On the giant list of problems facing the country compiled daily by stressed-out Americans, the fate of the 2020 census is perhaps not near the top. But no matter what you care about, if it has to deal with politics, you should care a bit about the census — especially since Census Bureau director John Thompson just resigned last week. Also, the agency's staff are expected to debut an entire new digital system for collecting the data without spending any more money than they did for the 2010 census. In short, things aren't looking too good at this exact moment for those hoping that the census paints an accurate portrait of America, as the framers of the Constitution demanded.
A bad census (i.e., one that started with massive snafus that called into question the data, or that obviously under- or overcounted certain populations) could have huge implications for what policies do or don't get implemented after the 2020 election. On top of that, if the census goes completely awry, we'll have an even more imperfect idea of what our country looks like at a time when there’s already much disagreement about its contours.
So … how bad do things look? Is the Census Bureau going to be ready to start counting us all up on April 1, 2020? We called up Kenneth Prewitt, who directed the Census Bureau before the 2000 census, to see where things stand.
[This interview has been edited and condensed for clarity.]
MTV News: To start off, let's get the big question out of the way. Exactly how much trouble is the 2020 census in right now?
Kenneth Prewitt: There are two issues that make it very troubling. One is the leadership vacuum, which can be fixed, but needs to be fixed intelligently. The second issue is the budget. Congress instructed the Census Bureau to conduct the census in 2020 without spending more money than they did in 2010. That's unprecedented. Nothing had been dumped on the census like that before. The population grows, people are harder to count, they are less trusting and more uncooperative, there are many more people for whom English is a second language — all things that make the next census harder to count than the prior one. The bureau went to work, investing heavily in new technology that would allow it to try and keep the census cost down. And with new technologies, you don't know until you've tried them if they're going to work.
Was it ever within the realm of possibility that the 2020 census could have cost the same as the 2010 one?
Prewitt: No. The 2000 census was budgeted at close to $7 billion. It actually came in under budget — we were pleased with that. It was a healthy budget. The 2010 one was nearly double [that]. And then to be told, not counting for inflation, not counting for anything, that you have to maintain that same level of funding … I should say that [former Census Bureau director] John Thompson is a man quite capable of doing that. He's enormously competent, a good manager and statistician, and knew what new technologies had to be put in place. As [recently] as 10 days ago, I had lunch with him. He was feeling very good about where we were. He was feeling we could have a healthy census.
Has there ever been an online government project that needed to handle an entire country using it?
Prewitt: Well, no. Obamacare came as close to that in recent history. There's no historical analogy to this. That doesn't mean that we didn't face new challenges before. The whole growth of urbanization in the early decades, and the rush of a new immigrant population [between] the late 1880s and early 1900s created new conditions, and the bureau did a good job on all of that — but it had the full backing of the United States Congress.
I saw somewhere that you said that if there were a bad census, you can imagine that discussions about the private sector getting involved might begin.
Prewitt: Yes, I think that will be one of the responses. Indicators of a not-very-good census would be something like a technological flaw that when you roll it out, people can't get [online] and it doesn't work — as in the Obamacare rollout. Remember, it took weeks to sort all that out.
I can imagine at the end of the 2020 census, [there will be] a kind of unhappiness with the whole effort and an even greater use of private-sector agencies. People saying, "Google could do the address file." Well, yes, but Google would have to do it as thoroughly as the census does it — and in places that aren't on the internet. I would think that this is a public responsibility, it is a public good, it's one of the biggest things the government is responsible [for], and I think a movement away from that would not be a healthy development for democracy.
In what ways would a severe undercount affect certain populations that are hard to count?
Prewitt: The census is the sample frame for all [government] surveys, private-sector surveys, etc., for [the following] decade. We adjust it during that decade, but you start with that census. Let us say — and this is an exaggeration, it won't happen — but let us say that the census is upcounting the American population in 2020 as 52 percent female and 48 percent male. All other surveys that are calibrated against the census would be calibrated as if that is fact. It would be a mess.
We wouldn't allow that to happen, because we'd know that was wrong, but more quietly, if it's the undercount that is spread across otherwise marginalized population groups, they will be invisible not just in the census, but they will be invisible in any other survey — labor statistics surveys, transportation surveys, health surveys — which are calibrated to the census framework. That would be damage that's not visible, but it would work its way into our entire federal statistical system, and then work its way into policy.
One other issue that's come up a lot this year: Many LGBT advocates have tried to get the Census Bureau to start asking questions about sexual orientation and gender identity. Ultimately this year, the questions were turned down. Why do you think the census hasn't asked these questions?
Prewitt: It's a tough call. There's certainly no doubt in this particular environment that counting LGBT [individuals] would have created a backlash among those who are already very angry about the legitimacy of same-sex marriage. For the bureau to have counted LGBT [individuals], there would have to be a strong reason that a given agency, perhaps the Department of Health, needed to know [information about them] because it had consequences for public health policy. It was put in play as an identity issue, not a policy programmatic issue.
Now, if it could be shown that having that count would allow us to reduce discrimination against the LGBT community, now that would be a rationale for it. But that would have to be an argument that the Department of Justice made. The Census [Bureau] doesn't just invent questions. It responds to the U.S. Congress. The Constitution says, the census shall be done as the Congress does declare. The Census [Bureau] is not an independent agency that can just add whatever questions it wants. There's a deep, complicated process going on right now in respect to whether the Middle Eastern/North African category will be added to the race/ethnicity question.
If the next census is bad, it's going to give us an especially unhelpful portrait of the youngest Americans, who might not have been alive during the 2010 census, or who were at least living with their parents. This happens to also be the most diverse segment of the population. As far as counting young Americans — especially those who have left home — is it really hard for the census to track them down?
Prewitt: Yes. Unrelated people living together is a very difficult group to count, because they themselves don't know who's going to leave. Someone's going off on a trip with their friend to Europe — are they coming back or not? Do you count them or not? If they don't think the census has anything to do with them, they won't answer. We actually don't have a good sense of the cooperation we'll get out of that population group. And that's why the advertising campaign is so critical. And the follow-up strategy is so critical.
We do have a whole new population that's grown up in a high-tech world. The Census Bureau is working hard to figure out how you get the attention of those people and make them feel like the census is something they ought to do or is fun to do. If you don't have an advertising budget, that's a population you're less likely to count at 100 percent.
And right now, there's no advertising money.
Prewitt: It was in the budget, but it was not funded.
It seems like the dress rehearsal is especially important this time because they're trying out an entire new digital system.
Prewitt: If the design were exactly like the one used in 2010, which was largely the design used in 2000, then this would be less worrisome.
As someone who has conducted a U.S. census, what kinds of problems did you find during the test?
Prewitt: One of the biggest issues that we needed to spend additional money on in 1999 was a flood of new building. The economy was strong, and there was an enormous expansion of the housing stock in Las Vegas, around Phoenix, in the Southwest overall — and these were buildings that weren't even there when we [compiled the address file for that census]. In the last year, we had to adjust the address file to take into account that surge of housing-stock growth. And we did. Had we not done that, the Republicans would have rightly complained. Because we would have missed all those households, which tended to be disproportionately [in] Republican voting districts. And they would have been right.
Whatever happens between now and 2020, it has to feel like it is of zero benefit to one party or the other. If [one party does benefit more], there will be a withdrawal of confidence. And that will spread to the entire federal statistical system. If parts of the public have no confidence in the numbers [used] to govern it, whether it's health or education … we have never been there and I hope we never get there. I'm not saying we will get there, but a bad census is a very public event.