of Representatives and the votes in the Electoral College. And the other piece of it was to allocate tax responsibilities that the states would pay to the federal government. That latter piece of it never really worked out, but the representational part became an absolutely core factor in the structure of the American state.

Anne Kim: And we’ll come back to that in a little bit. But I want to fast-forward to today, because census data is embedded in every interaction that citizens have with the U.S. government. On a practical level, can you talk about a few of the ways in which federal agencies use census data today that affects us as ordinary citizens in our everyday lives?

Margo Anderson: I’m going to start early in the sense that once the Framers asked for this to be done, Congress every decade would have to sit down and write a law about how to do it. As they did that, they realized there were other things that they could learn and use as they did the normal apportionment process. Over time they would try new questions, add questions, change them around, and it’s been quite successful. It basically grounds what is called the federal statistical system. 

Because we know population patterns, we also know where people live, we know what kind of work they do. We know distributions by age and sex and race and ethnicity, and all that helps to benchmark everything else that we want to know: unemployment information, fertility and mortality, public health data. The census sort of sits underneath that as a framework for almost all policymaking.

Anne Kim: So Medicaid, for instance, the so-called FMAP percentage, the percentage of federal money that goes to states for Medicaid, is dependent upon population measures, right? And there are a lot of formula grants that are based upon how big a state’s population is.

Margo Anderson: And we started doing that in the nineteenth century. So we’ve been doing it for a long time.

Anne Kim: But the census is a lot more than a statistical exercise, right? I want to broaden the lens again because I think it’s important context for understanding why the changes that the Trump administration have proposed are so important and potentially very damaging. 

You have written the definitive social history on the census, and I have to say that I found this book completely fascinating. It’s called The American Census: A Social History, and it’s a textbook, but it does not read like a textbook. 

One of my big takeaways from reading your work is that people should not think of the census simply as a collection of data. It’s a powerful tool that’s really shaped how we see ourselves as a country. One of the lines you wrote that really stuck with me is that the census is “a crucial marker for American history.” Can you talk a little bit about what you mean by that?

Margo Anderson: By the way, what you’ve just described was something of a surprise to Americans in the late eighteenth and nineteenth centuries. They were new to this process. They thought this would work for their political problems, which were how many seats does Virginia get versus Rhode Island versus Massachusetts versus Ohio and so forth. But what they discovered was exactly what you’re describing. What they learned was that the American population grew very rapidly: 30 percent a decade until about 1880.

Think about that. Every time we counted, it meant that there were 30 percent more people than there had been the last time. What they also discovered was diversity: that Americans are a very complicated people. We’re different by generational age, by sex, by race and ethnicity, and by what kind of work everybody does. And as the process developed, they realized that connected to issues that were before Congress at the time. So for example, and this wouldn’t be so controversial today, but [one question before Congress] was, “Do we give pensions to Revolutionary War veterans?” 

Anne Kim: Well, yes, we should.

Margo Anderson: But where are they? And how many of them are there? And what are the financial implications for that? They discover that the census will help, right? They’re interested in issues like, well, do we have enough blacksmiths? Do we need to build more roads? Where do we need to build more roads? All that emerges in time.

This is a quote from Census Director Barbara Bryant from 1990: “The census moves power and money.” In other words, it tells you where to put the public resources that American taxpayers pay. And what the people discovered quickly is that the more people you have, the more members of Congress you got. There were kind of winners and losers, just the way there are in elections. 

This realization then meant that the census, once it was understood, which was by the mid-nineteenth century, Congress said, we really have to pay attention and do this well. That led to a lot of technical innovation and the political understandings [we have today].

Anne Kim: I want to ask about the political uses of the census over time. And in particular, there seems to be a chapter in history that’s very relevant today: the use and abuse of the census right before the Civil War. Can you talk a little bit about that and the politicization of the census then?

Margo Anderson: By the third or fourth census, by the 1820s, there’s a heavy, furious debate going on about what we are going to do about slavery. 

It developed into two positions. The South basically said, no, it’s fine, we have a slave labor economy, and the northern states began to say, no, no, no, no, this is not good. We need to figure out systems of abolishing slavery, emancipating the slaves, and then integrating the former slave population, which is 15 to 20 percent of the total at the time, into American life as free people. 

That continued well into the 1850s, and I argue in the book that one of the things that the southern states did when they saw the 1860 census result was to realize that politically the gig was up. They were going to lose more because they weren’t growing as quickly as the rest of the country, and so they seceded. Then the Reconstruction Congresses had to decide how to abolish slavery, integrate the slave population into the United States and the electoral system, and then protect voting rights for what were then called the freedmen. That sort of seared in the public consciousness how important the demographics of the country are. And it echoes later through other debates about separate but equal, Jim Crow, immigration restriction, and so forth.

Anne Kim: So this finally brings us up to Trump and what he’s proposing. So in some sense, you have to give the Trump administration some credit for zeroing in on the census. It is this immensely powerful tool, but it also seems boring enough that people maybe aren’t paying attention. On the surface, what the Trump administration has proposed may seem to a lot of people like just technical changes in how the census is conducted. But there is a lot more at stake here, which is what your piece is getting at. What is really going on underneath the radar?

Margo Anderson: What the administration has proposed in a Federal Register notice that was issued on September 10, which is what you’re talking about, is to change the concepts that underpin the enterprise: who gets counted, where they get counted, and when they get counted. 

We [currently] count people by household at what is called their “usual place of abode.” They’re changing that term and saying that people who don’t have what they they’re calling “the durable permission of the sovereign to live in the country”—namely people that they think are illegally in the country—shouldn’t be counted. So that’s one big piece of it. There have been debates about counting what are either called “illegal aliens” [or] “undocumented immigrants”—the terms shift—for the last half century. They think they have a new methodology to do that. The second issue is to change the questionnaire to take off the race and ethnicity questions that have been asked since 1790. 

Now, to do all that, they need a lot of technical work done between now and April 1, 2030. One of the questions that I raise is I don’t think they’re ready to do it without making a complete hash of the technical problems. The Federal Register notice raises all those issues but doesn’t solve them. 

We’re in a period where the public can respond, and yesterday when I looked at the number of responses to the Federal Register notice, there’s over 50,000. Most of the overwhelming number saying, no, no, no, no.

Anne Kim: Wow.

Margo Anderson: And this includes some people saying, look, some of these ideas are interesting, but we should take some time to think about this more carefully before we rush headlong into dramatic changes.

Anne Kim: Well, this administration has not been afraid to rush headlong into a lot of dramatic changes without thinking things through, unfortunately. So they will make a hash out of it if they can. I want to get into a little bit of the details about both pieces. 

Let’s take first the administration’s proposal to basically only count citizens and people with green cards. So just to make that clear, we’re talking about excluding students who are here on visas, refugees, asylees …  Is there any historical rationale for doing things this way? 

Margo Anderson: There has been litigation and debates about this since the 1970s, but they have never made it to the point of changing the language or challenging the language that’s in the 14th Amendment that says we count the “whole number of persons.”

Anne Kim: Right. Not citizens, persons. 

Margo Anderson: And the Census Bureau, to their credit, basically says, look, our job is to count the population, right? The other agencies of the federal government do a lot of other demographic and economic measurements as a part of their administrative responsibilities. So for example, when we pay our taxes, we write down where we live and we file them every year. There’s often household information in that. What is now Homeland Security has records of the foreign-born immigration population. 

Starting in the 1970s, because of the way they changed the laws, Congress walked us into this problem of what is called “the undocumented.” Namely, they changed the immigration statutes but didn’t put in an administrative structure to monitor it. The immigration law passed in 1965 for the first time said that migrants from the Western Hemisphere, Canada, Mexico, for example, would come under immigration quotas. (Europeans and Asians and so forth already were under quota systems.)

Anne Kim: My parents were under the quota when they came in.

Margo Anderson: But the people who were from the Americas, the Western Hemisphere, literally could cross the border, and there was no record keeping of who was going back and forth. Congress did not put in a procedure at that point that stopped every person going back and forth across the border and asking, “where are you from? Why are you here? Should you be here? Are you allowed?”

That was already the case for anybody who was coming from Europe, Africa, Asia, and so forth outside of the Western Hemisphere. Because there had been porous boundaries at the northern and southern borders and Caribbeans, people kept coming and going despite the change. By the late ‘70s and early ‘80s, we had a population of people who were “undocumented”—the technical term is “entered without inspection.” 

Anne Kim: How did that affect the census and how these people are counted or not?

Margo Anderson: The census counts everybody and asks people where they were born. And it turned out more people claimed they were born overseas in a different country than the immigration officials had in their records. So that leads to debates about, well, what do we do about that? 

[In the past, we said] people who entered without inspection but have been good residents of the country, if they can prove when they came and what they’ve been doing, will be allowed them to stay. We passed legislation in the mid-1980s to do that. More recently we’ve passed legislation to create the DACA program: Deferred Action for Childhood Arrivals. 

The other side of that, of course, is people who say none of these folks should be here, and we’re going to crack down and identify people and deport them. This administration has moved very strongly toward revoking privileges to be in the country that already existed. That’s the policy position underneath what the census changes are about. 

The argument they’re trying to make is that the Framers meant that all along, so they cite a lot of eighteenth-century laws about residential status and sovereignty and who gets to make this decision. But none of that has been tested in law or litigated yet.

Anne Kim: The practical effect of excluding all these people from the census will be to change apportionment to some extent, but also to shift a lot of, as you say, money and power, right?

Margo Anderson:  We went through a similar discussion in the 1960s when people recognized that the census was missing substantial numbers of people in what was called the “differential undercount.” Cities were being undercounted compared to their suburbs. Some remote rural areas were being undercounted relative to others. There has been a 50-year effort for the census to count better.

Anne Kim: There’s some symbolic value in this, right? To say that every person counts for the purpose of the census sends a message that every person in this country counts. And now we’re sending the message that some people count, and some people don’t matter because they don’t count. There’s a moral underpinning to all of this.

Okay, so let’s turn to the second set of major changes that the Trump administration has proposed around questions of race, ethnicity, and sexual orientation. I think I heard you say early in this conversation that we’ve been asking about race since the 1790s. What’s the history of this set of questions?

Margo Anderson: The language that is in the Constitution was lifted out of a debate about tax responsibilities in the Continental Congress a few years earlier. Then they were debating how to raise money for a national government. At that point the colonies, now states, were independent sovereign states. There was no executive branch; there was no national judiciary. They needed money to pay for George Washington’s army, to send diplomats overseas to try to convince the French to join their side and then eventually to negotiate a peace treaty with the British. But where does that money come from? We didn’t have any taxing authority nationally. 

What they did is debate what taxes are based on wealth and economic capacity. 

And of course it’s an agricultural society and they said, well, it’s either land or people. So they go back and forth, and their original solution is to measure land, because as soon as they start talking about counting people, they confront the question of slavery. It would have meant admitting that the slaves were an incredibly important part of the capacity of the society. 

So they try land, and it turns out that they never quite get it done. When they come back in 1783 after winning the Revolutionary War, they say, no, it’s got to be people. And they put in what becomes the “Three-Fifths Compromise” that says that the slaves are not as productive as free people of any age, sex, or condition. That meant once it was lifted into the Constitution, they had to separate out the slave population and the free population. The slave population is designated as “colored” or “black,” so we asked that basic race distinction right at the outset. 

The second question they asked was, what do we do about the Indigenous population? There were still lots of Indians living in the 13 states. And they said, well, if they pay taxes locally—and at that point there were what were called “head taxes” or “poll taxes” or “capitation taxes”—we count them because they pay taxes to the state. They contribute. 

So the two distinctions that go into the Constitution are white—which is free white—and slave, or “colored,” and we don’t count Indians that are not taxed. It evolved over time as America wrestled with its racial identity. Congress wanted to know whether the free black population and the slave population were the same people. They wanted to know whether there is what they called at the time a “mulatto” population, a mixed-race population. 

After the Chinese show up on the West Coast during the Gold Rush, they wanted to find out if the Chinese are a nationality or a race. What they decide is that they’re a race, not a foreign-born person. So we end up evolving the categories in the context of very difficult issues about what “all men are created equal” means and then embed them into the system such that we have a very complicated framework right now.

Anne Kim: You have this terrific line in your piece about the 2030 census where you say “the nation has counted to discriminate and has counted to liberate.” You’re talking about the evolution of all these different categories, and now the Trump administration wants to wipe all that away. Does that take us to a place where we discriminate, or is that going to take us to a place where we’re “liberated,” as the Trump administration claims?

Margo Anderson: They’re arguing that that because “all men are created equal,” the constitution should be “colorblind.” Now, interestingly, most of the race and ethnic categories aren’t, “color” categories. The “color” categories were shorthand for the fact that we had slaves, we have Chinese, we have Mexicans, we have Koreans. At one point they had a category called “Hindu” in the census. 

They’re saying we have to just stop all this, but it’s patently ridiculous to do that without [understanding] the implications of doing that. Are they saying that there is no more discrimination in the United States on the basis of race and ethnicity? 

Anne Kim: Take the two changes that are proposed together, the citizenship question and the race question, and let’s assume for a second that the administration is competent enough to carry this out. What do you see as the implications for what the 2030 census will find and how that’s going to reverberate throughout policy and the way that we Americans see each other and ourselves?

Margo Anderson: So first, it’s still only year six and we have a long way to go before we hit year zero, namely 2030. We have elections between now and then. Technically, Congress can move in and completely change all this stuff somewhere between now and 2030. That’s the first thing. 

The Trump administration clearly thinks that this will benefit—in the immortal words of the redistricter who was behind the debate about putting a citizenship question on the 2020 census—that this will benefit Republicans and non-Hispanic whites. 

The Supreme Court didn’t let the [first] Trump administration put a citizenship question on the 2020 census, so we don’t know what that would have produced. We’ve never had a question on citizenship on the full count census.

Anne Kim: It’s possible—probably—that there will be unintended consequences and that whatever happens may not actually be what the administration thinks it’s going to get.

Margo Anderson: Most large corporations and so forth have demographic shops in their marketing operations. They need to know where to put the shopping center. They need to have a sense of what the local area is for the Kmart or the Target or whatever. If the data becomes unreliable, that doesn’t help. There are people who say, well, we can get it from other sources. Well, yes, you can, but the census, it turns out, is also the anchor that sits underneath a lot of the methodology for those “other sources.” You could change a lot of stuff, but it would mean changes all the way up and down the way data analysis has been done for the last century.